Digitized by the Internet Archive
in 2010 with funding from
University of Toronto
http://www.archive.org/details/generalelectricr23gene
I '5^3 ;
o
THE
General Electric Review
VOLUME XXIII
1920
PUBLISHED BY
GENERAL ELECTRIC COMPANY
SCHENECTADY, N. Y.
T'K
I
TWO DOLLARS PER YEAR
TWENTY CENTS PER COPY
GENERAL ELECTRIC
REVIEW
VOL, XXIII, No. 1
Published by
General Electric Company's Publication Bureau,
Schenectady, N. Y.
JANUARY, 1920
AUTOMATIC ARC WELDING MACHINE
In Operation on a 14-in. Shaft, Increasing the Diameter 's in.
by Means of Self-feeding Wire Electrode
/~\NE of the few things not affected by
^-^ the World War or by the conditions
which have followed it, is "NORHfl' Qual-
ity. Prices of materials and labor have
changed. But in all the essentials that
define value and determine serviceability,
"NORf!\fl" standards have been inflexibly
maintained regardless of cost.
See that your Motors
are "NORMA" Equipped
THE NORMA COMPANY OF AMERICA
ANABLE AVE.. LONG ISLAND CITY, NEW YORK. N. Y.
Hall. Roller. Thrust, Combination Bearintis
General Electric Review
A MONTHLY MAGAZINE FOR ENGINEERS
Manager, M. P. RICE Editor, JOHN R. HEWETT -^^^^'-"^ Ed.tors B M. EOFF and E. C_^SAXDERS
In Charge of Advertising. B. M. EOFF
Subscription Rates: United States and Mexico, $2.00 per year; Canada, $2.25 per year; Foreign, .$2.50 per year; payable in
advance. LibroTy and Student Rates: United States and Mexico. $1.50 per year; Canada, $1.75 per year; Foreign, $2.00 per year;
payable in advance.
Remit by post-office or express money orders, bank checks, or drafts, made payable t3 the General Electric Review. Schenec-
tady. N. Y.
Entered as second-class matter. March 26. 1912, at the post office at Schenectady, N, Y., under the Act of March. 1879.
Vol. XXIII. No. 1 „, Ge,^r^^&\ru'Z,.pa,.y JANUARY, 1920
CONTENTS Page
Frontispiece: William LeRo)- Emmet 2
Edison Medal for 1919 is Awarded to William LeRoy Emmet .3
Some Developments in the Electrical Industry During,' 1919
By John Liston
Thermostatic Metal 57
Bv Hhxky Herrman
Electric Propulsion of Merchant Ships GO
Bv W. L. R. EiMMi-T
Im]jroving the Mazda Autdmoliilc Headlight Lamj) . . . 07
B\- L. C. PORTliR
An Absolute Method for Determining Coefficients of Diffuse Reflection .... 72
B^" F. A. Benford
WILLIAM LEROY EMMET
who has recently been awarded the Edison Medal lor inventions
and developments of electrical apparatus and prime movers
EDISON MEDAL FOR 1919 IS AWARDED TO WILLIAM LeROY EMMET
The Edison Medal Committee of the
American Institute of Electrical Engineers
recently announced that the Edison Medal
for the year 191!) had been awarded to
William LcRoy Emmet "for inventions and
developments of electrical apparatus and
prime movers."
This is a signal honor for Mr. Emmet —
an honor right well deserved as reward for his
valuable work in the electrical industry, and
for the courage and masterh- ability which
he displayed in evolving the steam turbine
from an obscure embryonic stage of dc^'clop-
mcnt to the most highly improved and satis-
factory prime mover known; and in later
years for his advocacy and development of a
system of electric propulsion for ships of
the navy and other large vessels which has
been phenomenallv successful in application
and which has every indication of being epoch
making.
Mr. Emmet was born at Pelham, N. Y.,
July 10, 1S59, son of William Jenkins and
Julia Colt (Picrson) Emmet, grandson of
Robert and Rosina (Hublcy) Emmet, and
great-grandson of Thomas Addis Emmet
(q.v.), the first one of the family in America.
The latter was the distinguished Irish patriot
and leader in the Society of United Irishmen
in 179S, and an elder brother of the ideal
patriot of the Irish race, Robert Emmet, who
was executed in Dublin in ISO.'j.
He was educated at schools in Canada,
New York, and subsequently entered the
United States Naval Academy, where he
was graduated in ISSl. He served as a cadet
midshipman until 1SS3 at Annapolis and on
board U. S. S. Essex, and re-entered the Navy
as junior lieutenant in 1898, serving as
navigator on the U. S. S. Justin during the
period of the Spanish War.
Mr. Emmet first became associated with
electrical work in 1887 when he entered the
employ of the Sprague Electric Railway &
Motor Compan}'. He later went with the
Buffalo Railway Company as electrical engi-
neer, and soon afterwards accepted a position
with the Edison General Electric Co., in the
Chicago district. His association with the
General Electric Company began with its
organization in 1892.
Prior to his achievements in the steam
turbine field, he attained prominence through
his work in developing the general use of
alternating current, and a number of inven-
tions, which since have come into universal
use, stand to his credit. Among his more
important electrical inventions arc the oil
switch and varnished cambric cable; he also
invented several types of transformers, several
different forms of insulation for alternators,
and many devices that are employed in con-
nection with the Curtis steam turbine. His
most brilliant accomplishments, however,
have been more in the nature of an institutor
of new methods and ideas than an inventor,
and a great deal of his most useful work could
not be patented nor perhaps even classified
as invention. His qualifications have specially
fitted him. for finding new scope for the talent
and facilities of the Company's organization.
Mr. Emmet is the author of "Alternating
Current Wiring and Distribution" (1894),
and of ntunerous important papers presented
before the American Institute of Electrical
Engineers and other engineering societies.
He is a member of the American Philo-
sophical Society, American Institute of Elec-
trical Engineers, American Society of Mechan-
ical Engineers, Society of Naval Architects
and Marine Engineers and of the Naval
Consulting Board of the United States. He
is also a member of the University and
Engineers' Clubs of New York, the Mohawk
Golf, Tobique Salmon, Mohawk, Edison,
and Schenectady Boat Clubs. He received
the honorary degree of D. Sc. from Union
College.
lanuary, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. I
Some Developments in the Electrical Industry
During 1919
By John Liston
Plblication Bureau, General Electric Company
Mr. Liston's annual review has been a feature of our January issue for a number of years. It is always
an instructive and interesting summary of the recent developments in the industry, and this year the author
has more to tell than ever before. The year 1919, in spite of a bad start, was one of the best business years
that we have known and was chock-full of new enterprise. — Editor.
With the termination of hostilities, the
Electrical Industry-, as the result of its pre-
vious intensified efforts in research work and
constnictive production for war purposes,
found itself possessed of a rich heritage of
scientific achievement, much of which could be
practically applied to meet commercial needs.
Even at the beginning of the first year of
peace, the readjustment to a peace basis,
which had begun promptly after receipt of
the news of the signing of the armistice, was
well advanced in many lines. Thus, projects
which had perforce to be abandoned during
the war, were again carried forward, and,
combined with the urgent requirements of
reawakened industries for electrical appara-
tus, resulted in an unprecedented volume of
output by the end of the year, despite the
unfavorable conditions at its beginning.
Among the many ])rominent dcveloijm.ents,
one of the most important is not electrical,
although it is a direct product of the electrical
industry. This is the \-ery marked increase
in the equipment of inerchant shijjs with
geared-turbine dri^e, as the result of the
favorable operating records made by ships
so propelled, some of which ha\-e been in
sen-ice for a period of more than three years
(Fig. 1) without replacement of, or repairs
to, their turbine installations.
Another important item is the adoption,
for the first time, of electrical propulsion for
large cargo boats and the standardization of
this method for the larger ships of our navy.
In aviation work, the supercharger has
given every indication of present and ])oten-
tial value in this special field, while the radio
developments referred to constitute a com-
\)\^ite new system which is susceptible of wide-
spread application on a commercial basis.
As in previous articles on this subject, the
electrical ajjparatus, turbines, etc., referred
to are all i)roducts of the General Electric
Company, but references to their develop-
ment will serve as an indication of the ten-
dencies in design and construction as well
as the general trend of i)rogress in the electri-
cal manufacturing industry as a whole.
Fig. 1. S.S. Hanna Nielsen Propelled by a 2S00-h.p. Two plane Type Marine Geared Turbine. This ship has an active
service record of over four years without any repairs or replacements beins required for its turbine equipment
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919
Turbines
The removal of the pressure on turbine
production which characterized the period of
the war, the release from government control
and the consequent rescheduhng and cancella-
tion of turbines on order, together with con-
ditions which had arisen due to shortage in
skilled labor and other warti:ne handicaps,
had the effect of temporarily arresting turbine
development at the beginning of the year.
In a short time, however, the situation
changed radically: new parts were manu-
factured to replace those which, due to war-
time conditions, were below the high standard
which in recent years has been demanded in
turbine construction, and in addition an
average of two turbines per month of from
Some interesting facts in regard to modern
turbine economies are found in figures pub-
lished during the year showing the overall
station economy obtained in two moderate
sized stations. The New Cornelia Copper
Company, with 75()()-kw. turbines, operating
in conjunction with a spray pond, produced
under average operating conditions a net kilo-
watt hour from 18,S37 B.t.u. This corresponds
to a thermal efficiency of 18.15 per cent.
In comparing the economy of this cooling
pond station with that obtainable in tide
water plants, allowance should be made for
the obtainable vacuum under operating
conditions, and the increased pumping head
due to the spray nozzles. This point is illus-
trated by the results obtained at the plant
Fig. 2.
25,000-kw.. 1800-r.p.m., 17-stage Turbine Direct Connected to 25,000-kw., 60-cycle Alternating Current
Generator. This set is typical of the large turbo-generators produced during 1919
lo,000-kw. to 35,()00-kw. capacity were pro-
duced, as well as full output in smaller
units.
Due to the high price of fuel, economy in
steam turbine operation is an increasingly
important consideration. New development
has, therefore, been along the line of im.-
proved economy, even in comparatively small
ratings.
A number of the large single cylinder tur-
bines (Fig. 2) which were designed priirarih-
for stations where economy is essential were
specially successful in actual service. They
are ver}- carefully proportioned so that at
the point of maximum economy the velocity
ratios and steam areas will be conducive to
higher efficiency than has hitherto been
possible.
of the Arizona Power Company, where with
a (iOOO-kw. turbine and cool water, a net
kilowatt hour was produced from 18,628 B.t.u.
or a thermal efficiency of 18.3 per cent.
While these economies are rendered possi-
ble by the turbine design, they are also due
in considerable degree to intelligent operation
and careful selection of all auxiliary station
apparatus.
In smaller capacities improved economies
have been secured by refinements in design.
In addition to driving generators (Fig. 3)
these turbines are used for driving, through
gearing, large pum_ps for municipal pmnping
plants. In this class of work economy is of
the utmost importance, as the units operate
at full load for practically 24 hours a day and
3()o days a vear.
Januarys 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 1
Marine Geared Turbines
On November 1, 1919, there were 24.5 cargo
boats in ser\'ice which were equipped with
G-E marine geared turbines having an aggre-
gate rating of 612,500 h.p. Of these, 130, with
a total horse power of 320,000, were installed
after January 1, 1919.
The turbine sets, comprising five-stage
ahead and two-stage astern units and one-
plane double reduction gears, were built in
horse power ratings of 1800, 2400, 2500, 2600,
2800, 3000 and 4000, and are specifically
designed for propelling modern cargo boats,
oil tankers and refrigerating ships. The cur\-e.
Fig. 4, shows the very rapid increase in the
number of units installed since 1915.
Practically all ships equipped since 1916
have the so-called one-plane type; i.e., the
turbine and gear shafts all lie in one hori-
zontal plane; this inherently simple design
being adopted owing to the necessity for the
speedy production of propulsion apparatus for
ships to carry on the world war trade.
The pre-war 2500-h.p. geared turbine of
the two-plane type is shown in Fig. 5. This
design was adapted to the units required
for emergency shipbuilding during the war,
and gave practically 100 per cent of con-
tinuous service without replacement; this
while in the hands of the ever changing crews
260
W15
I9I6
1917 1918
Years
1110
Fig. 4. Marine Geared Turbines in Service in
Merchant Ships up to Nov. 1, 1919
which manned the ships both during and
succeeding the war.
A very large percentage of the gears of
the pre-war two-plane type have been in
service for nearly four years with negligible
replacements. With the liberal increase in
Fig. 3 3000-kw.. 3600-r.p.m., 40-stage Turbo-alternator with Vertical Condenser
SOME DEVELOPAIENTS IX THE ELECTRICAL INDUSTRY DURING 1919
size and consequent lower tooth pressure,
and with the immense amount of experience
gained in design, installation, lubrication,
care and maintenance, a practically indefinite
life can reasonably be predicted for this type
of apparatus.
Four of the twelve 6000-h.p. tmits for the
propulsion of the "B" type of ship being
built at Hog Island were shipped, and the
balance are now being finished at the rate
of one complete turbine equipment every
two weeks.
These units (Fig. U) consist of a six-stage
ahead and two-stage astern cross-compound
turbine divided into a high-pressure and a
low-pressure section, and double reduction
twin drive one-plane type gears; each turbine
section developing one half of the full power.
The design of the steam connections is such
that in case of necessity one section can
propel the ship independently of the other
section. The maneuvering is easily accom-
plished through the operation of one double
levered throttle valve.
Torpedo Boat Destroyers
The last three of six turbine driven destro}"-
ers, viz., the McKean, No. 90, the Hard-
ing, No. 91, and the Gridley, No. 92, had
their trial trips and went into ser\nce early in
the year.
Each propulsion unit, of which there are
two for each destroyer, consists of one 800-
h.p., lS.36-r.p.m., seven-stage cruising tur-
bine, immediately forward of and connected
Fig. 5.
Pre-war Type 2500-h.p., Two-plane
Marine Geared Turbine
by a flanged coupling to a 13,300-h.p., 3-197-
r.p.m., 16-stage ahead and one stage astern
main turbine (Fig. 7) driving through twin
single reduction gearing. The propeller
shaft revolves at
full speed ahead.
4.52 r.p.m. when running
Fig. 6. Testing a 6000-h.p. Cross Compound Turbine Set for Driving B Type Cargo Boat
8 Jamian-, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 1
Fig.
7. Destroyer Type Cruising and Main Turbines and Reduction Gear
Cruising Turbine. 800 h.p. ; Main Turbine, 13.500 h.p.
The total speed reduction from the main
turbine speed to the low-speed gear, or pro-
peller speed, gives a ratio of 7.74 to 1.
These turbines operate at 2.50-lb. gauge
steam pressure and 2S-in. vacuum exhaust :
the cruising turbines delivering power up to
185(j r.p.m. (or about 20 knots), with the
main turbines only delivering power above
this speed. The design of the m_ain turbine
permits steam to be admitted to either the
first or fourth-stage nozzles, depending upon
the speed required.
The last of eighty propulsion units similar
to those referred to above were com])letcd in
May, 1919, for forty destroyers (Nos. 29()-:«o),
thirt\'-six of them having been completed be-
fore the armistice (Fig. N).
Lighting and Power Sets for Ships of the U.S. Navy
Many sets, consisting of a multi-stage high-
speed steam turbine connected through flexi-
ble speed-reducing gears to a multi-polar
direct-current generator in capacities of 300
and 400 kw., were supplied to the government
for use on battleships and batten,- chargfing
.submarine tenders.
The turbines are built in both condensing
and non-condensing types, are mechanically
strong and simple in design, and embody
the necessary emergency, back pressure and
circuit breaker devices thoroughly to protect
them in service against the possibilities of
damage even when in charge of inexperi-
enced operators.
A larger unit for this specific sennce, of
oOO-kw. capacity, was developed in 1919 to
meet the rigid specifications of the navy
department. The turbines are of either the
four-stage condensing or non-condensing type,
with a speed of oOOO r.p.m. when oi)erating at
2.50-lb. gauge steam pressure and vacuum
exhaust for the condensing unit, or when
exhausting against 10-lb. gauge back pressure
for the non-condensing machine.
Fig. 8. U. S. Destroyer Robinson in Santa Barbara Channel
SOME DEVELOPMENTS IX THE ELECTRICAL INDUSTRY DURING 1919 9
This design will, in a great man}- cases,
replace the 300-kw. sets where additional
demands have been created for light and
power, and the smaller units are no longer
of sufficient capacity to meet the require-
ments. The 500-kw. size
will also be installed on
new battleships and battle
cruisers.
Electric Propulsion
A great many turbine-
generator propulsion sets
for installation aboard mer-
chant ships are under con-
struction at the present
time and one equipment is
now being installed on the
Powhatan of the West Coast
Steamship Company.
The turbines are rated
3150 h.p., have (S stages,
operate normally at 3000
r.p.m. and are direct con-
nected to 2350-kw., 50-cy-
cle, 1150-voIt alternating-
current generators. Energy
for excitation purposes is
furnished by two high-speed
turbine generator sets.
The propulsion is accomplished by one 60-
pole, 3000-h.p. motor which is connected to
the propeller shaft and revolves at 100 r.p.m.
at full speed ahead.
The turbines are designed for admitting
exhaust steam from the ship's auxiliaries to a
lower stage, and every advantage is taken to
obtain a high over-all operating efficiency.
Maneuvering is normally accomplished by
two levers mounted on the main operating
panel which is located close to the turbine-gen-
Fig. 9. 300-kw. Lighting and Power Geared Turbo-generator Set
erator set ; the speed lever being mechanically
connected to the turbine governor, and the
motor control lever being electrically connected
to the main exciter contactors. In this manner
the handling of the ship for ahead or astern
operation is easily and simply accomplished.
..%**-
Fig. 10. Airplane View of U S.S. New Mexico. The first electrically-propelled battleship
10 January, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 1
U.S. Battleships
During the year many sen-ice investiga-
tions were made aboard the U.S.S. New
Mexico* the first capital ship of the Ameri-
can Navy to be electrically propelled by
Mi^^
Fig. H. Scout Cruiser Salem, the first ship to use spring thrust propellor shaft bearings
turbine-generator sets (Fig. 10); and much
valuable data have been thus obtained which
will be applied in the design and operation
of future electrically driven ships.
The apparatus for the new battleships
California and Maryland is practically com-
pleted, and that for the West Virginia is well
under way. Many improvements and inno-
vations are embodied in these units based on
the experience gained in connection with the
New Mexico's sets.
For the battleship two units with a total
of 60,000 h.p. will be required, and for the
battle cruisers four sets will be installed,
giving a total capacity slightly in excess of
180,000 h.p. to give a maximum speed of
33 knots.
The battleships will each be propelled bv
four 15,G00-h.p., 2-phase, 5000-volt, 221-r.p.m.
motors, while the battle cruisers will each
have eight propelling motors with a unit
rating of 22,500 h.p., 3 phase, 5000 volts at
330 r.p.m.
Spring Thrust Bearings
An interesting detail of the equipment of
some of the electrical!}' propelled ships is the
use on the propeller shafts of spring thrust
bearings having characteristics similar to
those which were originally developed for
use as suspension thrust bearings on vertical
shaft waterwheel-driven generators.
These bearings have been
in use in hydro-electric
plants for several years
and, as they have success-
fully carried the thrust of
rotating loads up to 550.000
lb. in this ser\-ice, they
were adopted as a standard
equipment for G-E vertical
shaft generators, and their
recent adaptation to meet
the requirements of marine
service was a logical devel-
opment.
The pioneer installation
was made in March, 191S.
on the 20-knot twin screw
scout cruiser Salem (Fig.
11) which is driven by
two 10,000-h.p. turbines
through reduction gears.
The thrust exerted by each
propeller shaft when rotat-
ing at 3S0 r.p.m. is 80,000
lb., and the thrust bearings, which are similar
in design to that shown in Fig. 12, are each
located in a housing bolted to the reduction
gear cases, thus forming an integral part of
the turbine equipment.
*See General Electric Review. April, 1919, and G-E
Booklet Y-1307.
Fig. 12. Spring Thrust Bearing Tor Marine Propellor Shafts
After six months of active sen-ice, tiiese
spring thrust bearings, upon examination,
proved to be in as good condition as when
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 11
they were installed, and, in fact, indicated no
appreciable wear.
During 1919 bearings of this type were in
production for different classes of electrically
propelled ships, among them being four cargo
boats, two coast guard cutters, and a fishing
trawler.
The cargo boats are eleven knot single
screw craft, each of which will utilize a 3000-
h.p. motor to drive the propeller shaft at
100 r.p.m., giving a thrust of 57,000 lb. A
spring thrust bearing is located at the forward
end of the motor shaft and its housing consti-
tutes a part of the bearing bracket or end
shield of the motor; this being the first marine
application of a combination motor end-
shield and thrust bearing.
Lubrication of the thrust bearing and motor
iournal bearings will be supplied by a pump,
driven from the motor shaft.
On the 16-knot, single-screw coast guard
cutter the thrust . bearing will be located aft
of the 2600-h.p. 130-r.p.m. driving motor
and will be subjected to a thrust of 33,000 lb.
It will be lubricated by the turbo-generator
oiling system.
The 10} 2-knot single screw trawler with a
400-h.p., direct-current driving motor has a
self-oiling thrust bearing (Fig. 13) located
aft of the motor and sustains a thrust of 7500
lb. with the propeller revolving at 200 r.p.m.
It should be understood that these spring
thrust bearings are all of the single-collar
type, are self-aligning, and show greatly re-
duced friction losses as compared with the
rigid multi-collar type heretofore used for
propeller shafts.*
Radiator Cooling
Among the latest features considered in
the development of the turbine-driven alter-
nator is that of a closed system of ventilation.
Under certain conditions long air ducts of
large cross-section for the inlet and outlet of
cooling air required by the alternator are
highly objectionable, not alone from a con-
sideration of the loss of the valuable space
which they might occupy, but sometimes on
account of the openings being in a region of
poisonous or injurious gases.
In such cases a closed system of ventilation
is desirable. With this arrangement, how-
ever, it is necessary to remove the heat of
the generator losses from the cooling air,
which is used over and over again. Generally
an air washer may be utilized for this purpose
•Article General Electric Review, February, 1919, by
H. G. Reist.
to the best advantage, but there are important
cases where the use of an air washer is impos-
sible because of the character of the available
cooling water.
In order to meet such special conditions,
it is now considered quite advisable, as a
Fig. 13. Spring Thrust Bearing Installed on
Steam Trawler
result of nmnerous tests, to utilize a water
cooled radiator of the fin and tube type whose
function would be the reverse of that of an
automobile radiator.
It may be surprising to know that a radiator
having a core of 100 cubic feet would have a
good margin in capacity for cooling the air
from a 25,000-kv-a. turbine alternator.
That the use of a radiator is quite feasible
from a consideration of space, resistance to
air flow, rate of heat transfer, etc., will be ex-
plained by an article in the Februan,' issue
of the General Electric Review. The
advantages of this system are specially ^'alu-
able on shipboard.
The Supercharger
In connection with turbine research work,
there was developed a turbine-driven super-
charger for airplane engines, utilizing the
energy of the exhaust gases of the engine to
drive a small centrifugal compressor which
can supply air, at sea-level density, to the
engine intake at high altitudes.
The importance of this airplane auxiliary,
which makes it possible to maintain high
engine _ efficiency at high altitudes, can be
appreciated when it is understood that en-
gines not provided with a supercharger de-
12 Januan-, 1920
GENERAL ELECTRIC REVIEW
Vol XXIII, No. 1
liver only 50 per cent of their sea-level energy
at 18,000 ft. elevation, while at 25,000 ft. the
reduction is about 75 per cent.
The supercharger was first developed in
the laboratory', and after factor}' tests was
taken to the summit of Pikes Peak, Colorado.
of 137 miles per hour as compared with its
best previous performance, under similar
conditions but without the supercharger, of
96 miles per hour.
The turbine rotor and the light weight
impeller of the compressor are mounted on
a common shaft (Fig. 15)
and normally rotate at
about 22,000 r.p.m., a sim-
ple valve in the exhaust
piping of the airplane
engine permitting the pilot
fuU control of its opera-
tion. The weight of the
entire equipment, includ-
ing all necessars' piping, is
about 100 lb.
Fig. 14. LePere Biplane Equipped with G-E Supercharger in Te»t
Flight Above McCook Field, Dayton, Ohio
where for several weeks it was subjected to
long continued operation tests in connection
with an airplane engine. Its satisfacton-
performance under these conditions gave
even.' assurance of its safet\- and practical
value, and on August 2, 1919, the pioneer
flight of an airplane equipped with a G-E
supercharger took place at McCook Field,
Dayton, Ohio.
A LePere biplane (Fig. 14) driven by a
12-cylinder Libert}^ Alotor was used, and at
an elevation of 18,400 ft. it attained a speed
Electric Traction
Activities in the electric
traction field were largely
confined to the installation
of automatic substations
and the purchase of the one-
man, light-weight safety
cars. Electrification proj-
ects in the course of con-
struction progressed satis-
factorily, but few new de-
velopments were initiated.
Because of the lack of financial resources,
there was ver>' little addition to power equip-
ment by electric railways, with exception of
the automatic substation.
Electrification
The principal electrificatioh in the United
States, and without doubt the most important
project of its kind, is the Chicago, Milwaukee
& St. Paul Railway, which has completed a
217 mile extension across the state of Wash-
ington from Othello to the cities of Seattle
Fig. 15. Biplane Equipped with a G-E Supercharger which can be seen back of the upper blade of the propetlor
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 13
and Tacoma. The substations and the trans-
mission lines are now ready for operation,
and the overhead construction is completely
installed. Locomotive deliveries were com-
pleted by the end of the year and electrical
operation over the entire distance should be
an accomplished fact early in
1920. Two freight locomo-
tives were placed in operation
in October, 1919, on the 2.2
per cent grade west of the
Columbia River, releasing for
other service five steam en-
gines ordinarily used for
pusher service on this grade.
The five bi-polar gearless
passenger locomotives (Fig.
16) are now completed, the
first unit having been shipped
on November 1, 1919.
The electrification of the
government owned steam
suburban lines radiating from
the city of Melbourne,
Australia (Fig. 17), progressed
materially during the year.
About 200 miles of line are
now operating electrically,
the major part of the work-
having been completed since
the war, which held up all
construction work.
The contracts for this work were awarded
some time ago, and included orders for 400
complete motor car equipments. Each equip-
ment consisted of four GE-239, 750/1500-
volt motors, multiple unit control, air com-
pressors, and other accessories. Orders were
later placed for four 1500-kw., 1500-volt
synchronous converters which are now in
operation, and control equipment was also
furnished for 400 trail cars. This 1500-volt
direct current electrification is apparently
operating with all the success which was
predicted for it, and work is being pushed
rapidly on the completion of the remaining
suburban lines
Fig. 17.
1500-volt Direct Current Multiple Unit Train on Victorian
Railway, Melbourne, Australia
The Salt Lake, Garfield & Western Railway
completed its change-over to electrical opera-
tion during the summer and is now operating
about 21 miles of line with multiple unit
equipment and G-E automatic substations.
Another electrification which will begin
operation during the present year is the
Hershey Cuban Railway, which will operate
a line about 60 miles in length between
*»"
■>-wv,.5.i,--:
'rjrit
tV>«J^'Si»fW »»»»«,;
.?,V.«J-»^>">i^^
.?"-.»,*,- ;„ .,,
Fig. 16. 265-ton Gearless Passenger Locomotive in Test for Chicago, Milwaukee & St. Paul Railway
14 Tanuarv, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 1
Havana and Matanzas, Cuba. About one
half of this road is now operating; with steam
engines and the remainder will be new con-
struction.
motor-generator set now being installed for
supplementing the power supply to the Michi-
gan Central R.R. operating electric locomo-
tives through the Detroit river tunnel.
Autom.atic substation equip-
ment was sold in Cuba. Aus-
tralia and Xew Zealand, and
inquiries were received from
other foreign countries. At
the end of 1919. there were
approxim.ately 50 G-E equip-
ments in operation and about
twenty more either under con-
struction or being installed.
Fig. 18.
lOOO-kw. Automatic Railway Substation Installed for Pacific
Electric Railway
The electrical equipment furnished includes
G-E power and substation machinery, over-
head line material, motor-car equipment and
locomotives. The trolley po-
tential will be 1200 volts direct
current and the substations
are of the standard automatic
type, each containing two
obO-kw. 600-volt synchronous
converters operating in series.
A third unit will be installed
in each station as a spare.
There are seven electric
locomotives on order, each
weighing 60 tons. These will
be used for hauling raw sugar
and other freight to ports for
shipment. The passenger traf-
fic over this line will be hand-
led by multiple unit motor
cars, fifteen of which will be
used, each equipped with four
GE-263 motors and Type PC
control.
Automatic Substations
The continued popularity
of the automatic substation
(Fig. 18) is shown in the increased orders for
various types and sizes up to 2000-kw. and
15G0-volts direct current. The largest unit
so far constructed is a 2000-lav. synchronous
Safety Cars
One of the most popular
outfits in the electric railway
field during the past year was
the light weight safety car,
large numbers of which were
utilized to replace heavier
equipment. Results obtained
from the operation of these
cars indicate that their use
will, to a certain extent, assist
operating companies in keep-
ing down the cost of operation without
the necessit>" of resorting to increased fares.
Practically all safet\- cars employ the type
Fig. 19. Light Weight Safety Cars for Eastern Wisconsin Electric Co-
equipped with GE-2S8 motors and K-63 control
K-63 controller, which was specially designed
for this service. For motive power, the
GE-258 motor has continued to be most
popular (Fig. 19).
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 15
In order to meet the preference of some
railways for a motor of this size with standard
sleeve bearings instead of ball bearings, the
GE-264 railway motor was designed and a
considerable number of these
are now in service. Approxi- ':'':'■
mately 1400 of these two
motors have been sold dur-
ing the present year. Other
equipment for the light weight
safety car includes the CP-25
air compressor and straight
air brakes with safety devices
(Fig. 20).
Equipment for Metropolitan Rail-
ways
On account of the lack of
financial means for adding
to their electrical equipment,
there was little activity in the
purchase of equipment for
heavy city, subway and ele-
vated lines. Orders were
placed, however, for 600 more
GE-248 railway motors for
the Brooklj'n Rapid Transit Company, mak-
ing a total of approximately 1600 of these
motors now in service. The Boston Elevated
Railway Company is also putting in ser\'ice
200 electro-pneumatic air brake equipments
and has placed orders for 232 GE-259 motors
with tapped fields for replacing motors of an
obsolete type now in use.
Electric Locomotives
Aside from the five 265-ton passenger
locomotives which were completed for the
new Chicago, Milwaukee and St. Paul
electrification,* there were also two 70-ton
switchers for the same line, which are now
in operation on the Rocky Mountain Division.
Work is nearing completion on the 50-ton,
1200-volt locomotives ordered some time ago
for the South Manchurian Railway in China,
and on four 60-ton, 1200-volt locomotives
for the Cienfuegos, Palmira and Cruces Rail-
way in Cuba.
jDther locomotives under construction in-
clude seven 60-ton units for the Hershey
Cuban Railway previously mentioned, and
a 30-ton switching locomotive for Harlowton
Mills at Lawrence Mass., and other indus-
trial types.
Miscellaneous Railway Equipment
The power limiting and indicating system
installed along the lines of the Chicago,
* See article by W. D. Bearce in December, 1919, General
Electric Review.
Milwaukee and St. Paul electrification has
shown interesting possibilities. By the use
of this scheme, the Company has been able
to maintain an unusually high load factor.
Fig. 20. Safety Car Handling Rush Hour Traffic at the Indiana Steel Co.
Terminal, Gary, Ind.
thus securing very nearly the minimum power
rate provided for in the contract with the
Montana Power Company.
High-speed Circuit Breakers
_ The intensive study of methods to protect
direct current machines, particularly the 60-
cycle 600-volt synchronous converter, from
flashing, was continued during 1919 with im-
portant results and there were developed
three graduated forms of protection which
will now give immunity under all operating
conditions.
First, commutating poles of a high reluct-
ance type with much stronger field windings
than those previously used were designed,
tested in service and adopted as standard.
This insures ample protection where operating
conditions are favorable.
Second, where greater protection is required
the machines can be provided with a type of
flash barrier which has fully demonstrated its
value in railway service on lines where severe
short circuits are of frequent occurrence.
Third, the highest degree of protection
includes the use, with the foregoing, of a
newly developed high speed circuit breaker
(Fig. 21) which has fully demonstrated its
value under tests of much greater severity
than those imposed by the most unfavor-
able conditions encountered in actual serv-
ice.
16 Januar>-, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo 1
Standard circuit breakers have alwa}"s been
much too slow in operation to prevent flash-
overs on heavy short circuits. Repeated
tests have indicated that a circuit breaker, to
prevent flashover. should operate, stop the
current rise and reduce it below the flashing
Fig. 21. High Speed Circuit Breaker, showing Limiting
Resistance and Connections Used During Test
value in something less than the time required
for a commutator bar to pass from one brush-
holder to the next. On a 60-cycle machine
this means a speed of appro.ximateh' eight
one-thousandths of a second, whereas the
ordinary circuit breaker operates in about
eight to ten onc-hundredths of a second.
A simplified diagram, illustrating the
principal features of this breaker and the
connections for the negative side of a gen-
erator, is shown in Fig. 22.
Fi and F-> represent a laminated field
structure something like that of an ordinarv
alternating current magnet. The poles of
Fi and Fi are bridged by a very light armature
A pivoted at P. which is held in contact with
the field b>- a shunt coil 5i energized from an>-
convenient constant voltage source, such as
the exciter circuit or the main bus.
The series bucking bar S-z, which electro-
magnetically trips the breaker, is located be-
tween the poles of the field magnet in a plane
perpendicular to the plane of the lamiiiations
and in ver>- close proximity to the armature,
so that a given current flowing in it produces
a maximum change in the armature fivix with
a minimimi change in the flux interlinking
the shunt winding Si.
The tension spring attached to the armature
provides a means of adjusting the breaker
and also gives the high speed opening of the
contacts. The main contact tips Ci and d
are of the solid copper type used so success-
fully on railway- contactors. The blowout coil
S is of the series type and designed to g^ve a
ver\- intense magnetic field at the contacts.
Several hundred short circuit tests were
made of this circuit breaker on the Chicago,
Milwaukee & St. Paul aoOO-volt motor-gen-
erator sets and also by short circuiting the
trolley conductors at various distances from
the substations, and there were no cases of
failure in the protection afforded. Five of
these circuit breaker equipments have already
been provided for locomotives and eight for
the substations on this system.
At the Railwav Convention at Atlantic
City, Oct. 4-10, lit'lO, a 3()0-k%\-., 600-volt, 60-
cycle synchronous converter protected by a
high speed circuit breaker was subjected to
'^a'r/rf WxAJ
Fig. 22. Connections of JR High Speed Circuit
Breaker, Schematic Diagram for Negative
Side of Generator
short circuit from two to three times per
hour. This public test was continued with
entire success throughout the five days of
the convention.
Automatic Generating Stations
The success of the automatic generating
station at Cedar Rapids, Iowa (.Fig. 23), and
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 17
the man}' railway automatic substations in
sendee has encouraged the development of
other generating and substations along auto-
matic lines.
An automatic hydro-electric generating
station was developed for the Blue River
Power Company at Seward, Neb. This
plant consists of a 240-kv-a., 120-r.p.m.,
6()-cycle, 240()-volt waterwheel-drivcn gen-
erator, and three SO-kv-a., 24,00()/24(X)-volt
transformers, and is the first of several
stations to be installed on this system.
A second installation is for the Ontario
Power Company at Ontario, Cal. This con-
sists of one 500-kv-a., GO-cycle generator
direct-connected to a Pelton waterwheel.
Instead of being entirely automatic this
station is controlled by pilot wires from a
manually operated station a few miles below
on the same stream.
When the operator desires to start the re-
mote controlled plant, he closes the control
circuit, which opens the nozzle to the Pelton
wheel. When the machine is up to speed,
he synchronizes it and then increases the load
to any desired amount by a further opening
of the nozzle.
The machine can be shut down at the will
of the operator by closing a second control
circuit, but in case of necessity, due to over-
load or hot bearings, this generator will shut
down automatically.
as a spare. It feeds three 3-phase feeder
circuits and three single-phase feeder circuits
which are controlled by automatic oil circuit
breakers with a so-called notching relay.
This relay (Fig. 25) will close a circuit
breaker if it has tripped out, and reclose it
. 23. Automatic Hydro-electric Generating Station, Iowa Railway & Light Company
General view showing three 60-cycle. 500-kv-a., 60-r.p,m., 2200-voIt
automatically-controlled generators
Fig. 24. Temperature Relay, with Cover Removed, used for
Protection against Hot Bearings in Automatic
Generating Stations
if it trips out the second time within a brief
period. If the short circuit has not cleared
itself by this time and the cir-
cuit breaker trips out a third
time, the circuit will remain
open until the circuit has been
cleared and the switch closed
by an operator. Thus the
notching relay automatically
performs the usual duties of
an operator when a switch is
tripped out in a substation.
One of the 3-phase feeder
circuits leads to several con-
stant current series lighting
transformers. The switch in
this circuit is opened and
closed by means of a Warren
time clock, and in addition is
under the control of the notch-
ing relay.
Automatic Distributing Stations
An automatic distributing station was
developed for the Maiden Electric Company,
Maiden, Mass. This substation equipment
comprises one 3000-kv-a. transformer plant
fed over one 22,000-volt line, with a second
Radio Communication
It was demonstrated during the war that
practically continuous day and night trans-
oceanic radio service could be effectively main-
tained. It is now a matter of history that
radio was largely used for communication
IS Januan-, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 1
between the United States and the armies in
Europe and that the great war was brought
to a close by negotiations conducted by radio.
The work of the year 1919 was directed to
adapt the system of radio communication
which had beenldeveloped, to the increasing
Fig. 25. Notching Relay, Single-pole, Single-throw, 600 Volts. Contacts are
opened and remain open until reset by hand, if three separate impulses
are given on the coil circuit within a definite time of each other
demands of commercial communication in
peace times. An analysis of conditions of the
radio art as left by the war shows that while
radio has proven efficient and reliable, the sys-
tems which were used were inadequate to meet
the great volume of international commercial
traffic which is reasonably to be expected.
The practical and commercial aspects of
these new demands are being met by the
construction of a large number of 200 kw.
* See article by J. R. Hewett in Gener.\l Electric Review,
August. 1919.
tSee paper by E. F. W. Aleianderson in A.I.E.E. Proceedmgs
October, 1919.
radio frequency alternators with auxiliaries
for equipping radio stations in all parts of
the world.
These high power radio equipments (Fig.
26) are of the type which were used dur-
ing the war in the Xaval Radio station
at New Brunswick, N. J., which was
depended upon by the Na^•^■ for com-
munication with Europe during the
war and for the peace negotiations.*
It was also through this station that
transatlantic radio telephonic mes-
sages were sent, and the telephone
communication established between
officials in Washington and President
Wilson's ship at sea.f
The new equipments which are
being constructed embody a number
of features which will make possible
increased traffic capacity. The meth-
ods that have been proposed for this
purpose are:
Closer spacing of wave lengths, making
possible seven commercial wave lengths
within the range now occupied by one
station.
2. Increasing the speed of transmission
from 20 words a minute to 100 words a
minute or more.
3. Improvements in the receiving device
whereby several communications can be car-
ried on with the same wave lengths, b>-
adjusting the receiver so that the messages
are intercepted only from one direction with-
out interference with other messages which
1.
Fig. 26. 200-kw. High Frequency Induction Motor-driven Alternator of the Type Adopted for Commercial Radio Service
SOAIE DP:VEL0PAIENTS IX THE ELECTRICAL INDUSTRY DURING 19 li) 19
are carried by the same wave lengths in
other directions.
The increase of the number of radio
stations in the world by the closer spacing
of wave lengths is made possible by the use
of the high frequency alternator, and par-
ticularly by the accurate method of speed
regulation which has been developed. Thus
different alternators may be operated to
transmit different messages at speeds and
frequencies differing only by 1 per cent,
whereas the speed of each alternator is
regulated to within one tenth of 1 per cent.
Although the alternators are driven bv
ordinary induction motors from commercial
power supply, this accurate regulation of
speed has become possible by the new type
of saturation regulator (Fig. 27) which "has
been developed for this purpose. This reg-
ulator is a choke coil containing iron, and is
connected in series with the power leads of the
induction motor. The inductance or choke
effect of this coil is controlled by a vibrator
regulator of the ordinary power station type,
which again is controlled by a sensitively tuned
high frequency circuit. In this way an in-
duction motor driven from an ordinary'
power supply can be made to operate at a
speed varying not more than one tenth of
line per cent, although the load on the motor
is varying continuously in accordance with
the dots and dashes of the telegraph code.
The increase of the transmitting speed to
one hundred words per minute or more for
HIljH FREQ. SET
MOTOR ALTER.
POWER GO
TO ANTENr>IA
SPEED REGULATOR
ALTERNATING CURRENT PRtVE
Fig. 27. Arrangement for Speed Regulation of Induction Motor
Driven by High Frequency Alternator
telegraphy, and in fact the complete control
of the radiation by the human voice, has
been accomplished by the magnetic amplifier
(Fig. 28) which is a device based upon satu-
* Paper on Simultaneous Sending and Receiving, Proceedings
of I.R.E., August, 1919.
ration of an iron core and operated analo-
gously by the saturation regulator controlling
the power flow to the induction motor. In
order to commercialize high speed sending
new types of relays have been developed
which control the saturating current of the
Fig. 28. Relation of Magnetic Amplifier to Other Parts of
Radio Transmission Equipment
magnetic amplifier for telegraphy at speeds
considerably more than one hundred words
per minute.
For the reception of high speed signals a
new type of photographic recorder has been
developed. The recorder is a highly sensitive
type of oscillograph which prints on a sensi-
tized strip of paper the dots and dashes of
the high speed telegraph messages. This
machine, which has been developed in com-
mercial form, not only exposes the photo-
graphic tape but develops and dries the tape
so that it comes out of the machine ready for
translation on the typewriter.
The method of high speed transmission
and reception of messages not only makes
possible a larger volume of radio traffic but
makes the messages for practical purposes
as secret as the messages over telegraphic
wires.
The increase of radio traffic by several
communications on the same wave length
is made possible by the development of the
"barrage receiver" (Fig. 29), and other
methods of simultaneous sending and receiv-
ing.* The barrage receiver is constructed on
the principle that the messages are received
on two or more antennae which have dif-
ferent_ sensitiveness to signals from different
directions. The instrument can be adjusted
so that a signal from any one or from several
directions can be neutraHzed by bucking
the wave received on one antenna against
the wave received on another antenna,
20 January, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 1
whereas the signal from the desired direction
is not neutralized but in fact intensified by
the simultaneous action of the two or more
antennae.
Thus a receiving station may be located
close to a transmitting station which trans-
Fig. 29. Radio Receiving Set with Barrage Section
mits on the same wave length and yet be
insensitive to the signal transmitted by that
station, whereas it receives signals from the
other side of the ocean.
A receiving device operating on a similar
principle was installed on President Wilson's
ship, the George Washington, thereby making
it possible to speak through the radio trans-
mitter of the ship and at the same time
listen to the signals from shore. By the use
of this apparatus two-way conversation was
held successfully between the ship at sea and
officials in Washington, who were speaking
over the telephones in the War and Navy
department, connected with the high power
radio station at New Brunswick.
The Research Laboratory
In the realm of pure science there was
brought out a new theory of atomic structure,
which, it is believed, constitutes one of the
most important advances in theoretical chem-
istry that has been made for many decades.
The applicability of X-ray S])ectrum an-
alysis to the study of atomic stnicture was
greatly extended by the discovery that single
large cr\-stals are not required, but that the
material may be studied in powdered form.
This work has indicated the possibility of a
new method of not only qualitative but of
quantitative analysis by means of X-rays.
Early in the year a new X-ray tube designed
specially for dental work was produced and
is now available commercially.
A new portable X-ray outfit, including a
new tube, can be operated from an ordinar\-
lamp socket, and thus makes it practicable,
for the first time, to take X-ray plates of a
patient in his own home.
The war-stimulated development of radio
sets using vacuum tubes was continued. The
highest power \'acuum tube set e\-er installed
on shipboard was placed on the George Wash-
ington last spring for the President's use.
The range of pressures over which the
ionization gauge can be used was considerably
extended. This gauge still constitutes the
best known means of measuring an exceed-
ingly high vacuum.
A water-japan was developed in the lab-
oratorv' during the war. but was not placed
in production until 1919. Its characteristics
are such that it bids fair to replace ordinar>-
japan to a large extent, since it gives an
equally good coat, and at the same time
completely eliminates the element of fire risk.
Alternating Current Machines
The maximum unit capacity for synchro-
nous condensers will be doubled, as compared
with existing units, bv the installation of a
;5U,000-kv-a., 600-r.p.m.. GGOO-volt, .■3-phase.
.lO-cycle condenser which was under construc-
tion and nearing completion at the close of
the year.
This machine (Fig. 30) will be located at
Los Angeles, Cal.,in the Eagle Rock substation
of the Southern California Edison System
where, in combination with two lo,0()0-kv-a.
condensers already installed, it will be utilized
to maintain constant voltage on a l'iO,tMI(l-
volt transmission line about 240 miles in
length. It is provided with a loO-kw. 2.">()-volt
direct connected exciter.
On account of its unusual size and the
exceptional stresses which will be imposed on
the revolving parts, the rotor is built up of
laminated steel discs in place of the usual
cast spider. The total weight of the machine
is about .322,000 lb., the stator weighing
120.000 lb., the rotor 170,000 lb. and the base,
bearings, etc., about .32,0(M1 lb.
Ventilation will be provided by means of
air conducted from the basement, through
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 21
the machine, and discharging vertically. The
guaranteed losses for this record capacity
machine are less than 3 per cent.
The previous maximum voltage for syn-
chronous condensers was also exceeded by the
construction of two 12,500-kv-a., 500-r.p.m.,
22,000-volt, 3-phase, 50-cycle units (Fig. 31)
for the Andhra Valley Power Supply Co., of
Bombay, India*
In view of its exceptional rated voltage,
special insulation was required, and under
test the coils successfully withstood a potential
of 50,000 volts.
The 32,500-kv-a. 12,000-volt waterwheel-
(Iriven generator which was referred to in
last year's article was completed and shipped
during the year (Fig. 32). It is now being
installed for the Niagara Power Co. and rep-
resents the maximum capacity for machines
of this class.
Among the large alternators under con-
struction is a 26,700-kv-a., 300-r.p.m.,
6600/1 3, 200-volt generator which will be
direct driven by a 1.300-h.p. induction motor.
As this machine is intended solely for testing
oil switches, it is designed with very low
reactance so as to secure the largest possible
current values under short circuit.
There is also an exceptionally large motor-
generator set designed for use as a frequency
converter. The motor is rated 12,000 kv-a.,
7500 volts, 3 phase, 30 cycles, and the gen-
erator 15,000 kv-a., 5000 volts, 60 cycles.
The set will be reversible in operation and
will be installed at the Battle Creek, Mich.,
station of the Consumers Power Co., where
it will tie in the 30 and 60-cycle systems.
A complete new line of small vertical
shaft waterwheel generators was placed in
production. These represent 170 different
ratings ranging in capacity from 30 to 1000
kv-a. with speeds of from 360 to 100 r.p.m.
for potentials of 240, 480, 600 and 2300 volts.
These machines (Fig. 33) were designed
specially to meet the demand for an efficient,
moderate priced generator, particularly for
low head operation on relatively small
streams, and to permit the economical ex-
tension of the small automatic generating
station.
An important feature of the standard
equipment is the plate type combined sus-
pension spring thrust and guide bearing with
which all sizes are equipped. These bearings
as well as the lower guide bearings are all self
oiling.
♦Article by M. C. Olson in General Electric Review,
November, 1919.
22 January, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 1
A noticeable tendency during the \ear ^\■as
the increased use of synchronous motors for
driving air compressors, ice machines, flour
mills, rubber mills, pulp grinders, Jordan
engines, etc., which is indicative of a growing
appreciation of this type of motor for me-
chanical work combined with power-factor
correction and a better understanding among
industrial engineers of the true economy of
reasonable expenditures to inaintain high
power-factors on their lines.
During 191S the increased use of synchro-
nous motors in this way was more than 100
per cent over preceding years, and a further
increase of about 65 per cent was shown by
installations made in 1919. The capacities
in greatest demand ranged from 125 to 600 h.p.
shaft, and is automatically retarded and
brought to rest, stopping accurately- in the
dump and at the loading chute.
A similar installation has been in operation
several years for the Inspiration Consolidated
Copper Company, but at a rope speed of only
750 feet per minute, whereas the present equip-
ment with a rope speed of 1500 feet per minute
represents a considerable advance in the prac-
tice of automatic control of mine hoists.
These installations indicate the possibilities
of a more general adoption of automatic
features of operation in mine hoist ser\'ice.
The Oliver Iron Mining Company placed in
operation ten induction motor-driven mine
hoists at the Xorrie-Aurora Mines, Ironwood,
Michigan. This installation consists of five
Fig. 31. One of Two 12,500-kv-a., 22,000-volt Synchronous Condensers, which has a higher
voltage rating than any previously built machine of this type
Mine Hoists
An important installation placed in service
in the latter part of 1919 is that of the C, B. &
Q. R. R. at its mine operated by the Valier Coal
Company in southern Illinois. The drive con-
sistsof a i350-h.p. direct current motor (Fig. 34)
direct connected to a single cylindrical drum
on which two ropes wind for balanced opera-
tion of two self-dumping skips. The motor is
served by a 1000-kw. flywheel motor-generator
set, operated by public service power.
The most noteworthy feature of this in-
stallation is its semi-automatic operation.
The trip may be started either by an operator
on the hoist platform, in the usual manner,
or by the skip tender at the bottom of the
S75-h.p. 36()-r.p.m. motors (Fig. 35) for ore
hoists, and five 400-h.p. 360-r.p.m. motors for
man hoists, all operating on 2200-volt, tU)-
cycle circuits with liquid rheostat control.
These hoisting equipments arc noteworthy
by reason of the carefulh^ worked out system
of safety devices and also as representing
what is probabl\- the largest aggregation of
induction motor-driven hoists in a single
mining system.
Among the larger hoisting equipments
under construction is one for the Homcstake
Mining Company, Lead, South Dakota, com-
prising a 1400-h.p., 63-r.p.m. direct-connected
hoist motor with a 1 100-kw. fl\-whecl set and
66,000-lb. flvwheel
SOiME DEVELOPMI-XTS IX THE ELECTRICAL INDUSTRY DURING 1919 23
5 I
- a.
O -o"
■Z >•
E
o
I O
IS
■ -o
.£ £
I. o
° O
O u
c >
qS
24 January, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 1
Toward the close of the year an order was
received for two 5000-h.p. mine hoists for the
Randfontein Central Gold Mining Company,
Ltd., of South Africa. When these are com-
pleted they will be of considerably greater ca-
pacity than any existing electrical mine hoists.
The apparatus comprising each of these
hoisting equipments includes the following:
two 2500-h.p., 106-r.p.m., 600-volt direct-con-
nected motors supplied by a 375-r.p.m. motor-
generator set consisting of a 5000-h.p., 2000-
volt slip-ring induction motor dri\'ing two
2000-kw., 600-volt generators with a 60-kw.
exciter. Ward Leonard control will be used.
These hoists will serve a 5000-ft. shaft and
will carry five tons of ore per trip at a rope
speed of approximately 4000 ft. per minute.
Roller Bearing Hoist
The one-ton Sprague hoist which has been
on the market for a number of years was
redesigned to include roller bearings through-
out. This detail change is of considerable
practical importance as it solves the lubrica-
tion problem which has always been a difficult
one with hoists of this capacity on account of
the lack of attention which frequently char-
acterizes their use.
The roller bearings are packed in grease.
which does not need to be renewed more than
two or three times a year.
Electric Winch
An improved totalh- enclosed vertical
winch (Fig. 36) was brought out, equipped
with a vertical motor which can be cither of
the series wound direct-current t>-pe, or the
slip-ring alternating-current type.
This winch is now produced in a number
of ratings, up to 12,000 lb. pull, at 2o ft. per
minute. A large number of them with direct-
current motors and with a rating of 8000 lb.
.50 ft. a minute, arc being installed along the
route of the New York State Barge Canal,
where they will be used for warping canal-
boats up to the docks.
Electric Shovels
A radical departure from previous practice is
found in a recenth" developed electrical sho\-el
equipment in that it eliminates all rheostat
losses and the possibility of heavy peak loads
and requires no overload relays or other form
of protection for the electrical machinery.
The equipment comprises a four-unit syn-
chronous motor-generator set (Fig. 37) with
a direct connected exciter. There are three
generators, one for supph'ing current direct
to the two 170-h.p. hoist motors and one
each for the 60-h.p. swinging and crowding
motors, so that each of these three motor
circuits is supplied by an indi^-idual generator.
The hoist generator is rated at 2.50 kw. and
the other two at 50 kw. each. A master con-
troller is provided for each circuit and the con-
trol is effected entirely by voltage variation.
Due to the length of the motor-generator
set and the possibility of mechanical strains
to which it might be subjected when installed
on a platform which must frequently be
moved over rough ground in ser\-ice, the
set is divided into two pairs of machines,
each pair being mounted on a separate base.
The two shafts are united by means of a
flexible coupling.
There is also a motor for operating the
dipper trip, which is rated at 50-lb. torque
and is thrown in or out of circuit b^' means of
a push button switch located in the handle
of the crowder motor controller. This small
motor is energized from the 1 lO-volt alternat-
ing-current lighting circuit. The other motors
are all 230-volt direct current.
This unique set was constructed for use on
a coal stripper shovel similar to that shown
in Fig. 38, but its extreme simplicity renders
it readily adaptable for the operation of any
size or type of electric shovel.
Steel Mills
During the year there was added approxi-
mately 40.000 ii.p. (normal continuous rating)
to the existing capacity of main roll drives
installed by the General Electric Company.
The electrical equipment driving the 40-in.
reversing blooming mill at the Sparrows
Point plant of the Bethlehem Steel Co. was
put in ser\'ice in April antl has been in suc-
cessful operation since that time. This equip-
ment has a double unit reversing motor having
a normal continuous capacity of 5000 h.p.
(Fig. 39) at .50 r.p.m. and a momentary torque
capacity of approximately 2.()00,00() lb. at
one foot radius at any speed frona zero to
50 r.p.m. Power for this blooming mill motor
is derived from a fl_\-vvhecl motor-generator
set (Fig. 40), consisting of two 2000-kw.
generators, one .3000-h.]).. (iCiOO-volt induction
motor, and one 50-ton flywheel, operated at
a speed of 375 r.]).m.
The layout of the Sparrows Point mill is
such that the blooms, which are rolled from
ingots in the 40-in. blooming mill can bo
delivered without reheating to a 24-in. six-
stand continuous billet mill which is driven
by a 4000-h.p., S3-r.p.m., (UiOO-volt induction
motor (Fig. 41).
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 2;
The product of this billet mill can be de-
livered direct to an IS-in. six-stand, continuous
billet and sheet bar mill which is driven bv a
3250-h.p., 94-r.p.m., (lOdO-volt induction
motor.
tion motor, and one oU-ton flywheel operating
at a speed of 'MiU r.p.m.
The structural and bar mill consists of one
stand of three-hish 28-in. rolls driven by a
250l)-h.p., .S2-r.p.m., (560U-volt motor, and
Fig. 36. Sprague Vertical Dock Winch
with Self-contained Electrical
Equipment
Fig. 38. Electric Stripping Shovel, Piney Fork Coal Company, Smithfield. Ohio
During the early part of the year the new
mills at the Fairfield Works of the Tennessee
Coal, Iron & Railroad Co. were put in opera-
tion. The first mill to be started was the
United Engineering & Foundry Co.'s 36-in.
by 110-in. plate mill, which is driven by a
4000-h.p., .S2-r.p.m., 6600-volt induction motor
with direct connected flywheel (Fig. 42).
The 45-in. reversing blooming mill has an
electrical equipment which is the largest
reversing blooming mill equipment in this
country. It is driven by a double-unit revers-
ing motor having a normal continuous capac-
ity of 5600 h.p. at 5.5 r.p.m. and a momentary
torque capacity of 2,300.000 lb., at one foot
radius, at any speed from zero to .50 r.p.m.
Power for this reversing motor is derived from
a motor-generator set consisting of three 2000-
kw. generators, one 40(IO-h.p.. (KiOO-volt induc-
three stands of 2()-in. rolls driven by a 3000-
h.p., 6600-volt motor with double range
modifled Scherbius speed regulating set, by
means of which the motor sj^ecd may be
varied from 130 to 155 r.p.m., the synchronous
speed being 144 r.p.m.
The most revolutionary installation in the
Fairfield Works is the electric drive for the
hydraulic intensifier for the 1250-ton bloom
and slab shear in the blooming mill. For
reasons of mill layout it was necessary for a
single shear to be used, powerful enough to
cut 12-in. by 44-in. slabs and fast enough to
keep ahead of the mill on 8-in. by <S in. or 6-in.
by 6-in. billets to be cut in comparatively
short lengths.
To obtain this combination of power and
speed a hydraulic intensifier was decided
upon which would nrdinarih- be steam
Fig. 37. Four-unit Synchronous Motor-generator Set for Current Supply to Electric Shovel
26 January, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 1
i i
.= * >.
Si° o
S 6
= 2
o
O CI
S"
a »
g
_ £
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 27
driven. But as this would ha\-e required a
boiler plant for this drive alone, the builder
of the shear proposed an Ilgner-Ward Leonard
reversing drive with rack and pinion to replace
the steam cylinder of a steam intensifier.
This drive comprises a direct current motor,
700 h.p. continuous capacity, <S6 r.p.m., with
a momentary capacity of 24.50 h.p., a fly-
wheel motor-generator set and a special
control system. The electric drive has shown
itself ample for all demands upon it and has
thoroughly justified its selection.
There in now under construction an elec-
trical equipment for a 40-in. reversing bloom-
ing mill, for the Tata Iron & Steel Co., at
Sakchi, India. This equipment is similar to
that installed on the 45-in. blooming mill at
the Fairfield Works of the Tennessee Coal,
Iron & Railroad Co., already described,
except that the induction motor driving the
flywheel set is a 5()-cycle instead of a (JO-
cycle machine.
For the Superior Sheet Steel Co., of Canton,
Ohio, construction work was started on two
lOOO-h.p., oOO-r.p.m. induction motors each
to drive a 30-in. sheet mill. This represents
a departure from standard sheet mill practice
in the use of a 300-r.p,m. motor which necessi-
tates a gear ratio of approximately 10 to 1.
The majority of installations in the past have
used gear ratios of approximately S to 1 .
One of the most interesting applications of
induction motors with modified Scherbius
speed regulating sets is that at the Riverdale
Plant of the Acme Steel Goods Co. The mill
is a 10-in. continuous hoop mil], and consists
of a roughing train of six horizontal rolls and
a vertical edging roll driven by a 900, ooO-h.p.,
325/197-r.p.m. induction motor with modified
Scherbius speed control; an intermediate
stand driven by a 50/100-h.p., 300/ OOO-r.p.m.
direct current motor, and a finishing train
consisting of five stands of horizontal rolls
driven by a 1800/1200-h.p., 240, I60-r.p.m.
induction motor with modified Scherbius
speed control.
Extensive developments at the plant of the
Buffalo Bolt Company, Buft'alo, N. Y., in-
cluded the replacement of a direct current
motor driving a merchant mill b}- an induction
motor with modified Scherbius speed regulat-
ing set.
Printing Presses
A new type of combined predetermined
speed and full-automatic control for printing
machines was developed. The equipment
(Fig. 43) consists essentially of a combined
CEMF and current limit type self-starter,
with dynamic brake and vibrating field rela}'.
A separate field rheostat, operated by pilot
motor, is under the control of the push-
button stations.
The speed can be changed from the push-
button stations through a range of 3:1, or the
field rheostat setting can be left at any point
and the motor will automaticallot' accelerate to
the speed corresponding to this setting.
An exceptional jirinting press equipment
produced in 1919 is shown in Fig. 44. This
is a direct-current, full-automatic printing-
press control for handling a large newspaper
press driven by two 100-h.p. motors, each
motor being equipped also with a 10-h.p
starting-motor for obtaining the slow motion
and threading-in speeds. The two panels
are arranged for ])arallel operation, the pilot
motors being coupled together so that the load
between the two 100-h.p. motors will be
equally divided by a proper division of the
armature or field resistances.
The printing press which these motors
drive is one of the largest in existence.
Dynamometers
In order adequately to meet conditions
brought about by the increasing size of
aviation and marine gasolene engines, a new
electric dynamometer (Fig. 4.5) of exceptional
size was constructed for testing them.
During the war a considerable number of
dynamometers were built for testing Liberty
motors. These were rated at 400 h.p. at
1700 r.p.m., whereas the new machine has a
rating of 600 h.p. at 1200 r.p.m., and has a
maximum speed of 2000 r.p.m.
Fractional Horse-power Motors
The outstanding feature in regard to frac-
tional horse-power motors, that is, motors
rated at from ?4 h.p. down to 1/200 h.p., was
the enormous increase in production achieved
during 1919. As compared with 1918, this
increase represents an advance of about l.'iO
])er cent, and as compared with 1916 of over
300 per cent.
During the ten years in which these small
motors have been in standardized quantity
production, the field of their utility has
steadily broadened, and, whereas they were
originally used on a very limited number of
household devices and other light weight'
machines, they are now applied to more than
100 diff'erent classes of standard devices.
Originally the motors were applied as an
auxiliary to the device which they operated.
28 Jauuar>-, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 1
but with the steady growth of their popu-
larity, the driven machine has in many cases
been modified so that now the motor drive
often constitutes an integral part of the
complete outfit. The motor designer, on the
other hand, has modified the mechanical
Fig. 43.
Sprague Automatic Prcdetermincdl
Speed Control Panel
details of the motor from time to tim.e so as
to render possible the most compact arrange-
ment of the completed device, combined with
the most efficient application of the electrical
energ}-.
Ever}' type of these small motors is designed
with as much care and manufactured with
as great accuracy and attention to detail as
the huge turbines or waterwheel generators
which supply the large central stations and
transmission lines of the country, and the
sam^e care is also exercised in their test. Each
unit, no matter how low its rating, is given an
individual shop test before shipment. This
policy has resulted in a reliability in operation
which has recomjrended itself, not only to
the engineer, but also to the users of the motor
driven machines.
By means of electrical exhibits, demonstra-
tions by central stations and sales agents,
and by advertising, education in regard to
the adai)tability of electric motors has been
carried on, with the result that in all house-
hold devices such as washing machines,
vacuum cleaners, small pvunps and air com-
pressors, etc.. where power application is
required, the fractional horse-power motor is
now generally recognized both by the manu-
facturer and the general public as gi\'ing
entirely dependable ser^'ice.
Transformers
A number of small transformers designed
to be used between one line and neutral of a
66,000-volt, 3-phase grounded neutral system
have been operating successfully for several
months. They are especially suitable for use
as small town lighting transformers or con-
trol transformers in automatic sub-stations,
but some of them have been used with
entire success for the much more severe
sen-ice of "pulling" oil wells.
These transformers have one end of the
winding permanently grounded to the core and
tank and are provided with a terminal (Fig.
4(j) for connecting to the grounded neutral
of the system. The insulation is graded from
the grounded end to the line end which passes
through the single cover bushing.
The cost of high voltage bushings is a
large perbentage of the total cost of these
Fig. 44. 2100-h.p. Sprague Control P«nel Arranged
for Four-motor Control. Chicago Tribune
small transformers, and the saving thus
effected by the use of only one bushing makes
it possible for small communities near high
tension transmission lines, but remote from
the usual central station or sub-station
facilities, to secure economical electric service.
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 29
The demonstrated advantages of the four-
part "distributed core" (Form K) used for
many years for small lighting transformers
have led to the extension of this construction
(Fig. 47) to larger units, and during the past
year it was standardized for single phase
transformers up to 1000 kv-a. at 33,000 volts.
This represents a very considerable increase
in both voltage and capacity as compared
with maximum rating of previous years.
The use of alternating current for arc weld-
ing has led to the development of a special
transformer for supplying this current. Alter-
nating current arc welding requires an
operating potential of from 25 to 30 volts
across the arc, while to strike and hold the
arc with the ordinary bare metallic electrode
an open-circuit potential of about 100 volts
is required. Added to this, varying operating
conditions require that the welding current
be adjustable through a considerable range,
generally from 100 to 200 amperes.
The transformer designed for this special
service consists of a primary and a secondary
coil assembled on the center leg of a five-
legged core. The secondary coil is generally
placed at the bottom of the core and firmly
secured. The primary coil is placed above
the secondary and attached to a suitable
mechanism hv which it m.av be raised or
Fig. 45. 600-h.p. Sprague Electric Dynamometer
lowered, varying the gap between the pri-
mary and secondary coils, thus giving a
means of adjusting the welding current. The
transformer is enclosed bv a metal screen
and is mounted on casters so that it may be
readily moved from place to place.
The low voltage winding of furnace trans-
formers, those supplying synchronous con-
verters, or any in which high current is
required, is usually divided into a number of
Fig. 46. IS-kv-a., 34,500
60.000 Y-115/230 Form
KD Transformer with
One Side Grounded
Fig. 47. lOOO-kv-a. Form K
Transformer, High
Voltage Side
*"A New Form of Tank for Static Transformers." by W. S.
Moody, Gener.\l Electric Review, October. 1919. page 7.56.
multiple circuits, each consisting of a helical
coil of several turns interleaved with the high
voltage winding (Fig. 48).
This type of coil consists of a number of
strands of rectangular wire one above the
other, wound about a form, each turn or disc
separated from the adjacent one by an oil duct
(Fig. 49). The discs are braced by spacers
located radially across the face of the coil.
The great mechanical strength of this
winding as well as the excellent thermal and
electrical characteristics have led to the
extension of its use to higher voltages (Fig.
50), and its application to the concentric
type of winding. It has been found adapted
to voltages ranging from 2300 to about
15,000, depending upon the capacity of the
transformer.
The oil conservator* which has re-
cently been described in detail has so com-
pletely met the need which brought about its
development that during 1919 its use was
greatly extended and it is now recognized as
standard equipment for transformers of
500 kv-a. and over for use on high potential
circuits of SO, 000 ^•olts and above.
30 Januan-, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 1
One of the developments that might be
classed as a refinement, yet one which will
appeal to all those who have anything to do
with the installation of transformers, is the
new trunnion-shaped lifting lug (Fig. 51).
These lugs replace the hooks previously
Fig. 48. Internal Arrangement of Arc Welding Transformer
riveted or welded to the tank band for lifting
the complete transformer.
Being circular in section this type of lug
is as well adapted for lateral as for vertical
stresses and the continuous shoulder pre-
vents the rope or sling hook froni slipping,
no matter at what angle it may be neccssar}'
to exert the lifting force.
Another detailed improvement consists of
a combination dial thermometer and dia-
grammatic name plate (Fig. o2). In addition
to its attractive appearance, it is an advantage
to the operating engineer to have the con-
nection diagram and the thermometer both
located at the most convenient jjoint on the
tank surface, and at the correct height to be
most easilj' read b\- the a\-erage man.
High Voltage Bushings
Development work was completed on two
companion lines of high voltage bushings for
current transformers and metering outfits.
These bushings are now in production and
are designed for voltages of 2o,0{)() to 7.'5,l)(Mi
inclusive. One line is for use with trans-
formers of 200, 400 ampere capacity (Fig.
53) , and the other for transformers of 400/ SOI »
ampere capacity (Fig. 54).
The general construction of these bushings
follows the design of previous standard bush-
ings of the solid type, consisting of a paper
insulated tube, the upper end of which is
enclosed within a petticoated porcelain shell.
Four cables through the center metal tube
connect with the two sections of the double
ratio ^^nnding of the transformer, and ter-
minate at the top of the bushing in a series-
multiple connection board, by means of
which the ratio of the transformer may be
changed simply by changing the connections
at the top of the bushing. This connection
box is weatherproof and the bushings are
designed for outdoor installation.
In the case of the 200 400 ampere bushings,
the line conductors are brought out of outlets
on opposite sides of the connection box. In
the 400 800 ampere bushings, both line con-
ductors arc brought out through one opening
in the connection box in order to eliminate
the heating in the iron box by neutralizing
the magnetic effects of the opposing currents.
For power transformers, a complete line of
double conductor bushings (Fig. 55) for a
maximum operating potential of 7500 volts
in current carrying capacity up to .'{(Km
amperes is now available. Two leads from
the transformer winding are brought through
a single opening in the co\'er by means of a
double^outlet porcelain. The principal ad-
Fig. 49. Helical Coil with Offset Turns for Hi«h Current
Interleaved Disc Coil Transformer
vantage of this construction consists in the
elimination of the heating in the cover and
bushing support, which is ])resent when a
single conductor carrying high currents is
passed through an iron cover or bushing
holder.
SOME DEVELOPMENTS IX THE ELECTRICAL INDUSTRY DURING 1919 31
In the double conductor bushing, the in-
coming and outgoing leads pass through a
single opening and the magnetic effects of
the currents which are flowing in opposite di-
rections are neutralized, so that heating of the
cover and bushing support is entirely elimi-
nated. This permits the
use of iron supports and
bushing holder at a much
reduced cost as com-
pared with non-mag-
netic alloys.
A second advantage
is a reduction in the
number of cover open-
ings to accommodate
the incoming and out-
going leads. A single
round opening is all
that is required for this
double conductor bush-
ing. This reduces the
munber of joints which
must be made weather-
tight, and sometimes oil-
tight, and provides a
Fig. 50. Helical Coil Con- more compact arraugc-
centric Winding mcnt of parts on the
transformer cover.
A line of solid type interchangeable bush-
ings from 15,000 to 73,000 volts inclusive,
and for current capacities up to SOO amperes
transformers, potential transformers, oil cir-
cuit breakers and lightning arresters.
Fig. 52.
Supporting Plate for Dial Thermometer and
Name Plate
The bushing proper is identical for the
different classes of ser\'ice, and is accom-
modated for the different uses (Figs. 56 and
51. Transformer with Trunnion shaped LilLin^
Lugs and Combination Dial Thermometer
and Diagrammatic Name Plate
#^
Fig. 53. 73,000-volt Fig. 54. 25,000-volt Fig. 55. Double Conductor
Current Transformer
Bushing of 200,400
Ampere Capacity with
Cover Raised to Show
Series-multiple Con-
nection Board
Current Transformer
Bushing of 400 800
Ampere Capacity
with Cover Raised
toShowSeries-multi-
pleConnection Board
Bushing for Transformers.
Working Voltage 7500 Cur-
rent Capacity, 2500 Ampere
was fully standardized, and hij^her current
ratings are being added to the present line.
These bushings are designed for use on power
57) by exchanging the terminal parts and
other accessories, all of which are detach-
able.
32 January, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 1
This interchange of detachable parts can
be made readily by the user, who is thus
afforded the advantage of standardization
of bushings on the several classes of appa-
ratus, and a reduction in the number of
spare bushings which must be carried for
replacement purposes. All of these bushings
are for outdoor as well as indoor ser\-ice.
The standardization of bushings for operat-
ing voltages above 73,000 and up to 2oO,OOU
has been completed, and these bushings are
now in regular production for power trans-
formers, potential transformers, oil circuit
breakers and lightning arresters.
These bushings (Figs. oS and 59) are of the
oil filled type and are designed for both out-
■pgr
py
Fig. 56. Interchangeable High Voltage Bushings, as Used on
Constant Potential Transformers
Fig. 57. Interchangeable High Voltage Bushing*, aa Used oo
Oil Circuit Breakers
Fig. 58. Interchangeable High Voltage Bushings. 400 Amp..
Equipped with Various Terminal Accessories
Fig 59. Filled Type. Flnngr Clamped Porcelain High Volmge
Bushings for Transformers. Oil Circuit Breakers, and
Lightning Arresters
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING lltlO 33
door and indoor service on the various
classes of high voltage apparatus. Their
interchangeability* is accomplished by ex-
change of the detachable terminal parts and
accessories, which adapt the bushing proper
to the class of apparatus on
which it is to be used.
Feeder Voltage Regulators
The development of feeder
voltage regulating apparatus
progressed normally during 1910.
No radical changes were made
in any of the designs but they
were improved and extended to
meet ever increasing require-
ments.
A 600-kv-a. 3-phase, self-
cooled, automatically-operated
regulator of the outdoor design
(Fig. 6U) was completed during
the first part of the year and a
duplicate unit is now nearing
completion. This is one of the
largest capacity self-cooled regu-
lators of the outdoor tvpe ever
built.
For certain classes of work,
specially electrolytic and fur-
nace control, which require a very appreciable
voltage range, the standard design of regu-
lator is comparatively large and costly. A
combination of a regulating switch, connected
to taps of the transformer supplying the load,
and an induction regulator for gradually vary-
ing the voltage between the steps of the, switch
and also eliminating the breaking of any ap-
Fig. 61.
Fig. 60. Automatic, Oil Immersed, Self Cooled, Outdoor
Polyphase Regulator
Regulating Switch, (Front View with Oil Tank and
Casings Removed for Inspection)
preciable current by the regulating switch,
was suggested many years ago and various
designs and combinations have been built.
Only recently, however, has an apparently
satisfactory combination been developed, one
design of which is shown in Fig. 61. This
illustration shows only the switch, as the
regulator is of standard design but arranged
with slip rings so as to allow of continuous
rotation.
With this combination of switch and reg-
ulator, the latter may be any percentage
of the kilovolt-ampere capacity of a single
regulator which would otherwise be required
to give the same voltage range, depending onh-
on the number of taps it is feasible to bring ou c
of the transformer supplying the load.
The advantage of the combination is its
high efficiency and power factor compared
with a single regulator, its only disadvantage
being that it is more complicated. The
switch shown was designed and built to
l^roduce a range of from 40 to SO volts on the
secondarjr side of a 1000-kw., 11,000-volt
transformer, the voltage regulation being
obtained by a gradual voltage change between
the successive transformer taps.
♦"Interchangeable High Voltage Bushings." by E. D. Eby,
General Electric Review. .\'ovember, 1919.
34 January, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 1
Voltage Regulators
A new counter electromotive force voltage
regulator (Fig. 62) was developed for the con-
trol of direct current generators. The princi-
ple upon which it operates is as follows :
A small motor (Fig. 63) is used with its
armature in series with the field of the direct
Fig. 62. Voltage Regulator for One Direct Current Generator
(Front View;
current generator to be regulated. The field
of this small motor is energized from the
armature of the generator, but is controlled
by means of a set of contacts carried on a
ver\- sensitive control magnet which, in turn,
is connected to the generator busbars.
Assuming that the voltage of the regulator
tends to drop, due either to load conditions
or to changes in the speed of the i)rim.e mover,
the control magnet will allow a spring to
close the contacts, short circuiting the field
of the small motor. This will immcdiateh'
cause its voltage to drop, allowing more field
■ to be applied to the generator which will
tend to raise the voltage.
The voltage is in this way immediateh"
restored to the point where the contacts
start to open, and at this point they will
continue to vibrate, thereby holding an aver-
age field on the small machine. The periotl
of time during which these contacts are in or
out of engagement is determined by the
tendency of the voltage to rise or fall.
Assuming that the voltage of the generator
tends to rise, the contacts of the control
magnet will open, allowing full field to be
applied to the small motor, which will
force it to generate a higher counter electro-
motive force, which will in turn immedi-
ately cause the voltage of the main generator
to decrease until it reaches normal. At this
point the contacts will again start to
^•ibrate, holding an average field on the small
machine necessar\- to obtain the proper
voltage to meet the new condition of speed
and load.
Referring to the diagram (Fig. 64), it will
be seen that there is a resistance in series
with the field of the small motor, its purpose
being to limit the current when the contacts
on the control magnet close and short circuit
the field of this motor. In addition, the
motor is supplied with an eddy current brake
which is excited from a coil connected in
series with the generator field. This brake
is necessar>- to prevent excessive speed of the
small motor.
In addition this brake is supplied with an
adjustable air gap so that the speed may be
kept to a safe value. A double-pole, double-
throw switch is supplied so as to permit cut-
ting this motor out of the circuit when it is
desired to operate with hand regulation.
Fig 63.
Voltage Regulator or One Dirc\t Current Generator
(Back View I
Static Condensers
During the past year, a numlier of impor-
tant changes were made in the design of static
condensers. The capacity of the indiviilual
condenser unit itself (.Fig. 6o) was increased
from 2 kv-a. to 5 kv-a., thereby reducing the
number of units required to make up equip-
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 35
ments and the amount of floor space required
by the assembled condenser.
In order to obtain the increased capacity
per section, the number of couples has been
increased, and also the number of paper
laniinations forming part of each couple.
For 23t)()-volt service, the condenser imits
are designed for direct installation in the line,
but for 220, 440 or .joO-volt circuits they are
designed for 1200-volt operation, and an auto
transformer is therefore furnished for step-
ping up the supply voltage.
With the earlier equipments, an auto
transformer was furnished to step up the
supply voltage to SOO volts, but, inasmuch
as the capacity of the condenser varies as the
square of the applied voltage, it will be readily
appreciated that the active material is now
being used more economically, and that a
considerable saving has been effected which
has practically counterbalanced the increases
in cost which would otherwise have been
incurred if these changes had not been made.
The auto transformer furnished with three-
phase equipments is provided with a lead
from the neutral connection, so that the
.p <r 3ijs af^s
D^^" .S*v.
/=-^o/vr y/£ys
Fig 64. Connections for Voltage Regulator with One Arrange-
ment of One Direct Current Generator
neutral may be grounded, and any possibility
of an abnormal voltage being impressed on
the supply system is entirely eliminated.
In order to provide as great a factor of
safety in the design of the units as formerly,
the laminations between the condenser plates
have been increased and the clearances in
each unit have also been increased. An addi-
tional terminal has been provided on each unit
to permit read>- grounding to the rack struc-
ture, thereby preventing any danger of a
difference of potential existing between the
condenser unit and the rack.
Fig. 65. Detail Assembly of Static Condenser Units Showing
Bus and Fuse Arrangement
The construction of the racks upon which
the condensers are assembled (Fig. 60) has
been changed from pipe framework to an
angle iron supporting structure. The numer-
ous pipe fittings have been replaced by angle
iron supporting braces which have not only
reduced the weight but have materially
simplified and strengthened the entire struc-
ture. Provision has been made for mounting
the various auxiliaries, such as discharge
resistances and disconnecting oil circuit
breaker, directly upon the rack, thereby
making the complete outfit self-contained.
Lightning Arresters
When the oxide film lightning arrester
discharges, experience has shown that the
conversion of the lead peroxide into an in-
sulating plug is so rapid that the arc rises only
very slightly on the gap and, in consequence,
sphere gaps alone are used. These permit a
compact construction and made it possible
for the first time to provide protection of the
spark gap from the weather as an integral
part of a high tension outdoor lightning
arrester equipment.
A number of arresters similar to that
shown in Fig. 67 were prepared for service
during the year. With this housing, the
sphere gaps on the outdoor oxide film light-
ning arresters can be more closely adjusted
than if exposed gaps are used and their pro-
36 Tanuar\-, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 1
Fig. 66. Typical Static Condenser Showing Arrange-
ment of Rack, Condenser Units and Switch
tective value is thereby rendered equivalent
to that of an indoor type.
For more than three years the o.xide film
type of arrester has been in operation under
sen^ice conditions on circuits up to GO,0(Hi
volts, and as it does not require daily charj^-
ing, as does the aluminum cell type, it has
proved to be not only effective but also
economical in the cost of attendance and has
been installed in m.any places where the need
of daily charging the aluminum cell type
would ])reclude its use.
The mechanical details of the arrester have
been simplified and improved during the past
year and subjected to a standardization of
parts as the result of commercial experience
and continued experim.ent. The structure
of the improved form is indicated in Fig. 6N.
Electric Welding
A new direct
produced (Figs,
current directly
voltage without
welding outfit was
70) which delivers
current
Uit and
to the arc at the required
the use of an\- form of
ballast resistance or external regulating de-
vice.
This result is obtained by means of a dual
magnetic circuit, one section of which gen-
erates constant potential in part of the
armature by means of a shunt field receiving
excitation from this ])art of the armature
winding while the armature reaction and a
differential series field cause a varving volt-
Fig. 67. Oxide Film Lightning Arrester for Three Phase
Outdoor Service, IS. 000-25. 000 Volts, showing
Shielded Hemisphere Gap
age in the other part of the armature winding.
The constant potential is 30 volts, while the
other component ^■aries from positive 30 volts
on open circuit to negative 30 volts on short
circuit.
The generator rating is 200 amperes, no
exciter is required, and either a-c. or d-c.
motors or belt drive can be used. The outfit
is self contained, including a control panel.
Fig. 68. Oxide Film Lightning Arrester for Indoor Service
on Three Phase Circuit 15,000-25,000 VoJts
SOME DEVI'LOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1!)1'J 37
and is compactl\' mounted on a
structural iron sub-base so that it
can be readily moved about. It
weighs about 1300 lb.
A new automatic welding ma-
chine, for work varying from 25
mils to one quarter inch, was also
produced. It consists essentially
of a pair of feed rolls (Fig. 71j
which are driven at ^'arying speeds
by a small direct current shunt
wound m.otor. The rolls deliver
the electrode wire to the working
face, and, when the welding arc is
drawn, the field and armature of
the motor are instantaneously in-
fluenced by the voltage across the
arc and respond by increasing or
decreasing the rate of feed of the
wire, thereby regulating the length
of the arc to the value for which
the machine is adjusted. Above
the feed rolls, wire straightening
rolls are provided to insure accu-
rate feeding of the wire and the
proper location of the arc.
This machine may be operated
from any direct current welding
circuit and will use any size of
electrode, up to its mechanical limits, with
equal precision in operation, as neither of
these factors enters into the question of the
rate of feed control, which is governed solely
by the voltage across the arc.
Fig. 71. 14-in. Shaft with Fit Increased -^s-in- in Diameter by
Automatic Arc Welding Process
The rate of increase in the use of electric
welding, which was greatly stimulated by the
rapid production and repair requirements
of our various industries during the war.
has been well maintained. This is indicated
Fig. 69.
200-ampere Arc Welding Generating Set with
Control Panel (Front View;
Fig. 70.
200-ampere Arc Welding Set
(Back View)
Theequipmentincludesapanelboard on which
relays, regulating switch, etc., are mounted for
the control of the motor. The complete outfit
is very compact, having a length of 4 ft., a
width of IS in., and is about fi in. in height.
by the fact that in 191 S the number of elec-
tric welding outfits sold was more than double
that of any preceding year, while 1919 in turn
gave a further increase of 100 per cent over
191S.
38 January, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 1
Industrial Heating
A new form of electric solder melting pot
was developed in which the heating is self-
regulated. The pot (Fig. 72), which can be
operated at any voltage from 100 to 125.
consists of a substantial iron casting, on the
Fig. 72. Self Regulating Electric Melting Pot
sides and bottom of which the heating units
are clamped. The sides and bottom of the
pot are jacketed with corrvigated asbestos
board protected by welded sheet steel. The
leads are brought out through insulating
bushings in the bottom plate.
The material of which the units are made is
of such a nature that the temperature of the
solder cannot rise above a predetermined
point.
When the pot is cold and the current is
first turned on, the electrical resistance of the
units is low. This allows the maximum cur-
rent to flow (Fig. 73) and gives quick initial
heating. As the temperature rises the resist-
ance of the units increases, thereby reducing
the amount of current used. When the
maximum temperature is reached, it remains
constant and the flow of current is maintained
at a minimum value. This means that the
solder will ne\'er reach a temperature at
which it will appreciably oxidize and form
the usual heavy coating of slag.
For light and intermittent service the self
contained electric soldering iron is very satis-
factory, but for moderate and heavy duty
work it is frequently desirable to use the
ordinary soldering copjjcr which must be
heated in a furnace. When fuel fired furnaces
are used for this service, they have man\-
disadvantages, such as noxious fumes, ex-
cessive heat, a high fire risk, etc., and to
obviate these there was de^'eloped an elec-
trically heated two compartment muffle
furnace (Fig. 74).
It consists of two special steel alloy muflSes,
wound with nichrome wire on insulators
made of a compound which retains its elec-
trical resistance at high temperatures. Two
specially moulded nonpareil bricks jacket
the muffles and are protected by a two-piece
sheet metal casing. Four tie rods from front
to back hold the furnace together and the
terminals are brought out through a bushing
in the bottom of the furnace.
The 110-volt furnaces provide three heats
by means of a three way switch. On high heat
the muffles are in parallel giving the maximum
wattage (1500 or 2000), equally divided be-
tween the m^uffles. On intermediate heat
one muffle across the lining is heated, the
other by its proxim.ity being kept at a low
temperature for holding an iron warm. On
low heat, both muffles are in series, providing
sufficient heat to keep the furnace at a work-
ing temperature when not in use. The 220-
volt furnaces can only be operated on one
heat, and the mutlles are in parallel.
An electric muflle furnace was also de-
\-eloped for tool room work, where it is
essential to obtain temperatures as high as
S.50 deg. C.
The furnace proper (Fig. 75) consists of an
arched muffle, approximately Sjo in. wide,
4^ in. high and 15 in. long (inside dimen-
sions) around which is wound a spiral coil
heating unit, covered with a compound to
1 1 J 1 1 I 1 1 1 1
'
^
1
1
^JJ
isrr!T'^^'--v°!"i
jJiW-"
-.>t.r< Oldlsl
■^ J -H-
-r-
■'in
x' ^'' X
l»C0^.»OO^M-P-
J^" >=^ 1 111'
^ tft R" T — 7
^ ^"^-LtAd imp«r«tur«IO0Ve(tS
' /■-'
1
J » o \\ ^ > '
^ -+- -*~
1
_
1
i:oc »oc I'c XVoA/r-*
locc :co «« _ y/y-C!^
-..._,___j^ ^- Watt^lnp»Jl"OVo't»'
SCO 100 ^"-5^/
, .TTTJ .1-1 M : 1 -
«"» » "ii^fh
1 li li It 2 2
^.
i :
V^.
Fig. 73. Characteristics of Self RegulatinK Melting Pot
protect the wire from injury. Around the
entire muffle there is 2' 2 '"• of nonpareil
heat insulating block. The outside casing
is constructed of sheet steel, firmly riveted
and mounted on legs approxiniatei\- <i in.
high.
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 39
The door is hung on a rod supported b\-
hinges and may be operated by a handle on
either side; a ball weight is also provided to
hold the door securely closed or open as
desired. To eliminate heat losses around the
door there has been mounted approximately
2 in. of insulating block on the door itself.
On the top of the furnace there is a control
panel, provided with pilot lamp, line switch,
and a triple-pole double-throw switch for
obtaining high and low heat in the furnace.
The overall dimensions of this furnace are
233^ in. wide, 2 ft. lO^g in. high, 2 ft. :->^i in.
long.
When starting thu furnace it is only nec-
essary to close the main line switch, and
the control switch should then be thrown
to the blades marked high heat. When the
furnace is on high heat the pilot lamp will be
lighted.
It requires approximately two hours to
bring the furnace to maximum temperature
and when this temperature has been reached
the switch should be immediately thrown to
blades marked low heat and should be left
on low heat as long as the furnace is con-
tinuously operated.
The furnace consumes 4 kw. on high heat
and l.S kw. on low heat, and on the basis of
power being supplied at 2 cts. jjer kilowatt
hour, it m.ay be operated one day at a cost for
power of apjjroximately 4o cts.
discomfort in a small room, as it does not
radiate heat.
The cartridge type of heating unit was
first developed about thirteen years ago. The
units which have been used until about a
vear ago contained a lava core on which
Fig. 75. Electric Muffle Furnace for Tool Room Work
Fig. 74. Electrically Heated Soldering Iron Furnace
There is practically nothing about this
equipment to get out of order and the furnace
itself is so built that the heating element is
protected from harm. One other feature is
the fact that it may be operated without
Fig. 76. Magnesium-
Oxide Insulated
Cartridge Unit
heating coils (nichrome wire)
were wound. This was in-
serted in a brass tube lined
with mica which acted as an
electrical insulator and a very
close fit insured high ther-
mal efficiency. One end was
plugged with a brass disc,
over which the end of the tube
was swaged. The terminals
were brought out through the
other end through a brass
washer with two holes, the
leads being insulated by
smaller mica washers.
Recently this unit has been
modified, in that pure magne-
sium oxide is applied in place
of the mica heretofore used
as an insulator. This oxide,
in powdered form, is vibrated
through the unit (Fig. 70) in
such a manner as to com-
pletely surround it with a
40 Januan-, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 1
uniform insulating la^-er which will not break
down under extreme temperatures or exces-
sive vibration.
The units are used in shoe and cigarette
machines, glue pots, soldering irons, water
boilers, paper box machines and in heating
Fig. 78. Safety Enclosed Unit Removable Truck Type Switch
board, showing Bus and Cable Compartments
moulds; in fact, wherever localized heating is
needed the cartridge unit will usually solve
the difficulty, but they are not designed for
operation in the open air or <liroctly immersed
in liquids.
Switching Apparatus
As in preceding years, jirogress was con-
tinued during 101!) in the development of
safety enclosed switching ap]jaratus and the
line of truck tyi)e safet\- panels which i)rcvi-
ously co'vered feeder'circuit control onh", was
extended to include generator, synchronous
motor and lightning arrester panels. In addi-
tion to their essential safety features these
panels (Figs. 77 to 79) have several other
notable advantages:
The removable truck t>"pe is deliv-
ered "knocked down" as far as the
compartments are concerned, but the
removable trucks are completely as-
sembled. The work required then
consists only of placing them in posi-
tion and leveling the housings and
tracks for the trucks. The high grade
fitting work and the adjustment of
parts to insure interchangeability have
all been attended to at the factory
before shipment. The installation
therefore is reduced largeh' to the
assembly of fitted parts and does not
require a staff of skilled mechanics to
insure a finished job.
After the switchboard is in opera-
tion, if repairs are required, the
particular truck affected is pulled
out and a spare truck substituted.
Any necessary repairs to the trtick
can be made conveniently, quickly, thor-
oughly and safely in a suitably equipped
work sho]) to which the truck can be wheeletl.
Existing units can be increased in capacity
very advantageously, as the disconnecting
devices are identical for all capacities up to
(iOO amperes. The compartments have the
same dimensions for the same type oi breaker,
as do also the removable trucks. This means
that a larger capacity lireakcr of the same
type ma\- be substituted in the same tnick
Fig. 77. 76in. Safety Enclosed Unit Removable Truck Type Switchboard
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 41
and comijartment, and connection copper in-
creased to give a circuit of the desired
increased rating.
The vital factor of safety in handling is
assured, for with the truck element in place
for operation all the live parts are completely
enclosed, while with the tnick rolled out for
inspection, changing oil or repairs, all truck
parts are accessible, and at the same time,
dead electrically.
The new stationary type of safety enclosed
switchboard follows closely the design of the
ordinary open type board, except that it is
built of metal throughout. Steel front panels
(Fig. SO) are used, upon which can be mounted
an\' equipment required.
All switching equipment, including oil cir-
cuit breakers, field switches, lever switches, or
air circuit breakers are of the back-of-board
type so that a dead front, and therefore a
safe switchboard, is assured. The back and
ends are enclosed by grille (Fig. SO) so that no
Fig. 79. Safety Enclosed Unit Panel Removable Truck
Type, showing Circuit Breaker Equipment
live parts are accessible except to an author-
ized operator.
*Motor operated oil circuit breakers are
now made in removable truck form so that
* "Recent Developments in Circuit Breakers." by J. W. Upp.
General Electric Review. November, 1919.
either the entire breaker or any individual
pole may be easily and quickly removed for
inspection, adjustment or replacement.
The motor mechanism is mounted on top
of the cell (Fig. SI) in the usual manner and
an interlocking arrangement possessing several
Fig. 80. Safety Enclosed Unit Panel Stationary Type,
250 Volts with D-C. Motor Starter and Rheostat
safety features is part of the standard equip-
ment.
The swingout type of safety enclosed panel
(Fig. 82) is an ingenious self-contained unit
which obviates the need of providing separate
locations and supports for the instruments,
instnuTient transformers, disconnecting switch,
oil circuit breaker and conduit end bells, or
other devices for bringing the leads to and
from the breaker.
It is especially suitable for use in exposed
positions in niills and factories, as there are
no live exposed parts with which workmen
can come in contact.
The panels occupy small space, can be set
up singly or in groups, and can also be moved
readily from place to place if desired.
They can be swung in and out, as shown
in Fig. S3, and when a panel is out all the
apparatus mounted on it is dead and fully
accessible.
An interlock between the housing and the
oil circuit breaker prevents the panel from
42 January, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 1
being swung out, except when the oil circuit
breaker is in the "off" position, as shown by
an indicator on the panel, and the disconnect-
ing device is therefore carr\'ing no current.
Similarly, the interlock prevents the panel
from being swung back into operating posi-
Fig. 81. Removable Truck Oil Circuit Breaker Bottom and
Back Connected. Oil Tanks Mounted in Parallel
tion when the oil circuit breaker is held in
the "on" position. It can be locked in either
the open or the closed position.
Safety requirements in switching apparatus
are not limited to those controlling high
potential circuits or to the use of oil circuit
breakers. It is essential also that air break
devices be so enclosed that accidental contact
with live parts be prevented. In view of this,
the development of enclosed apparatus was
continued also for the air circuit breaker and
switch. Complete lines of both were pro-
duced either as individual devices or to ionm
component parts of a safety enclosed switch-
board for the control of direct current gen-
erators or feeders.
Safety enclosed and dead front air break
circuit breakers (Fig. 84) are equipped with
magnetic blowout instead of carbon break.
The breaker and its slate base are supported
from a front panel or plate upon which the
operating lever is mounted in an inverted
position.
When used in a switchboard this front plate
is a part of the switchboard panel. MTien the
breaker is mounted in a box as an individual
device it forms the front of the box. Links
extend through the panel to a roller, which
butts against a special casting on the breaker
proper when the lever is pulled out to close the
breaker. The lever and mechanism are
mechanically free from the breaker proper
(Fig. 85), which makes it possible to remove
Fig. 82. Safety Enclosed Uiut Swingout
Type Panel in Operating Position
Fig. 83. Swingout Type Panel Swung Out
for Inspection of Circuit Breaker
the front plate with the lever and inspect
the breaker parts.
A trip rod which engages the trip button
on the breaker extends through the front
plate, thus allowing the breaker to be tripped
manually without removing the front plate
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 43
and a window is pro-
vided in the front plate
through which the
breaker contacts are
visible. An indicating
device showing the open
or closed position of the
breaker is also moinited
on the front plate.
The safety enclosed
lever switches (Fig. 86)
may be mounted in a
box (Fig. S7) or used in
conjunction with other
apparatus to make up
safety enclosed panels.
The capacities of these
switches as used either
on a panel or in a box
are limited by the sizes of the 250 and GOO-
volt enclosed fuses approved by the National
Board of Fire Underwriters. This at present
is GOO amperes.
Fig. 84.
Dead Front Circuit Breaker
(Closed)
Fig. 85.
Dead Front Circuit Breaker
(Open)
cover and switch shall be so interlocked that the
fuses are accessible only when dead electrically.
Under this latter class comes the new Type
LM-4 enclosed switch (Fig. 88) mounted in
86. Dead Front Lever Switch,
Double-pole, Single-throw
Enclosed lever switches are divided by the
underwriters into two classes; viz, Class B,
which includes simi^ly a lever switch enclosed
in a box and Class .4, which specifies that the
Fig.' 88. Enclosed
Lever Switch
Fig. 89.
Disassembled View of Enclosed
Lever Switch
Fig. 87. Dead Front Lever Switch, Triple-pole,
Single-throw, Mounted in Steel Box
a rectangular box and cover, both made
from punched parts, and therefore light, and
at the same time strong and durable. A
front-connected lever switch movmted on an
insulating base is located inside and
operated by an external handle at the
side of the box. This handle is at-
tached to a "U" shaped shaft which
engages a hook shaped punching
(Fig. 89) mounted on the cross bar
of the switch. The handle is so
interlocked with the cover that the
fuse compartment cannot be opened
unless the switch is open (Fig. 90),
and conversely, the switch can-
not be closed unless the compart-
ment cover is closed. "On" and
44 Januan-, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 1
"Off" positions of the switch are indicated on
the box.
This switch is built double and triple pole
in 30, 60 and 100-ampere capacities, and is
rated at 250 volts a-c. or d-c. It can be
mounted with safety directly on a machine
Fig. 90. Enclosed Lever Switch with
Fuse Compartment Open
tool for the control of the individual motor,
or can be used on any feeder circuit within
its rating.
Grouped together or with instruments and
rheostat, on a suitable frame, with connec-
Very nearly identical with this switch
(tj'pe LAI-4) is the type LM-5 switch de-
veloped for use in connection with a compen-
sator for motor starting. This switch is
rated at 500 volts a-c. in 30, 60 and 100-
ampere capacities. It varies in construction
from the LM-4 only in that greater spacings
are required on account of the increased
voltage rating, and the addition of terminals
on the hinge clips which are connected to the
starting throw of the compensator, thus
shunting the excessive current at starting
around the fuses. Both types are arranged
for conduit or open wiring.
Several new types of relays were placed in
production during the year. The mechanically
balanced differential relays (Fig. 91) are
intended for the protection of parallel trans-
m_ission lines against unbalanced current in
the phases, such as would be occasioned by a
fault in one of the lines. As the current
increases in the lines, the difference in current
must also increase before the relay will oper-
ate. This compensates for a normal inherent
difference in impedence in the two lines. Fig.
92 shows the results obtained on outgoing
lines.
Fig. 91
Mechanically Balanced Differential Relay with Circuit
Closing Double-throw Contacts, 5 Amps.
tions between switches and other apparatus
installed in conduit, a complete jsanel can be
built which is compact, totally enclosed, and
safe. Such panels can control generator or
feeder circuits within the switch rating.
In operation the relay trips the line carry-
ing the greater current. It may be used,
therefore, for outgoing lines, or, providing
there is some other source of power to insure
that the injured lines will carry the greater
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 45
current, for incoming lines. The simplicity
of this relay is a valuable characteristic for
the use referred to. This relay operates on
current supplied by current transformers
only. It has a noteworthy advantage, in the
case of short circuits, over relays operated by
potential coils, as the latter lose their effective-
ness when the potential of the circuit falls off.
The relay consists of three solenoids (Fig.
93) , the two smaller outside solenoids tending
to hold down the moving mechanism, while
a differential current passing through the
larger center solenoid will tend to raise it.
When the difference becomes sufficiently
great to overcome the weaker of the two
small solenoids, the contact mechanism will
operate on the side to trip the breaker carry-
ing the heavier current. So long as a balanced
condition exists within the operating values,
the relay will not trip either breaker no matter
how high the current may be in the two.
Where differential protection is used for
alternators, each circuit should be equipped
with a device for opening automatically the
field circuit of the alternator after the oil
circuit breaker connecting this alternator to
the busses has been opened. This require-
ment demands either solenoid operation for
the field switch, or a manually operated field
switch equipped with a shunt trip coil.
A circuit closing auxiliary switch should be
provided on the oil circuit breaker, to insure
the breaker opening before the field switch.
,
-^^^
/
«s
7
i. o
/
S4
/
E5
/
/
f
.
^
«jo
<
^i
,^
^^
y'
'
- » ' ^
>
■'
'
,-*
L. t*
O -
'~
0 10 20 !0 40 50 60 70 80 90
Current in Restraining Coil in Amperes
ILowcrofthe Currentsin the Two Lincst
Fig. 92. Operation Characteristics of Mechanically Balanced
Differential Relay
With the breaker open there will be less
liability of damage to the field circuit, due
to the high voltage which would be induced
if it were opened when heavy currents were
passing through the armature. Opening the
field last also reduces the possibility of the
alternator falling out of step with the re-
mainder of the system, and thereby increas-
ing the disturbance on the system. It is, of
course, evident that under none of these
conditions is the difficulty entirely overcome
by the opening of the oil circuit breaker first.
Fig. 93. Diagram of Connections of Differential Relays in
Combination with Definite Time Limit Relays
on Two Parallel Lines
The trouble is, however, sufficiently reduced
to consider it the preferable method.
For this service Type PQ-6 relays (Figs.
94 and 95) with hand reset contacts are
used to insure tripping the circuit for the
field switch, after the main circuit breaker is
opened. By resetting the relay contacts the
field switch may be reclosed with the main
circuit breaker still open.
Hesitating control relays have been in use
for a long time where the electrically operated
circuit breakers controlled are equipped with
an auxiliary' switch automatically to break
the coil circuit of the control relay. The
relay closes instantaneously but "hesitates'.'
about one second after being deenergized
before the contacts open again. By this
time the breaker will ha\'e been positively
latched closed.
An earlier type of hesitating control relay
made use of an oil dash pot, but in the new-
design (Fig. 96), the time delay is obtained
by means of a heavy copper tube surrounding
the relay plunger and inside the operating
coil. When the coil is energized, the plunger
46 January, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xe. 1
is raised and the contacts closed. When the
circuit of the operating coil is broken, the
usual inductive "kick" starts up a heavy
current in the copper tube, which in turn
tends to maintain the flux. As a result the
flux dies always slowly, and in approxi-
Fig. 94. Instantaneous Hand
Reset Relay
Fig, 95. Type PQ-6 Instan-
taneous Hand Reset Relay
with Cover Removed
mately one second the plunger falls again
and opens the contacts.
A development that promises to assume
large proportions is the application of elec-
trically energized position indicators in con-
nection with systems of remote electrical
control.
The operation of this type of indicator is
based on the fact that two machines of the
induction type (Fig. 97) which are excited
from the same source and suitably inter-
connected, will rotate in synchronism.
Position indicators operating on this prin-
ciple were originally used on the lock control
boards of the Panama Canal, for showing at
the board by means of miniature replicas the
positive and progressive movements of miter
gates, chain fenders, water valves and water
levels. A subsequent ai)i)lication was made
in the form of signal pedestals for the trans-
mission of orders between the switchboard
and the generator rooms at the Keokuk
hydro-electric plant of the Mississippi River
Power Company.
At the present time there is being built a
lock control board for the New Orleans indus-
trial canal, which connects the Mississippi
River with Lake Pontchatrain. The lock
sj'stem, control board and position indicators
are similar to those used at Panama.
A further application of position indicators
under consideration is in connection with a
number of floating dr\' docks, designed for
the emergency fleet corporation. These dry
docks are of the multiple pontoon type (Fig.
98) and the proposition is to control all the
operations of listing, submerging and raising
from a central station, as opposed to past
practice of local control under the direction
of the dock master.
The scheme provides for a control board,
located at the head of the dock in such a
position that the operator has a complete
view of the dock itself and performs all the
manipulations that are required by merely
pushing a button or throwing a switch lever.
For the purpose of orientation, the top of
the board is laterally divided in the same
number of sections as there are pontoons in
the dock, and each section contains the neces-
sarv devices for the control of its pontoon.
Fig. 96. Hesitating Control Relay
The board is equipped with individual
indicators (Fig. 99) for reading the water
levels in the six compartments into which
each pontoon is divided. Flood valve indi-
cators are used to show at all times the actual
position of the main flood valves which are
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 47
used to fill the pontoon for submerging. A
list indicator will always show the divergence
from horizontal alignment. This can be con-
trolled to a nicety by the manipulation of
flood valves and pumps and with the aid of
the water level indicators.
geared. The pendulum will naturally main-
tain a vertical position and the listing of the
pontoon will therefore cause the transmitter
to rotate. This rotation is in turn transmitted
to the indicator on the board where it is
readily interpreted into degrees of list.
SHflFT
Fig. 97. Wiring Diagram, showing Interconnection of the Two
Electrical Units of the Position Indicator
The indicators are actuated by suitable
transmitter machines which can be briefly
described as follows :
The water level transmitter (Fig. lOOj
consists of a float attached to a chain which
passes over a sprocket wheel and has a
counter weight at its other end. Thus the
variations of the water level rotate the wheel
the shaft of which is geared to a position
indicator machine (See Fig. 97), which in
turn transmits its position to the water level
indicator. The transmitter is mounted on the
deck of the pontoon and is protected by a
weatherproof housing. It is further provided
with a dial for local reading of the water level.
The list transmitter (Fig. 102) consists of a
weatherproof casing, also mounted on the
deck of the pontoon and containing a posi-
tion transmitter to which a pendulum is
or
m
OOfTTOOr^
Fig. 98.
Plan and Section of Multiple Pontoon Dry Dock,
showing Location of Transmitters
The flood valve transmitter is connected
to the rising stem of the flood valve machinery,
the position of which is transmitted to the
Fig. 99 Arrangement of Water Level, Flood Valve and List Indicators
48 Tanuarv, 1902
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. i
flood \-ah'e indicator on the board. Six con-
trol switches and flood valves and two push
button switches for the pumps complete the
equipment for each section.
A further development in which position
indicators will figure is in the form of rudder
position indicators and revolution and direc-
Fig. 100. Water Level Transmitter
tion indicators for ships. In fact there are
numerous new fields for the application of the
principle.
Lighting
The total sales of incandescent lamps
(excluding miniature) in the United States
during the year 1919 is estim.ated to be 17o
millions of lamps, a decrease of 1 1 millions
(about 6 per cent) from the previous year.
The number of lamps sold each year from 1S9()
to date is shown by the cunes in Fig. 101.*
The sales of Gem lamps, which have a carbon
filament, are included in the sales of carbon
lamps. The Gem lamp is no longer on the
m.arket, its m.anufacture having been discon-
tinued in the early part of 1919. The tanta-
lum lamp was on the market from 19tl7 to
1912, but on account of the relatively small
numbers sold, it is not shown on the curves.
Of the total sales in 1U19 it is estimated
that 1(32 millions (923^2 per cent) are tungsten
filament lamps, a decrease of 4 millions from
the previous year's sales, and lo m-illions are
carbon lamps (T'o per cent), a decrease of 7
millions. It will be noted that the relatixe
* Since this estimate was made the actual increase in the
number of lamps suM has been so great that the corrected fiK-
ures for 1919 will ;ictuill\- exceed those for 191.S.
n
Fig. 102. List Transmitter and Local Indicator
SOME DEVELOPAIENTS IN Till': I-:LE("rRICAL INDUSTRY DURING 1919 49
number of tungsten filament lamps has in-
creased oxev the previous year, as is shown by
the curve (Fig. 103), which gives the number
of tungsten filament lamps in per cent of all
lamps sold in the United States from the com-
mercial introduction of this lamp in 19(17 to
date.
The white Mazda lamp (Fig. 104) was put
on the market during the past year. This is
a 50-watt Mazda C lamp for 110 to 12o-volt
service, the bulb being made of milky white
glass to diffuse the light. It should be seen
lighted to fully appreciate the beautiful soft
light it gives.
The milk-white diffusing glass of the bulb
is translucent rather than transparent, being
of sufficient density to protect the eyes from
the brilliancy of the filament when lighted,
}"et radiating practically all the useful light
rays, softened and evenly diffused. The white
Mazda lamp actually gives approximately the
same quantity as, and better diffused light
than, the ."lO-watt clear Mazda lamp.
Mazda B lamps (Fig. 105) specially de-
signed to withstand rough usage, were also
recently put on the market. They are known
as Mill Type Mazda lamps, and are made in
25 and 5(3-watt sizes for 110 to 125 and 220
to 250-volt service. The mount supporting
the filament is flexible, connected by a steel
wire "shock absorber" and every other fila-
ment loop is anchored in the middle to prevent
one leg of the filament overlapping another
and becoming short circuited.
There has been quite a change in the past
few years in the channels through which lamps
are sold, central stations largely giving up the
handling of lamps, as indicated by the fact
that the direct sales of a large Lamp Works
to central stations is now relatively, in pro-
jjortion to sales, about one half that of five
years ago.
In regard to lighting practice, the year 1919
may be characterized as one of unusual
activity. With the removal of war restric-
tions, conditions rapidly returned to normal,
and in many classes of lighting higher stand-
ards were reached. While the progress was
perhaps most rapid in store lighting, the
general advance in industrial lighting was
probably the most remarkable. One of the
lessons taught by the war was the importance
of good lighting as a means of increasing
manufacturing production.
Authoritative tests had shown production
increases as high as 20 per cent, due to im-
proved illumination. State industrial com-
missions and compensation insurance com-
panies ha\-e been advocating better lighting
for safety. Improved reflector equipment had
become available, and a foot-candle meter
permitted quick and easy surveys of light-
ing intensities. These and other conditions
working together stimulated interest in in-
dustrial lighting, and gi\'e promise of even
/SO, 1— T— T
1
—
~'
— t"
_
_
_
_
1
^
-
/ao-\
_
__
„
T
—f^ — 1—
_
_
/TO—] ^~
— \—
„
H
1 1
2
1
NvMBCR or Lapqc Lamps Solo
,N rne l/ivircD Statcs
1690 -I9t9
■*/
1
~~i —
1
;»-l
'^
i
--
1
k-
.—
— (
1
'^-
_J-_4-.^. -
P
--
_J
it
-
5 no
-—
—
~
~
\ —^
i
I'f. 1
\
V\
'
1 «-
1
\\
^
/
1 '
.
y
^7'
f
!
5; *
1
V
1
-J-
-
l^
70
w
f-'y
s
/-ia*
*
/
1"
#^.,
y
•/
•^
?J
1
IS
^
Y
i .*
\*
§ Vis
■
-f
^i
'13,
k'
f
^Y
-4P
"."«:
i»
-Ui'^rf^J:
c^^
—
_
+
K-
V
—
« ^^
^-\-
—
S
—
—
~
—
t it
A
s s s
3) s S
1 i
1
1
1
0.
1
1
Fig.
101. Number of Incandescent Lamps Sold in the United
States (excluding miniature lamps)
wider use of better lighting in the future. The
industrial lighting codes (Fig. 106) have
begun to influence factory lighting in the
several states, where adopted, to which were
recently added California and Oregon. Sev-
eral other states appear to be on the verge of
similar action.
The automobile headlighting problem re-
ceived considerable attention, and improved
regulations for safer driving lights were
adopted in several states, including California,
Connecticut and Pennsylvania. Additional
tests by committees of the Illuminating Engi-
neering Society and Society of Automotive
Engineers have confirmed the specifications
already prepared.
Among the new types of lamp accessories,
it m.ay be mentioned that the RLM Standard
dome reflectors (Fig. lOS) are now being
50 Januan-, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 1
made and strongly recommended by practi-
cally all of the leading manufacturers of
steel reflectors, thus helping to assure good
industrial lighting.
Quite a number of new fixtures and acces-
sories have become available during the year.
TuN&STEN Filament Lamp Sales in
Percent of Total Sales
90
£ 80
4 70
i 60
^
?V
-^
^\
f
\
f
f/
^ 40
kJ
O 30
u
a, 20
10
A,
{
i^
<f
4\
r
/
^
^
/
Year
Fig. 103. Tungsten Filament Lamp Sales in Per Cent
of Total Sales
Among these may be mentioned the "Ace."
which consists of a combination reflector,
diffuser and enclosing globe (Fig. 109) made
of one piece of glass.
A dense opal coating on the upper portion
reflects a large percentage of light down-
ward, but still provides suitable illumination
on the ceiling. A lighter opal on the lower
part conceals the lamp filament and diffuses
the direct light. A clear section between
avoids loss in transmitting the reflected light.
The whole forms a large diffusing light
source which gives a decided downward direc-
tion to the predominating light. It is large
enough so that the lamp can be tightly
enclosed without overheating, and dust can
thus be excluded from the interior. It finds
application in stores, offices, drafting rooms,
and the better class of workrooms.
The Duplexalite fixture (Fig. 110) has been
widely applied during the year in residence,
hotel and commercial lighting. This is a fix-
ture of the semi-indirect type, so arranged as
to minimize the light projected horizontally.
While the light controlling part of the fixture
is standardized, it lends itself to decoration
by means of silk shades, thus permitting the
expression of individual taste. Many vcr\'
attractive installations have been made,
including room lighting in leading hotels.
The incandescent lamp for moving picture
projection is receiving more extended appli-
cation. Quite a number of improvements are
still being made in the lamps and optical
systems, and it is probable that incandescent
lamps will soon prove applicable for larger
screens and longer throws than were originally
contemplated.
The new developments in street lighting
were not either radical or revolutionary in
scope but tended toward simplification
of apparatus and maximum utilization of
light.
In the pendent unit for series Mazda lamps,
the most important addition was the com-
bination of Holophane dome refractors and
stippled or rippled outer globes (Fig. 112).
The refractor when used alone has not been
universally satisfactory, although its use has
been ver\- general. The collection of dust
and dirt is still serious in some communities
and this deposit can collect on three sur-
faces, viz., the lamp bulb, and the inner and
outer faces of the refractor. Where dust and
smoke are prevalent and where the glassware
is not cleaned frequently, this condition may
account for a fift\' per cent absorption of
Fig. 104. White Muda
Lamp
Fig. 105. Mill Type
Masda Lamp
light. By enclosing the dome refractor in a
stippled globe, only one surface is exposed.
In these globes the diffusion is obtained by
protuberances and depressions in the surface
of clear glass which breaks up the light but
does not interfere with the directional effect
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 51
of the refractor; the absorption being practi-
cally that of clear glass.
There is an attractive installation of these
units at Niagara^Falls, N. Y.
SAFETY STANDARDS
INDUSTRIAL BOARD
PENNSYLVANIA DEPARTMENT OF
LABOR AND INDUSTRY
MJUmNTIO 18
INDUSTRIAL CODE
LIGHTING
OPERATIVE ON AND AFTER JUNE i
Lighting of Factories and Mercan-
tile Establishments
STATE OF NEW YORK
DEPARTMENT OF LABOR
STATE INDUSTRIAL COMMISSION
■UREAU OF INDUSTRIAL COOL
INDUSTRIAL UGHTINC CODE
FACTORIES. MILLS, OFFICE AND
OTHER WORK PLACES
l>J0U5nil*L COMMBBON OF
Code of Lighting
Factories, Mills and Other
Work Places
DEPAHTmENT OP LABOR
CODE OF UGHTINC
Factoda, Milb ud Other Wok PIko
Ifilbn Wh* Smrimm
/^^^l^^rf CtmmUtUi^Cmmntll Hall
Gjde of Lighting
for
FKtories, Mills and other Work Places
RepOT^ of [hvuiona] Conf
mjttee on Lixhtins. S«cbon
on SuiilsQan. Committra
wWeUueWoik
OOMMTTItEON LABOR
Fig. 106. Some Typical Lighting Codes, Indicative of the
Growing Appreciation of the Importance of
Adequate Industrial Illumination
The single light of high candle-power has
proved more pojjular than the clusters of
lower candle-power lamjjs from a standpoint
of economy, efficiency and appearance. The
standards themselves are becoming slender
and unobstrusive, while the new globes are of
Fig. 107, Lighting Fixture Using One-piece Porcelain Combined
Radial Wave Reflector and Refractor Holder
more graceful shapes with smaller top and
bottom openings and with as low absorption
as is consistent with perfect diffusion. Recent
forms (Fig. 113) show a serious effort to
harmonize the architectural features of the
l)ole, casing and top.
In an effort to reduce manufacturing
operations and to increase the safety factor,
there was produced a unit (Fig. 107) moulded
of one piece of porcelain, combining a radial
wave reflector and a refractor holder. This
should simplify the question of maintenance
and replacement as it increases enormously
the insulation of the unit.
To increase the actual illumination secured
with luminous arc lamps (Fig. Ill), the ingre-
dients of the electrodes were compounded
under great pressure. This permits a high
efficiency mixture, giving 30 to 40 per cent
more light than the standard electrodes with
at least equal life. If the standard intensities
are satisfactory, the compressed electrodes
will'yield an increased lifeof 30 to 40 per cent.
Januan-, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 1
Fig. 108.
Photometric Test of RLM Standard Dome Reflector, showing
Distribution of Candle-power in a Vertical Plane
With this electrode, it should be possible
to use a glass mirror internal reflector which
adds a further improvement to the effective
illumination. In the luminous arc 60 per
cent of the light is above the horizontal. The
present porcelain reflector cannot be made
particularly eflScient and becomes of less
value as it is discolored by fumes. The glass
reflector will be initially of higher efficiency
and will be more easily cleaned.
Pendent luminous lamps have been regu-
larly furnished with clear globes. With the
Fig. 109. The "Ace" Combination Reflector. Difl'uscr
and Enclosing Globe
increased light from the new electrodes it is
now desirable to use a blown rippled globe
which gives an appreciable degree of diffusion
with no greater absorption than the clear glass.
The increased efficiency electrodes also per-
mit the use of a lower wattage adjustment on
each lamp without reducing the effective
lighting. This saves 40 watts per lamp and
increases proportionately the capacity of the
rectifier. In Detroit for instance, they are
operating 90 to 94 low wattage lamps on each
75-light rectifier.
Unusual interior lighting effects were se-
cured at the Chicago and Buffalo Electrical
Shows in 1919; in each case the required
illumination being combined with a unique
artistic scheme of decoration consistently
adhered to throughout.
Chss
diffusing
Fig. 110. The Duplexalitc Fixture
At the Coliseum in Chicago (Fig. 114) the
display was called the "City of Aladdin"
anti represented a Chinese market place.
illuminated by Chinese lanterns.
In the center there stood the "Palace of
Aladdin," a structure 50 feet in height, stud-
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 53
Fig. 111. Series Luminous Arc
Lamp Equipped with Clear
Rippled Outer Globe
Fig. 112. Pendent Novalux
Unit with Stippled Globe
and Dome Refractor
dec! with glass jewels and utilizing panels
and plates of a new form of painted mirrored
glass, flood lighted by concealed searchlights.
At the Buffalo show an entirely different,
but equally striking effect was obtained by
using an ultra-modern decorative scheme (Fig.
115) in which 4500 illuminated discs were
distributed among the roof girders for the
purpose of neutralizing or camouflaging these
girders. The discs were covered with geo-
metric designs in metallic paints and the
color tones from the sides of the building to
the center line of the roof were graded so as
to give an effect of height. Thirty stage spot-
lights were used to illuminate these discs.
All other lighting at the show was produced
by 5000 white Mazda lamps which were, in this
way, displayed in quantity for the first time.
As the electrical industry devoted all its
energies from the beginning of hostilities to
the intensified production of apparatus and
supplies which were utilized either directly
or indirectly in National service, it also, with
the coming of peace, developed forms of
artistic illumination especially designed to
serve as a visible welcome to our returning
soldiers and sailors. This special application
of illuminating engineering has been generally
referred to as "Victory Lighting."
Fig. 113. Typical Ornamental Single Light Street Lighting Units
54 January, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII. No. 1
Fig. 114. "City of Aladdin"— C licago Electric Show
Fig. lis. White Mazda Lamps— BufTalo Electric Show
SOME DEVELOPMENTS IN THE ELECTRICAL INDUSTRY DURING 1919 55
Fig. 116. "The Jeweled Portal for the Victorious Army," New York City
Fig. 117. The Jeweled "Altar of Victory," Chicago, III.
50 Januan', 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 1
An excellent example of Victor}- Lighting
is the "Jeweled Portal for the Victorious
Army" (Fig. 116) in Xew York City. The
two vertical shafts of the portal, SO feet in
height, are each surmounted by a sunburst
of glass jewels and between them is sus-
pended an ornamental jewelled curtain in
geometric design; the complete portal con-
taining about 30,000 of these jewels. The
lighting of the portal was accomplished
by means of 24 IS-in. arc searchlights
which were provided with color screens
to gi^•e variety to the lighting effects pro-
duced.
In Chicago the corresponding display took
the form of a huge "Altar of Victory," the
central feature being a curtain of jewels
(Fig. 117) suspended from two candelabra
each'90 feet in height, which were also studded
with glass jewels. The 30,000 jewels which
were utilized in this manner were illuminated
by 60-in. army searchlight projectors located
at a distance from the altar supplemented by
a number of small flood lights located on the
platform at the base of the altar and at other
nearby points.
A less elaborate but very attractive exam-
ple of "Victory Lighting" is shown in Fig. 1 IS.
It consists of a sunburst shield containing more
than IGOO o-watt Mazda lamps with which the
required yellow and red effects were produced
by the use of color caps on the lamps. The
background was painted as a sunburst so
that even by daylight the shield presented an
attractive appearance.
The central figure is the flag, which is seen
flashing in waving motion against the steady
brilliancy of the shield of victory-.
Fig. lis. The Incandescent "ShielJ of Victory,"
Schenectady, N. Y
57
Thermostatic Metal
B}- Henry Herrmax
Metals Division, Fort Wayne Department, General Electric Company
For use in the temperature controlled devices it manufactures, the General Electric Company sought in
vain to find on the market a thermostatic metal of the same high standard of quality as the remaining parts of
the devices. It naturally set about to strengthen this weak point and in consequence developed the metal
described below — a duplex metal superior in characteristics to any domestic or foreign product The attention
of manufacturers of devices influenced by temperature is directed to the usefulness of this material, as its field
of application has alread}' proved to be far greater than originally contemplated. — Editor.
Thermally responsive devices have been
taxing the inventive minds of the world for
many years. Various combinations of
materials having widely different coefficients
of expansion have been used in various man-
ners to get a thermally responsive device.
The thermostatic element of these early
devices consisted of separate strips of zinc
and steel riveted together; hard rubber rods
operating in connection with glass ttibes;
separate strips of brass and steel attached at
the outer ends in such a way as to form a
bellows-like combination; and many similar
adaptations along the same lines, all designed
to produce motion as the temperature varied.
All of the combinations tried were found lack-
ing in one very important respect; that is,
the permanency of position or what may be
called pennanent zero. Upon
material changes of tempera -
ttire the deflection which took
place strained the connection
between the two materials
with the result that they
shifted slightly with regard
to each other, and therefore
any pointer or mechanism
which they controlled as-
stuTied a new position when
the basic temperature was
again reached. This common
defect in the earlier thermo-
static elements lead the Ger-
mans to develop a duplex
metal with an absolutely
homogeneous joint, the two
dissimilar metals being ma-
nipulated so that at the
surfaces of connection they
were completely alloyed. In
this way a thermostatic
metal having permanency
of position was obtained.
A secondary part of the problem lay in the
choice of metals having coefficients of expan-
sion sufficiently different to give a large
deflection. For use in the thennostatically
operated devices of its own manufacture, the
General Electric Company developed a bi-
metallic thermostatic material which is similar
to the German material in that the two
metals are permanently united, but which is
superior in many characteristics; for instance,
its deflection per degree is 18 to 20 per cent
greater than that of the foreign product.
The combination of metals used is actually
the one which gives the greatest deflection
per degree of temperature change that can be
obtained with reliability.
In its manufacture a bar of metal having
the lower coefficient of expansion is placed in
a mould and heated by an oxy-acetylene flame
to a bright red. A fluxing agent is then
sprinkled on the surface and the bar is brought
to its fusing temperature. On this fluxed sur-
face a laver of the metal havine the higher
—I—
J.
i """X"""-L ^
± ^. T . 2 -.- -
1 / \- - ^r-
^ i-v - -
" - ^7
"' T _..___. T ~^/^-
"-Ll-L ^^ J^
Ii Ip ^ . A ^^
^^ ■ ■ IT . . \ A
AS "XiL ■ ^ -i- ->-L -^ -V-
T t- ^1 '^ '
" —4- y^ -1--^
a.-i ~r -r-p-pT^^/ , T "] ■ --
1 ] -^ ^'
=±=
Fig.
l^n^Lf of A!etai strips n y/>: V.'
Deflection Obtained from Thermostatic Metal Strip Subjected to Different
Temperatures. Size of strips 4 in. long, -|\ in. wide, 0.030 in. thick
coefficient of expansion is melted b\- the oxy-
acetylene flame. The surface of this layer is
then brought to a molten state by the oxy-
acetylene flame and molten metal of the same
composition is poured on to build up the de-
58 January, 1920
GENERAL ELECTRIC REVIEW
Vol XXIII, Xo. 1
■
1
!
1
/
/
y(
f' \
/
'
/
^.^O
/'
'^ 35
/
/
\0A
/
/
■§ 03
/
/
^
/
^
;
§
/
1
■??'"
/
1^
--
-
-
-
-
/■
-
-
-
--
-
-
-
"+■
-■"
*^
'
/'
/
/
/
1
1
/
/
1
"~a 20 40 60 80 !00 fZO f40 f60 f80 200 Z20 Z40 260
Fig. 2. Deflection Obtained from Thermostatic Metal Strips
of Different Lengths When Subjected to a Tempera-
ture Change of 100 Deg. F. Size of strips
A in. wide, 0.030 in. thick
'
1
1
■
1
—I —
1
,'=?
'
fl
1
■
1
1
fo
.
;
1
1 1
\ 1
\
'
'
■"
V
\
o
\
'
V.
^
v-
!
.
2
1
•^
1
\
^.'d
V
i
1^7
\
'
\
1"
03
\
1
'
\
s
\
0-1
\
1
\,
1
03
jN
s
1
1
'^L
az
1
~.^
—
-^
~-^-
Of
■
J aOW 0020 0030 0.040 0.^$OJ06OQO?0 0aOO0O3C OfOO OrO J^tOJ'iC
Vy-c/fr^ss a'' •'<*- j. -Scrp m^^c^^
Fig 3. Deflection Obtained from Different Thicknesses of
Thermostatic Metal Strip Subjected to a Temperature
Change of 100 Dcg. F. Size of strips 4 in. long
\ in. wile
'
1
1
-!—
■
'
1
1 1
1
1
1 1
J6
'
'
'.
^_7
(_
'
^
'
1
tl
z:
ZII '
^
/
/
1
f'^
/
'
,/
i
/
% -/,
'
/
/
,5
/
/
^' -r
/
^
1
1
1 '
-
/
1
f
/
'
1
/
!
7
/
/
/
/
.
/
\
/
!
1
/
!
1
^
'
/
.
/
!
i^-^
n
--
' 1
■7 co/0 C020 aox> 304^0050 oosff aa?3 aax? oo90 o.vo a.-io ;?/so oao
Fig. 4. Force Exerted by Thermostatic Metal Strips of
Various Thicknesses for tOO Dcg F. Temperature
Change. Size of strips 4 in. long, \ in. wide
'
/
I
1
'
1
,
f
/
/
/
/
/
/
S CI
I
, L-,
^
1
f
*.^
1
1?
f
/
I
1
f
/
,
,
/
i
/
'
/
^
1
/
1
,/
/
*-■
/
T
>
1
^
7 /M
VI at
WlOA
i
s-.;.'
T-i.-,
~l : !
.. 1 L
Fig. 5- Force Required to Obtain Permanent^Sei of Thermo-
static Metal Strips of Various Thicknesses. Sixe of
strips 4 in. long, \ in. .wide
THERMOvSTATIC METAL
59
sired thickness. The resullinj,' composition
ingot is shaped, cleaned, rolled, and annealed.
After the thorough superiority of this do-
mestic made duplex metal had been demon-
strated, the Company placed its manufacture
on a larger scale in order that the metal would
Fig. 6. Photomicrograph oi Thermostatic Metal,
showing Union Between Components
be available to other manufacturers of devices
who can use it to good advantage in their
products. In some of the applications of this
thermostatic metal, it is used for the purjjosc
of temperature indication, in other applica-
tions it is used for temperature control, while
in still other applications it is used to com-
pensate for or neutralize errors in apparatus
due to changes of temperature. These avail-
able applications have already resulted in the
use of the metal in :
Oven thermometers
Electric heaters
Ice machines
Refrigerators
Thermostats
Scientific instruments
Automobile ignition control
Battery charging control
Electric signal control
Carburetors
Computing scales
Speedometers
Steam radiators
Sterilizers
Gas range valves
Automobile shutters
Heating pads
Flatirons, etc.
It is of interest lo note the laws which
express the characteristic action of this metal.
Careful laboratory tests show that
1. Deflection upon Temperature Change Varies
Inversely as the Ihu-kiicss
As the square of the loii^th
Not affected by changes of n'idih
Directly with decrees temperature change.
2. Force Exerted upon Temperature Change Varies
As the square of the thickness
Not affected by changes in length
Directly as the width
As the square of the degre- temperature
change.
3. Weight Required Jor Permanent Set Varies
As the square of tlie thickness (between
narrow limits)
Inversely as the length
Directly as the icidth.
The curves in Figs. 1, 2, .'5, 4 and .5 illustrate
the action of the metal under certain varying
conditions.
Under the sclerescope the yellow of the
metal shows a hardness of 23, and the white
side a hardness of '.ic>. based upon a one per
cent carbon steel (hard) as 100.
Fig. (i shows a photomicrograph of the
union of the two metals and indicates how
impossible it is for one metal to slip upon the
other as temperature changes cause the metal
to bend. So perm.anent in fact is the union
found to be that no amount of bending or
twisting will separate the two metals; there-
fore it is possible to form the combined
metal element into various shapes and to
anneal it after the forming operations. Even
on heating the bond will not be broken below
the melting point of ti e metal which has the
lower melting temperature, and it is found
that the metal can be used safely at any
temperature below .")()(! deg. F.
The force that the thermostatic metal is
capable of exerting without taking per-
manent set is dependent upon the thick-
ness of the strip. Fig. 4 shows that the
metal on bending with temperature change
will exert considerable force for the
mechanica' operation of various devices
without taking perrranent set.
Another feature of interest is the fact that
both materials used n forming the ther-
mostatic meta' are \-ery resistant to cor-
rosin, so that ii can Ite used in any location
reasonable for the use of metal without
deterioration nr change in its operating
characteristics.
60 Januan-, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 1
Electric Propulsion of Merchant Ships
By W. L. R. E-MMiiT
Lighting Department, General Electric Company
In view of the successful applications of electric drive to large naval vessels, and the unforeseen limita-
tions met in the turbine gear drive of cargo vessels, the author of this article shows how suitable is electpc
drive for ships of the merchant tvpe. The article was read as a paper before the Society of Naval Architects
and Marine Engineers, New York Citv, November 1.3 and 14, 1919.— Editor.
The use of electricity for propelling ships
was first advocated in the case of large war-
ships in which it affords particular advantages
in the matter of cruising economy through
change of speed ratio, interchangeability,
space distribution, etc. The first application,
however, was made in the case of the U. S.
collier Jupiter, which is in most features a
ship of the merchant type. The demon-
stration of geared-turbine propulsion came
after the first serious proposals of electric
drive, and the advantages which have been
attributed to the geared method have sus-
discontinued such activities in this direction
as had been planned.
Two or three >-ears ago the writer was of
the belief that the geared equipments then
being made aft'orded a solution of the problem
which in cost and results would probabh"
prevent the commercial success of electric
drive in merchant ships, although it was
realized that the margin of possible advantage
was small. Since that time improvements in
electrical designs ha^•e been developed, and
limitations of gear possibilities have appeared
which put the question in a different light;
Bou»r
/o /3- ZO af 30 3S ■*« **■ so SS «« •*■ 70 73- SO if J»
Larr^it t/a/no f Sect Jon
Shot^ingElecCr-ic Orife
Fig. 1 . Diagram showing Arrangement of Apparatus in an Electrically Propelled Merchant Ship
pended such activities as were considered in
this countn,' in the direction of electric drive
for merchant ships, while, in the case of war-
ships, electric drive activities have been
uninterrupted. In the meantime certain
electrically driven ships built in Europe, and
operated with very high degrees of superheat,
have shown wonderful fuel economy, and
many more such ships arc being equipped.
The larger American shipbuilders, having
their own facilities for machinery construc-
tion, have, not unnaturally, been opponents
of electric drive; and the Emergency Fleet
Corporation, which for some time has repre-
sented ownershi]), has for various reasons
and it is now the writer's belief that electric
drive is justified in all large ships and that
it will very soon develop a wide application
notwithstanding the great efforts of skill,
organization, and capital which have been
given to the introduction of the gear drive
for vessels of all classes.
The discussion of this subject is largely a
matter of comparison with other methotls.
and the purpose of this article is to make
clear what is proposed in a specific case and
to suggest comjiarisons which may affect
relative value.
The case selected is that of a vessel of
8S00 deadweight tons, length 424 feet.
ELECTRIC PROPULSION OF xMERCHANT SHIPS
61
Boiler
Room Floor
1 1 n 1 1 1 M )[ 1 1 1 1 1 1 1 1 1 1 1 1 1 [ 1 1 1 1 1 1 1 [ 1 1 1 1 1 f ] I f 1 1] 1 1 1 1 1 1 1 n 1 1 f n ' 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 |TiTi|'i 1 1 1 1 iTTij
O ^ /O /S ZO ZS 30 35" '^O ^S SO SS^ &0 <^S 70 7*" SO SS SO >9^ /&0^
Long ftu£::/ina/ Sect /on
Fig- 2, Diagram showing Arrangement of Equipment in Merchant Ship Propelled by
Triple Expansion Steam Engine
Sv^'tcMboard.
a
X4n. I
SOOOH.f?
A C Turho Gen /SSHin/.
Q C Turbo C^n.
■--^''■' * Center L me
,-■ s ^, '.yc.Ofi/nieNOZ
Condenser
for
DC. 7urbo&.
p0s}t/on i- ooH/nty /I ft at F/~ame AJo. SS
Fig. 3. Details of Turbo-electric Propelling Equipment
30 BZ SO SIS'
E/et^otiorr Looking to ^o/^t
LOOhing /Jft oi Fr-ame Na 85
ao as so
Elevation Loohing to Port
Shelter QacM
Beam for Lifting C&Qr ^
Coal Chute.
Bridge Qa.ck
Fig. 4. Details of Triple Expansion Engine Equipment
62 January, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 1
beam 54 feet, having a cubic capacity of
460,000 cubic feet and capable of making
11.5 knots with 2500 shaft horse power
delivered to a propeller operating at 100
revolutions per minute. Figs. 1 and 3 show
an electric propelling equipment applied to
such a ship; and, for comparison. Figs. 2 and
4 show an equipment with triple-expansion
engines. It will be obser\-ed that the motor
s placed as far aft as convenient, affording
space for disassembling and for removal of
the tail shaft. The generating unit and con-
trolling equipment are placed near the boilers
in such a manner as to afford a m.aximum
convenient saving of cargo space, the con-
denser being suspended below the turbine in
the same compartment with boilers. The
auxiliaries are distributed in convenient
locations in the turbine room and in the
space below near the condenser.
The weight of thisequipment. including gen-
erating unit , motor, controlling mechanism and
direct-current exciter, will be about 67 tons.
Auxiliaries
In connection with such equipments it is
proposed to use, as much as possible, elec-
trically driven auxiliaries. It is necessan-
to maintain an electrical supply independent
of the main generator for purposes of excita-
tion and lighting. The losses involved in the
operation of larger auxiliary generating equip-
ment are relativch' much less, and there is
no increase of complications. With such an
equipment it is proposed to install two 150-
kw., turbine-driven, direct-current auxiliar\-
generating units, one being required for
serv^ice and the other installed as a spare.
Excitation and lighting will only amount to
40 kw., leaving 110 kw. available for any
possible auxiliary uses. A little more than
half of this should be sufficient for normal
conditions. Whik- the ship is at sea, for rea-
sons of simplification and economy, it is pro-
posed to exhaust the auxiliary generating
unit into one of the lower stages of the main
turbine at a pressure somewhat abo\-c the
atmosphere, so that some of this exhaust
steam will be available for feed water heating
if that obtained from steam-driven auxiliaries
is insufficient. In jjort. these auxilian*- units
would be exhausted into an auxiliary con-
densing plant which would be idle while tlie
ship is at sea.
Moior Compartment
The motor carries ihe thrust bearing and
is also equii)ped with a simple, slow-moving
oil pump which maintains automatic lubrica-
tion in the motor compartment. This lubrica-
tion can be arranged with a storage tank and
with an emergenc\- drip supply to the low-
speed bearings contained in the after com-
partment, so that, even if the oil pump
should fail, many hoiu-s might elapse before
injury could result to any of the bearings.
With such an arrangement the self-lubrication
of this compartment becomes entirely simple
and safe, and with occasional inspection it
should be operated without an attendant
and without any passage connecting it with
the engine-room; in fact, there is nothing that
an attendant need do in this compartment,
and there would be quite as much reason for
keeping an attendant on the truck of an elec-
tric locom.otivc where the electrical and lubri-
cating conditions are far more complicated.
Space Saving
It will be observed that the omission of the
shaft alley and the diminution of space re-
quired for the engine-room m-atcrially in-
creases the cargo space and simplifies its
shape. This increase amounts to something
over 12,000 cubic feet, nearly three per cent
of the total capacity of the ship. The omis-
sion of the shaft alley, shafting, and support-
ing bearings effects a weight saving of about
60 tons, and there will be an additional
weight saving in the machinery itself since
the electrical equipment will weigh about
nine tons less than the engine equipment for
such a shi]>.
Economy
If this equipment is oix-rated with 2LK>
pounds steam pressure. 200 deg. F. superheat,
and a vacuum of 2S.5 inches, the steam con-
sumption per shaft horse-power hour, not
including auxiliaries, will amount to 9.5
pounds. Under nonnal conditions at sea.
with most of the auxiliaries driven electrically,
this should give a steam consum])tion for all
puri)oscs not greater than 1 1 pounds per
shaft horse-jjowcr hour. Such a steam con-
sumption will require at 'east .JO per cent
less fuel for all purposes than would be re-
quired by a goo(i reciprocating engine equip-
ment operating without superheat, and even
if an equal superheat were used with a
reciprocating engine equipment, the gain
would still be over 20 ]>cr cent.
In this connection it must be considennl
that large numbers of American ships arc now
being equipped with recii^rocating engines
and without superheat, although it has been
amply demonstratetl abroad that the use of
high superheat is practical and economical.
ELECTRIC PROPULSION OF MERCHANT SHIPS
63
If such a ship were in operation 250 days
in a year between CaHfornia and AustraHa,
burning fuel oil at $1.00 per barrel, the saving
in fuel over a similar engine-driven ship
operating without superheat would amount
to about $17,000, and the increased freight
capacity leaving California would amount
to 585 tons, which is 7J^ per cent of the dead-
weight tonnage.
Reliability
A study of the records and uses of such
electrical apparatus as is applied in this case
will show that the equipment is less liable to
interruption of service than any other form
of single-screw equipment which is applied
to vessels. With such an equipment, how-
ever, arrangements could easily be made by
which the ship could be navigated about half
speed with the main generating unit out of
service. This could be done by providing a
motor-generator set or rotary converter so
arranged that the power of the auxiliary
generating units could be delivered to the
main motor. In an electrically propelled
ship, electricity is produced simply for one
definite purpose, and the arrangement is
simpler and more reliable than shore appli-
cations where power is taken from large
distributing systems. It is also possible to
provide automatic means which, by inter-
rupting excitation, guard against the pos-
sibility of serious damage through possible
accidents or insulation failures. Such elec-
trical apparatus of the type used in ships is
very easily repaired, and even when damaged,
can generally be temporarily connected so
as to be operative. The knowledge necessary
for such repairs is very easily imparted and
is constantly being practiced in our industries
b\' persons who have had little or no electrical
training.
Reliability of Gearing in Ships
To make comparison of such an equipment
with a gear-driven ship is much more difficult,
since a great variety of arrangements of
turbines and gears have been applied to ships
of this type. In the matter of reliability, as
has been said, the electrical equipment is
entirely beyond question, while many evi-
dences of serious trouble and deterioration
have developed in geared ships of most types
which have been produced. Gears have been
verj^ successful in many warships, but these
are subject to only occasional short periods
of high-power service. In some merchant
ships, gears have been very successful; and
in others, most serious trouble has been
encountered. Variations of results in similar
equipments in different ships illustrate some
of the possible uncertainties. Parsons' orig-
inal gear applications operated with a single
reduction, very small diameter pinions, and
a large diameter gear on the propeller shaft.
Some of these have been reported to be very
successful, but the gains in economy shown
in Parsons' publications are nothing like so
great as those accomplished by high-speed
turbines with double reduction gears. There
have, however, been inany cases of failure
with gears of this type. In fact, there seems
to be no type of gearing with which trouble
has not been experienced after long service
in cargo vessels.
Recent production of so-called "Standard
English ships" shows that they are being
equipped with double-reduction gearing; and
at the same time that this change of method
is being adopted in England, the use of single
reduction is being extensively advocated and
applied here. Although all the original Am.er-
ican equipments in merchant ships were
double reduction, the writer has seen a solid-
gear, double-reduction equipment of American
make in which the gears were badly worn
and pitted after 17,000 miles of service, and
in this case the proportions of gears are closely
equivalent to those which have been adopted
in the new standard English ships, and the
conditions of design and manufacture were
quite as good.
These indisputable facts and many others
certainly indicate that gearing for ships has
not yet reached a state of finished develo]3-
ment.
One of the tuicertainties of gear operation
in ships is illustrated by the very great differ-
ence in durability of gears in ship propulsion
and in shore uses. In trials on shore, gears
have borne without blemish, for equal periods,
loads equivalent to approxim.ately four times
the average loads which have caused bad
destruction of similar gears at sea. This is
illustrated by the photograph in Figs. 5 and
G, and the -data given in the titles are char-
acteristic of many other similar experiences
which have developed.
The reasons for these astonishing differ-
ences have never been adequately explained.
Fig. 7 shows a record taken from a torsion
coupling on a cargo vessel operating in ballast
in a moderate seaway. This record shows
that the torque on the propeller shaft varied
from zero to approximateh- 73 per cent over-
load under certain wave conditions. The
effect of bad weather on the endurance of
gears has often been observed, and it is quite
64 Januan-, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 1
possible that variations much greater than
that here shown may at times be experienced.
In this case the ship was pitching only four
degrees. Part of the small, quick variations
shown in this record were caused by a trans-
mitting ring which ran slightly out of true
in the instrument, but otherwise the condi-
tions were such that the record must be
substantially correct.
Another matter of uncertainty in gearcd-
turbine equipments is that of the temperature
in the turbines. The operation of turbines
in the reverse direction occasions large tem-
perature variations, and temperature varia-
tions constitute a fruitful source of danger
to turbine stnictures. Fig. 8 shows a record
for temperature taken by a pyrometer situated
indicate that such effects may be serious and
should, if possible, be avoided. A turbine
which is kept running in its normal direction
is not subject to any large temperature \-ari-
ations
The economies incident to the use of sup-
erheat on shipboard are very great and cannot
long be neglected, although there have been
few applications of superheat to American
ships. The following extract from a letter
from Van Xievelt, Goudriaan & Co.. Rotter-
dam, Holland, illustrates the superheat possi-
bilities in engine-driven ships:
"We are using during the last five years, in our
multitubular boilers .S'j-inch tubes, verj- high tun-
nels and Diamond blowers, and have no trouble at
all in getting sufficient steam. We have practically
no leakage at the connections of the pipes and
Fig. 5. Low Speed Ship Gear and Pinion. Pinion diameter 11.44 in. Pitch 3'-.' in. Normal load 11 SO
pounds per inch face. Tooth speed 1272 feet per minute. Time run, about 400 hours at sea
between the nozzle and bticket of the last
stage of a marine turbine while the turbine
was being operated at normal speed in the
reverse direction. It will be observed that
the high temperature shown by that record
was produced in an extremely high vacuum
by the introduction of small amounts of
steam.
A turbine when operated in the reverse
direction has a friction loss something like
ten times as great as when it operates in a
normal direction. In the General Electric
shops it has been discovered that revers-
ing wheels of marine turbines turn blue
with heat when operated at normal speed in
a vactium of 20 inches. While no definite
information can be given concerning the
possible effects of high superheat in reversals
of a marine turbine, the facts here given
boxes. The original pipes are still in usi-. The
capacity of a 20-ton evaporator is sufficient for
supplying feed water."
"Three of our steamers have been running half
a year without superheaters with a coal constitiiption
of 24 to 25 tons. After fitting superheaters the
consumption was about 22 tons, making a saving
of at least 10 per cent."
In electrically driven ships the gain is
quite as great as is here shown, and no
practical difficult ies can result even from
degrees of superheat wliich would Ik- in\-
practicable with rcci])rocating engines.
Efficiency of Transmission
The selection for conijiarison of a ship of
low ])ower is unfavorable to electric tlrive in
the matter of transmission efficiency, the
conditions being better for this inelho<l in
ships of higher jiower. The gonerat'T de-
ELECTRIC PROPULSION OF MERCHANT SHIPS
65
signed for this case has an efficiency of 9o.6
per cent and the motor 95.9 per cent, making
the transmission efficiency, including cable
loss, etc., 91.6 per cent. In machinery
designed for certain high-power ships, the
efficiency is as high as 94 per cent.
Shaft
Horse-power
R.P.M.
Loss of
Gearing
Efficiency of
Gears
2400
1420
87
77
125 h-p.
80 h-p.
95.0 %
94.7 %
Fig. 6. Experimental Gear Disc and Pinion. Pinion Diameter
7.28 in. Pitch 4 in. Load carried 3000 pounds per inch
face. Tooth speed 7000 feet per minute. Time run
263 hours in Schenectady. Has made 8 times as
many tooth engagements as above with a
pressure which, considering smaller pin-
ion diameter, is relatively four
times as heavy
To determine the efficiency of gear trans-
mission as compared with the figures just
given, very careful tests have been made at
Schenectady. A 2400-horse-power ship tur-
bine was connected through two sets of
double-reduction gearing to a generator, and
the steam consumption was tested at various
degrees of load and speed; then the same
turbine was connected to the same generator
without gearing, and tests were run with the
same conditions and the same degree of steam
flow. All this was done on a testing stand
where conditions are uniform and accurate,
the gears ran with perfect smoothness, and
all conditions were favorable. Since the
comparison gives the loss of two gears, the
differences are considerable and the deter-
mination should be very close to the correct
value. This test showed that the performance
of a single gear is as follows:
In addition to these gear losses, we must also
consider the loss in friction of the reversing
turbine, which is estiinatcd from reliable data
to be 2S horse power, and we must also con-
sider the bearing losses on about 100 feet of
shaft, which in perfect alignment will be 8.5
horse power. These additional losses reduce
the transmission efficiency to 93.5 per cent,
leaving only 1.9 per cent advantage to the
gearing. With the shaft more or less out of
line and the gears operating under sea condi-
tions, it is probable that the losses given
would be greatly increased. Noise is an
indication of loss, and most marine gears are
at times noisy, while the gears in this test
were almost silent. The gears tested in this
case were of the General Electric Alquist
type, and it might be claimed that other
kinds of gears would be more efficient but
it is obvious that, under fixed load and with
similar gear speeds and diameters, there
could be no advantage in any other type
even if it ran with equal smoothness.
Cost
In the present condition of prices, it is very
difficult to compare costs, but the cost esti-
mates of the General Electric Company on
electric equipments for cargo boats and
geared equipments of recent design indicate
that the electric is slightly cheaper. If we
consider savings in shafting support, shaft
alley, oiling system, etc., the saving with
electric drive should be as much as 20 per cent
of the cost of the driving machinery.
Propeller Speeds
111 ships requiring less than 3000 horse
power, there is some practical disadvantage
in using propeller speeds below 100 revolu-
tions per minute because of the large number
of motor poles required if a high-speed turbine
is adopted. Studies recently made by the
Navy Department and elsewhere have indi-
cated that there is practically no disadvantage
in using a propeller speed of 100 revolutions
per minute on an 11-knot, 2400-horse-power
ship, but in all cases of electric drive the mat-
ter of propeller speed should be carefully
studied. In ships of higher power it is not
desirable to use extremely high turbine speeds,
and therefore there can be no difficulty about
propeller speeds. Even in low-power ships,
G6
Tanuar\-, 102(1
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 1
7 S€.C.-^
3SOG S.H.R
ZOOO S.H.R
Fig. 7. Record from Torsion Spring Coupling on S.S. Jebsen in Ballast m a Moderately Rough Sea.
Average r.p.m. 78. Average shaft horse power about 2000. Part of the smaller fluctuations
shown came from an untrue collar in the instrument; otherwise record is correct
lower turbine speeds could be used if expedient
but this is disadvantageous to a small tur-
bine, and the relative advantages and dis-
advantages should be duly considered.
Operating Force
The history of the electrical industry has
repeatedly shown that persons who have not
used electrical apparatus assume that its
operation requires a high order of skill and
expert knowledge, and of this assumption
we have alread>- heard much in connection
with electric drive for shijis. A vast amount
of experience has reix-atedly shown that this
assumption is the direct reverse of the truth,
and a little thought as to the conditions in
electrical apparatus should make the reason
obvious. Conductor circuits are much simpler
mechanically than pipes and mechanical
motions, and electrical machinery is simply
a combination of electric circuits with motion
of rotation. The connections are easily
shown by diagrams, and little mechanical
skill is required to make them. The work of
insulation can be so done that, under such
conditions as exist in ship installations
troubles which might involve difficulty of
repair by unskilled persons are very improb-
able. In all the extensive uses of electricity
in mills, mines, railways, and other indus-
tries, it has seldom failed to become poi)ular
immediately with the operating forces. In
no case has this been more marked than in
the ships which have been tlriven electrically.
Large electrical apparatus is generally simpler
than small, and the machinery used to propel
a ship is in many resi)ects simjiler than that
which is used to light it. Instead of introduc-
ing difficulties to the operating force, the adop-
tion of electric drive will eliminate them and
make ships much less deix-ndent ujjon the skill
and resourcefulness in their crews.
/OOOi
—
' —
V-
I
1
-
I
I
1
1
1
1
1
1
1
'
1
1
1
soo
1Z,2%of full load
steam admitted to
first staQ€ nozzles
r
-
J
eoo
;
\
i
/
/
¥ -rrvi
i TOO
_L
1 eoo
-
TSfoOf full load
steam admitted to
first stage nozzles.
I
\
-
1
f~
^
/
/
,
"«j
-
i
-
/
e
3
J
■ft "'^
1
-r
/
1
«; 300
-1
r '
Tcmoeraturc test of 25O0 HPfives toQ*
ship turbine runninq inrttmrst direction
at fullspKd36O0RPM. mthandwilhout
steam admission Pijrxyrtet^r rmodinq b*-
inq tahen bettveen the nozzle end and
the bucMets of the fifth stao«.OocJtf
/
j
/
200
/
'
1
'
~
too
*
*
;
1
1
0
— t—
1
0
t
0
2
0
T
0
rime - Minutes
F.g 8
67
Improving the Mazda Automobile Headlight Lamp
Bv L, C. Porter
Commercial Em.inkkk, MiNiAiUKii Department, Edison Lamt Works oe the
General Electric Company
Lamps that are intended for use with lens systems or parabolic reflectors must have the light source
located along the axis, tisually at the focal point, and it is readily seen that manufacturing methods must be
adopted which will insure uniform dimensions in lamps of a given type, as only with such lamps can replace-
ments be made quickly and satisfactorily. This article describes the manufacture of Mazda automobile
headlight lamps stage by stage, and indicates that every effort has been made to secure a product as nearly
uniform in dimensions and performance as is possible. — Editor.
Dtirinj.,'' the war the greatest need in the
manufacture of headlights was quantity pro-
duction with miniinum labor and time. Now,
however, the pressing need for minimum
labor is past and more attention can be given
to perfecting the qtialit>% and uniformity of
product is one of the thing.s that is being
given special attention. Since the armistice
was signed an engineer has been appointed to
study the manufacturing methods and sug-
gest changes which will result in more perfect
automobile headlight lamps. He has insti-
tuted a system of gauges and insi^ections
which is proving remarkably effective. It
may be interesting to describe in detail the
present method of manttfacturing a typical
lamp, say the G-S-volt, G-12 bulb, 21-c-p.
Mazda headlight lamp and to point out
recent improvements and call attention to
the inherent variations required for practical
manufacturing in large quantities.
The stem of the lamp is made from straight
glass tubing received in long pieces. This
tubing is gauged for thickness and diameter.
It is then cut up into short pieces of the
proper length for the stem, by jnishing the end
of the tube against a stop and then moving it
to a rapidly revolving emery wheel, see Fig. 1 .
A
Cut t ing Wheel — ►
VI
Tubing
5top'
Fig. 1. Tubing for Lamp Stems Being Cut
to Proper Length by Emery Wheel
The next operation is to put these short
tubes into an automatic flare machine, which
heats one end of the tube by means of a
Bunsen flame and then spreads it out with a
rotating metal plunger (see Fig. 2). Until
recently this flaring was done by hand'
resulting in a considerable variation in the
spread of the flare and consequently in the
length of the stem. With the automatic
Ste-m
Fig. 2. Flaring End of Tubing for Junction with Bulb
after Assembly of Filament
L eadinq
in Wire
Hole
Fig. 3. Method of Inserting
Leading-in Wires
machine, however, the i:)lunger enters a certain
fixed distance and, being of constaiit diameter,
there is less chance for variation in length. In
order, however, to see that everything is work-
ing properly a certain percentage of all stems
68 January, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 1
are gauged for length and the results of these
tests are recorded.
The next operation is the insertion of the
leading-in wires (see Fig. 3) in the stem. The
glass tube is placed in a machine, flare end up,
and the two wires stuck down into holes in
machine makes a constant number of turns,
then allows the coil to pass along, leaving a
short piece of straight wire before making
another coil. The result is a series of con-
nected coils wound on a steel wire, as shown in
Fig. 7.
Uneven
Leads
^4]
Fig, 4. Uneven Lengths of Lead- Fig. 5. Leads of Uniform
ing-in Wires — Often a Result
of Cutting Leading-in Wires
to Exactly the Required
Length
Length Insured by
Trimming After
Sealing In
the rod holding the glass. Occasionally one
of these wires will be slightly bent, or a
particle of dirt will get into one of the holes,
thus preventing the wire from going in its full
depth. After the wires are in place the glass is
rotated and heated by Bunsen flames until
soft and then the glass is pinched together to
make the seal. Formerly the leading-in wires
were cut to just the required length and when
it happened that a wire did not go to the very
bottom of the hole, the result was leads of
uneven length on the mount (see Fig. 4).
Now, however, all leads are made two milli-
meters longer than necessary, and after being
sealed in the stem are trimmed off by a semi-
automatic machine which fixes the distance
from the top of the flare to the ends of the
leading-in wires (sec Fig. 5) and assures leads
of uniform length. This means a considerable
waste of nickel, but the high cost thus entailed
is warranted by the more uniform product
obtained.
The next process is to bend the leading-in
wires preparatory- to welding on the filament.
This used to be done by hand, but is now
accomplished semi-automatically by a metal
plunger, the stem being held against the top
of the flare to keep the overall length con-
stant (see Fig. 6). The stem is now ready for
the filament, which is made as follows:
The lamp under consideration being a G-S-
volt, 2} 2-ampere lamp, it will be gas-filled and
will require a coiled filament. To make this
the tungsten wire is automatically wound
around a fine steel wire. The winding
Fig. 6. Leading-in Wires Being
Shai>ed Ready for Auacb-
mcnt to Filament
Filament Steel Wiru
Fig. 7.
Forming Filaments by Winding
on Fine Steel Wire
These coils are then cut apart with a pair of
pliers, this work being done by hand and there
must necessarily be some slight variation in
the length of straight wire beyond the coil,
though a standard of two milli-meters has been
Weld
S
P
-2 ^7^
Fig. 8. ShoM-ing Shape and
Position of Filament After
Being Electrically Welded
to Leading-in Wires
Fig. 9. Permissible Limits
of Variation in Assem-
bly of Lamps
set. These little coils are next bent by hand
into a V shape, the operator picking out. as
nearly as practical, the center of the coil as the
bending point.
The filament coils arc then put into a hot
acid bath which dissolves out the steel core
IMPROVING THE JMAZDA AUTOMOBILE HEADLIGHT LAMP
69
on which they are wound. The filaments are
then heat treated at 1000 deg. C. in hydrogen
gas, which removes all impurities and leaves
them ready for mounting.
The filament is now laid on the leading-in
wires and is semi-automatically electrically
welded to them (see Fig. S). In this process it
can be seen that there must necessarily be
some slight variation in the length depending
on the angle at which the filament is bent and
the exact position in which it is laid on the
leading-in wires. The variation in overall
length of the m.ount which has been set to
take care of these conditions is plus or
minus one half millimeter measured from
the bottom of the flare to the point of
the filament (see Fig. 9). This is gauged
by a sliding rule (see Fig. 10), the mount
being placed on the slide and brought up
until the point of the filament touches the
upper stop. The mount is now ready for
sealing in the bulb.
The bulb is blown in a mould, but, strange
as it may seem, the diameter of the bulbs will
vary somewhat. In the mounting machine
the bulb is held in a ring and the height of the
bulb will varv slightlv with its diameter (see
Fig. 11).
In order to set the bulb properh- a steel
ball, of the exact size of a correct bulb, is made
with a hole in it which allows the rod that
holds the mount to rise exactlv the right
and the flare of the stem are then melted
together by rotating them in a Bunsen flame.
After this is done i t is necessary to " work "the
glass at the joint a little to prevent cracking.
This used to be done by removing the bulb
and inount while still hot and drawing the
Fig. 11. Showing How Height of Bulb in Mounting
Machine Varies with Bulb Diameter
5tee/ So//
Fig. 10.
Method of Gauging Overall
Assembly Length
Fig. 12. Means Employed for Mounting Filaments
in Center of Bulb
distance to bring the filament center in the
center of the bulb (see Fig. 12). This device
is used in setting the machine. The mount is
then put on the rod and the bulb placed down
over it (see Fig. 13). The bottom of the bulb
latter down a little by hand, thus stretching
the glass. Now, however, it is done by blowing
compressed air in through the exhaust tube and
stretching the joint by expanding it. There
must necessarilv be some variation in length
70 January, 1920
GENERAL ELECTRIC REVIEW
Vol. XXin, Xo. 1
in that process and also the shape of the neck
of the bulb will vary somewhat. The bulb
and the mount are now ready for the base.
The two leading-in wires from the stem are
stuck through the holes in the bottom. These
holes are at 90 deg. from the pins on the base
neck determines to some extent the distance
that the base will go up on the bulb before
striking the glass.
It is easy, therefore, to see that with con-
ditions as pointed out it is practically impos-
sible to keep the light center length of the
Exhaust
Tube
Center
I unction of Bu/b ondM)uni
Fig- 13. Bulb in Position Over Filament and
Mount, Ready for Melting Together
Fig. U, End View of
Lamp Base showing
Position of Leading-
in Wires with Re-
spect to Pin
Fig 15.
Chart for Determining Position of
Light Center
and thus determine the plane of the filament
(see Fig. 14). The base is filled with glue and
set into a heater to harden the glue. In this
process if the wires happen to be slightlx" bent
or twisted, the plane of the filament will ^■ary
somewhat from 90 deg. from the jjins, and
filament (distance between the nearer edge
of the pins on the base and the center of the
filament ) to an absolute figure. The allowable
tolerance in the light center length is .'J '-Vl in.;
i.e., if the center of the filament is not more
than .'> .'52 of an inch above nor more than
Fig. 16. Optical Device for Testing Light Center Length of Lamps
even if the operator twists it around to 90
deg. the wires are liable to cause it to spring
back before the glue hardens. There must,
therefore, be some leeway allowed in this
respect. The slight change in shape of the
'.\ 32 of an inch below the l'4-in. distance
from the top of the ])ins, it is acceptable. The
filament must also lie entirely within "i (14 in.
of the axis of the lamp jiassing through the
center of the base and the tip. A certain
IMPROVING THI-: MAZDA AUT( )Mr)BILI-: HEADLIGHT LAMP
71
percentage of every run of lamps is tested
for these variations.
The device for testing the light center
length consists of an optical projector which
throws an image of the filament on a cali-
brated screen, see Figs. l(j, 17 and IS. From
this it can be easily determined just where the
center of the filament comes. The center of
the filament is taken as the central point of
the triangle formed by the two filament legs
and a line joining the points where they are
four had light center lengths of 1 7 32 in.
(1'4 in. 1 32 in.) with some part of the
filament 3 ()4 in. off of the axis (Square
C) and one lamp had its filament outside
of the light center length Hmit of 3-32
(Square D), etc.
Records are kept of all the tests and inspec-
tions during the entire process of manufacture
and these are ])lotted as curves so that the
engineer in charge can see at a glance just
how the production is running and whether
Light Center Test
-Dote-
r r 5- 1- 5
. 3-
M 32 6a 16 6t
*3Z
A
*A-
B
II
*Te
*^2
1
III
*f.
L ighi Center
1
It
1
/Ixis
/•
1
1
III 1
r
II
nil
iZ
3-
/
'61
r
lb
£•
/
ea
1
-y
1
32 lo
61 iZ 6-1
16 61
Fig. 17. Test for Variation in Light
Center Length
Fig 18.
Test for Variation of Filament
from Axis
welded to the leading-in wires. The testing
device enables the lamp to be rotated !•() deg.
so that the filament image can be inspected
as to its axial positon.
In recording the test of a batch of lamps the
form shown in Fig. 19 is used, which shows
that of 2(5 lamps tested, one had a light
center length of 1 11/32 in. [\\i in. 3/32 in.)
with some part of its filament 1 ()4 in. off
from the axis (Square A); two had light
center lengths of 1 5, 1(3 in. with some part of
the filament 3 t)3 in. off the axis (Square B) :
the various parts of the lamjjs are becoming
more or less uniform.
The lamps which do not come within the
specifications are opened and the defect cor-
rected where possible. In cases where this is
not practical the lamp is destroyed.
New methods of construction and tests are
continually being taken advantage of and
every possible means is used to make the
Mazda automobile lamps, as well as other
tvpes, the most unifom and best miniature
lamps on the market.
il
Januarv, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 1
An Absolute Method for Determining Coefficients
of Diffuse Reflection
By F. A. Bexford
Illlmixatixg Engineering Laboratory, General Electric Company
In this article the author describes an exceptionally accurate method which he originated and developed
to measure the coefficient of reflection of a diffusing surface. In addition to the method being more accurate
than Its predecessors, it is simple, is independent of the color of the standard lamp, and furnishes absolute
not relative, values. — Editor.
Character of Method
The test method herein described has
several features that are perhaps unique;
and, because of the extreme simpHcity of the
photometric work, this method is Ijelieved
to offer possibilities for precision determina-
tions of the coefficient of reflection of diffusin.t;
surfaces. It is an absolute method because
no photometric standards are involved; the
brightness measurements may be made with
uncalibrated lamps and unknown instrument
constants, the only condition being that the
lamps and other accessories maintain their
constancy during the test.
Outline of Method
The brightness of the interior surface of a
spherical integrator depends upon three
factors: (1) the quantity of light received
from the light source, (2) the coefficient of
reflection, and (3) the solid angle of the
spherical surface, if an incomplete sphere.
It is by taking advantage of the last factor
that there are obtained two separate equations
relating brightness, with flux, coefficient, ami
solid angle; and in the solution of these
equations all factors except the two readings
of comparative brightness, the solid angles
corresponding to the two brightness readings,
and the unknown coefficient of reflection are
eliminated.
The test equipment consists of a sphere
that has one or more removable sections,
leaving sections of known solid angle, and
whose interior surt'ace is coated with the
substance under test. A lamp, i^referabh-
with a concentrated filament, and a lens to
project a sharp beam of light through an
opening into the Sjihere, furnishes the light
for the test surface. The surface that is to
serve as a working standard of comparison
may be either the milk glass of a brightness
photometer, or a diffusing surface with
constant illumination if a spectrophotometer
is used. The section of the test surface under
observation must not receive any direct light
from the entering beam.
Derivation of Equation
Let the unknown, but constant. quantit\
of light entering the sphere be indicated \r:
Fa, and let the brightness determination W
Bi when the part-sphere has an area .4,, tli.
total area of the complete sphere being >
If the coefficient of reflection is A', then then
is reflected from the section under direc".
illumination the quantity
F = KFo lumens
on the first reflection.
It is the basic property of the integratim;
sphere that light reflected from any point
on the diffusing surface is uniformly dis-
tributed over the entire surface. If, now, an\
section of the sphere is missing, the amount
of light that escapes is directly proportional
to the area of the opening, or, inversely, the
light received on each reflection is pm-
portional to the surface present. Wc ma\
then write for the light received from the
first reflection
/'"i = ^ KFo lumens ( I
Upon the second reflection there is the useful
flux
Fi = '-^ KfJ -^ ''^ ) = ( V ^^ l^" '""i^^ns ( 2 '
and each succeeding reflection is less by the
which is thus the common ratio
in the convergent infinite scries
factor '-;^ K
■^KF,
•+(>''0''-"+('''0"'
(IV
-\ lumens
the sum of which represents the total usi^fui
light due to the infinite series of reflections.
Calling this total /•". we have
AJ<Fo
/•" = -
(,-^.)
lumens
(4)
If Ai and >" are expressed in square centi-
meters, we ma\- find the illumination in
AN ABSOLUTE METHOD FOR DETERMINING COEFFICIENTS
phots by dividing tlic useful flux by the area
illuminated
J, ■4i/\Fo KFo , ^ ...
t. = -^ \ ^= — :i -. ^ photos (.))
.4iSf 1
and the brightness of the point under obser-
vation is
A'- F
lamberts (6)
~^? ~A \
The photometer reading will be propor-
tional to B, and by using a proportionality
constant A' we may write
1\~ F
XRi = — 7 ~ — ^ lamberts (7)
<.-^'0
Rearranging, we get
NRiS - NR,A i/v = IC-Fo lumens (S )
Selecting another solid area A^, we get another
brightness reading Rn and
NR2S - XR.2A2 = A'-'Fo lumens (9)
Soh-ing for K, we get from (8) and (9)
.YS(Ri - R.) - NK(RiA 1 - R.A,} = 0
,. SiR.-R,)
(10)
Equation (11) is the desired expression for
the coefficient of reflection. If Ri and R2 are
ta,ken with a brightness photometer, then A'
must be defined as the average coefficient for
light from a lamp at T° as modified by the
projection lens. If a spectrophotometer is
s
o.l o.Z 0-3 0-4 o.S 0-& cT at
k
-f
>
-s
\
llr
-c
0
IV
-7
t
kj
-a
>^
-9
'
-^^J
^
■^
^R-' !^
N
V#B
^
^
^
■\
^
^
\
>
y
\:
t
\
\
Y
\^
\
\
\
\
\
\
,= -'-
ao Cs-zt/Kj/s-^ziOfc-/) p^i.
•^LAlfS-Zt^^J-CA^CS-A/ K}] Cent
Fig. 2. Curves of Errors in the Result Due to +l',tj Error in
Determining the Ratio R =-^ for a Coefficient of 0.50
5:
s
c'
V °f °:' °-* V °:
♦ «
7 C.B 0.9 /.
-/
^'.
H^->
jr»
^^
o
S*n
\
f
N
tf?
^
\
\
-4
-J
^
V
■
\
1-
-7
-«
-9
-/o
- £
/
= — —
A
\
-AiK
\
7-0
rc-/;
A,K)
r
ecf,
V7--
Fig. 1. Curves of Errors in the Result Due to +1' t Error in
R"
Determining the Ratio R = ^ for a Coefficient of 0 90
Dividing both numerator and denominator
by 5, we get the spherical areas expressed as
parts of a complete sphere
i?,-A,
A' =
R\-^ — R-i^
num.eric
(11)
used, then A' is independent of the light
source but it is defined as the coefficient for
wave lengths Xi, X2, etc.
Selection of Best Working Conditions
From equation (11) it appears that an>-
two sections Ai and ,4 2 may be used without
regard to their relative size. This would be
true if the photometric quantities R\ and Ri
could be determined without error; but, as
such is not the case, it is proposed to determine
if for a given error in the ratio -j^ there is any
Ri
particular selection of .4i and ^42 that will
give the most favorable working conditions.
Effect of Errors in Photometry
As Ri and R^ are merely proportional to the
brightness, it will be sufficient if we determine
the result of an error in their ratio-^.
Denoting the incorrect ratio bv
and the error bv
«'=4:
E =
K'-K
K
(13)
(14)
74 January. 1920
GENERAL ELECTRIC REVIEW
VoL XXIII. Xo. 1
we get the expression
(S-A,K) (S-A,K) (C-1)
£=--
K[A,(S-A.K)-CA,(S-AiK)] ^^'''*
after using equations (14). (11). (13), and (7)
in the order named.
As this last equation is too complicated to
analyze by inspection, it has been plotted in
Figs. 1 and 2 for various combinations of
.4i and .42 and for A' =0.9 and 7\ =0.o.
In spectrophotometric analysis in particular
it is important to have ver\- bright test
surfaces, and in order to obtain this condition
the test sphere must be small. After making
allowance for an opening to admit the light
and another for observing, the remaining
S
^
^
S O/' 0.f O.J O
4- o^ o.£ ex7 o^ o._ y.c
- f
^^
■*■*>..,
^-
-
>i^-.
i^.
1
"^o
\
fe
\
\
-4
W '
>
^
-.\
1
- 4
\
\
\
-a
1
\
1
\
\
1
1
I
-n
1
-9
■ /o
^ IOO/^,(S-/)2X)(C-/)
— f' P£l?C£fJT
—
A:
s
Fig. 3. Curves of Errors in Result Due to -f-l'i Error in
Measuring A\: Full Lines of Coefficient of 0 90,
Dotted Lines for Coefficient of O.SO
surface cannot greath- exceed 9.') per cent of
the complete sphere. With this value for .4i
(in the remainder of this article .4i and .4o are
.4i 4-.
used in place of '-^ and '-^) and /\ =0.9, the
choice of .4; is seen from the curves to lie
between zero and O.N 5. In this entire region
a photometric error of 1 per cent in the ratio
of R^ to Ki will give an error in the result of
less than three-tenths of one per cent. It is
under these conditions that the method may
properly be called a precision method.
Taking .4i =0.755, the best value for .4:
lies between zero and ()..')05 and the error
rises to 0..") and O.N i)er cent. If .4i is 0..j.}5,
the best value of .4; will give an error of
more than one per cent in the final result.
From these data it is evident that the largest
possible value of .4i gives the most favorable
working conditions.
The selection of .42 will in practice probably
always be near the upper limJts mentioned.
If too small a section .42 is used, the results
are more liable to be affected by stray light
or b}- light from the sphere being reflected
back from the surroundings.
In Fig. 2 the error equation is plotted for
the same values of .4i and .42 as before, but a
coefficient of 0..5 is assumed for the test
surface. A comparison of the two sets of
cur\'es for A" = 0.9 and A' = 0.o shows that the
latter gives very inferior accuracy, which
means that the method is best suited for
Ai
S
/oo^r/-^//fUc-/J
* B
^lO-AzXJ-CAz (l-A/K)
1
1
* ^
/
;
; ;
/
;
/
/
1
(
/
\
1 f
//
k 1 k
' A^
■1
'
1 h 1
^b / 1
i-
lb
/ ^
'M / / y J
^
.^i^S::^<-^ll^-r^ %
J £ o J O •#
s
0.7 o.B o 9
Fig. 4. Curves of Errors in Result Due to *1' . Error in
Measuring A-\ Full Lines for Coefficient of 0.90.
Dotted Lines for Coefficient of O.SO
surfaces having high coefficients, anil if the
coefficient is much below 0..) the method may
fail through multiplication of the usual pho-
tometric inaccuracies.
Effect of Errors in A\ and A:
In general, the sphere can be built u|} of
sections whose areas are known with a high
degree of accuracy. The high accuracy
obtainable when only the photometric work
is in error is due to the fact that R\ and R^
have symmetrical positions in the numerator
and denominator of equation (11). both of
which are increased or both decreased, thus
minimizing the effect u])on the result. With
.4i and .42 the case is different, and in Fig. .'{ is
shown the result t)f using a value for .4i
AN ABSOLUTE ArP/niOl) FOR DirPERMINIKC COEFFICIENTS
10
one per cent high. The equation for determin-
ing the error is given. From the solid cvirves
(for /v'=U.9) it is evident that the angular for
surface) measurements .should be made with
great care, because the resulting error in the
final result is always larger. The results
when testing a surface having a coefficient of
(l..iO are less accurate, but for low values of
.42 the difference is not great.
An error of one per cent in .42 has much
less effect upon the results than the same
error in Ai. The error in K actually
approaches zero as A« decreases, and here is
found the best reason for selecting .42 as
small as possible. There is a practical limit
fixed by the necessity of having one section
of the surface for receiving all the direct
light from the lamp and another surface for
observing, i.e., illuminated only by reflected
light. These two surfaces are not necessarily
continuous or joined. A ]jraetical consider-
ation that bears on the selection of .42 is the
return of light escaping from the part-sphere.
This return light increases at a greater rate
than the openings 5 — .4i and 5 — ^42; and
unless the conditions are highly favorable for
quenching all the reflected light, the error
(iue to this cause may overbalance any
supi:>oscd advantages of a small value for .42.
Another lower limit is fixed by the difficult}'
of measuring a small section as accuratch'
as a large section.
In general, it would seem that the rules for
])ractice arc:
(1) Make .4i as large (in jiroportion to 5)
as i)ossiblc.
(2) Make .42 one half to three fourths as
large as ,42.
(.3) Quench all stray light from the i.iart-
sphere.
(4) Use the method \\-ith caution when K
is less than ()..i().
^
^^J^^J
±.
^^^^
^""Wl
^^'^
«MH^nMi'>BI
W^y^im
Fig. 6. Photograph of the Testing Set-up L _ .
the Coefficient of Reflection of Magnesium Carbonate,
Fio =; hv the Part-snhere Me
up Used in Determining
. ui x^ciici- iivjii ^i .Yiagnesium I..
Fig 5, by the Part-sphere Method
Coefficients of Magnesium Carbonate
By the m.ethod just described, a test was
made of the reflection coefficient of mag-
nesium carbonate. The material was obtained
at a dmg store and the package bore the
inscription " Silk -finished Magnesia Carbon-
ate for technical use." The blocks were
/oo
0.93
§ 0 97
i
.V
o
—
—
'^o
o
"
^
o ^
/
K
U o.S/
0.90
■
o«o o-V a** a*^ a*e aso oSK oS* aS» OSB aso oti oA* o.s*oM oto
\-twjr\- Oi i>£ •(• ifKSe/^ — ^ r [. — o«m/mr? — ^ fo 1
Fig 5. Spectrophotometric Test of Coefficient of Reflection of Magnesium Carbonate
by Part-sphere Method
76 January, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 1
2 by 2 by 2J^ inches and nothing was known
about their chemical purity. A 4.5-deg. bevel ■
was cut around one end of each block, and
the end was then turned out to a concaA"e
spherical surface with a radius of 1.125 inches.
When five of these blocks were assembled as
shown in the photograph of the test equip-
ment, Fig. 6, they formed five sixths of a
sphere, and the removal of one block left
two thirds of a sphere. In order to have
definite edges, the bevels and sides of the
block were blackened with draughting ink,
which was found to stay on the surface and
not penetrate. Owing to the extreme fragility
of the carbonate, it was found impossible to
get perfectly sharp edges all around, and there
was a little blackened area visible at all
points. From measurements on the cracks,
it seems that the assumed areas .4i and .4>
are about one per cent too high. The true
values for K thus would seem to be slighth-
greater than given in Fig. o, if no account is
taken of the effect of stray light. A few
readings with three and two blocks showed a
progressive decrease of several per cent for
the lower values of A<; and as it is believed
that the stray light just about compensates
for the errors in .4i and .-is mentioned above,
the results are given as found, and they are
probably within one per cent of correct at
all points except possibly at 0.43^. where the
intensity was near the lower reading limit.
The lamp and focusing lens for the part-
sphere are shown near the top of the photo-
graph, and the comparison lamp is in the
metal sphere at the extreme right-hand side.
During the test the part-sphere was sur-
rounded by screens and the white sides of
the blocks were covered. The color com-
positions of the two beams entering the
collimators were quite different, the readings
at different parts of the spectrum ranging
from 50.3 to 73.9; and there was considerable
unevenness due to the lens and the paint
and diffusing glass of the metal sphere.
These variations of course canceled out as
they affected Ri and /?•> alike. Ten readings
were taken at each point, and the apparent
high photometric accuracy shown by the
agreement of the points with a smooth curve
is in line with the data of Fig. 1.
Regarding the high coefficients, it can onh'
be said that the figure A' = O.SS published a
number of years ago is obviously too low, as
it has been found possible to get this value
with a reflcctometer. which is known tu
read low. Correcting the reading of this
instrument for the equivalent spherical area
of the nickel band, and using for the nickel
its test coefficient of 0.53, the coefficient
of magnesium carbonate is 0.970 which is
evidence in favor of the accuracy of the
data here given.
A large number of experimenters have at
various times used magnesium carbonate as
a standard of reflection, and exact knowledge-
of its coefficient is of considerable importance
in several types of photometry-.
TWO DOLLARS PER YEAR
TWENTY CENTS PER COPY
GENERAL ELECTFIC
REVIEW
VOL. XXin, No. 2
Published by
General Electric Company's Publication Bureau.
Schenectady. N. Y.
FEBRUARY, 1920
il
Large modern alternators must be driven by waterwheels or steam turbines because only these types of prime
mover are of sufficient caliber for the purpose. This illustration shows one 20,000-kv-a. and two 16,300-kv-a. water-
wheel-driven alternating-current generators, and our frontispiece shows a 45,000-kw. steam turbine-driven generator
A Group of Articles on
SYNCHRONOUS GENERATORS AND MOTORS
( PATENTED)
It is not alone mechanical friction within a
machine itself, that "NORmfl" Hearings minimi/c.
Their trouble-free performance also minimizes
friction between machine buyer and machine
builder. A "NORfflfl" equii^ped machine is a bet-
ter, more serviceable machine. Its builder may
justly claim it to be so, because its performance
will surely prove it to be so.
See that your Motors
are "NORmfl Equipped
IRE m^mm/^ CPil^F/^IMY
W^'W Y®ii°lk
Ball. Roller. Thrust and Combination Bearings
General Electric REVIE^A/■
Manager, M. P. RICE
A MONTHLY MAGAZINE FOR ENGINEERS
Associate Editors, B. M. EOPF and E. C. SAXDERS
Editor, JOHN R. HEWETT
In Charge of Advertising, B. M. EOFF
Subscription Rales: United States and Mexico, $2.00 per year; Canada, $2.25 per year; Foreign, $2 50 per year; payable in
advance. Library and Student Rates: United States and Mexico, $1.50 per year; Canada, $1.75 per year; Foreign, $2.00 per year;
payable in advance.
Remit by post-office or express money orders, bank checks, or drafts, made payable to the General Electric Review, Sche-
nectady, N. Y.
Entered as second-class matter, March 26, 1912, at the post office at Schenectady, N. Y.. under the Act of March, 1879.
Vol. XXIII, No
Copyright. ISiO
hy General Electric Company
Febriarv, 192U
CONTENTS
Frontispiece: A Modern Alternating-current Generator
Editorial: The Status of the Synchonous Generator
Early Days in Alternator Design
By W. J. Foster
Investigation of Water-air Radiators for Cooling Generators and Motors
By H. G. Reist and E. H. FREiBrRCHorsE
Steam Turbine Generator Ventilation
By Geo. !Moxsok
Mechanical Design of Large Turbo-generators
By M. A. Savage
The Behavior of Alternating-current Generators When Charging a Transmission Line
By W. 0. Morse
Synchronous Motors
By W. T. Berkshire
Magnetomotive-force Diagram of the Synchronous Motor
By E. S. Henningsex
Oscillating Frequency of Two Dissimilar Synchronous Machines
By R. E. Doherty
Some Mechanical Features of Svnchronous Machines
By A. P. Wood
Parallel Operation and Synchronizing of Frequency Converters
By O. E. Shirley
Motor-generator Sets
By G. H. Tappax
Svnchronous Condensers
By E. B. Plenge
Large Horizontal Alternating-current Waterwheel-driven Generators and Synchronous
Condensers
By M. C. Olson
Measurement of Losses and Efficiencv bv Temperature Rise of
By Wm. F. Dawson
Bearings and Lubrication for Vertical Shaft Alternators .
By T. W. Gordon
A LTnique Design of Waterwheel-drivcn Alternator .
By A. E. Glass
Belted Alternating-current Generators
By A. L. Hadley
Sine Wave Testing Sets
By E. J. BuRXHA.M
111 Mcmoriam: Timothv S. Eden
Ventilating Air
Page
79
SO
91
99
105
109
112
122
125
130
136
140
143
147
153
102
1(30
171
177
ISl
z ;
< ■;
a. =
Z c
o -
* £■ ■
8 t •
5
o
z I
< -
i- >,
K -
o -
< "
a; ■=
z -
^O
^ c
K H ,
E ; ■=.
2 =
U c:
* ;:
^ :
General Electric Review
THE STATUS OF THE SYNCHRONOUS GENERATOR
Alternating-current generation and trans-
mission are the cornerstone of the electrical
industry. Only in a few instances where the
energy is used directly on the spot, as in the
manufacture of aluminum, is direct current gen-
erated on a large scale. That this would ever be
the case, however, was directly contrary to the
belief of most electrical men in the early 'SOs.
Following on the heels of the invention of
the incandescent lamp by Mr. Edison, in-
ventors and manufacturers busied themselves
with the development and sale of direct-
current dynamos for operating the lamp.
Although Professor Elihu Thomson had
succeeded in building some of the most
satisfactory direct-current generators then on
the market, he foresaw the wonderful oppor-
tunities offered by alternating current for long
distance transmission and set to work to
improve the alternating current system and
overcome the strong prejudice which existed
against it. His inventions made possible the
safe distribution of low voltage alternating
current, and once this fact had been demon-
strated, the development of alternating-current
apparatus progressed with astonishing rapid-
ity— the inherent superiority of alternating
current as a medium for moderate and long
distance transmission had been established.
Mr. W. J. Foster, in his article in this issue,
traces the progress in alternator design from
the early machines of Elihu Thomson to the
inception of the polyphase revolving field
machines of today.
Often in the development of alternating
current generators, as with most engineering
development, there has appeared, sometimes
only a step ahead of current achievement, a
formidable barrier to further progress — some
limiting factor that seemed to block the way.
In the early days it was a question of volt-
age and insulation, then of prime movers of
adequate capacity. As the size of machines
increased limits were obviously being ap-
proached in the definite pole revolving field
construction because of centrifugal stresses:
and almost simultaneously the problem, of
adequate cooling and ventilation thnist itself
into consideration. Each of these limits has
been pushed farther and farther along the
road of progress as it threatened to become a
serious factor; new insulations were devised,
the steam turbine replaced the reciprocating
engine, salient pole construction surrendered
to the smooth core rotor, and force draft suc-
ceeded to natural ventilation.
In the steam turbine generator we have
arrived at the 45,000 kilowatt unit and have
under consideration still larger machines.
On the electric end at least we do not appear
to be confronted with any new limitations —
we are merely overtaking some of our old
problems in new guise. Centrifugal stresses
in the rotor were for a time brought within
safe limits by the elimination of salient poles;
but with the great axial length and massive
field coils that are required in the largest
machines, these stresses are again approaching
the limit of safety. Also, it is becoming
exceedingly difficult to dissipate the heat that
is generated in these solid rotors — air veloc-
ities through the air gap are now in the
neighborhood of 12,000 feet per minute. In
the stator, likewise, ventilation is becoming
a serious factor; also, it is becoming increas-
ingly difficult to support the end windings so
that they will resist the tremendous forces
that act to wreck them under short circuit.
In waterwheel-driven alternators the re-
quirement of being able to withstand double
speed introduces difficulties in the design of
the revolving element, as the larger diam.eters
have made it impracticable to adopt the
sm.ooth core rotor for these machines.
The largest waterwheel generators have
been of the vertical type, and one of the
greatest problems has been to provide a
satisfactory thrust bearing to support the
entire revolving element, including the water
thrust. Considerable trouble from scored and
burned beai'ing surfaces was experienced with
the older types of thrust Ijcaring, and it was
obvious that the maximum allowable weight
for this design had been reached. The spring
thrust bearing and an efficient oiling system,
however, have shifted the limiting factors in
the size of waterwheel generators to otjier ele-
ments in the construction.
Closely related to the synchronous generator
are the synchronous motor, synchronous con-
denser and frequency converter, and the sig-
nificant position of this class of apparatus
with respect to the future expansion of elec-
tric power generation and utilization empha-
sizes the value of the special series of articles
appearing in this issue. B. M. E.
so February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 2
Early Days in Alternator Design
By W. J. Foster
Alternatinc.-current Exgineerixg Department, General Electric Company
The author takes us back to the days when some of the first experimental work on alternating-current
generators was being performed by the predecessors of the General Electric Company. Professor Elihu
Thomson's inventions and his appreciation of the advantages of alternating current were largely instrumental
in overcoming the early prejudices against the commercial use of alternating current, and some of his work is
illustrated and described in this article. The evolution of the alternator is traced from Professor Thomson's
early experimental machines, through the first commercial machines of the Thomson-Houston Companv and
the mono--yclic generators of Steinmetz, to the early polyphase generators built between the years 1894 and
1900. During this period the first turbo alternator was built and tested. Some remarks in explanation of
the great range of frequencies that were to be found in the early days are of interest. — Editor.
Ti
'HE Alternating
Current Engi-
neering Department
of the General Elec-
tric Company, as now
constituted, may be
said to date from
January-, 1S94. Hence,
the quarter-centurj'
mark has been passed.
For the origin of the
Department we must
go back to the con-
stituent companies
that merged into the
General Electric Company in 1S92; in fact,
it is necessary to go even further back, to the
time when there were no organized com-
panies manufacturing electrical machinery.
The real origin is to be found in the work of
Prof. Elihu Thomson; first, in the Phila-
delphia High School; second, in the American
W. J. Foster
Fig. 1. Sketch from Drawings of a Patent Application Prepared
January 13, 1879, showing Induction Coils in Parallel as
Arranged and Tested Out by Elihu Thomson
at Franklin Institute
Electric Co., New Britain, Conn.; and third,
in the Thomson-Houston Electric Co., Lynn.
The fundamental patents were taken out
about fort\- years ago by Prof. Thomson.
These have been added to from time to time
by him and by others connected with the
Department.
The alternator had no commercial value
until means for delivering the current, and
apparatus for applying it to some useful
purpose, had been devised. Naturally, light-
ing was the first use made of the alternating
or ' ' reversed ' ' current . Some idea of the part
taken by Prof. Thornson and his associates
in this work may be obtained from Figs. 1, 2
and 3.
Fig. 1 is reproduced from Fig. 5 in the
drawings of a patent application by Elihu
Thomson, prepared January 13, 1S79, contain-
ing the following description of the illustra-
tion;
►±er/Ki^
#VMt. &#U* t*t^ L^-JU~a^t^A.s C. ,«/^»;7t-..«^L
Fig. 2. Taken from Personal Notes of Elihu Thomson
Fig. 5 shows the method of employing a vibrating
lamp where a single undulatory or reversed current
is employed to operate a number of lamps, the main
circuit remaining unbroken. // H II is the main
circuit from the machine furnisliing reversed cur-
rents, it being branched at 1, 1,2, 2, S, S through the
EARLY DAYS IN ALTERNATOR DESIGN
81
primary wire of induction coils
N, N', N". The secondary
wire of each induction coil has
its terminals connected to the
terminals of the lamps L, L', L" .
The variations or reversals in
the primary circuit H H H
cause variations or reversals in
the secondary circuits, of which
the lamps L, L', L" form a
part. Since the secondary cur-
rents traverse the coils C of the
lamps, vibrations are thereby
imparted to the electrode E as
before."
Fig. .3 shows sketches
and explanations of the
sketches taken from the
note-book of Prof. Thom-
son's assistant, Mr. E. W.
Rice, Jr., now President of
the General Electric Com,-
pany.
Fig. 2 is reproduced from
Prof. Thomson's own notes
of a later date.
Fortunately, some of the
first machines built by Elihu
Thomson are still in exist-
ence and have been photo-
graphed for this article.
Figs. 4 and 5 show the
dynamo built in 1S78 and
operated in theFranklin In-
stitute, Philadelphia, before
the days of the incandes-
cent lamp, or of am' of
the many applications of
electricity to the useful arts
as we know them today.
This particular machine.
C-ffH-^^A*^^
Figs. 4 and 5.
ui'^^aA^-^ J)'^'^- '^'^
AXO 'OO^CaZ^.C^ p/v^^u.^^t»-^, iJi-^tity^V- 0^a*^'^«^ttu/3 .
Fig. 3. Taken from Notes Made in January, 1884. by Elihu Thomson's
Assistant, E. W. Rice, Jr.
Alternating-current Dynamo Built by Elihu Thomson i
Front and Back Views
as shown in the illustrations, had a
re^'olving armature and windings
connected to a commutator for fur-
nishing direct current, as well as wind-
ings connected to slip rings for alter-
nating current. The immediate use
that was made of this alternator was
the operation of an arc light. Fig. 6
shows two of the transformers in-
vented by Prof. Thomson in order to
make use of the alternating current,
and Fig. 7 one of the lamps for fur-
nishing light at the Franklin Insti-
tute.
During the next few years Prof.
Thomson's energies were occupied
largely in the development of direct-
current dynamos, arc machines, arc
lamps, etc. During this tim.e Mr.
Edison developed the carbon-filament
82 Februarv, 1920
GEXER.\L ELECTRIC REVIEW
Vol. XXIII. Xo. 2
incandescent lamp. Other inventors and
manufacturers developed and put on the mar-
ket direct-current dynamos of suitable poten-
tial to operate the lamps. The operation of
incandescent lamps from the direct-current
dynamos proved to be a success. When it
Fig. 6. Transformers Made in 1878
was proposed to make use of altematinjj
current with stepdown transformers, in order
to distribute over a wider range, objection was
made on the ground of danger to life. This
was the situation when Prof. Thomson again
attacked the problem of distribution at
higher potentials. His efforts resulted in
inventions, patented about 1SS5, that made
the distribution perfectly safe. He then took
up actively the design of commercial alter-
nators. Fig. S shows his first alternator,
built in Lynn, in ISS.^. It was a revolving-
field separately-excited machine, and was
used for commercial purjjoscs in the shops of
the Company. Fig. 9 shows an early type
of transformer having a ring core, the inven-
tion of which made possible the adaptation
of the alternator for lighting one of the
Company's factories.
It is interesting to note that of the two
constituent Companies that merged into the
General Electric, the Thomson-Houston de-
veloped alternating-current m.achines and
apparatus, as well as direct-current, while the
Edison General Electric Company confined
its operation to direct current alone. Dia-
metrically opposite policies prevailed in these
two Companies in their attitude toward
alternating current. The Edison, with its
three-wire system of distribution for lighting,
took the position that such a system was the
best possible. Moreover, this Comjiany
opposed the use of the alternating current on
the ground of the danger to life and limb.
In the early days it was a sort of superstition
that alternating current was far more danger-
ous at any given voltage than direct current.
The adoption of alternating current by some
of the states for inflicting capital punishment
added greatly to the pop-
ular prejudice against it.
It is hard for us to realize
at the present time how
deep seated was that
prejudice. This com-
mercial warfare con-
tinued for some time
after the inventions of
Prof. Thomson had made
the alternating-current
low tension perfectly
safe from the possibility
of contact with the high
tension of the primary
distribution.
The first alternators
sold by the Thomson-
Houston Company and
installed for furnishing lights to customers
were built early in 1SS7. Two of them
were tested and shipped in May of that
year; one to the Lynn Electric Light Com-
pany, and the other to Xow Rcchelle,
Fig. 7. Vibrnting Lamp Used in 1879
X. V. The latter was first in ojicration,
having been started up by one of the engi-
neers from the factory. Mr. A. L. Rohrer.
The only photograph we have of it (Fig.
10) was taken by Mr. Rohrer: there was
no Photographic Department in those days.
EARLY DAYS IN ALTERNATOR DESKiN
83
Fig. 8.
First Alternator Built by Elihu Thomson
in 1885
Fig. 9. Transformer Used to Light Factory
at Lynn in 1886
The picture was taken in the basement of
Factory B of the old Thomson-Houston
Works, West Lynn. The alternator stands
on a hand-truck at the left. It was of the
"revolving armature" type, was single-phase
of about 900 volts, had 0 poles occupying
a horizontal position, and ran at a speed of
1250 r.p.m. It was self-excited, the excitation
being furnished by a separate winding con-
nected to a six-part commutator on the same
Fig. 10. View in Basement, Factory B. Thomson-Houston Works. Lynn, in May. 1887.
showing the First Alternator Built for a Customer
84 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 2
shaft. Its capacity was 200 16-c-p. lamps
and it was named the " A2." Several of them
were built and sold.
A larger alternator, the " A4," with capacity
for 400 16-c-p. lamps, was the next pro-
duction of the Thomson-Houston Company.
co/e.
Fig 11. Compensating Coil of the A6B Alternator of
Thomson-Houston Co. Four-wire Distribution at
330 Volts Without Stepdown Transformers
This alternator had radial poles, eight in
number, and ran at 1500 r.p.m. The next
in order was a 10-pole, 1500-r.p.m. machine,
known as the "A6" and capable of lighting
600 lamps. There was a modification of this
alternator known as the "A6B," or the
"Compensated." The term "compensated"
in this case did not have the significance of
maintaining constant potential for changes
in load, but obtained its significance from the
feature of distribiition. The system of distribu-
tion was a four-wire circuit with lamps con-
nected three in series between the several
distributing wires, thus minimizing the drop
of potential on the intermediate wires. Inas-
much as 110-volt lamps were used, the alter-
nator was wound for .'5;5() volts as shown in
Fig. 11.
A larger alternator of the same periodicity
as the "A6," viz., 125 cycles, was soon
developed and was known as the "A12," or
"1200 lighter."
The next development consisted of a line of
machines with rectifying commutators, but of
the same general type and construction, in the
naming of which a different significance was
given to the numeral. These were the " AlS,"
"A35," "A70" and "Al(i5," in which the
numeral signified the kilowatts. About 1N!)1
this line was changed from smooth-core
armatures to toothed armatures. Modifi-
cations were made in the shaft and bearings,
thereby increasing the ratings to "A25,"
"A50," "AlOO" and "A240," respectively.
These machines were still bought with
reference to the number of lamps that could
be carried; the "A25" for 500 lamps; the
"A50," 1000 lamps, etc. When it was
definitely settled that a 16-c-p. carbon-fila-
ment incandescent lamp required over 50
watts and that there were transformer and
line losses to be allocated to the lamps, the
ratings of the alternators were raised 20 per
cent; hence the line became "A30," "AGO,"
•■A120" and "A300."
The first three sizes were in great demand
and gave good satisfaction. They had two
windings on the fields; one excited from an
outside source and the other, or the composite
field winding, excited from the rectifying
commutator, a two-part affair standing at
the middle point of the armature winding.
Fig. 12 is a photograph of one of these alter-
nators.
Fig 12. A.35 Alternator of Thomson-Houston Co
During the years 1SS7 antl is.ss Prof.
Thomson built the first experimental induc-
tion motors, including a single-jihasc motor
having a commutator. The first induction
motors built b\- the Company, with reference
to conuncrcial use, were brought to test early
EARLY DAYS IN ALTERNATOR DESIGN
85
in 1S92. As a result of the tests it was
decided to proceed with the development of
motors of two or three sizes. The necessity
then sprang up for polyphase generators, or
some system of operating polyphase motors
from single-phase circuits. The monocyclic
generator was conceived by Mr. Charles P.
Steinmctz, as something that would supply
the want.
Monocyclic Generator
The monocyclic generator in its conception
was a dynamo from which could be obtained
single-flow energy and polyphase potentials.
Fig. 1-t is a cut of one of the commercial
machines as built during the '90s, with
diagrams of the armature and field windings
Fig 14. A Standard Revolving Armature
Monocyclic Generator
shown in Figs. 15 and 1(3. All of the earlier
types and sizes, either belted or direct con-
nected, had revolving armatures with rectify-
ing commutators and compounding windings
on the fields. It is interesting to note that the
first field coils of copper strip wound on edge
were made for one of these monocyclic re-
volving armature generators that had SO
poles and was driven by a Corliss engine at
90 r.p.m.
Many revolving field monocyclic generators
were designed and built in the years 1897 to
1900, in both belt-driven and direct-connected
units. At least ten or twelve sizes, ranging
from 50-kw. at 300 r.p.m. to 1500-kw. at 90
r.p.m., were developed for direct connection
to steam engines. These generators were
entirely separately excited.
Polyphase Generators
The demand for generators for power pur-
poses increased so rapidly during the period
1894 to 1900 that numerous three-phase and
Diagram of Armature Connections of Revolving
Armature Monocyclic Generator
St-at-ionao' Shunt.
Or Compounding Rheostat
^-Collecrt-OT- PJir-igs
Connmut_al,or
Fig. 16. Field Windings of Monocyclic Generator
two-phase generators and motors were devel-
oped. As a result, lighting circuits were run
more and more from polyphase circuits.
At the same time, the rating that could be
86 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 2
given to a machine was much higher if it were
polyphase than monocyclic. As a conse-
quence, the manufacture of monocyclic
machines was eventuallv abandoned.
Or Compounding Rheostat
43 5 261
Fig. 17. Connections of Three-phase. Revolving
Armature Compound Generator
During this same period it became apparent
that revolving-field alternating-current
machines had decided advantages over those
of the revolving armature type, as higher
voltages were possible and also greater capac-
ity. The characteristics of these machines
in the matter of inherent voltage regulation
were of great importance. Generators were
being used for both power and lighting pur-
poses. Power applications were chiefly
through induction m_otors; consequently, the
power-factor of the generator was too low to
permit of m.aintaining a steady potential on
lighting circuits, no matter how alert was the
operator at the switchboard. The customer's
specifications regularly called for 6 or S per
cent inherent regulation. By this was meant
that if full non-inductive load were thrown off
the generator, the potential on open-circuit
would not be m.orc than 6 or S per cent
greater than under load. The situation was
relieved by the development of field regulators
that controlled the excitation, m.aintaining
constant voltage by regulating the field of the
exciter.
Compensated Polyphase Generators
Nearly all the first revolving-armature
three-phase generators had two field wind-
ings, one excited from an external source,
usually a small direct-current generator
driven by belt from a pulley on the shaft of the
alternator. The other field winding received
its excitation from a rectifying commutator,
the excitation varying directly as the current
output. The connections arc shown in Fig. 17.
The commutator was a three-part affair,
consisting of three castings each having one
segment for every pair of poles. The com-
mutator was located mechanicallv immcdi-
Fig. 18. Three-phase, Revolving Field Compensated Generator with
Exciter Mounted in End Shield of Stator
EARLY DAYS IN ALTERNATOR DESIGN
87
atcly adjacent to the three collector rings on
the shaft inside the bearing. It was located
electrically in the "Y" of the winding, the
continuity of the winding being established
outside, or through the field coils and the
stationary shunt. There was also a closed
connection between the inner ends of the
three windings through the rotating shunt,
which was located mechanically inside the
armature spider. This shunt had non-induc-
tive windings as it carried alternating current,
a small percentage of the total current.
Compensated Generators
For the purpose of compensating for volt-
age drop, due to load, in revolving field
generators, exciters were developed with taps
with geared exciters, as shown in P"'ig. 19.
The connections shown in Fig. 20 are very
complicated. This fact, together with the
skill required in properly adjusting the posi-
tion of the magnet frame of the exciter, fre-
quently resulted in dissatisfaction. These
generators operated well, provided the at-
tendants were well informed and careful
in handling the machines. The application
of this method of compensating for po-
tential drop was applied to single-phase
and two-phase as well as three-phase gen-
erators.
Turbo-generators
The first work in steam turbine driven
generators was done in 1S9G, several years
Fig. 19. Three-phase Compensated Generator with Geared Exciter
brought out from equidistant points in the
armature winding to collector rings, similar
to those in synchronous converters. The
voltage generated by the exciter was varied
in accordance with the load by use of series
transformers feeding into the windings of the
exciter through its collector rings. The
compensation for different power-factors was
regulated by adjusting the mechanical posi-
tion of the exciter, and the amount of com-
pounding for any given power-factor by nieans
of taps on the series transformer. Fig. IS shows
one of the belt-driven three-phase generators
of which a complete line was developed and
built for a period of several years.
There was also developed a line of low-
speed engine-driven three-phase generators
before the original 500-kw. Curtis turbine
unit was developed.
The turbo-generator built in 1S96 was a
single-phase 50-cycle 500-kw. 2500-volt 3000-
r.p.m. horizontal inductor-type alternator.
It had a single armature with totally enclosed
slots and armature coils of the pancake type
grouped and connected up to make four poles,
as shown in Fig. 21. Excitation coils were
mounted on both ends of the armature or
stator. The magnetic circuits closed through
shells at the two ends into the shaft and
thence back into the rotor proper, which had
two polar projections diametrically opposite,
as shown in Fig. 22. This generator was
brought to test in October, 1896, and pro-
claimed its periodicity by emitting acoustic
88 February, 1920
GENERAL ELECTRIC REVIEW
Vol, XXIII, Xo. 2
waves of great intensity that illustrated in a
startling manner the nodes and antinodes of
the text books. With injunctions staring us
in the face, we made haste to change the body
of the rotor into cylindrical shape by the
addition of brass filling pieces so shaped as to
produce a true cylinder. It was gratifying
to have the electrical tests prove satisfactory'.
The first commercial Curtis turbine unit
was designed, built, and tested in 1901. It
was installed and operated for several years in
the Power House of the Schenectady Works.
The generator was a three-phase 40-cycle 500-
kw. 1200-r.p.m. horizontal-shaft machine with
a salient-pole rotor. A 1500-kw. 60-cycle
type of rotor was designed in order to have
form-wound field coils that could be easily
assembled or replaced, and with the expecta-
tion that the cost would be less than for the
radial-slot type where the coils had to be
assembled turn by ttun with much of the
insulation applied in place. It was soon found
that machines with such rotors could not be
rated as high as those of the same diameter
with radial slots of the increased number
possible and, furthermore, the potential wave
was inferior. Hence, the cylindrical rotor
with radial slots, as built today, soon super-
seded in all designs the parallel-slot type of
rotor.
Collector Rin^S^ B ♦
Commut.at.or
Exciter Armature ^evolv.n^ Field
Corinection Board -^
ITlain Terminals
Fig. 20. Connections of Three-phase, Revolving Field Compensated Generator
horizontal-shaft generator of the same type
was next developed. Then followed vertical
units, the first of them with salient pole rotors,
which were eventuallv built in sizes up to
7500 kw.
Cylindrical rotors with distributed field
windings were first designed in 1903 and 1904.
The two-pole machines had radial slots;
the first of the four and six-pole rotors
were made with two field coils per pole
assembled in parallel slots. This construction
resulted in the inner of the two coils having a
greater depth and, consequently, a greater
number of turns than the outer coil, since the
two coils bottomed on a chordal lino. This
Periodicity
Neither voltage nor periodicity was
seriously considered in the earliest days; in
fact, there were no instruments for measuring
voltage when the first experimental alter-
nators were built. Even at the time the first
commercial machines were brought to test,
the potential was measured by the incandes-
cent lamp and not by a meter.
It probably was a matter of chance that a
periodicity of 125 cycles was established for
the first line of alternators built by the
Thomson-Houston Electric Company. The
first alternators shipped freim the Works were
6-pole 125l)-r.p.m. machines — hence. (')2',>
EARLY DAYS IN ALTERNATOR DESIGN
S9
cycles. From design considerations, the
number of poles was increased to eight and the
speed to 1500 on next larger size — hence, 100
cycles — then to ten poles at same speed for
still larger sizes — hence, 125 cycles. This
periodicity gave way to lower periodicities
when other use than lighting became general
for alternators, and when it became desirable
to direct connect alternators to prime movers.
It is interesting to note that the first com-
mercial induction motors were 50-cvcle
machines, this periodicity having been decided
upon as a good one after the first tests. The
polyphase machines, both generators and
motors, that were first installed, viz., those
in Southern California and in the state of
New Hampshire, were 50-cycle machines.
However, other manufacturers were develop-
ing 60-cycle machines and that soon became
the prevailing frequency except for railway
work, involving rotary converters, where 25 or
30 cycles had been decided upon as proper
periodicity.
The advantage to be gained by using a
smaller number of poles than that required
for 60 cycles in connection with certain engine-
driven units, in sizes from 500 to 1000 kw. at
speeds of 100 and 120 r.p.m., led to the decis-
ion to use 40 poles for 120 r.p.m. and 4S for 100-
Fig- 21, Armature of SOcycle. Single-phase, SOO-kw., 3000-r.p.m.
Turbo-generator Built in 1896
r.p.m. This started the use of 40 cycles.
One of these installations was in a cotton mill
in New England. Induction motors for this
periodicity were developed at the same time.
Other mill owners became interested and
some of them ordered duplicate equipments,
etc. In a short time 40 cxxles had become
quite strongly entrenched in certain local-
ities. It was at one time thought b}' many
interested parties that 40 cycles would prove
the prevailing periodicity, as it would prove
satisfactorv for both arc and incandescent
Fig. 22. Rotor of 500-kw., 3000-r.p.m. Turbogenerator
in Balancing Device
lamps and at the same time would be suitable
for synchronous converters, for which GO
cycles then appeared to be almost impossible.
Near Periodicities
Many installations in different parts of
the country were made of periodicities that
were not exactly those for which machines in
general were being built. Some conspicuous
cases of this kind were the following;
Hydraulic development near Portland,
Ore., about 1S92, was undertaken for 200-
r.p.m. vertical-shaft waterwheels. Twenty-
pole generators were decided upon as giving
a good design for the capacity of the water-
90 Febraarv, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 2
wheel that was contemplated. This gave a
periodicity of 33j^ cycles. As a consequence,
one of the two large sj^stems of the Portland
Railway, Light & Power Company grew up
at 33 cycles.
Two or three years later, due also to the
speed of the waterw'heel, a frequency of 34^
cycles was started at St. Anthony Falls.
Minneapolis. As a result, the Twin City
Rapid Transit Com.panj^ has a large amount
of 35-cycle apparatus.
The speed of the waterwheel installed at
Chambly Falls, Quebec, gave to the Stanley
inductor two-phase generator a periodicity
of 66^3 cycles. A little later, the speed of
175 r.p.m. of the waterwheels installed at
Lachine Rapids gave a periodicity of 5SJ^3
cycles.
Frequency changers have been responsible
in certain places for periodicities not exactly
those considered standard. To change from
exact 25 to 60 cycles in a frequency changer
restricts the speed to 300 r.p.m., conse-
quently, the costs in small sets are ver\' high.
Numerous frequency changers have been
built with 4 and 10 poles, respectively, instead
of 10 and 24, resulting in a periodicity of
62 3^ cj'cles when connected to 25-cycle
systems, or of 24 cycles when connected to
60-cycle systems. There have also been
developed frequency changers with 6 and 14
poles, giving 5SJ^ cycles when driven from
25-cycle systems.
In this connection it is interesting to note
how admirably adapted to frequency changing
is 300 r.p.m. This speed gives exactlv correct
changes from 25 to 60, 25 to 50, 25 to 40, 25 to
35, 25 to 30. 40 to 60, 40 to 50, or 50 to GO
cycles.
Engine-driven Units
During the years 1S90 to 1902. inclusive,
there was great activity in developing engine-
driven generators; the size of the units
increased by leaps and bounds. It became
popular toward the end of that period to build
the flywheel type. Considerable mechanical
ingenuity was brought into play in designing
some flywheels with rotor rims of laminations
or riveted plates and others with steel castings
in sections bound together by links, the entire
rim mounted on a cast-iron spider. Stators
usually were designed to have sufficient
strength and rigidit>" to hold the airgap within
allowable limits. Other stators were of the
trussed type. These generators were made in
sizes up to 5000 kw., 5C per cent overload —
7500 kw. maximum — at 75 r.p.m., the overall
diameter of such a generator being of the
magnitude of 33 feet.
The scope of this article does not permit
of any discussion of the problems that arose
in connection with the production of suitable
magnetic steels for cores, the insulating
of the sheets, and the mechanical construction
of armature and field cores; nor of the prob-
lems encountered in the development of suit-
able insulations for windings of low and high
voltages, proper mechanical protection and
supports for the windings, etc.
The author has confined himself to the
consideration of the part jilayed by the
General Electric Company in the iiroductinn
of alternators.
91
Investigation of Water-air Radiators for Cooling
Generators and Motors
By H. G. Reist
Alternating-current Engineering Department, General Electric Company
and
E. H. Freiburghouse
Turbo-alternator Engineering Department, General Electric Company
The satisfactory solution to the problem of ventilating large capacity electrical units in land stations, by
means of ducts, air washers, etc., is not as readily applicable to the cooling of electrical apparatus where
extreme compactness is essential — -as for instance in the propelling machinery of a ship. The use of the custom-
arily large ventilating air ducts would tend to offset one of the great advantages of electric marine drive;
viz., the enlargement of the cargo carrying space. The following article reviews an investigation made to over-
come the difficulty by applying a closed system of air ventilation with water-air radiators for removing the
heat from the circulating system. — Editor.
Special Conditions of Ventilation
'NDER favorable
U"
conditions the
common arrangement
of ducts and air wash-
ers, which permits of
a complete control of
the ventilation for
turbo-alternators and
stations, is believed
to be best. However,
cases sometimes exist
where it is very de-
sirable to eliminate
the numerous long air
ducts of large cross-
section. This is especially apparent in the
case of electrically driven vessels where lack
of space and other reasons make ducts objec-
tionable. In order to meet such special con-
ditions, it is essential that a closed system
*See "Steam Turbine Generator Ventilation," by G. Monson,
page 99 in this issue.
H. G. Reist
be employed for the
circulation of the cool-
ing air.*
Heat of Losses Must be
Removed from the Air
For seagoing ves-
sels, a closed system
of ventilation is de-
sirable. However,
with such a system it
becomes necessary to
cool the circulating
air through the same
range of temperature
E. H. Freiburghouse
Fig. 1. Photograph of Truck Radiator Fitted with
Special Header Tank for Test Purposes.
For dimensions see Table I
Fig. 2. One of the Radiator Sections in the Air Tunnel
Sxirrounded by Baffles. Baffles at front
and sides of section not seen
92 Februarv, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, Xo. 2
Fig. 3. Side View of Air Tunnel and Water Piping of the Radiators
that it is raised in passing through the gen-
erator or motor it cools.
The quantity of air which passes through
a turbo-alternator is so related to the electri-
cal and windage losses of the machine that
the rise in temperature of the cooling air, or
the range through which it must be cooled if
used again, is approximately 20 deg. C. As-
suming the standard temperature of the in-
going air to be 40 deg. C, as estabHshed bv
the Rules' Committee of the A.I.E.E., the
temperature of the air which leaves the gener-
ator is 60 deg. C, and 50 deg. C. is the mean
of the ingoing and outgoing air of any cooling
device in the ventilating circuit.
Cooling Methods
The apparatus used for the removal of heat
from the air will depend somewhat upon the
character of the cooling water. L'nder most
conditions, however, with a plentiful supply of
fresh water, the air can readily be cooled by
the use of the spra>" form of air washer now so
generally emplo>-cd for the cleaning and cooling
of air. The cooling water in this case is discarded
or, if necessan.- , recooled and used again . As the
spray washer is an apparatus in which the air
comes in contact with finely di\'ided water par-
ticles, formed by spray nozzles, the character
of the water supplied will determine whether
an air washer would be applicable.
On ocean vessels, sea water is the only avail-
able heat conveyor from the cooling device of a
closed system of ventilation. It is quite doubt-
ful whether it is safe to have the cooling air pass
continually through a si)ray of salt water, on
account of the liability of small quantities of
salt being carried into the windings.
Fig. 4. End View of the Air Tunnel and D-sch^ir^c Diu't
INVESTIGATION OF WATER-AIR RADIATORS
93
In such cases two courses are open: either Tests of Radiators
to recirculate a relatively small quantity of The tests conducted upon this small
fresh water through the air washer and cool radiator having one quarter inch tubes con-
this water outside of the washer through the firmed the opinion that the scheme was
intermediation of a cooler, or to install some feasible; in fact, they indicated results better
form of radiator in the ventilating air circuit than were expected. The reduction in the
and use salt water directly in the radiator diameter of radiator tubes has a double
tubes. This latter scheme was considered by effect in accelerating the transmission of
one of the writers nearly one year ago as a heat, since it not only reduces the travel of
feasible solution of a ventilating problem the water particles between successive con-
where it was very difficult to provide external tacts with the tube wall, but also increases the
ducts of sufficient size, and where an air area of contact surface of a unit weight of the
washer was impossible because of the char- water. Although it is beneficial to use tubing
acter of the cooling water. An investigation of small diameter for these reasons, it was
was started which led to tests from which the nevertheless considered advisable to use
results given in this paper were obtained. tubing of larger diameter than one fourth
The function of the automobile radiator is inch in order to reduce the liability of clogging,
to cool the circulating water and to heat the This justified a more extensive series of tests
air passing through the radiator core or, in which were made upon radiators having tubes
other words, to transfer heat from the water of greater diameter and fins of different design.
TABLE I
Tanks were constructed of No. 18 B.&S. copper.
Shaped 4j| by 3 by 23 J-^j inches.
Core had six rows of 29 tubes 174 tubes
Diameter of tube (outer) 0.375 in.
Diameter of tube (inner) 0.339 in.
Length of tube 23iJ in.
Length of tubing per section 246 ft.
Fins outside 0.75 by 0.75 by 0.007in.
Spacing of fins 0.165 in. pitch
Spacing of tubes front and back 0.8125 in.
External surface of tube per linear foot 13.55 sq. in.
Surface of fins (total per linear foot of tubing) 67.3 sq. in.
Total cooling surface per linear foot of tubing 80.85 sq. in.
Equal to 0.5615 sq. ft.
Free air section (between tubes and fins) of frontal area 2.02 sq. ft.
Frontal area of core 3.90 sq. ft.
Weight of core empty per cubic foot 37 lb.
Weight of core plus water per cubic feet 45.9 lb.
to the air. For the purpose under consider- These tests were made to determine the
ation, this heat transfer process would be relative dimensions of radiators, the heat
reversed since the heat is to be transferred transfer as functions of the speed of the air
from the cooling air of the generator to the and the water, the amount of cooling surface
water in the tubes. for a given duty, and the resistance of the
radiator to the flow of the air and the water.
Selection of the Radiator Since the frontal area of a radiator within
In those fields of service where radiators a restricted space would be limited, thus
are given hard usage, the fin-and-tube type determining the minimum speed of the air
seems to be almost universally used. Clog- through the core, it was realized that the
ging of irregular water channels by foreign practical application of radiators for this
matter is the most common cause of trouble purpose greatly depended upon whether the
in radiators. Obviously, the round tube offers air could be forced through the necessary
the least trouble from clogging and, for given depth of core without prohibitive pressure,
weight, is of the strongest construction as a Five large radiators, intended for truck
conductor of water. service and fitted with special headers, were
It was for these reasons that a small obtained through the courtesy of the G. & O.
fin-and-tube type radiator was selected, and Manufacturing Co., of New Haven, Conn,
tests made to determine the rate of heat Fig. 1 shows one of these sections as described
transfer as a function of the rates of flow of in Table I. The radiators were installed
water and air. in an air tunnel in series with the air and water
94 February, 1920
GENER.\L ELECTRIC REVIEW
Vol. XXIII. Xo. 2
flow as indicated by Figs. 2 and 3. An end
view of the entire testing equipment used is
shown in Fig. 4.
Design of G. 86 O. Truck Radiator
The header tanks at each end of the core
were specially designed for test purposes.
ft) 0 M
\ ^ t
i ::":::-:J:::::::::::^::::
\ 4
« t
Cj _ '^
^7 A- ' 7
^ it^ t EEt
^ jn r ,'- it
-F^ ' ^
< 6 \ — ■ y-| — \ — ' — p
- i±--:-::-^+^zt:^
1 T±7 ^-Jv
b ^^ jzr i^zz
1 ^ lb x
S - + X# -^7
L-— ^-x-±:::ij=:5^_
r — ^ ht-^=^i^ - -
1 __:t__ 7
K ^ J. ^^ j^ ^
Yz z :::: /^^ Z- __ _ T
1 __X____T XX^
^^ ii- ^
"^ .xx
^ Ja-'X-'' i
1 :: /-^^ ^
\' — y—y----x
^ ^^^ X
fc "z-^ it
"'^ -^^: ±- J
o 500 /ooo /soo ^ooo asoo jooo
Fig. 5. Impact Pressure in Inches of Water to Force Air
Through the Radiators as a Function of the Velocity
of the Air in the Free Air Section of the Core
Curve I required by five sections in series
Curve II required by three sections in series
A motor driven blower of known speed,
pressure, and volume characteristics delivered
air through electrically heated grids to the
radiators.
Air Flow
The air flow through the radiators was
varied from 0.61 to 1.9 pounds per second
per square foot of the frontal area of the core,
which corresponds to a range in velocity of
930 to 2880 feet per minute through the
minimvun free air section of the core.
The quantity or mass flow was determined
by means of a pitot tube and inclined mano-
meter, also by the rise in temperature of the
air and the electrical input as the air passed
through the heating grid.
Air Pressure
Tests were made to determine the fan
pressures required to deliver air through
the five-section radiator. Ordinarily a radia-
tor for sen-ice could be so designed that the
depth need not exceed that equal to three of
the sections tested, also the frontal area could
be such as to limit the air velocity through
the radiator core to less than 2000 feet per
minute, which would add a resistance in the
air circuit equal to about 2.5 inches of water.
If the speed of the fan and the resistance of
the air circuit are fixed, the volume of the air
and the air pressure developed by the fan are
also fixed. Moderate increases in the resist-
i:i55
i 1 i 1 ' ■
~_.^..lJl^4;.-X 1 X. J_
E -^ IL--^ X:^'!- • " X
-1- ^ ■ — j-TT ' .^^I
x:t . - Z'^-'- ■- %
i - ^^ -T-
■' ■ ^ ■ ^ > " J-- - +
I ' ^'''jir^'^
a-^---
t'/« ^ '
^M ---
" ,^^^^- ■ ^■^-^■T-XT-'t^
^ j^dl-^ ^-c^T_ ■ 4 X
rpy^^ .3'^' ^ 1 i T^ I
%^- - ' \if>'un • ri-i-T-i • I • 1 1 1 ! 1 M 1 1
^J£ - - ■ ^
■^\ i>^' 1 ■ 1 ■ 1 ■ 1 ■ 1 I 1 1 1 1 1 ' 1
^ ir JC^ ^^ T r : I : 1 : I ; 1 : 1 1 1 1 1 : 1 r 1 1 1 1 1 1 1
^05-- + -- -^^
<: , ,. t;
^/v« 3.'
; t ±
.. :r ^ -T
^«^ "
.. _. .x
. «, .. ...
. _ ... X __- -
^a^
u
-, . -, . - -
*■ /,,
i « .::: :. _
: X^x^li-i-.x.:
'tco aoc /zoo /eoo S'ooo p*oo 2900 jjoo
Fig. 6. Rate of Heat Transmission as a Function of the
Speed of Air Through the Free Air Section
of the Radiator Core
Curve I. For an approximate speed of water in the
tubins equal to 10 ft. per minute
Curve II. For 30 ft. per minute
Curve III. For 40 ft. per minute
Curve IV. For 50 ft. per minute
\
■J
'"iix:::: x: x._xx^tTT^
"" - "=---5xx±E:
,1 1
4 M . J 1 J/L 1 4I1U
i ^- — ~~
EE p^:EEE EEEEEEEE
a /6 14- X 46 -M
tVATCe re OtV //V OALLOA^S fCK All/^t/Tt.
Fig. 7. Difference in Water Pressure in Pounds Per Sq. In.
Obtained on Four Sections of the Core as a
Function of the Water Flow
INVESTIGATION 01' WATER-AIR RADIATORS
95
96 Februarv, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 2
ance of the fan circuit do not appreciably
affect the pressure developed by the fan as
only part of the fan pressure is used in passing
the air through this additional resistance. For
moderate changes in resistance the volume
of the air may be taken as proportional to the
square root of the ratio of the pressure avail-
able to pass a given quantity of air through
the generator alone under the two conditions.
For instance, consider a case where a certain
machine has no external duct and the fan
produces a pressure of 13 inches of water
while passing 60,000 cubic feet of air per
minute through the machine. Assuming that
an air duct and radiator are added in which
the total head lost is three inches, then the
quantity of air under the second condition
is closely given by
Q^^
13-3
13
X 60,000 = 52,700
cubic feet per minute.
Since long air ducts having numerous bends
or changes in cross-section are eliminated by
the use of the radiator and short ducts, it
will be possible in many cases to obtain the
necessarj' quantity of air without the use of
external blowers.
The relation of air pressure to air velocity
through the radiator cores is indicated by Fig. 5.
Beyond a certain air velocity the loss in air
pressure required to force the air through
RADIATOR DESIGN ESTIMATED FROM INFORMATION SECURED FROM THE TESTS
(1) Power dissipated 000 k\v.
(2) Heat per minute dissipated 34,200 B.t.u.
(3) Temperature of air into radiator - 55 deg. C.
(4) Temperature of air out of radiator 37 deg. C.
(5) Temperature of air mean 46 deg. C.
(6) Temperature of air drop 1<S deg. C.
(7) Temperature of water into radiator 25 deg. C.
(8) Temperature of water out of radiator 28 deg. C.
(9) Temperature of water mean 26.5 deg. C.
(10) Temperature rise of water 3 deg. C.
(11) Mean air minus mean water 19.5 deg. C.
(12) Velocity of water in tubes 50 ft. per min.
(13) Velocity of air in free air area of core 1200 ft. per min.
Cooling Surfaces
(14) External surface of tube per linear foot 13.55 sq. in.
(15) Surface of fins per linear foot of tube 67.3 sq. ft.
(16) Total surface per linear foot of tube S0.S5 sq. in.
(17) Total surface per linear foot of tube .5615 sq. ft.
(18) Total cooling surface per sq. ft. of frontal area in one layer of tubes 8.3 sq. ft.
(19) Total cooling surface required under conditions (12) and (13) is from SX
0.0614=600 kw S= 9770 sq. ft.
Estimated Weights, Volumes, and Surfaces
(20) Weight of air required per minute based upon (6) and (2) 4310 lb. per min.
(21) Volume of air from (20) 57,500 cu. ft. per min.
(22) Free air area required from (20) and (13) 47. S sq. ft.
(23) Weight of sea water based upon (2) and (10) 6670 lb. per min.
(24) Number of tubes required to carry water based upon bore of tube (0.083 sq. in.)
also (12) and (23) 3602 tubes
(25) Spacing tubes 0.8125 in. apart or 14.75 per ft.
(1
(2
(3
(4
(5:
(6
(7
(8:
(9
(10
(11
(12
(13
TABLE III
RADIATOR DESIGN FOR DIFFERENT CONDITIONS FROM THOSE OF TABLE II
Power dissipated ; 600 kw.
Temperature of air ingoing 50 deg. C.
Temperature of air outgoing 32 deg. C.
Temperature of water ingoing 25 deg. C.
Temperature of water outgoing 27 deg. C.
Quantity of water at 50 ft. per minute 9800 lb. per min.
Width of core 19.94 ft.
Height of core 4.82 ft.
Volume of core 117.5 cu. ft.
Weight of core empty 4340 lb.
Weight of core plus sea water 5400 lb.
Velocity of air in core 1 1,50 ft. per min.
Resistance to air flow in inches of water 1.0 in.
I
INVESTIGATION OF WATER-AIR RADIATORS
97
the radiator makes the
prohibitive.
gain in heat transfer
Temperature of the Air
The temperatures of the air were obtained
over the sections of air flow at the front and
back of the heating grid and over each of
the five sections of the radiator by means of
extensive temperature coils and numerous
thermometers. These furnished data for
determining the mean temperatures of the air
for the various sections of radiator.
Water Flow
The flow of water through the five sections
of the radiator in series was varied from S.25
to 40.5 gallons per minute, corresponding to
a range in velocity of 10.1 to 49.6 feet per
minute in the tubes. Entering at the bottom
of the last or fifth section of the core, the
water flowed upward in all of the 174 tubes
of this section, then alternately downward
and upward through the remaining sections,
leaving the radiator from the top header
tank of the first section. The air flowed
through the cores of the five sections from
front to back, thus moving through the core
in the direction opposite to the progression
of the water.
In determining the curves shown in Fig. 6,
it is seen that the observed points for the
higher flows do not lie as regularly on the
curves as do those for the lowest flow. The
curves were detennined from the temperature
rises of the water, which were lower for the
greater quantities of water and thus made a
greater percentage error in the observed
temperatures. Nevertheless, it is shown that
for an increase in the water flow up to a
certain rate there is obtained an increase in
the heat transfer for a fixed difference of the
5TEEL Rc-lNFoecm^
STSJP
Copper Fins
^\\\\\\\V.-k^\\V^\\\\^'^
^WflTER OUTLET
Fig. 10. Details of a Radiator Section
mean temperatures of the air and water, after
which little benefit is secured by a further
increase in the velocity of the water in the
tubes.
Fig. 9. Indicates the Ventilation System Obtained with a Turbo-alternator and a Radiator for Cooling the Air
98 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 2
Temperatures were obtained in the water
circuit on each side of each of the five sections
of the core, which made known the relative
utility of the sections throughout the depth
as a function of the air speed through the core.
The cur\'es shown in Fig. 6 were estimated
from data taken on the first three sections of
the core, since it was realized that this would
be the approximate depth in the direction of
air flow on most radiators.
The five sections of radiators were con-
nected by L25-inch water piping provided
with relief valves, pressure gauges, and U-
tubes. Probably most of the drop in water
head registered on four sections of the radiator
was produced in the small connection piping.
The curve shown in Fig. 7 obeys the "square
law" approximately, and indicates turbulency
of flow.
Arrangement of Radiators with Generator
Figs. 8 and 9 indicate possible arrange-
ments of radiators between the sections of the
stator frame, or entirely outside of it with the
necessary inlet and discharge ducts for the air.
As indicated by these diagrams, the radiator
would be divided into a number of sections or
units, thus facilitating handling or repairing
a section while the remainder of the radiator
was in service. .
Fig. 10 indicates an arrangement of the
tubes in a flexible tube sheet, to which is
secured the header tanks. The flexible tube
sheet has been found most satisfactory' to
prevent leaking tubes. Radiators intended
for use with sea water should be constructed
of the metal which will best resist the cor-
rosive action.
In order to agree with the \-alues of Table II,
the width, height, and ntimber of layers of
tubes must be found b}" trial or by solving
through simultaneous equations relating the
unknowns.
Using the latter method:
Let x = width of core in feet.
Let y= height of core in feet.
Letz=number of layers of tubes from front to
back of core.
Equations for:
Frontal area 47.8 =0.518 x y
Surface 9770 = 8.3 x y z
Number of tubes 3602 = 14.75 x z
x = 19.15 ft. width permitting 283 tubes.
y=4.82 ft. height,
z =12.75 layers.
Anticipating a reduction of capacity due
to clogging or higher temperatures of ingoing
water, the radiator may be constructed with
the front divided into 14 sections each having
a width of 21 tubes, while in depth the core
ma\- be made in two divisons each having a
depth of nine tubes. The frontal width will,
therefore, be 294 tubes and the depth will be
eighteen tubes or 5292 tubes total in core.
Substituting back through the above equa-
tions and computations there will be obtained
the approximate relations shown in Table III.
99
Steam Turbine Generator Ventilation
By Geo. Monson
Alternating-current Turbine Engineering Department, General Electric Company
This article reviews the development o£ forced air ventilation for cooling large generators. On the early
machines salient poles were used and these served to draw in the air and expel it through the ducts in the
stator laminations. The next step was to provide passages through which air from the outside of the building
could be obtained. The accumulation of dirt in the windings led to the introduction of air washers, which
while removing the dirt, at the same time reduced the temperature of the air. As the size of generators
increased, fans were introduced to increase the circulation of air. Separate blowers were used in some cases,
but later designs embody the ventilating fans as part of the rotor structure. The latest system employs
the closed air circuit and conserves space by mounting the humifying equipment and air dryers under the
generator foundation. The advantages are the elimination of numerous station ventilating ducts and
quantities of dirt that are always drawn in when outside air is employed. — Editor.
The principal departure was the inauguration
of the cylindrical type of rotor having two
or more field coils per pole. The rotor con-
struction varied; some had laminated disk
bodies with radial slots for receiving the field
coils and others had parallel slots.
In generators of larger capacities than
5000 kw., laminated poles were dovetailed
into steel plates which were shrunk on fluted
cast-steel spiders, and these in turn were
mounted on the shaft.
At first, the air M^as drawn in at both the
top and the bottorn of the generator by the
action of the rotor spider and forced out
through the rotor and stator ducts. Part of
the ventilating air also passed through the end
windings and then through apertures in the
frame to the room. On later machines, the
ventilating apertures in the stator frames
were eliminated, the air being taken in at the
top of the generator and expelled to the room
through openings in the generator base. The
ventilation for this type of generator was very
easily accomplished; in fact, in many cases
more air passed through the generator than
was needed for good economy, hence baffie
plates were inserted to reduce the quantity
of air and thereby the windage losses.
George Monson
Introduction
IV/TANY articles
-^^■*- ha\'e been writ-
ten on the necessity
of ventilating turbine
generators to remove
the heat and to
obtain increased out-
put. It may, there-
fore, be of interest to
review the progress
that has taken place
in the ventilation fea-
ture of these gener-
ators.
Review
About twenty years ago the General Electric
Company entered the steam turbine business,
and decided upon vertical type units. One of
the first commercial machines designed was a
two-stage 500-kw. four-pole ISOO-r.p.m. unit.
The generator was patterned in general after
the then prevailing waterwheel and engine-
driven types, having salient pole construction.
The rotor body, with the field poles, was built
of laminations riveted together in sections
and forced on the shaft. This rotor had
sufficient blower action to ventilate the
generator in the following manner :
Air was drawn from the room through the
top and the bottom of the generator into spaces
between the poles, and then expelled directly
into the room through the armature ducts and
apertures cored in the frame for that purpose.
Several other units of different capacities
were designed. Four-pole generators with a
speed of 1800 r.p.m. up to 14-pole machines
at 514 r.p.m. followed similar lines of con-
struction and ventilation.
As the rotative speeds and capacities
increased, the number of generator poles was
correspondingly reduced and changes in the
generator design were made to meet the new
requirements of stresses, ventilation, etc.
Use of Hoods
It was found that in some stations the
heated generator air exhausted from the base
openings did not mix freely with the station
air, but instead moved up along the stator
frame and returned to the air inlet opening at
the top of the generator, thereby causing
undue heating in the generator. To prevent
this a hood made either of steel plates or cast-
iron was placed over the air intake shield.
This hood had an opening which could be
placed in any direction, e.g., could be turned
toward the station windows which were kept
open when the temperature inside the build-
ing was high. By this arrangement the air
intake was further removed from the rising
100 February, 1920
GENERAL ELECTRIC REVIEW
Vol XXIII. Xo. 2
column of heated air, and cooler air was
drawn into the generator. This hood served
its purpose very well on those machines.
Innovation of External Ventilating Ducts
Up to that time air had been taken from
and discharged into the engine room. The
power stations had been originally laid out
and built for a certain kilowatt output with
a given number of units, but this output was
soon outgrown and increased capacities were
required. The simplest and cheapest scheme
was to replace existing units with new units of
increased capacity. In some cases, due to the
substitution of new units, the amount of
power was several times that originally con-
templated. The increased quantity of heated
air from the generators could not be removed
from the engine room by the old method.
Therefore, in order to keep the temperature of
the generator within safe limits, and to secure
a more comfortable room temperature, it
became necessary to install ducts through
which cool air could be drawn from outside the
station building to the generator, and thence to
the discharge outlet, also located outside.
Air Washers
It has always been more or less dangerous
to allow dirt to accumulate on the windings,
since it clogs the air passage and introduces
a heat insulating material on the surface
exposed to the air, thereby causing excessive
heating. Yet it has been difficult to prevent
such accumulation. Although only a ver}^
small percentage of dirt carried by the
ventilating air may be deposited in the
machine, the quantity of air passed through
the machine is very large and therefore the
amount of dirt deposited is considerable.
Some conception of the magnitude of the
quantities involved may be had from the fact
that a 30,000-kw. machine requires approxi-
mately 6,000,000 lb. (81,000,000 cu. ft.) of
ventilating air to pass through it during 20
hours of operation. The rapid deposit of
dirt under such conditions makes frequent
cleaning necessary. In order to reduce the
frequency of cleaning, which is a slow, expen-
sive process requiring the dismantling of the
machine, air washers were installed in the
air intake duct. In addition to removing a
large percentage of the dirt, these washers also
ser\'e the purpose of cooling the air, thus
permitting higher generator output, especially
during hot weather. The problem was thus
partially solved. A complete solution, involv-
ing a closed system of ventilation, is described
later in this article.
First Use of Fans
The construction used in some of the earlier
vertical alternators caused air pockets to be
formed around the top and bottom ends of
the windings, be\-ond the armattire core.
These produced eddies which prevented the
proper flow of air. To overcome this diffi-
culty a straight bladed fan was mounted on
the top of the rotor for forcing part of the
incoming cool air through, and also around
the top end windings, thus overcoming
the eddies. This fan directed the air down-
ward between the stator core and frame to
the outlet apertures in the base. In later
designs most of the generators had fans
mounted also on the bottom part of the rotor.
Dampers in the Air Ducts
It was recommended from the start that
station air ducts be introduced, and that
these be furnished with dampers for regulating
the quantity of air. Later, doors were made
in the ducts inside the station so that all the
air, or part of it, could be taken from the
room, or from outside the building, whichever
was desired. Similar dampers were placed in
the exhaust ducts, and in case of fire in the
generator these dampers could be closed, thus
impeding the progress of the fire.
Changing from Vertical to Horizontal Units
The largest vertical steam turbines built
were 20,000-kw., 750-r.p.m., 25-cvcle and
18,750-kw. (25,000-kv-a.) 720-r.p.m.^ 60-cycle
units. With the demand for further increased
kilowatt capacity and rotative speed, it soon
became obvious that the limit was about
reached in the design of the vertical machines.
It was advisable from an engineering stand-
point to change to the horizontal type of
machine. This line of steam turbine gener-
ator sets was started at the Schenectady
Works with the 300-kw., 4-pole, 1800-r.p.m.,
00-cycle units and continued to date with
capacities of units up to 30,000 kw., 1800
r.p.m., GO cycles: 3.5,000 kw., 1500 r.p.m..
2.5 cycles; and 45,000 kw., 1200 r.p.m., m
cycles.
The first horizontal sets up to 5000 kw.,
1500 and ISOO r.p.m., had laminated cylin-
drical rotors with fluted shafts for blower
action. The ventilating air was taken from
the room through openings in the generator
end shield, and discharged at the top of the
stator frame to the engine room. Later
machines were changed so that the discharge
could be made upward or downiward according
to preference.
STEAAI TURBINE GENERATOR VENTILATION
101
Latest Construction of Generator
The turbine rotative speeds have now
reached a maximum of 1.500 r.p.m. for 35,000
kw., 25 cycles; ISOO r.p.m. for 30,000 kw.;
and 3600 r.p.m. for 6000 kw., GO cycles.
The limit of output capacity with respect to
speed is dependent upon the rotor stresses
and the relation of the critical speed of vibra-
tion to the normal speed. To make such
generators possible the solid forged rotor
construction was adopted, .and has proved sat-
isfactory. With this construction, the ventila-
tion became a still more difficult problem.
small diameter and a great axial length. With
the air pressure available, the air gap in such
machines did not afford adequate space for the
passage of the large quantity of ventilating air
required. A new departure in ventilation was
therefore devised.
Double bladed fans were provided on each
end of the rotor. Part of the air passes into
the air gap in accordance with standard
practice and part flows through tubes arranged
outside the armature core leading to a
central air pressure chamber where it passes
through a number of armature core ducts
m
i
I
X^TAf^E fitQi^ c^^vf >9»(*'r o^
5Cȣtnep oPci'tO'
WfTLL.
o>re^L£ TO
TO Bt CLOItO IN
\ cfne Of rime
ofTMpem ,
Fig. 1. Diagram of the Ventilation System for Turbo-alternator
Fans Mounted on the Rotor
For ventilating some of the first of these
generators separate blower outfits were used.
In some cases blowers were mounted on the
end of the rotor shaft; but this practice was
later abandoned and since that time it has
been standard practice to mount fans on each
end of the rotor body. The fans force the air
at both ends, around and through the end
windings into the air gap, and from there
along the rotor body, and out through the
armature ducts and frame as already explained. ,
Double Air Flow System
With the ever-increasing capacity of units,
it became necessary on account of stresses
to build the larger machines with a relatively
radially inward toward the rotor and then
axially along the air gap and out through the
armature ducts, in the usual manner. By
this arrangement it was possible to introduce
air into the gap at four places and thereby
force the necessary amount of air through
the generator.
Exhausted Generator Air for Use Under Boilers
It seemed logical for power stations to use
the exhausted generator air under the boilers
for heat conser\-ation. Ducts were therefore
installed for the purpose, but this arrangement
did not prove to be as successful as anticipated.
During cold weather considerable condensa-
tion was caused in the engine room by air
having high humidity coming in contact with
102 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 2
the cold ceiling. Under unusually severe con-
ditions fog would be present in the room, caus-
ing dampness and dripping of water from the
ceiling. When this condition developed it
became desirable to take air from the engine
room through the air washer and to discharge
s^e^siv M^trsf .^^'Jorj
Fig. 2. System of Ventilation in Which Inlet and Outlet Air Ducts and Air
Rectifying Equipment Are Located Under Generator Foundation
it back to the room during cold and moist
weather, and to regulate the room temper-
ature by doors or openings to the boiler room.
This arrangement is shown in Fig. 1. After
the air passes through the generator and its
temperature is raised approximately 20 deg.
C, its ability to absorb moisture is greatly
increased; and when discharged into the
station toward the ceiling it reduces the
relative humidity. The temperature of the
station was regulated by opening the win-
dows. This is the usual ventilating arrange-
ment at the present time, but there are con-
ditions where a closed air s\stem is preferable
as mentioned earlier in this article.
Closed Air Circuit System
In some cases insufficient consideration
was given to the installation of washers and
they could not be operated in freezing
weather. In other cases the washers were not
kept in good condition and failed to clean the
air properly. These conditions were detri-
mental to the generator and a better venti-
lation arrangement was desirable. It is
believed that this has been found in the
closed air circuit system.
The closed air circuit system operates in
the following manner: The air enters the
rectifier cooling chamber directly after leaving
the generator, then flows to the speed reducing
chamber where the water particles are segre-
gated from the air flow, then through an
eliminator chamber where the last vestige of
water is removed from the air before it
re-enters the generator. This system uses
the same air continuously, and can be m_ade
to occupy a relatively small
and compact space as com-
pared with the present sys-
tem.
The foundation directly
under the generator is
usually provided with a
space which is occupied by
the inlet and outlet air
duct pipes. This space can
be utilized to advantage
for mounting the air recti-
fier as shown in Fig. 2. The
cimibersome station venti-
lating ducts, including the
room occupied by the air
washing apparatus, will be
eliminated, thereby sim-
plifying the station con-
struction considerably and
saving space.
As stated before, the air washer is used
for cleaning and cooling the incoming air.
By using the same air continually, and with-
out any possible way for it to mix with impuri-
ties, the problem of cleaning is eliminated.
Fig. 2 shows a view of the rectifier installed
with a generator; the complete construction
consisting of a steel tank having an inner
compartment where the cooling water spray
nozzles are located; a separating chamber
where the heated water is collected from the
air and discharged; and the eliminator com-
partment through which the air has to pass
before re-entering the generator. The oper-
ation, simplicity in design, economy of space,
and positivcness of action can readily be
understood.
Noise
All high speed machines are more or less
nois\- when the air is discharged to atmosphere,
but in the closed air system this should be re-
duced to an unobjectionable tone and volume.
Amount of Water
The necessary quantity of water for remov-
ing the generator losses from the heated air is
approximately four gallons per kilowatt-
minute for one deg. C. rise of water, and
varies inversely as the allowed temperature
rise of the water. Suflicient data are not now
STEAM TURBINE GENERATOR VENTILATION
103
available to determine how many degrees centi-
grade the cooling water should be allowed to
rise to obtain the best results from the gener-
ator; bvit there is a considerable saving in the
quantity of water used, as compared with the
present air systems, when the water from the
washer is heated only about 2 dcg. C.
Regulating the Generator Heating
As a rule turbine units are seldom operated
at their maximum output and it would there-
fore be inefficient to operate the air rectifier
continuously with the amount of water cor-
responding to the maximum load of the
generator. Under such load conditions, and
especially when water is scarce, an automatic
water regulator should be installed for con-
trolling the number of spray nozzles in
operation. There are several ways to accom-
plish this: by electrical and mechanical
arrangements for opening and closing the
valves of the water supply pipes to the spray
nozzles at prearranged temperatures; by the
use of thermostats placed in the heated air
leaving the generator, or in some other
suitable location, etc. With such automatic
regulation the generator will operate at prac-
tically constant temperature for different
outputs and be entirely independent of sur-
rounding atmospheric conditions. This can-
not be accomplished with present systems
because the air washers used are built for a
combination of cleaning and cooling; i.e.,
any reduction in the quantity of water may
adversely affect the cleaning of the air.
Air Free from Water
All the air washers the writer has seen in
connection with turbo-generators have the
flow of air through the washer in a horizontal
From Outs/de
23° c.
Fro/n Generator
Fig. 3.
Arrangement of Dampers Taking Air from
and Discharging it to the Outside
direction so that the drops of water after
leaving the nozzles must fall to drip-pans at
right-angle to the air flow, which has a
velocity of 500 to 1000 feet per minute. In
some cases where the washers are not of liberal
dimensions, small particles of water are carried
along with the air flow and may enter the gen-
erator. In the system shown in Fig. 2 the
air flow is vertically downward after leaving
the spray chamber so that the water has to
drop straight with the air flow. The direction
of the air flow is then reversed to vertically
sb'c
From Generator
To Generator
Fig. 4.
Arrangement of Dampers Forming a Closed
Circuit for Ventilating Air
upward at a speed of 300 to 500 feet per
minute before it passes the eliminator plates.
The low velocity, vertical flow, and eliminator
plates effectively remove all particles of water
from the air before it reenters the generator.
Dampers and Fire Protection
In the present system, the air intake or
outlet ducts or both arc generally supplied
with dampers installed for modifying the
incoming air temperatures, and to prevent air
circulation in case the armature winding
should take fire. The object is to let the fire
.smother itself; but on account of the diffi-
culty of preventing leakage past the dampers,
this method has not been entirely satis-
factory for the purpose.
In the closed system dampers are unneces-
sary since the amount of water controls the
temperature. In case of internal fire the
water valves may be closed which would
prevent the air rectification and by this means
the fire would be soon extinguished. In case
a quicker action is desired, steam or pyrene
may be injected in the system or water can be
sprinkled directly on the windings by pipes
installed for this purpose.
Danger Signal
The closed air system cannot operate with-
out cooling water being supplied to the recti-
fier any more than bearings can operate with-
out lubrication. It will, therefore, be prudent
to install an alarm device which will annotmce
danger in case the water cooler does not
function properly, or stops entirely; and
which will also serve as a tell-tale should the
generator happen to operate above normal
temperatures from other reasons such as
heavy overloads, internal fire, etc.
104 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 2
Such an alarm system can be readily
installed and may consist of a thermostat in
the outgoing air duct, in cormection with a
bell that will ring at a predetermined tem-
perature; or of an electric bell connected in
circuit with one or more of the stator tem-
perature coils; also temperature coils may be
mounted in the outlet air duct in the same
manner as the thermostat.
Other Applications of the Closed System
The closed air circuit system is useful not
only for ventilating turbine generators and
other electrical apparatus, but should prove
of value in places where fumes from acids,
carbon and steel dusts, or other injurious
At the time the test was started the
machine had been running all night under
approximately full load, with the dampers in
the position shown in Fig. 3 (i.e., the air being
drawn from the outside and discharged into
the room) and the temperatures were constant
with ingoing air at 23 deg. C. and outgoing air
at 45 deg. C, or a rise of 22 deg. C. (for
about 9000 lew.). The temperature coils and
thermo-couples in the armature of the machine
averaged 60 deg. C. actual, or 37 deg. C. rise.
At 7:45 a.m. the dampers were thrown to
the position shown in Fig. 4, forming a
closed circviit for the ventilating air, and a
heat run lasting till 4 p.m. was made under
these conditions to determine the efficacv of
A-y-a
II 000
°C.
10 000
100
9 000
90
8 000
SO
70
60
SO
40
30
20
10
1
1
1
1
1
I
1 1 !
'
o«V
\
^~^
/
■^T
^**^
_
_
.»
V-
^^
,
\
-4Loa<l-
/-,
.
==
^
r'^
! Ml
iCCoj
*cu/at
^^
^^
.nt-
-Km Loss.
; i
i Temp Coi
^
— —
^
■
■*
■>.
^
_—
-^
_
-/-
___.
,_
.— -. ..
._
—
^.
._
..
_
-
.-.-.
,___
,.^-.
■ _M
1
/\
Kw LOSS (/■ rom Mist in wocer) i
1
i
1
^
1
1
1 1 1
Air Out -
1
!
=i=
^
^~^
-
r"
!
1
^^^
1
Z3000 CuFt Air per M in.
~ 175 Gol. WoterperMif? ~
— ^ —
1
'
_
Airlrt
„__
_ .—
J^
1
-
-
<*
ater
Out \
— ^
=^=H
iA/Qt£rIn
^~
^
~^
i
\ 1 1
'^ir From Outside Shut Off at 7=45 j | |
1
'
1 1 1
8: 00
9:00 10:00
A.M.
11:00
12 00
Time
1:00
2:00 3:00
PM
4:00
Fig. 5. Curve Showing Results of Test on lQ,000-kv-a. Turbogenerator
Cooled by Ventilation from a Specially Designed System
substances prevail as it will safeguard men and
machinery from their effects.
Heat Run Made with Closed Circuit Ventilation
System.
Tests have been made on a closed air circuit
system utilizing the same washer as installed,
which washer was designed for cleaning the air
when it was taken from the outside atmosphere.
The tests were made on a 10,000 kv-a. turbine
generator unit that is fitted with a special
design of air duct which, by means of dampers,
permits of various schemes of ventilation.
Air may be taken from out of doors and
exhausted into the room, or the air may be
circulated and used over and over, the heat
being removed each time by means of water
passed through an air washer.
this ventilating system. During this time the
load on the machine was maintained as nearly
constant at 10,000 kv-a. as conditions would
permit. Fluctuations in the load cur\-e were
caused by a variable demand made by the
shops being supplied with power and by
changes in steam pressure. Evcr\- half hour
readings were taken of the load. tcmiK^rature
coils, air, and water. At the end of the run
and at a load of about 10,000 kv-a., the tem-
peratures were fairly constant with inlet air
at 33 deg. C. and outlet air at .iS deg. C or
a rise of 25 deg. C. in passing through the
generator and a corresponding drop of 25
deg. C. in passing through the air washer.
Water at the rate of 175 gallons per minute
was passed through the spray nozzles. of the
air washer, with an average rise in tempera-
MECHANICAL DESIGN OF LARGE TURBO-GENERATORS
10.5
ture of G?2 deg. C. From these data the loss
in the machine was calculated to be 302 kw.
with a vohime of air of 23,000 cu. ft. per min.
The discrepancy between this value of 302
kw. and the theoretically calculated loss,
based upon the design, which averages about
340 kw., may be partly accounted for by
radiation of heat from the generator and air
ducts, and by small errors in reading the
temperature of the cooling water.
Air temperatures, both ingoing and out-
going, were read very accurately by means of
copper resistance coils wound on wooden
frames in such a manner as to obtain the
average temperature over the entire cross-
sectional area of the air ducts. These tem-
peratures were also checked by means of
thermometers and thermo-couples. Thermo-
couples placed in the armature of the gener-
ator checked the standard temperature coil
within 2 deg. C. plus or minus.
A record of the entire run is shown by the
curves in Fig. 5.
The air washer referred to is fitted with
nozzles covering a cross-sectional area of
approximately 36 sq. ft.
The results of the test indicate that the
closed circuit system of ventilation is practical
in every respect, and that the air washer
installed at present is entirely adequate as
regards size. Further tests will be made
using varying amounts of water, and it may
develop that by permitting a larger increase
of temperature in the water the rate of flow
can be reduced materially without seriously
aftecting the temperature of the generator.
Mechanical Design of Large Turbo-generators
By M. A. Savage
A-C. TURBO-GENER.'iTOR ENGINEERING DEPARTiMENT, GENERAL ELECTRIC COMPANY
The turbine generator of today is the essence of compactness. It is the result of a persistent effort to obtain
the maximum usefulness from every pound of material employed in its construction. One pound of material
in the present 5000 kw. turbo set does the work required of five pounds in the first 5000 kw. set built in this
country. Stator construction has reduced itself to the simplest form, as the requirements of rigidity, light
weight and flexibility of design are best fulfilled by such a design. However, high speeds and increased capac-
ities have introduced real difficulties in the construction of rotors. The centrifugal stresses have necessitated
the use of a solid forged rotor, and in the largest machines it has been desirable to use a three-piece rotor because
of the great length and weight of a solid one-piece forging. Ventilation, which is of prime importance in the
modern turbo generator, is briefly referred to in this article and is more fully discussed in other articles in this
issue. — Editor.
t;
'HE early turbine
generator design
practice followed
closely that of engine-
driven machines with
respect to large diam-
eter and relatively low
speed. As these gen-
erators were invaria-
bly built with salient
poles and very little
attempt was made to
direct the air through
different paths, they
were extremely large
and heavy for their output as compared with
modern machines. For example, the first
5000-kw. turbine generator built in this
country weighed 225,000 lb. ; whereas the pres-
ent .5(in(i-kw. machine weighs approximately
47,000 lb. Most of this development has been
along the lines of increased speed and better
ventilation. Increased speed has made neces-
sary the employment of better materials,
and a more careful studv into the dutv to
M. A. Savage
be performed by every element in the
machine. Better ventilation has been brought
about largely by a more complete knowledge
of the source and location of the \'arious
losses, and a more careful direction of the
air over the surfaces where these losses
occur.
Stator Frame
The stator of these large units is made up
of a number of annular " I " sections which
are held together at the outer periphery by
thick boiler plate and at the inner periphery
by rolled steel ribs. Heavy steel foot plates
along each side of the stator frame are
bolted to feet cast integral with the circular
"I" beams, as shown in Fig. 1.
This construction possesses a number of
advantages.
(1) It is extreinely stiff in the direction
in which stiffness is required.
(2) It eliminates shrinkage strains in the
castings, etc. It requires but simple and
inexpensive patterns, and also it reduces the
space required to store patterns.
106 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 2
(3) It permits of a large reduction in
weight, and therefore results in lower ship-
ping charges. It also makes possible the
assembly of machines of larger capacity at
the factory.
Fig. 1. Large Turbine Generator Stator Frame Built Up of
Annular "I" Sections with Boiler Plate on the Out-
side and Rolled Steel Ribs Inside
(4) For a given diameter, the lengths
can be increased or decreased within certain
limits by adding or subtracting one or more
annular "I" sections.
Electrical Characteristics
Due to the high rotative speed of this type
of machine, the diameters are necessarily
small and the axial length great. On the
majority of machines of General Electric
manufacture, the length is usually one and
a half or more times the diameter. This
proportion results in a relatively small
number of slots. Naturally, the coils for
such long cores are extremely heavy and
present quite a problem in handling and
assembling in the factory. The flux per pole
of this type of machine is very large, con-
sequently the number of turns to generate
the required voltage is small. In most of
the later machines the number of turns
rarely exceeds three.
Since the number of circuits in which the
armature winding can be divided is limited
by the number of poles, two circuits being
the maxinaum for two poles, and four circuits
for four poles, the current per circuit is
extremely high and this results in very large
conductors. Great care, therefore, must be
exercised in the construction of these large
conductors to keep down the edd\- current
losses (commonly called load losses). This
is done by dividing the conductors into a
number of thin, narrow strips, each strip
being insulated from its fellows by a cotton
covering. The coil thus formed is then
twisted over at the V ends so that the strips
which form the bottom layer in one slot will
form an intermediate layer in the slot in
which the other side of the coil is assembled.
This gives a partial neutralization of the
group eddies and is usually effective enough
in the ordinary' type of machine.
Fields or Rotors
The carh- type of rotors was made by
assembling punchings on a shaft. Later,
due to larger capacities, the stresses became
so great that it was necessary to use a solid
forging.
In the very largest six-pole machines the
rotor is built up of three parts; the center
part, forming the main field body, is shrunk
on and bolted to two stub shafts which form
the bearing portions of the rotor. This
construction is illustrated in Fig. 2. The
bolts holding these parts together are given
an initial tension by heating them to some
known temperature and then forcing them
home. Ample keys are also provided for
taking the torque from the prime mover. A
rotor thus built up is as stiff as a solid struc-
ture. As no metal is needed at the center
of these rotors for carrying flux, the rotor is
cored out thus greatly reducing the weight
on the bearings. This hole in the center
increases the stress in the rotor body, but as
the angular velocity is low when compared
with a rotor operating at 3G00 r.p.m. the body
stresses do not become a serious matter.
In the two-pole machine the material at
the center is needed for carrying flux. This
is especialh- tmo in the 3lU)0-r.p.m. machines
where the diameters are small and the densi-
ties are often quite high.
Radial slots are milled in these forgings to
receive the conductors.
The copper strips which form the field
turns are wound on a machine which auto-
matically changes the length of each strip for
dirtVrent radii encountered as the slots pro-
gress toward the center. The turns are
insulated with mica tape ami are then as-
sembled by feeding turn by turn into the
slots, which are insulateil by a trough of
tough insulation. After the turns are once
assembled they arc cenn-iiied tos^ethor hv
MECHANICAL DESIGN OF LARGE TURBO-GENERATORS
107
applying heat and pressure. This prevents
the coils from moving in the slots when the
machine is started and stopped. The end
windings which project from the slots are
taped with mica and asbestos and a ring
insulation put over the whole.
There are certain points of superiority of
this type of rotor which it might be well to
mention briefly.
The first, and foremost, is ruggcdness. Of
the hundreds of rotors of this type which are
constantly in operation there have been
surprisingly few failures or faults of any sort.
This is in part due to the solid structure
surrounding the winding which resists any
rapid change of flux through it and thereby
resists any sudden rise in potential in the
winding itself. Electrical failures in the rotor
due to short circuits on the armature are
therefore cxtrcmclv rare.
Ventilation
Sufficient ventilation has been and is the
major problem in the design of the large
turbine generator. More time probably has
been spent on this subject than on any other
in connection with turbine generator develop-
ment. These machines are compactly built,
the surfaces are small, and the heat loss is
enormous when compared with the size of
those surfaces. It has, therefore, been
necessary to use forced ventilation at pres-
sures and volumes which were never dreamed
of for that purpose a few years ago. The
quantity of air is based on the kilowatt loss
in the generator and is usually so appor-
tioned that the air rise through the machine
will not exceed IS or 20 degrees C. This
condition makes necessary about 100 cu.
ft. of air per minute per kilowatt loss and
in the larger machines will necessitate some
DRIVIt/G K£y
Fig. 2. Three-part Rotor Construction Employed in the Largest Six-pole Machines
Second, the solid structure of the rotor
acts as a squirrel-cage damping device
which reduces the danger of oscillations and
greatly improves the parallel operation- of
the machine.
Third, the better control of the flux dis-
tribution. Since the slots for this type of
rotor are milled from a template, it is no
longer of manufacturing advantage to have
the slots uniformly spaced. They can, there-
fore, be spaced to give the best electrical
characteristics. These rotors when operating
in an armature with three slots per pole per
phase will give very nearly a perfect sine
wave, a feature of great importance as it
reduces the secondary losses and also the
likelihood of interference where the power
lines run near telephone or telegraph lines.
Fig. 3 shows the flux distribution of one
of these rotors; Fig. 4, the resultant wave
shape on the armature. This armature has
five coils per phase per pole.
GCGOO to 70,000 cu. ft. of air per minute.
Since the spaces through which the air is to
be forced are relatively small, the velocity
becomes extremely high. The air taken
through the air gap very frequentlv reaches
12,000 ft. per min.
Probably the greatest problem of turbine
generator ventilation is to keep the rotor cool.
Attention has been previously called to the
compactness of these rotors. The heat gener-
ated in the copper is first to be carried to
the outside of the insulation, then through
the iron forming the teeth, and to the air gap,
thence dissipated into the air. This means
that the drop in temperature between the
rotor surfaces and the air has to be small
or the temperature of the rotor becomes
prohibitive.
The data which have been collected on the
subject of ventilation give the designing
engineer a feeling of certainty that the appa-
ratus will meet the requirements for which
lOS February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 2
it is designed. ^luch, however, depends upon
the quality of the cooling air.
Quality of Ventilating Air
The quality of the air used in ventilation
is a large factor in the operation of the
machine. When it is remembered that one
Fig. 3. Oscillogram of the Flux Distribution of a Five-coil-
per-phase-per-pole Rotor
of the machines takes 70,000 cu. ft. of air
every minute, even though the air may be
relatively clean, it will in time pass a sufficient
amount of dirt and foreign matter to com-
pletely clog the air passages. Air washers
have therefore become almost universal for
the larger machines. These washers, how-
ever, do not remove all the dirt from the air
and their makers do not guarantee them to
remove more than 9S per cent of it. Even
with this small percentage passing through
the generators, a number of machines which
have been in operation a year or more have
been found to contain a considerable amount
of foreign material.
Humidity of the Air
The quantity of heat removed is very little
afl'ected by the humidity of the air in contact
with it, but in passing the air through the
washer the temperature is usually lowered
from ten to fifteen degrees which means tliat
the machine will run that much cooler.
Mechanical Stresses
The stresses to which these machines are
subjected should be divided into two classes:
First, the running stresses; second, the short-
circuit stresses. The first are occasioned by
centrifugal force; the second by accidental
or intentional short circuit on the machine.
As previously mentioned, the end windings
of the field are held by steel retaining rings.
When it is stated that these retaining rings
on the higher speed machines have to hold
a force of about 3,000,000 lb., it will be
seen that both the quality of the material
and the workmanship ha\-e to be perfect.
These retaining rings are shrunk on the
centering spider at a stress greater than that
which is expected to occur in ser\'ice, so that
when the centrifugal forces are exerted on
them they will still remain tight. The next
limiting stress which occurs in the rotor is
probably at the root of the teeth. This
stress often limits the depth of the slot and,
therefore, the output which can be obtained
for a rotor of given dimensions.
Short-circuit stresses are more serious as
regards the armature winding Enormous
forces are exerted on the end portion of the
coils, outside of the slot, which tend to distort
them and if they are not sufficiently supported
crack the insulation with the result that the
machine subsequenth' breaks down. To
overcome this, wooden blocks are inserted
between coils and the coils are laced down
to steel binding bands. Great care is exer-
cised to see that the shaft and coupling bolts
have sufficiently low stresses to withstand
the shock of short circuits.
Capacity
In conclusion, it might be of interest to
state that machines of 50,000 kv-a. operat-
ing at 1200 r.p.m. have been built and are
operating satisfactorily. Machines of 3S.SS9
kv-a. at 1500 r.p.m. have also been built
and are in successful operation. In the
ISOO r.p.m. class, there arc a number of
Fifc: ':- Ui- .:.. »;ram of the \S - - - S- , -: ^f the Anr- • i r
Resulting from the Flux Diitnbution Shown in Fis. 3
machines in commercial son'ico with ratings
of 31.250 kv-a. at O.S p-f., and 33.333
kv-a. at 0.9 p-f., while at the highest speed,
viz.. 3600 r.p.m., machines of 7500 kv-a.
arc in operation, and two of 9375 kv-a. are
under construction.
109
The Behavior of Alternating-current Generators
When Charging a Transmission Line
By W. O. Morse
Alternating-current Engineering Department, General Electric Company
This article is a very interesting discussion of the effect of transmission line capacity on the behavior of
alternating-current generators. When a generator is thrown on a transmission line of proper characteristics
it is possible for the generator voltage to build up to a value considerably higher than normal without field
excitation; while an alternator having different characteristics may be entirely unable to generate under the
same conditions. Just what will take place when switching an alternator on a transmission line may be deter-
mined by plotting together the line characteristic and the alternator volt-ampere armature characteristic; it
the armature characteristic lies above the line characteristic it is probable that the alternator will charge the
line without field excitation, while it the alternator characteristic is below the line characteristic it will be
impossible for the alternator to generate without field excitation. The e£fect of negative field excitation is also
discussed. — Editor.
w
W O. Morse
^HEN a gener-
ator with a
small amount of ex-
citation is thrown on
a dead transmission
line of proper charac-
teristics, it will build
tip in voltage until it
ultimately reaches a
point of stable opera-
tion. This phenome-
non is caused by the
transmission line act-
ing as a static con-
denser and supplying
leading or magnetizing current to the alternator ;
and if this magnetizing current causes the alter-
nator voltage to build up to a value higher than
the corresponding voltage of the line, the volt-
age and current will continue to increase until
a value of current is reached at which, on ac-
count of the saturation of the generator the
line voltage and the generator voltage are
equal. This is the point of stable operation.
It is quite possible that the residual mag-
netism of the generator will be sufficient to
start the phenomenon.
The behavior of an alternator when charg-
ing a line cannot be detennined from the
generator characteristics alone; the line char-
acteristics are also involved. Another point
to be noted is that some generators when
switched on a line of given characteristics
will build up in voltage, whereas other gen-
erators switched on the same line will not.
Whether the generator builds up depends upon
the relative slopes of generator and line char-
acteristics.
Before discussing the relation of these
characteristics, their nature will be considered.
The volt-ampere charging characteristic of
a transmission line is a straight line; i.e., the
charging current is directly proportional to
the line voltage. This charging current is,
of course, leading and practically wattless.
The alternator exciting volt-ampere char-
acteristic for the armature has the shape o
the ordinary saturation curve. It is, in fact,
the saturation curve of the machine when
excited by alternating current in the arma-
ture; but, due to the different disposition of
magnetic flux in the iron circuits in this case,
the knee of the curve occurs at a higher
terminal voltage than in the case of the ordi-
nary saturation curve. Under the conditions
where the generator is excited by armature cur-
rent, all the flux is effective in inducing voltage
in the armature conductors; whereas if the
generator is excited in the usual way, there
is a certain amount of leakage flux between
poles which does not interlink with the
armature conductors and hence is not
effective flux. However, it is essential at
this point to have in mind only the shape
of the curve; i.e., it is like the ordinary sat-
uration curve.
By reference to Fig. 1, it is obvious that if
the alternator characteristic lies above the
line characteristic along the straight portion
of the former, the leading charging current
of the line, at any point in that range, will
cause a higher alternator terminal voltage
than is required to produce that current on
the line. Hence the current and voltage will
both continue to increase until, by saturation,
the "volts per ampere" of the alternator
become the same as that of the line; i.e.,
reach the point where the alternator charac-
teristic crosses the line characteristic. This
is designated "stable point" in Fig. 1.
It is equally obvious that if the alternator
characteristic falls below the line character-
istic the alternator will never build up with-
out permanent field excitation.
However, if additional alternators are
available for charging the line, the problem
may be easily solved. Suppose two duplicate
alternators, each having a characteristic as
shown in Fig. 1, were switched on the line.
This would mean that the line voltage per
110 Febfuarv, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. -2
ampere for the pair would be one half that
shown for the one alternator in Fig. 1. In
other words, the combined alternator char-
acteristic would now be a cur\^e having
ordinates one half of those of the single
alternator curve, and would be as shown by
the dotted cur\'e in Fig. 1. The line char-
\
"T"
-
~
~
~
/
■
/
^
^
-^
■
/
1^
j
'
^
[
/>-
'
>
'F
//
'Stable Toint
/
1
c"
ChAFActc ristic of"
ASinqle flitcrnator>
/
^
r
>•-
1
/
/
'^'
1
■o
j
/
f
L
/
1
"
/
1
^
-
1
/
Line Charactens
ir
1
)
1
/
/
n-
f
f
m
i
f
_
.-
■ -r"
—
1
/
^' '
1
-
,
_
1
1
1
~
?
J
^
■
E
1
1
<
^
-
1
1
^
combined Characteristic of
two Duphcjvtc Alternators
-
--
^
/
f
^
/
1
'
/
1
*
-f-
/
/
1
i
1
I
J
1
1
^.-
'
'■■
1
~
_
!
1
wattless A mpcrcs Charging Current
Fig. 1. Voltage-current Characteristics of a Single
Alternator, of Two Duplicate Alternators.
and of a Transmission Line
actcristic, however, has not been altered by
placing the additional generator on the cir-
cuit. Hence, since the combined alternator
characteristic falls below the line character-
istic, the pair of alternators will not build uj)
without permanent field excitation.
Likewise, any altcrnatoror alternators whose
combined characteristic falls below the line
characteristic will not build up without per-
manent field excitation; and, conversely, if
the combined characteristic falls above the
line characteristic, the voltage will build up,
once it is started by a momentary application
of field excitation, or by the residual mag-
netism of the alternators. And in the latter
case the voltage will rise cumulatively until,
as already slated, the characteristic of the
alternator, or alternators, bends by saturation
until it crosses the line characteristic.
It should be noted that there is lit Ik-
relation in general between the alternator
and the line characteristics, since the former
is dependent largely upon the rating of the
generator, and to some extent upon other
considerations of design; while the line
characteristic is roughly a function of its oper-
ating voltage and length. Therefore, we may
find that a generator which requires per-
manent field excitation to charge one partic-
ular line may. on another line of different
characteristics, build up in voltage and current
to values far beyond the normal capacity of
the m.achine. In such a case, the problem of
charging the line w-ithout injun,- to the genera-
tor becomes a serious matter. It is not al-
ways feasible to obtain complete control by
modifying the design of a normal generator.
From the foregoing, however, it is clear that
adding more alternators (if it is possible to
keep them in phase during the process) will
accomplish the desired result. In at least
one instance, the paralleling of two alterna-
tors for the purpose of charging a line has
been found practical. However, additional
units are not always available.
There is another scheme involving the
manipulation of the generators, which also
has been confirmed b\' experience, but which
is limited in application. It is the use of
negative field excitation for neutralizing part
of the charging current, thereby lowering
the voltage and current. Its application is
limited by the well-known fact that with
increasing negative field current a point is
soon reached where the machine slips a pole,
thereafter intensifying what it is intended to
diminish. There are two factors which fix
this limit. One is the "reaction" torque of
the machine; i.e., the torque which tends to
hold the salient jjoles in line with the armature
rotating poles when there is no excitation on
the field, and which is of the same character
as the force which tends to hold the poles of
any two magnets in the position of minimum
reluctance. In other words, it is due to thi-
salicnt-jiole construction and would not exist
in a cylindrical rotor. It is this torque which
makes it possible for a synchronous motor,
without field excitation, to carry some load
in complete synchronism.
In the present problem it operates in the
following manner: When the field is reversctl.
the machine is equivalent to a synchronous
motor which has slipjied a pole. It is oi)erat-
ing ISO electrical degrees from the nonnal
no-load i)osition. In other words, it is
operating in an unstable position, the least
tlisplacement from which will protluce grt-ater
synchronizing torque tending to produce still
further dis])lacement. This involves a shift
THE BEHAVIOR OF ALTERNATIXO-CURRENT GENERATORS
111
of the flux relative to the pole and is therefore
opposed by the "reaction" torque. Large
negative field current means greater svnchro-
nizing torque. It also means less '"reaction"
torque because the voltage, and therefore the
flux, is decreased. When equality is reached
between these two opposing forces, the sny-
chronizing torque pulls the rotor into the
normal synchronous position and reverses the
action; i.e., tends to increase the voltage
instead of to decrease it.
The other factor which may limit the value
of negative fleld current that can be applied is
shown in Fig. 1 . The two characteristics inter-
sect at two points; viz., zero and the "stable
point." The application of negative field
current operates to shift the alternator char-
acteristic to the right, as shown in Fig. 2
(since the voltage induced in the generator
by the wattless charging current a is exactly
neutralized; i.e., reduced to zero by the equiv-
alent negative field excitation Oi). This
causes the two intersections to approach each
other, and ultimately they meet; i.e., the
line characteristic is tangent to the alternator
characteristic at this point. Any further
increase in negative field will cause the ma-
chine to reverse, since now the alternator char-
acteristic will be entirely below the line
characteristic. This point can be predeter-
mined with a fair degree of accuracy if the
line and the alternator characteristics are
l<nown, because the negative field current Oi
is equal in ampere-turns to a. In this partic-
ular case a negative field excitation Qi would
reduce the voltage corresponding to the stable
point to E, Fig. 2; i.e., to the neighborhood
of norm.al magnetic densities.
Another interesting possibility is suggested
by a study of Fig. 2. If it is required to hold
the voltage at a still lower value, say e, it is
obvious from Fig. 2 that a negative field cur-
rent equal to two thirds of ai would cause the
new displaced alternator characteristic shown
in part by the heavy dotted line, to cross the
line characteristic at two points; one at volt-
age e and another at voltage ei. At ei the
operation would be stable. At e the line and
alternator each require the same current, but
the condition is not stable. The least change
of voltage, either way, causes a change in
current which further augments the voltage
change. Yet it may be possible, under favor-
able conditions, to operate at this point b\^
the use of a voltage regulator. The line volt-
age might be started by a momentary appli-
cation of positive field excitation, or by the
residual magnetism of the alternator. Then
the generator could be quickh- thrown to the
voltage regulator which would attempt to
hold the generator voltage at the value e by
applying negative field current. It is of
course problematical whether the voltage
regulator, working with an exciter, could
respond quickh' enough.
-
—
-
-
~
5Ubl« Point*
?
^
I
^
^/
,-A
H.
f'-'
0
t
' ,''A\
-0
(7,
c
/a
/
« No
ma G«n, Volts
r°A',
.c
.
/
/
•g
,"/
/
*
/
/
0
.V
mi
7
7
"o" represents 1
maArmum possible
._ o_.
(V
'--^
t,
^oqativefiflldencitat'on
'f
1
1 1 1 1
_
c
^> .
• n I 1
g
t
Line
^Iharactanst.o
k3i
' t
L
■ i
-i-L
Wattless flmp
a n
t U
Charainq
Currant
■ —
-0|1±
rf/
+ -
/ u
/ 1
2 ^
r 1
I T
T
~f
1 '
"/ "
~ t
' t 7
-
/ /
/
V
z
z
c
T-.^ -=
^ ^
---
i" 1
1
1
Fig, 2. Diagram showing the Effect of
Negative Field Excitation
Another important factor to be considered is
the voltage at which it is necessary to charge
the line; for while the charging kilovolt-
amperes required at about normal voltage may
not be larger than the generator rating, the line
charging conditions may be such as to require
a reasonably small kilovolt-ampere output, but
at a low voltage and consequently high current :
this is necessary in order to give normal voltage
at the other end of the line, the voltage rising
with the length of the line from the generator.
This ma\- be an impossible case to handle
with the generator alone, but it may of course
be met by the use of some transforming ap-
paratus to obtain the proper ratio of volts
to amperes, or by temporarily changing the
connection of the generator. These artifices,
however, are subject to the obvious limita-
tions of complicated switching and high
cost.
A scheme which has already been used for
relieving the generator from excessive charg-
ing current is the use of shunt reactors across
112 February, 1920
GEXER.\L ELECTRIC REVIEW
Vol. XXIII, Xo. 2
the line, which absorb a portion of the leading
charging current. For normal operation of
the system, the reactors are switched off the
line. In order to avoid severe line disturb-
ances which are incident to switching the
reactors on or off the line, the use of saturated
core reactors has been suggested. These react-
ors would be designed with a comparatively
high reactance. When the cores are saturated
bv means of direct-current excitation, the
reactance is reduced to a relatively low value,
thus permitting a share of the charging cur-
rent of the line to be absorbed. When the
load is switched on the line, the direct-current
excitation would be decreased gradually, thus
minimizing the voltage fluctuations. Whether
this device would be practicable in an}' par-
ticular case would have to be determined by
a consideration of its cost and of the advan-
tages to be gained.
Synchronous Motors
By W. T. Berkshire
Alternwting-current Engineering Dep.\rtment, Gener.\l Electric Company
Because it has been the usual practice to install induction motors wherever power was required from an
alternating-current circuit, without regard to their collective effect on the power-factor of the circuit, many
systems are today operating at low power-factor and consequently low efficiency. The judicious employment
of synchronous motors in combination with the induction motors will correct the fault and thereby benefit
both the distributing companies and the consumers. For the information of the latter, the following article
describes the characteristics and qualifications of the synchronous motor for such applications. — Editor.
T'
W. T Berkshire
'HE desirability of
using the syn-
chronous motor as a
synchronous con-
denser, and as the
motor of a motor-
generator set or fre-
qticncy converter, for
neutralizing low pow-
er-factor has been
recognized for a long
time by all users of
electric power.
Within recent years,
however, considerable
impetus has been given to the general applica-
tion of synchronous motor drive to all classes
of industr\' because of certain advantages it
has over other forms of (Irive. Heretofore,
the induction motor has been more generally
applied to all classes of industrial drive.
For various classes of service the synchro-
nous motor has several advantages over the
induction motor, the recognition of which
has resulted in an ever increasing demand for
motors of the former type. These advantages
consist in better efficiency and power-factor,
and, particularly for low-speed machines, lower
first cost.
Efficiency
The efficieticy of the synchronous motor is
generally higher than that of the induction
motor even when operating at leading power-
factors as low as O.S. Particularlv is this true
of the more modem synchronous motors
designed for unity power-factor operation
whose high efficiency is practically the same
from full load to half load and is only slightly
lower even at one quarter load.
Power-factor
The synchronous motor can be designed to
operate at either unity or any leading power-
factor, thus improving the power-factor of the
system. With normal field excitation these
machines will continue to improve the power-
factor when underloaded. In this respect the
induction motor is always at a disadvantage;
its power-factor is always lagging and al-
though this power-factor may be high at
full load it becomes rapidly lower at partial
loads; consequently an underloaded induction
motor further impairs the power-factor of a
system.
Dependability
From the standpoint of dependability of
operation, the synchronous motor has a
mechanical advantage over the induction
motor by reason of its larger air gap which
varies from five to eight times that of the
induction motor. The operating character-
istics of an induction motor may be seriously
impaired by a slight change in air gap
due to a little wear in the bearings. Due
to the larger air gap of the synchronous
motor, the same change on account of bearing
wear will not materially affect its operating
characteristics.
SYNCHRONOUS MOTORS
113
Starting Ability
In making a comparison of the relative
starting ability of normally designed squirrel-
cage synchronous and induction motors, the
following points must be understood :
First: If a motor has a high initial starting
torque it must also have a low pull-in torque,
and vice versa. The high-resistance squirrel-
cage winding, which is required for high
initial starting torque, produces low pull-in
torque; whereas the low-resistance squirrel-
cage winding, which is required for high pull-
in torque, produces low initial starting torque.
Second: The induction motor cannot use a
high-resistance squirrel-cage winding on ac-
count of resulting high losses and low effi-
ciency under normal operation.
Third: The synchronous motor can use a
high-resistance squirrel-cage winding because,
when operating in synchronism, there is prac-
tically no loss in this winding.
Therefore in cases where high initial start-
ing with reasonablv low pull-in torque is
required, the synchronous motor has the
distinct advantage that the high-resistance
squirrel-cage winding, with its accompanying
high starting torque and low kilovolt-ampere
input, can be utilized.
In cases where the required starting and
pull-in torques are about equal, but of a com-
paratively low value, the synchronous motor
still has the advantage.
If, however, a high pull-in torque with a
correspondingly low starting torque is re-
quired, then the induction motor has a slight
advantage.
There are a few classes of service requiring
both high initial starting and high pull-in
torque. In such cases the double squirrel cage
or other means is used to obtain the required
torque; the double squirrel cage is also
used on some of the larger induction motors.
The starting of such loads by the synchronous
motor, however, is usually attended by high
current being drawn from the line. This is
often objectionable both from the standpoint
of the power company and the power con-
sumer. This starting current can be more
readily controlled by the use of the slip-ring
induction motor, consequently this type of
motor is better for driving loads requiring
both high starting and pull-in torque.
Limitations
Owing to certain starting torque or speed
requirements, there are three classes of
service for which the normally designed
synchronous motor is not suitable for direct
drive. These are tor service requiring the
motor to start under full load; service requir-
ing variable speed; and service requiring
frequent reversals in the direction of rota-
tion or requiring frequent starting and
stopping.
The first class includes flour mills, grain
elevators, or heavy line shafting where the
torque required to overcome the static fric-
tion equals and often exceeds the full-load
torque. In such cases the synchronous motor
should be directly connected to the shaft .
through a clutch, thus permitting the starting
and synchronizing of the motor before the
load is applied.
Where the service requires a variable speed
some mechanical means must be provided to
obtain such variation.
Application
Sj'nchronous motors have been successfully
applied for driving the following:
Motor-generator Sets
Frequency Converters
Air Compressors
Ammonia Compressors
Pulp Grinders
Jordans
Stone Crushers
Centrifugal Pumps
Plunger Pumps
Screw Pumps
Blowers
Fans
Convevors
Tube Mills
Flour Mills
Rubber Mills
Cement Mills
Line Shafting
Steel and Copper Rolls and
For Operationg and Synchronous Condensers
During the year 1918 alone, the General
Electric Company built over 500 synchronous
motors for various classes of service, having an
aggregate capacity of over 300,000 horse
power. This number does not include a large
number of synchronous condensers. Of this
number over 200 motors, having an aggregate
capacity of more than 90,000 horse power, were
built for air compressor drive alone.
The study of the synchronous motor, with
particular reference to its application to
various forms of industrial drive, has resulted
in many improvements in design that have
increased the efficiency of the starting ele-
ments, thus widening the field of application.
Inasmuch as the torque required at starting
and pull-in varies with the class of service,
whether it be for driving an air or ammonia
compressor, pump, cnisher, grinder, line-
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 2
SYNCHRONOUS .MOTORS
shattin^', motor-generator, etc.. great care is
given to the design of each motor in order that
these various requircm.ents may be fully met
in each ]iarticular case.
The squirrel-cage winding, in each case, is
\-ery carefully designed, both as to the ma-
terials used and their mechanical arrange-
ment, to meet the starting torque require-
ment for the class of sen-ice for which the
motor is to be used.
The use of a fractional instead of an in-
tegral number of stator slots per pole in the
design of synchronous motors eliminates the
possibility of dead points during starting.
It furthermore insures the maximum obtain-
able starting torque for every position of
the rotor; i.e., the torque will not be low for
one position of the rotor and high for another,
but will be uniform and a maximum.
Torque
The solid curve in Fig. 5 represents the
torque required during the starting of the
average synchronous condenser, motor-gen-
erator set. or frequency converter. The
dotted cur\-e represents the torque developed
by a synchronous motor normally designed
for this ser\nce when starting at reduced
voltage from a compensator as an induction
motor, i.e., with no excitation on the field.
Similarly Fig. 6 shows, by a solid cur\-e, the
torque required bv an average centrifugal
pump or blower during starting, the dotted
cur\-e again representing the torque developed
by the synchronous motor normally designed
particularly for this type of service. The
shifting of the m.aximum torque point of the
motor to different positions during starting,
as required by the various classes of ser\nce,
is accomplished by proper design as already
described.
^30
u 10
^ 0
^_J„.. — •-— >~ .. »^^_^
s "'"^'^^
—^t
10 10 30 4€ 50 60 70 30 90 100
Per cent Si^nchronism
Fig. 5. Dotted Curve Represents the Starting Torque of a
Synchronous Motor Designed for Starting Torque
Service Represented by the Full Curve
Referring again to Figs. 5 and 6, it will be
noted that in one case the maximum torque
required occurs near the initial start; in the
other, it occurs near the synchronous speed.
It will be further noted that there is a point
where the curve representing the torque
de\-cloped by the motor crosses that of the
required torque. It is at this jjoint that the
machine reaches constant speed when operat-
ing as an induction motor. This point
usually comes at approximately 95 per cent
BO
-
,
^'^
\
60
/
V
1
\
1
^50
,■'
_1
y
^5
■<l40
..'f
i
'-'
/\
1^0
^
--<
'A
,/
''
/
zo
y
^
— 1
y
^
10
V
<A
<
'
r
10 ZO 30 JfO so iO 70 90 90 WO
Per cent Sunchrvntsm
Fig. 6. Curves Corresponding to Those in Fig. 5, but
Applying to a Load of Different Character
synchronous speed and it is here that the
field excitation is applied, the motor thrown
from the compensator directly on the line
and the load pulled into synchronous speed;
i.e., it is at this point that the motor begins
to operate as a synchronous instead of an
induction machine.
The torque cun-es shown in Figs. 1 and 2
are representative of those of the various
loads to which synchronous motors are direct
connected or direct coupled. For properly
by-passed air and ammonia compressors, the
starting torque varies from l.j to 3.5 per cent
and pull-in torque from 15 to 25 per cent.
Some classes of pumps may also have similar
starting characteristics. A pulp grinder may
require a starting torque varying from 30
to 60 or 70 per cent of normal, the pull-in
torque being approximately 15 to 25 per cent.
All these starting requirements can be m_et
with normally designed synchronous motors,
starting at reduced voltage from a tap on the
starting compensator.
For special sen-ice, synchronous motors
have been built to develop 150 per cent
normal torque at start and 75 per cent at
pull-in, but such motors are of abnorm.al
design.
Pull-out or Break-down Torque
The "pull-out" torque is an important
factor in a synchronous motor. It \-aries in
116 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 2
SYNCHRONOUS MOTORS
Ii;
different motors, of normal design, from 150
to 300 per cent of normal torque, depending
on the particular class of service to which
the motors are applied. Statements regard-
ing the pull-out torque developed by a motor
based on formula instead of fact may lead
into difficulty where high pull-out torque
is required. It is important, therefore, that
the method of determining this torque should
be understood.
This torque can be easily determined from
standard test curves. In Fig. 1 1 are shown the
no-load saturation curve and short-circuit
characteristics; i.e., the "synchronous im-
pedance" curve of a three-phase synchronous
motor. The power delivered by the three-
phase motor at "break-down;" i.e., the pull-
out capacity at any terminal voltage e and
any field current / is approximately
P max =
\/ 3 e to
1000
kilowatts
where to is the current per armature terminal
corresponding to field current / on the syn-
chronousimpedance curve. Therefore, the pull-
out capacity varies directly as the terininal volt-
age and also directly as the field excitation.
It follows that at a constant normal terminal
voltage E„ and a field excitation Fi which cor-
responds to normal armature current 7„ on the
synchronous impedance curve, the pull-out ca-
pacity will be in kilowatts numerically equal to
the normal kilovolt-ampere rating.
Hence if Fa is the full-load excitation, the
pull-out capacity for this excitation will be
V 3 E„ Iq
1000
that is, it is equal to
~ =— times normal kv-a. rating.
In rx
F
The short-circuit ratio is A =^^ where
Fi
F = field current corresponding to normal
voltage on the saturation curve and
Fi = field current corresponding to normal
armature current on the synchronous
impedance curve.
F2 = Fi\/l-|-/l^ for 1.0 power-factor or
= F,\/l + 1.3 7v -f 1.2 /v2 (approximately)
for O.S power-factor leading.
Therefore if the short-circuit ratio K is
known, the pull-out capacity in kilowatts
is approximately equal to the rated kilovolt-
ampcrcs multiplied by
Vl-fA.'^ for a unity, or 1.0 power- factor
motor or by
Vl-f 1.3 A'-f-1.2 A'- for a O.S jiower-factor
motor.
\/
4|----
Fig. U.
r, r />
Saturation and Synchronous Impedance Curves
For example, a 200-kv-a. synchronous
motor having a short-circuit ratio of 1.41
would have a pull-out capacity of
200 X\/l-l- 1.412 = 200X1.73 = 347 ^^^
If this were a 200-kv-a. 0.8 p-f. motor its
jJuU-out capacity would be
200X^1-1-1.3X1.41-1-1.2X1.412 = 200X2.29
= 458 kw. While this method is approxi-
mate, its results are sufficiently accurate for
checking pull-out torque guarantees.
The following fonnula, which may also
be used in determining the pull-out torque, is
slightly more accurate, providing it is correctly
applied, but it is more complicated. The pull-
out torque in "synchronous watts" is
<
E"- cos\ mf-tair
£.
— ian ''— I where
>
X = synchronous reactance per leg of the
armature winding
R = resistance per leg of the armature winding
Z= VF2-|-A'2 = synchionous impedance
£i = applied voltage per leg of the arma-
ture winding
£2 = nominal e.m.f. per leg of the arma-
ture winding.
US February. Hi2 )
GEXERAL ELECTRIC REVIEW
Vol. XXIII. Xn. 2
•= u
S i
E <
c .
QS
£ =
SYNCHRONOUS MOTORS
11<»
By an incorrect use of this formula, the self-
inductive or leakage reactance of the armature
may be substituted for the value of .Y, instead
of the synchronous reactance, and the pull-
out torque may therefore appear to be from
400 to 500 per cent when it really is only
from 125 to 150 per cent of normal torque,
since the value of the synchronous reactance
is usvially from four to fi\-e times that of the
leakage reactance.
Unnecessarily hi^h pull-out torque is not
desirable as it usually impairs other operating
characteristics of the motor.
Power-Factor
The disadvantaj^es of low power-factor are
generally known to most users of electric
power. Low power-factor means unneces-
sarily large and more expensive generators and
exciters with poor efficiency due to increased
losses, increased cost of station, transforming
and switching cquijjment, and increased cost of
transmission line and distributing transform.-
ers. Furtherm.ore, it may mean underloaded
prime movers with decreased prime-mover
efficiency. It results in poor voltage regula-
tion. Because of these disadvantages, power
companies are already beginning to charge
higher power rates where the power-factor
of the consumer's circuit is lower than a
certain limit, consequently low power-factor
means increased m..otor operating costs.
Low power-factor is due to the lagging cur-
rent drawn from the line by inductive loads
such as induction motors, series and multiple
arc lamps, or even transformers sujjplying in-
candescent lamps.
For any given mechanical load on a syn-
chronous motor the current in the armature
is a minimum at a certain field excitation,
this current being neither lagging nor leading,
and the motor constitutes a unity power-
factor load on the line. In this case all the
current in the motor is energy current whose
function is to drive the load and supply the
losses of the motor. Now, if the field excita-
tion is increased the motor will take a leading
current from the line. This leading current
may be separated into two components: first,
an energy component as described, and,
second, a magnetizing current; i.e., a current
that tends to magnetize the generator fields
and is the so-called "wattless leading compon-
ent." This wattless leading component of the
synchronous motor may be used to neutralize
an equal amount of wattless lagging ■ com-
ponent clue to induction load on the system.
The excitation mav be further increased until
the motor current consists almost wholly o
the wattless leading component; the energy
component being only sufficient to supply the
losses of the motor, none being available for
driving mechanical load. The m.otor is then
operating purely as a synchronous condenser.
Fig. 16. Diagram showing the Improvement in Power-factor
by the Addition of a 1000-kw. Unity Power-factor Motor
Imjjrovement in power-factor can, there-
fore, be effected by the a^jplication of the
synchronous machine in either one of three
ways:
(a) As a unity power-factor m.otor; i.e., all
the input being used for energy.
(b) As a power-factor motor; i.e., part of the
input being used for energy and part for
furnishing wattless leading current to the
line.
c) As a synchronous condenser; i.e., all the
input being used to supply wattless leading
current to the line.
The following are examples of each of these
uses :
Assume a load of 1200 kw. at 0.(5 p-f. or
2000 kv-a. What will be the effect of adding a
1000-kv-a. synchronous motor to the system
(a), at unity power-factor; (h) at, say, O.S
power-factor, and (c) at zero power-factor,
i.e., as a synchronous condenser?'
(a) As a Unity Power-factor Motor
Referring to Fig. 10, the initial load of 1200
kw. at 0.6 p-f. is represented by the line AC,
the kv-a. being ,7,, = 2000 represented bv the
O.u
line AB.
The wattless lagging kilovolt-amperes rep-
resented by
BC = \/AB^-Aa = \/2()00= - 12002 = 1600 kv-a.
Now by adding a l()()0-kv-a. 1.0 p-f. load.
120 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 2
represented by the line CD, the total load
becomes
AC+CD = AD or 1200 + 1000 = 2200.
The wattless lagging kilovolt-amperes remain
unchanged; i.e., 1600 k\--&.=BC = DF. The
total kilovolt-amperes required, therefore, to
deliver 2200 kw. energy load will be
\/ A D'-+DI^ = V 2200'- + 1600^ = 27 20 kv-a. :
■^200
and the new power-factor will be ^hrr,-, =0.S1.
2 1 20
With the addition of this unity power-factor
motor it will be noted that two important
things are accomplished, viz:
First, the power-factor is raised from 0.6 to
0.81. Second, the energy load has been in-
creased from 1200 to 2200 kw., or 83^2 per cent,
the wattless lagging kilovolt-amperes = 1000;
and the total kilovolt-amperes required to
carr\- this load will be
AF = \/aD--^DF^ = \/2000= -1-1000= = 223.5 kv-a.
The new power-factor, after the motor is
9000
added, will then be ^;^— = 0.895 instead of 0.6.
It will be noted also that the energ\- load
has been increased 66^ per cent with an in-
crease of only 11^4 per cent in the generator
capacity.
(c) As a Zero Power-factor Motor; i.e., a Syn-
chronous Condenser
In this case the problem is: How much
will the power-factor of the system be in-
Fig. 17. Diagram showing the Improvement in
Power-factor by the Addition of a 1000-kv-a.
0 8 Power-factor Motor
/^JffHW.
Fig 18. Diagram showing the Improvement in Power-factor
by the Addition of a Zero Power-factor Motor
or Synchronous Condenser
with an increase of only 36 per cent in the gen-
erator capacity; i.e., from 2000 to 2720 kv-a.
(6) As a 0.8 Power-factor Motor
Referring to Fig. 17, the lines .4C, AB, and
BC represent the initial energi,- load, total
kilovolt-amperes, and wattless lagging kilo-
volt-amperes respectively the same as in Fig.
16. A lOOO-kv-a., 0.8 p-f. motor is now to be
added to the system. This motor will deliver
1000X0.8 = 800 kw. energy load represented
by the line MF = CD. It will also deliver a
wattless leading kilovolt-ampcre equal to
BM = x/bF'-MP = \/l000=-800- = 600 .
This 600 wattless leading kilovolt-ampere
will neutralize an equal amount of the 1600
wattless lagging kilovolt-amperes so that the
wattless lagging kilovolt-amperes after the
motor is added will be
DF = BC- BM = KiOO - 600 = 1000 kv-a.
The energy load will then be
AD = AC+CD = 1200 -f 800 = 2000 kw. ;
creased by the addition of a 1000-kv-a.
synchronous condenser, it being the intention
to keep the same energy load ? Fig. IS, as Fig.
16 and 17, again shows the initial energy load,
total kilovolt-amperes, and wattless lag^ging
kilovolt-amperes by the lines AC, AB, and
BC. The 1000-kv-a. synchronous condenser
(neglecting the energy required to supply its
losses which may be only 3 or 4 per cent) will
supply 1000 wattless leading kilovolt-amperes
= BM. The total wattless lagging kilovolt-
amperes will then become BC—BSI = CM or
1600- 1000 = OOO kv-a. and the total kilovolt-
amperes required to carry the same load will
become
\/AB'-\-CM^ = AM or \ 1200«-f-600' = i:Ml.6 kv-a.
The new power-factor will then be , " , ^.
1341.0
= 0.80.5.
Not only is the power-factor increased from
0.() to 0.89.5, but it will be noted that, by the
SYNCHRONOUS MOTORS
121
addition of this synchronous condenser, a
generator of only about two thirds the former
capacity is required to carry the energy load
with a corresponding reduction in the ca-
pacity of all transforming and switching equip-
ment, transmission lines, and exciters.
Unity Power-factor (1.0 p-f.) Synchronous Motor
Particular attention is drawn to the fact
that the highest efficiency, in the driving of
mechanical loads, can be obtained by the use
of the unity power-factor motor. It requires
less exciter capacity and has a considerably
lower cost than a synchronous motor designed
to operate at leading power-factors. Such a
motor, of course, constitutes a non-inductive
load on the line (i.e., it does not furnish any
wattless leading kilovolt-amperes when oper-
ating at full load), but as shown in Fig. 16, it
does improve the power-factor of the system
to a considerable extent. If, however, it is
under-loaded, with full-load excitation on the
field, it always operates at a leading power-
factor furnishing a certain amount of wattless
leading kilovolt-amperes to the line depending
on the degree of underloading. Fig. 19 shows
the leading power-factor at which a nonnally
designed unity power-factor synchronous
motor will operate at partial loads with the
normal full-load field excitation constantly
maintained. It will be noted that at 50 per
cent of the normal kilovolt-amperes input the
machine will operate at 0.73 p-f. leading, and
at about 33 per cent normal kilovolt-amperes
input it will operate at zero power-factor; i.e.,
purely as a synchronous condenser. In the
average motor this value of kilovolt-amperes
input at zero power-factor will vary from 20
to 33 per cent.
Flywheel Effect
Sjmchronous motors driving air or ammonia
compressors or other reciprocating apparatus
embody, in the most essential respects, the
same factors as engine-driven generators, so
far as synchronous operation is concerned.
Whether a synchronous machine is operating
as a generator or motor, it behaves in the same
manner as regards stability. That is, if its
rotor is pushed or pulled away from the stable
position in rotation — whether this be at no
load or full load — there is a force opposing
* A complete discussion of this phase of the subject is given in
the article, "Oscillating Frequency of Two Dissimilar Synchro-
nous Machines," by R. E. Doherty, in this issue.
such displacement that is practically propor-
tional to the displacement. Hence, if a vari-
able torque is imposed upon the shaft, either
positive or negative; i.e., either generator or
motor operation, there will exist the tendency
toward instability.
1 —
7.0
—
^
^
' — '
'
0.9
^
.^
/
0.8
/
/
0.7
—
-■
--
■-
--1
-■
--
■-
--
7
,/
V-
ko.6
/
/
%'■>
/
1.0.4
0.3
1
0.2
0.1
0 10 20 JO 40 SO 60 70 SO 90 100
■Per cent Nor ma J MVA . Inpu t
Fig. 19. Curve showing the Leading Power-factor at Which a
Normally Designed Unity Power-factor Motor Will
Operate at Partial Load with Normal
Full-load Field Excitation
Now, the air or ammonia compressor is a
reciprocating machine of relatively large mass,
and therefore has a turning effort curve which
in a general way resembles that of reciprocat-
ing engines. The variations are periodic and
of sufficient magnitude to cause the motor
to oscillate if proper precautions are not
taken.
The precautions consist in the use of fly-
wheel weight stifficient to limit the periodic
angular deviation to 33^^ electrical degrees, in
either direction from the position of uniform
rotation, and to fix the natural oscillating fre-
quency at a safe distance from the frequency
of the forced impulses of the compressor. The
two most important impulses have frequencies
equal to the revolutions and to the strokes of
the compressor and it is considered desirable
to keep the natural frequency of the motor
away from the revolutions or strokes by 20
per cent.*
122 February. 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 2
Magnetomotive-force Diagram of the
Synchronous Motor
By E. S. Henningsen
Alternatixg-cirrent Engineering Department, General Electric Company
The author shows that the magnetomotive-force diagram of the synchronous motor can be used to
quickly determine, with a sufficient degree of accuracy for most cases, the operating characteristics of a given
machine; for instance, the excitation required for any load, the phase characteristic curves, the pull-oat
torque and the wattless component for any excitation. The light errors in the method are due to the fact
that the saturation curve is assumed to be a straight line. A concrete case is assumed as an example, and
the method of applying the diagram to the determination of the several characteristics is worked out for
illustration. — Editor.
ALL of the operat-
ing characteris-
tics of a synchronous
motor that can be
determined from the
complete excitation
calculation can also
be determined with
reasonable accuracy,
and in a fraction of
the time, from the
m agnetomotive - force
diagram. That is,
E. S Henningsen kuOWitlg thC field CX-
citation required for
norm.al voltage open circuit and that for
normal current on short circuit, we can at
once determine approximately the excitation
required for any load, the phase characteristic
3C 40
Field Amperes
Fig 1. Saturation and Synchronous Impeiance Curves
Arm.aturc Reaction = 17!)() ampere turns
Armature Reactance =17.2 per cent
Armature Resistance = 1.0 per cent
Field Turns =220
cunes, the break out capacity, and the
wattless kilovolt-amperes for any excitation
by combining these values in the proper
phase relation. The method is of course not
exact, since it assumes that the saturation
curve is a straight line. Its inaccuracy there-
Fig. 2. Excitation Diagram, Unity Power-factor
fore \'aries on different machines depending
upon the reactance and degree of saturation.
However, the magnetomotive-force diagram
is a very useful tool because of the case and
speed with which the various characteristics
can be obtained.
For cxam])le. consider the field excitation
required at unity and also at O.S p.-f. leading
by a 125-kv-a. motor, the saturation and
synchronous imjiedance curves of which are
given in Fig. 1 . This also gives the reactance,
armature reaction and resistance, and field
turns so that the complete excitation diagram
as well as the magnetomotive- force diagram
may be constructed.
Fig. 2 shows the excitation diagram for
unity power-factor. The line current is /. the
temiinal voltage E, the internal induced volt-
age E, and the angle between / and E.\ is ^i.
From inspection Ei = [E — Ir)+jlx
where Ir is the resistance drop and Ix the
self-inductive reactance drop through tlie
armature windings. To produce the temiinal
\oltage E requires a flux <p and a field exci-
tation F. The internal voltage /ti requires a
MAGNETOMOTIVE-FORCE DIACRAM OF THE SYNCHRONOUS MOTOR 123
liux 0, and field excitation Fi. This field
excitation (Fi) can be obtained from the
saturation cun^e corresponding to the A-oltage
/ii. The armature reaction ampere-turns is
.4. Hence the full-load excitation Fo = (-4 cos di)
+j {Fi + A sinei). Calculating Fo from the
data given in Fig. 1 gives 47.50 ampere-turns.
■3..
No load Nor ma J Vo/taqe A t
Fig 3. Magnetomotive-force Diagram,
Unity Power-factor
of the line current I by the angled. Calculat-
ing Fq for cos 8 = 0.8 gives the load excitation
as 6240 ampere-turn's. Fig. 5 gives the
magnetomotive-force diagram for the same
condition, .45. The synchronous impedance
ampqre-turns, AB, has the same value as in
Fig. .3, but is of course di.splaced from its
No Load Normal VoJta^e AT. A
Fig. 5.
Magnetomotive-force Diagram,
Power-factor =Cos t)
Fig. o shows the magnetomotive-force
diagram for unity power-factor load. Since
increased saturation under load is to be
neglected in the m-agnetom.otive-force dia-
gram, it is not necessary to separate the
synchronous impedance into armature self-
inductive reactance (which is treated as a
voltage in the excitation diagram) and arm-
ature reaction. The field excitation required
for normal voltage open circuit is 0.4, the
field excitation for normal current from the
synchronous impedance test is AB. and the
field excitation for full load is OB. Con-
structing the diagram from, the data given
in Fig. 1 shows OB to be 4740 ampere-turns
or practically the same as the excitation
calculation. This is to be expected of course
because the increase in saturation under load
is slight for unity power-factor and the
magnetomotive-force diagram diff'ers from the
Fig. 4. Excitation Diagram. Power-factor =Cos (?
excitation diagram only in the neglect of this
factor.
Fig. 4 shows the excitation diagram for
normal kilovolt-amperes at a power-factor of
cos 6. The vectors are the same as in Fig. 2,
except that the term.inal voltage E is ahead
Fig 6- Magnetomotive-force Diagram for
Load Phase Characteristics
unity power-factor position in Fig. 2 by the
angle cos~^ = 0.8. Solving the triangle of
Fig. 5 by graphical construction, using the
values from. Fig. 1. gives F,j = .5,s40 ampere-
turns or about seven ]3er cent less than by
the excitation calculation. For normal kilo-
\'olt-amperes zero power-factor, the excitation
required is G940 ampere-turns by the excita-
tion diagram, and (i450 ampere-turns by the
magnetomotive-force diagram. A very close
approxim.ation can therefore be obtained
if, when constructing the magnetom.otive-
force diagram for leading power-factors, the
ampere-turns required for normal voltage no
load are increased from 10 to l.i per cent to
allow for increased saturation.
To obtain the phase characteristic cur\'es o f
a synchronous motor, it is only necessary to
construct a m.agnetomotive-force diagram
such as is shown in Fig. (i. The field excitation
required for no-load norm^al voltage from the
saturation curs'e is 0.4. The synchronous
impedance ampere-turns corresponding to
the line current at unity power-factor is AB
for the load considered. The required field
excitation is then OB. .^t any other value of
field current such as OC the line current is
AC/AB times the current corresponding to
AB. Values of field less than OB mean an
underexcited motor, therefore lagging arm-
124 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 2
ature current, and values greater than OB,
leading current.
Fig. 6 also gives values of wattless kilovolt-
amperes available for power-factor correction.
For instance, suppose AB represents the
synchronous impedance ampere-turns for
Fig. 7. Magnetomotive-force Diagram, Constant Field
Excitation, Increasing Load
normal load unity power-factor armature
current, and that the field excitation on the
motor is OD. There will then be a leading
wattless current in the armature equal to
DB/AB times normal. Or looked at another
way, if OD is the total field excitation, and 0.4
the field excitation corresponding to no-load
normal voltage, the resultant AD must be
balanced by a current in the armature large
enough to require the field excitation AD to
force it through the synchronous impedance.
Hence, by finding the armature current cor-
responding to the field excitation AD on the
synchronous impedance cur\'e, and measuring
the angle d, cosine of which is the power-
factor, the wattless kilovolt-ampercs can be
determined.
There are several methods of determing the
break-out capacity of a synchronous motor;
i.e., the load at which the motor will breaJc
out of synchronism. There is but a single
principle however; viz., that a synchronous
motor drops out of step when the watt-com-
ponent of the synchronous impedance ampere-
turns equals the ampere-tums on the field.
Neglecting saturation, the break-out capacity
varies directly with the field excitation. Fig. 7
shows the magnetomotive-force diagram,
assuming constant excitation and var\'ing
load until the break-out point is reached.
The full-load unity power-factor diagram is
OAB identical with Fig, 3, If we assume the
field excitation to be constant, Fo must for
any load terminate in the arc FDB of radius
OB. Suppose the watt load to be doubled.
Then the watt-component of the synchronous
impedance ampere-tums, which was AB for
normal load, becomes .4C = 2 ^4JS for double
load. Since Fo is constant there must be
a wattless component of current corresponding
to DC, or the magnetomotive-force diagram
becomes OAD and the power-factor cos 0.
Since by inspection, OD, the field ampere-
tums is greater than .4C the watt-component
of the synchronous impedance ampere-tums,
the motor will still sta>' in step according to
the principle stated above. If now the watt
load is still further increased until the watt
component of the synchronous impedance
ampere-turns equals AE, the motor will
break-out of step, because at this point AE
equals the field excitation OF. The power-
factor at break-out is cos 0. lagging, and the
wattless component of armature current is
proportional to EF. The value of load in
kilowatts at which break -out occurs is AE/AB
times normal load. The motor kilovolt-
amperes at break-out is the kilowatts break-
out load divided by cos 6. Apphnng this
diagram to the motor previously considered
gives the break -out capacity to be 250 kw. and
the power-factor 0,70 lagging.
125
Oscillating Frequency of Two Dissimilar
Synchronous Machines
By R. E. DoHERTY
Alternating-current Engineering Department, General Electric Company
In operating two or more alternating-current generators driven by reciprocating engines it is important
to avoid a condition that will produce "hunting," or a periodic oscillation of the revolving elements ahead of
and behind the normal position. In cases where the natural oscillating frequency of the alternator is near the
periodic variation in the torque of the driving engine, hunting may occur in such proportion as to throw the
generators out of phase, or produce violent flickering of lamps or other trouble on the circuit. In a previous
article published in this magazine the author discusses parallel operation of similar alternators with respect
to hunting, and in the present article extends the discussion to parallel operation of two dissimilar alterna-
tors.— Editor.
F = oscillations per minute
A' = r.p.m.
/ = electrical frequency, cycles per second
^=fantrvr rl^npnrlincr nnnn thf
' I 'HE natural oscil-
-■- lating frequency
of a synchronous
machine is like the
oscillating frequency
of a pendulum, or of a
weight suspended by a
spring, in this respect :
if the rotor of the
machine is momen-
tarily displaced from
its stable position* in
space, it will oscillate
R. E. Doherty ^t a definite frequency
just as the pendulum
or suspended weight. Within the range of
ordinary loads, the synchronizing force acts
upon the displaced rotor in the same way that
gravity acts upon the pendulum, or that the
stretched spring acts upon the weight. In the
synchronous machine the "stretch" occvirs in
the magnetic field.
The natural oscillating frequency is an
important factor in parallel operation where
the synchronous machine is coupled to recip-
rocating apparatus. The periodic varia-
tions in the torque of such apparatus may
cause "hunting, " or oscillation, if the natural
oscillating frequency happens to be nearly
the same as the frequency of the torque
variation. It is important, therefore, to
prevent coincidence, or even proximity of
these frequencies. This is done by choosing
the proper flywheel effect. The formulaf for
calculating the oscillating frequency is
35,200 I'Kf
r.p.m.
electrical trequency, cycles p'
0 = factor depending upon the synchro-
■niyincr fnmp
F = "
♦The "stable position" is that position in rotation where the
load torque on the shaft is just equal to the electrical torque
exerted on the rotor, consequently where there is no acceleration.
+ ' 'Parallel Operation! of Alternating-current Generators
Driven by Internal Combustion Engines; Factors Affecting
Generator Design," by R. E. Doherty, General Electric
Review, March. 1915. Equation (14).
X This proportionality does not hold for large displacement
angles, just as it does not for large amplitudes in pendulum
oscillation, but is practically correct for the angles involved in
this problem.
gravity 32.2
nizing force
n'/?- = flj'wheel effect in lb. ft.'-
This formula, however, applies only to a single
machine operating on a relatively large system,
or to two duplicate machines alone in parallel.
In this article equations are developed for the
case of two dissimilar machines in parallel.
The principle underlying the theory of
this case, as well as the previous one, is that
the rotor field poles are locked to the rotating
stator poles through the elastic meditrm of
the magnetic field. It will be shown that if
the rotor is displaced from its stable rotative
position, it experiences a torque which is
proportionalX and opposite to the displacement.
This is the definition of harmonic motion — ■
the motion of an oscillating weight or pen-
dultmi — hence the well known expression
given in equation (1) for the period of har-
monic motion can be applied.
It is necessary to inquire why in this case,
as well as in the case of a single alternator
connected to a large system, or of two dupli-
cate alternators in parallel, the restoring
force or the strain of the magnetic field is
proportional to the displacement from the
stable position; also whj^ the two machines
must oscillate at the same frequency. With
these points established, the frequency of
oscillation of either, and therefore of the
system, easily follows.
Consider the spring-weight analogy. In
Fig. 1 the weights or masses** M are suspended
by very long cords so that the eftect of gravity
is eliminated and thus the only accelerating
force is the spring tension, which in the analogy
corresponds to the synchronizing force
of an alternator. In Case A, Fig. 1, which
corresponds to two duplicate alternators, if
an oscillation is set up between the masses,
it is obvious that the arrow a, attached to the
126 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 2
midpoint of the spring, or rather at the joint
of the two equal springs, will be stationary.
That is, the case is the same as if the point a
were fixed: which would also represent the
condition of a single alternator connected
to an infinitely large system. In other words,
T t 4
_L
i
JMl^^MML
a
y=x
Ca.seA, oqual weights, equal springs
b
y=x
Case B, equal weights, unequal springs
O
i 4 4
c
y=x
Case C, equal springs, unequal weights
Fig 1. Three Diagrams Which Represent Two Weights or
Masses Connected by Two Springs, and Which Arc Analog-
ous to Two Alternators in Parallel- The amplitudes of
oscillation are indicated by x and y
the natural oscillating frequenc\- of an alter-
nator when operating in parallel with a
duplicate machine is the same as when
operating on a relatively large system. Obvi-
ously in Case .4 the restoring force, i.e. the
spring tension, is by Hooke's Law propor-
tional to the displacement of the weight and
therefore the motion is harmonic.
In Case B, Fig, 1, two equal masses are
connected by unequal springs. This cor-
responds to two alternators with equal
moments of inertia, but unequal synchro-
nizing forces. But two different si^rings
in series can be replaced by an equivalent
single spring, which would cause the two
masses to oscillate in the same manner
and through the same amplitude .v and y as
if they were connected by the two different
springs. The only difference is that with the
single spring the midpoint of the spring
would be stationarv-; with the two, the arrow
b would oscillate between m and w. This
corresponds to the phase shift in the line
\-oltage generated by the oscillating alter-
nators. However, the accelerating force on
each mass, i.e. the spring tension, is of course
the same at all instants, and is proportional
to the total stretch of the two springs in series.
This in turn is equal to the sum of the dis-
Iilacement of the two masses from the posi-
tions of zero spring tension. Therefore if,
as shown, the amplitudes .r and y are equal
and the force is proportional to x plus y.
the force on either mass is also proportional
to the displacement of that mass and the
motion is harmonic.
Case C, showing unequal masses connected
by equal springs, represents two alternators
with unequal moments of inertia, but with
equal synchronizing forces. Here also the
force on each mass is the same at all instants,
being the tension of the spring, and is pro-
jjortional as in Case B to the total stretch of
the springs. The motion of the masses is
therefore harmonic. The arrow c , represent-
ing the terminal voltage, will obviously
oscillate between the limits p and q, since
with the same force acting on them the large
mass will oscillate through a less ainplitude
than the small mass.
The foregoing considerations show that the
oscillations of any such combination of two
masses and two springs in series will Ix-
harmonic. It follows also that the frequencies
of the two masses are the same; because if
the accelerating force is the same on each
mass at all instants, the momentum (massX
velocity = force X time =// J/) also must be
the same for each at all instants. This means
■ that the velocities are inversely as the masses,
i.e., the ratio of the velocities is constant.
Hence, when the velocity of one is maximum,
the other will be also: when the vclocit\- of one
is zero, the velocity of the other will i>e .-^ero;
and ,so on. Thus by the spring-weight
analogy, which is com])lete within ]>ractical
limits, it is established that in two alternators
of different capacities, i.e,, different synchro-
nizing forces, and of different moments of
inertia, the aicclcration. xrlocily. umi dts-
placoiicnt of eoili rotor during oscillation
is a harmonic function of time, and the
oscillating frequency of each machitte is the
same.
OSCILLATING FREQUENCY OF TWO DISSIMILAR SYNCHRONOUS .MACHINES 127
The following well known formula for the
period of harmonic motion can therefore be
used :
= corresponding sim.ultaneous angle for
machine B.
Qa poles on -4
7 = 2
7r\ -
seconds
(1)
where, for linear motion
/=mass of oscillating body
a = ratio of force to displace-
ment = lb. force per foot
displacement.
For rotation
7 = moment of inertia
cr = lb-ft. torque per radian
displacement.
The natural frequency' is
a oscillations per
minute. (2)
qb
la
poles on B
moment of inertia of ^4
lb m.omxnt of inertia of B
—*■ Powerp
Z7 <^0 n -
stable Position
Fig. 2.
be
The problem is to determine the
value of cr and I for two alter-
nators which are dissimilar in
respect to speed, synchronizing
force and moment of inertia. In
relating the several factors in-
volved, the ultimate reference for
the determination of displacement will
taken as the "stable position" in rotation.
Consider the effect of different synchronous
speeds. The power (involved in oscillation)
which is given up by machine .4 as a gen-
erator is equal, neglecting losses, to the power
consumed by :nachine 5 as a motor, the
energ\- transfer causing A to decelerate and
B to accelerate. The power being the same,
it follows that the accelerating or decelerating
torque produced thereby is inversely as the
synchronous speed, hence directly as the
number of poles. Therefore instead of the
momentums of the two rotors being equal at
all instants, which is true for machines of
an equal number of poles, there is in this case
a constant ratio of momentums. This ratio
is obviously the inverse ratio of the synchro-
nous speeds, or the direct ratio of the number
of poles. There is, therefore, also a constant
ratio of velocities and displacements.*
These relations ma}- now be put in equa-
tions.
Let
6a = mechanical displacement angle of
machine A at power p, measured in
radians from the "stable position."
See Fie. 2.
Makchine A
Machine B
Diagram of Two Machines Having a Different Number of Poles. The
dotted arrows indicate the limits of oscillation about the "stable
position." The diagram shows the instant when A is ahead and
B is behind the stable position; hence power flows from ^ to B
♦This refers to the velocities and displacement involved in
the oscillation, not to the velocity and displacement of norma!
rotation.
0a = phase angle, in electrical radians, be-
tween the mechanical and magnetic
pole centers, i.e., the distortion angle
of A corresponding to the power p
(in kw.)
<^6 = corresponding simultaneous angle
of 5
4> = 'i>a-\-<t>b
Wa = angular velocity (of oscillation) of ,4
0)6 = corresponding simultaneous velocitv
oiB
Poo = power in kilowatts of A correspond-
ing to a distortion angle of one elec-
tric radian, i.e. <^a=unity
Pob — corresponding power for B
'''=p7b
(Ta = lb-ft. torque, exerted on .4 '5 rotor,
per mechanical radian displacement,
i.e., the ratio of torque to da
(Xb = corresponding ratio for B
i = torque in Ib-ft.
So = r.p.m. of A
S6 = r.p.m. of B
F = natural oscillating frequency in peri-
ods per minute
/ = electrical frequency in cycles per
second
ie: = gravitv = 32.2 ft. per second"
Il'/?--^='flv wheel effect in lb-ft.= = p/
128 February. 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 2
The momentum of .4, due to the velocity
of oscillation, is
0}„la
Of B, it is
(jt}bh
The ratio of momentums is the ratio of the
number of poles. Hence,
(jOala 03a
nq= — J = — m
Therefore,
(3)
Thus if A's angular velocity is always
n^-i-ni times B's, then A's displacement will
always be nq-i-nj times B's. Thus, consider-
ing both angles positive,
nt
'db
da+db=t
<-»:)
(4)
(5)
It is now necessary to relate 9 and <(>. Both
represent the distortion angle, i.e., the
"stretch" in the magnetic field. They are
different in two respects : da or db is expressed
in mechanical radians, and measures phase
displacement of the rotor pole from its
"stable position;" whereas <t>a or (pb is ex-
pressed in electrical radians,* and measures
the phase displacement of the rotor pole from
the magnetic pole, i.e., the time phase from
the line voltage. The stable position may or
ma\- not correspond to the time phase of the
line voltage; in other words, the zero reference
for 6a and <t>a may or may not be identical,
depending upon whether the line voltage
itself oscillates. In Fig. 1, it has been sho%vTi
under what conditions the line voltage,
represented by arrows a, b and c oscillates.
For two two-pole similar machines, obviously
6a=<t>a
and
ea+db=d=<t>a+(t>b=<t>
because electrical and mechanical angles are
identical, and the line voltage does not
oscillate. However, if the machines have
different moments of inertia or different
synchronizing forces, then 6a and <l>a arc no
longer equal; but it is of course still true that
for the same number of poles
e=<t>
because each measures the total distortion or
"stretch" of the magnetic fields of the two
*That is, mechanical radians multiplied by one half the
number of poles.
machines. This relation obviously holds also
for different numbers of poles provided the
electrical angles <pa and 4>b are reduced to
mechanical angles. Thus
where
/ 2 2
<t> = — <l>a-\ <^i>
qa qb
That is
6
qa qb)
(6)
The next step is to relate 6 and the cor-
responding force. Poa is the power which
would be delivered by A at unit angular
displacement ahead of the line voltage; or
received at unit angular displacement behind
the line voltage. Hence the power on A at
any angle <^a is
pa = <i>a Poa
Likewise the power on B is
pb = <i>b Pob
But pa and pb are identical during oscillation.
That is, taking directions consistent with the
assumption in equation (4),
pa=pb=p
whence
P
and
4>a =
4>b =
Poa
J_
Pob
(7)
(8)
Substituting (7) and (8) in (6)
\qa Poa qb Pob/
or.
9 />
(9)
Equating (5) and (9)
\ tlqj qa Pec
(l-f-«i nf)
Hence, the power exchange per unit dis-
placement of .4 from its stable position is
p _
<7u /',.,. V »,J UU)
Oa'
2 \+n<,»t)
The general
relation between power and
torque is
27r5/ = power (.U)
wliorc
S = r.p.m.
< = torque in lb. ft.
OSCILLATIXG FREQUENCY OF TWO DISSIMILAR SYNCHRONOUS MACHINES 129
and power is in ft-lb. per minute. To express
the power in kilowatts, (11) becomes
33,0011 ., <,^
and
< = 7040^ in lb. ft. (12)
Solving for p and substituting in (10), the
torque per mechanical radian displacement
becomes,
■ 5a 2 V^nJ
But,
Hence,
qa = 120-^
(l + tiq-np
f
Sa =
422,400
J) a
fP.
<-S)
(l + nqtit)
The moment of inertia of A is
"' p ' 32.2
(13)
(14)
Substituting (12) and (13) in (2), the final
equation for the natural oscillating frequency
of A, and therefore of B, expressed in periods
per minute is,
F =
.35,200 I Poo/
('^D
(15)
Sa \H^if-„ (! + ","/.)
Example :
Machine A
400 kv-a., 1.50 r.p.m. 60 cycles
l-ri^^ generator = 90,000
WB^ flywheel = 254,000
Total WB}a =344,000
♦This percentage, based on experience largely, is generally-
accepted as the necessary difference to insure satisfactory-
operation.
Poa = 1000 kw.
5a = 150 r.p.m.
/ =60 cycles
Machine B
200 kv-a., 720 r.p.m., 60 cycles
WR- generator = total II 'R' = 2765
Poh =410 kw.
/ =()0 cycles
Sb =720 r.p.m.
WK'a 344,000 , ,_, ,
m = = — r^T^- = 1 24
np-
nn =
' WRh
•Loa
qb
2765
1000
410
48
10
= 2.44
= 4.8
^ 35,200^ 1000X60
(-E0
150 > 344,000 1+4.8X2.44
= 142 periods per minute.
If machine A, with a speed of 150 r.p.m.,
were driven by a reciprocating engine, trouble
from "hunting," or oscillation, could be ex-
pected, since there would be an engine impulse
of 150 periods per minute acting upon a
system whose natural frequency of oscillation
is 142, i.e., a difference of only 5 per cent.
In such a case the obvious solution would be
to increase the WR- of B until there is a
difference between frequencies of 20 per
cent,* i.e., until F is reduced from 142 to
120. This means a 43 per cent increase in
B's WR\ or 3950 Ib-ft." instead of 2765.
It is obvious from equation (15), that if
there is a large difference in the WR- of the
two units and if ni is greatly different from
unity, as in the foregoing example, a given
percentage change in the smaller WR~ will
result in a greater change in F than if that
percentage change is made in the larger WR-.
130 Februarv, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, X(
Some Mechanical Features of Synchronous Machines
By A. P. Wood
Al.TERSATIXG-Cl'RRENT ENGINEERING DEPARTMENT, GENERAL ELECTRIC COMPANY
The immediately following pages are devoted strictly to a description of the mechanical features of the
present day design of synchronous machines. Illustrations are given of the improvements which have been
made in design and construction by the introduction of arc welding and spot welding. Descriptions are given
of the oil starting system and the oil circulating system, the coreless and the cored box type of stator frame,
and the rotor spiders of various types and speeds. — -Editor.
T F one would take
•^ time to look at
more than the surface
of one of our latest
machines, he would
discover those many
features of present
day design which
have improved the
efficiency and relia-
bility of the machine's
performance. He
would find that elec-
tric arc and spot
welding have played
an important part in producing a better
machine.
In the smaller or belt-driven generators
a further step is taken by having the pole tips
A P Wood
accordance with the A. I. E. E. Rules for
motor operation, no trouble would be ex-
perienced from this source.
Should the operator desire a direct con-
nected exciter at some future time, he will
find machined lugs on the bearing bracket
and a proper shaft extension to accommodate
the exciter.
As to size of alternators the limiting
feature seem.s to be shipment, tunnel clear-
ances, etc. Quotations have been given on
waterwheel generators as large as 4o.(J<Mi
kw. and there is now installed a 32.oll()-kw.
waterwhcel-drivcn unit. By sectionalizing
the stator and rotor, pieces 14 feet across
have been transported across the United
States.
In general, the alternator design should be
as compact as possible, especially as to
elevation, because each foot in height entails
a considerable c^tra outlay in the cost of the
station to hou.sc the machine.
Recently a large number of proposals
have been made to use outdoor generators.
A large unit has already been installed in
the West without anv station over it. To
Fig. 1. 7.5-15 and 2S-kv-a. Stlf-eicitfd Generators
Fig 2. Stator Core C; ,.
Combined by Arc Welding
indexed for future operation as s\-nchronous
motors. All that is necessary for such use is
to insert an amortisscur winding. This
feattire is desirable where a larger machine
replaces a smaller one, thus allowing the use
of the smaller machine as a condenser. The
field coils having been originally insulated in
protect the generator from the elements, a
temporary wooden stracture is erected over
it. In case of trouble or where inspection is
necessary, a gantry crane moves over the
generator.
Sixty-cycle synchronous machines ranging
from 7.5 to '2't kv-a. are of the self-oxcited
SOME MECHANICAL FEATURES OF SYNCHRONOUS MACHINES
llil
type and cmljody two machines in one. The
revolving armature has both the multiphase
winding and an independent direct-current
exciter winding in the same slots. Figs. 1
and 'S show a photograph and wiring diagram
of one of these machines. The stator lam-
inated field poles have the cast-iron frame,
standards, and base cast around them in one
piece to reduce the machining operations to
a minimum.
No external excitation is required.
The exciter winding, therefore, is not
similar to that of a synchronous converter
where taps are taken off the main armature
winding, nor does it depend on rectifying
commutators or series transformers for step-
ping down part of the armature current. This
arrangement has been thoroughly tried out,
both in test and practice, and has proven
thoroughly reliable in every way.
When operated as synchronous motor,
grids are placed on the poles which have
been punched to accommodate them.
Arc Welding
The combined clamping flanges and end
fingers used to secure tight stator cores consist
of pieces of sheared machine steel arc welded
together. Fig. 2 shows a perfect ventilating
path between the flange and the punchings
that is not obtained in any other type. The
punchings can be tightened in sections at any
time without stopping the machine by remov-
ing shims placed beneath the flanges.
This type of flange entirely supersedes the
old heavy cast-iron sectional flanges and is
Many large eastings are being reclaimed
by the use of arc welding.
Rotor spiders which have too large a bore
can be used without a bushing by slightly
increasing the shaft diameter by arc welding.
3<pUne
Fig 3.
Diagram showing Electrical Connections of Three-phase
Self-excited Generators
fast taking the place of the combination
segmental cast-steel flange and end fingers.
The delays in production due to making
expensive patterns, as well as the foundry
loss which runs as high as 2.^ per cent, have
been eliminated.
GENERAL ELECIffiCCCMraH
Fig. 4. 3000-kv-a. Synchronous Condenser, the Enclosing
Ventilating Shields of Which Are Constructed of Sheet
Iron and the Seams Arc Welded
The tool expense for machining generators
has been greatly reduced by welding high
grade steel cutting ends to cheaper grades
of body stock.
When positive ventilation is required, the
enclosing shields consist of arc welded pieces
of sheet iron as shown in Fig. 4. A more
finished appearance is now obtained by
flanging over the corners, thus
doing awa\' with the square
appearance. A set of enclosing
shields can be produced in a few
days. In the ease of castings an
expensive pattern is required
which sooner or later becomes
obsolete, and the foundry is
fortunate in obtaining 75 per
cent good eastings.
The danger of damaging the
stator windings when removing
east-iron end shields is elimi-
nated by the use of light sheet-
iron shields. The absence of
heavy machine work such as the
truing up and facing required by large
east shields facilitates production. Being
much lighter and made in either 90 or
120-deg. sections, the operator can readily
remove them for inspecting or repairing the
field or stator coils.
Armak/re
") Stationary
Held
132 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 2
The stairways for large vertical machines
are now being arc welded and present a
finished appearance.
Laminated field spiders where bolts are to
be tapped in the periphery*- have the separate
laminations arc welded together.
^:^^
Fig 5.
Stator of a 1400-kv-a. Synchronous Motor Showing
the Construction of the Cold Rolled Steel
Spot-welded End Shield
section cold rolled steel, spot welded to the
laminations. This method eliminates the
chance of space blocks becoming loose,
dropping out, and damaging either the rotor
or stator. These straight beam space blocks
allow a greater volume of air to pass through
the core than woiild be possible with blocks
of rectangular shape, which in most cases
are cur\-ed or bent to keep them from turning
over.
Advantage of this welded space block is
being taken to replace various types of
clamping flanges for retaining the punchings
in small stator frames. Should a loose core
develop, the end fingers would still remain
in place which would not be the case were
friction depended upon.
Amortisseur Winding for Synchronous Motors
In recent years the use of tin and solder
have been replaced by a method of making
better joints between bars and end rings.
Fig. 6 shows an amortisseur winding which
is practically indestructible. Fig. 7 shows a
method of increasing the cross section where
the bar enters the end ring. This construc-
tion has been used for several years with
]:>crfcct results. Brass, copper or monel
metal is used according to the class of sers-ice
the motor has to perform.
S>nchronous motors have been made with
multiple amortisseur windings and in some
cases with cast windings.
Experiments are now being made to extend
the use of arc welding.
Spot Welding
For several years the lengthy job of rivet-
ing has been superseded largely by spot
welding. The open-type shields consist of
strips of cold rolled steel held in jigs and spot
welded. Fig. 5 shows the finished appearance
of such an end shield. This shield also can
be produced in a few days. It appeals
strongly to the operator as it is unbreakable
and does not require removing when shifting
the stator. The expensive pattern shop work
and foundry- breakage always present where
castings are used has been done away with.
Where large machines are boxed for ship-
ment, the welded shields do not have to be
removed as is the case where cast-iron ones
are used. Also, if the welded shields do
become damaged they can be easily straight-
ened.
The space blocks for ventilation between
the stator punchings are pieces of I-bcam
FiE- 6. Synchronous Motor Rotor Having the Amortisseur
Winding Before the Laminated Poles Were Riveted
The amortisseur winding shown in Fig. G
was assembled before the laminated pole
was riveted, thus eliminating the drifting
of the slots where the windings are assombliHl
after the poles arc completed.
SOME MECHANICAL FEATURES OF SYNCHRONOUS MACHINES
133
Oil Starting Systems
For many years hand operated or motor-
driven high-pressure oil pumps have been
available for reducing the kilovolt-ampercs
required to start heavy rotating elements.
A film of oil is forced beneath the shaft until
the oil rings become effective,
after which the pump is shut
down. A motor-driven outfit
for this purpose is shown in
the lower right-hand corner
of Fig. S. The saving in kilo-
volt-amperes is obtained by
using a lower tap on the start-
ing compensator than would
be required were it not for the
high-pressure oil starting sys-
tem.
The oil for lubricating the
rotating element is obtained
from an equalizer pipe which
parallels the two bearing
standards and flows into the
pump plunger chamber. After
the pump is started and the
bearings are lubricated under
pressure, and the rotating
element is receiving oil from
the bearing oil rings, the
pump is shut down. The
pump plungers act as check
valves, otherwise the rotating
BrassCndfiing
'^/a'Brass Bar
shaft would build up a back pressure and
drain the bearings.
Oil Circulating Systems
A pump has been developed to replace
the gears and gear pump, Fig. 11, common to
Fig. 8. Motor-driven,
itate Starting the L,
>*£<
Brass rerrule driving Fit
HantJ f7iveted^oinC3
Fig. 7. Photograph showing the Method of Connecting Amor-
tisseur Winding Bars to the End Rings and showing the
Enlargement of the Bar at the End Ring
, High-pressure Oil Pump at the Lower Righthand, Used to Facil-
arge Horizontal Machine by Forcing Oil Beneath the Bearings
oil circulating systems now in general use.
The expensive split driving gears, so hard to
assemble on vertical machines, and the
inefficient check valves with screw adjust-
ments are all eliminated. The oil pan shown
in Fig. 11 is designed so that the pump can be
completely assembled on one half of the pan.
This arrangement allows for easy inspection
without interfering with or disconnecting the
piping system.
The piping is arranged to take oil from the
bottom of the oil pan to a point above the
lower guide bearing outside of the stator
frame, through a flow indicator, then to the
top of the guide bearing and after filling the
grooves in the guide bearing it overflows
back into the oil pan.
The lower guide bearing has an independent
oiling syste:n on the smaller vertical machines.
The upper guide bearing and spring thrust
bearing are designed as a unit, are self oiling,
and require no attention whatever.
Stator Frames
The old, familiar, heavy skeleton type
stator frame has been replaced with the
present box type. This latter frame is not
134 Februarv, 1H20
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. i>
only stronger and lighter mechanically, but
eliminates the errors in air gap which resulted
from improper alignment of the cross ribs in
the old type.
The stators are made of cast-iron and consist
of two types. One requires cores in casting
Fig 9.
Coreless Box Type Stator Frame Used
for the Smaller Machines
and the other does not. Due to its unique
design the coreless type, Fig. 9 (used on smaller
machines) when assembled acts the same as
clamping flanges to retain the laminated
punchings in place. The cored type. Fig. 10,
has separate clamping flanges for retaining
the punchings and is superior for larger stator
frames as any looseness in punchings can be
taken up by removing shims from beneath
the flanges.
Cast-iron is the most economical material
since steel castings cannot be obtained thin
enough to take advantage of the increased
strength.
The deflection allowed in calculating the
stifTness of frames varies with the diameter;
and the allowance, say for six feet in diameter.
is increased by increments for larger diam-
eters. In single-phase machines there is a
pulsating flux and consequently the stiffness
of the frames has to be increased consider-
able. In all cases the frame nnist be stilT
enough to prevent undue distortion in shop
handling, turning over, or boring mill work.
It is essential to have the stator cores
tight and this is obtained by the use of
individual combination clamping flanges and
end fingers, the result being a very tight core.
Field or Rotor Spiders
The larger percentage of the cast-steel
rotor spiders have been replaced by a lami-
nated type which cuts down the time of
production from weeks to days. These
spiders are produced by interchangeable
dies that supersede the expensive old com-
bination method wherein one set of dies was
required for each particular size of machine.
Very few cast-iron field spiders are used
now on account of the improved magnetic
circuit obtained by the use of steel. The
excitation required to give full voltage no
load has been decreased b>- replacing cast-
iron with cast-steel.
Rotor spiders are so proportioned that
shrinkage strains may be avoided, these
being overcome in steel castings by annealing.
For securing the poles to the spiders the
simple constrviction of putting bolts through
the rim into the poles is used on rotors
Fig. 10
Cored Box T>T>e Stator Frame Used
for the Larger Machines
having speeds up lo approximately 22.") r. p.m.
For higher speeds dovetails are useil.
The \' or wedge dovetail was usoil for
many years until the size and speed of
machines rapidly increased; then the T dove-
tail took its iilaee. As the neck of the V
SOME mechank;al features of syxchroxous machines
135
dovetail is increased to cover tension stresses
and the depth is increased to overcome
bending (this also applies to the rotor dove-
tail), we soon come to a point where the
laminations of the pole dovetail will buckle,
thus preventing its use. The ordinary T
dovetail has some bad features, such as
keying up the pole and field coil, also introduc-
ing the reluctance of an air gap between the
pole and rotor. This T type dovetail has
been improved so that the objectionable
features have been removed.
The stresses are calculated so that at
double speed of the rotor no stress will exceed
half the elastic limit of the material. Sup-
porting brackets are placed between the
poles where higher stresses of pole and copper
winding are encountered.
Some of the rotors are of too high speed
to depend upon a pressing fit on the shaft.
They must be shrunk on, otherwise at double
or runaway speed they would float on the
shaft.
Shafting
Carbon steel with a tensile strength of
approximately 75,000 lb. per square inch is
generally used. At maximum power trans-
mitted, a stress of 7000 lb. per square inch is
not exceeded. On special shafts nickel steel
is used. Shafts are designed so that the
deflection will not exceed a certain percentage
of the air gap when the rotor is assembled
in the stator.
Bearings
With ball-seat self-aligning bearings
lubricated by oil rings, pressures are used as
high as 130 lb. per square inch projected area.
Ventilation
The rotating field poles are purposely
designed to o^'erhang the field spider, and
for peripheral speeds of approximately SOOO
or 9000 ft. per minute the pole pieces them-
selves act as natural fans and have sufficient
blower effect to ventilate the stator. Holes
are also placed in the rotor rim between the
poles and are of great assistance in distribut-
ing the air properly. Bafflers are also placed
Fig. 11. Gear-driven Oil Circulating Pump Now in Common
Use on Large Vertical Machines
between poles in some cases to stop air
from passing directly through the rotor poles,
the result being a more even distribution
through the stator and a considerable reduc-
tion of temperatures.
The tendency is for higher speeds for large
output alternators and where the peripheral
speed exceeds 9000 ft. per minute, enclosing
shields are used, producing a much quieter
running m.achine. In many large installa-
tions, especially where more than one machine
operates in the same room., ducts are furnished
to carr\' cold air to the m.achines and other
ducts to allow the heated air to be led away.
Vertical machines present more difficult
problems in ventilation than horizontal
machines. Fresh air is generally taken from
the wheel pit. Approximately three cubic feet
per kilowatt rating is required.
136 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 2
Parallel Operation and Synchronizing of
Frequency Converters
By 0. E. Shirley
Alternating-current Engineering Department, General Electric Company
The operation of a single frequency converter between two systems is comparatively simple. However,
when more" than one converter is used the phase angle between the incoming generator and the bus intro-
duces a factor which must be provided for. The author gives a very clear explanation of the reasons for this
phase displacement between loaded and unloaded converters trying in the same systems. In modem fre-
quency converter sets the difficulty introduced by this condition is taken care of by means of a motor-operated
screw which shifts the stator of the motor or generator in its cradle to the required position. — Editor.
' I *HE tise of fre-
-^ quency convert-
ers has become of con-
siderable importance
since power systems
have expanded and
overlapped. Thus,
s>-stems of different
frequencies have come
together and it has
been necessary to con-
nect them through
frequency converters.
_ „ ^,_. , As the size of the
O. E. Shirley , . ,
systems has mcreased,
it has followed that the size of the con-
verters has also increased and these convert-
ers can now be built in capacities as high
as 1.5,000 kw. The operation of a single
converter between two systems is compara-
tively simple, for it is only a matter of
synchronizing the generator with its system.
When more than one converter is operated in
parallel between two systems, the matter of
proper paralleling becomes of importance.
The present-day design practice is- tending
toward machines of higher regulation, with
resulting higher efficiencies, lower short-
circuit stresses, and cheaper costs. These
characteristics, which are somewhat different
from those of the units of older design, are
liable to introduce operating difficulties unless
proper attention is given to the conditions
of operation and design when installing new
units. Conditions of load division are cjuitc
important in some cases, especially where
reversible operation is required, and an
understanding of the behavior of alternators
under load is qtiite necessary to insure
satisfactory operation.
Lag Angles
The power output from an alternator
operating in parallel with other synchronous
apparatus is dependent onh" on the power
input from the prime mover, and does not
depend on the field adjustment. The chang-
ing of the field current will change the reactive
component, but does not affect the energy-
component of the armature current.
It is now quite generally known that the
rotor of an alternating-current generator
moves fom-ard from its no-load synchronous
position as the load increases. In the same
way, the rotor of a synchronous motor drops
back from the no-load position. That is,
if the rotor of the generator is at a certain
place when the voltage of one phase reaches a
ma.ximimi with no load on the machine, it
will be several electrical degrees ahead of
that position when the voltage of that same
phase reaches a maximiun with the machine
under load. Similarly, the rotor of the motor
is behind the no-load position when the
machine is loaded. This angle varies for
different designs, but the usual value at full
load is from 20 to 40 electrical degrees. The
value of the angle depends on a number of
different design factors and is rather difficult
to calculate from the design of the machine.
Some of the most important factors are the
ratio of pole arc to pole pitch, the variation
of the air gap from the center to the edge of
the pole tip, the relation of the air gap to the
pole pitch, the relative values of field and
armature magnetomotive forces, and the
value and power-factor of the load.
In the case of frequency converters, the
angle to be considered is the resultant for
the two machines, referred to one or the
other of the imits. Assume that the con-
verter is a 2.")-to-t)0-cycle set, and the full-load
lag angle of the 2.")-cycle unit is 20 electrical
degrees, while that of the (iO-cycle unit is
2.") electrical degrees. The pole pitch of the
2r)-cycle unit is 2.4 times that of the (iO-cycle
unit, hence one electrical degree on it is equal
to 2.4 degrees on the (iO-cycle unit. When the
converter is transferring power to the (lO-cycle
system, the angle by which this system is
SYNCHRONIZING OF FREQUENCY CONVERTERS
137
behind its position at no load is the sum of
the two angles of the machines, allowing for
the factor 2.4. That is, the electrical lag
angle of the (50-cycle system will be 25 deg.
plus 2.4 times 20 deg. or 73 deg. Similarly,
if the converter is operating with the (iO-
cycle unit as a motor, the lag angle of the
25-cycle end will be 20 deg. plus 25 deg.
divided by 2.4 or 30.4 deg.
It should be noted that with a lO-and-24-
pole combination, such as is required for a
25-to-(i0-cycle converter, it is possible on
starting from the 25-cycle end to come into
step on any one of ten positions depending
on the pole that happens to be at the reference
position when the motor comes into'isyn-
the generator voltage will have any one of
five different angles.*
Parallel Operation
The parallel operation of frequency con-
verters made by different manufacturers or of
different dates of manufacture by the same
company may introduce some little difficulty,
but if the stator frame of all the units, or all
except one, are made adjustable it is usually
quite easy to get satisfactory operation by
shifting tile frames of the various units until
the proper load division is obtained. This
adjustment may be made once for all when
the sets are first installed after which the
setting of the frames need not be changed.^
Fig. 1. Frequency Converter with Exciter. To permit of adjustment for parallel operation with another
frequency converter, the stator of the near unit is mounted in a cradle to which it can
be clamped after the load adjustment has once been made
chronism. The rotor may be caused to drop
back a pole pitch by reversing the field. This
means that the 25-cycle unit will drop back
ISO deg. and the 60-cycle unit will drop back
2.4 times 180 deg. or 432 deg. This is equiv-
alent to one complete cycle and 72 deg.
more. On the fifth reversal of the field the
rotor will be back one complete cycle on the
60-cycle end and the last five poles will
simply repeat the cycle of the first five.
From this explanation it can be seen that the
motor may come into synchronism so that
* For a more complete discussion of this feature of parallel
operation, covering the more common combinations of frequen-
cies aside from the example given above, refer to "Some Fea-
tures Affecting the Parallel Operation of Synchronous Motor-
generator Sets." by J. B. Taylor, lA.I.E.E., 1906, Vol. XXV.
Page 113.
When the converters are operated to trans-
fer power always in one direction, that is non-
reversible operation, it is not necessary that
the lag angles of the sets be equal, as the
frames can be adjusted so that all the units
will take the maximum load for which they are
designed for continuous operation, and the
load division may come what it will at the
lighter loads. With sets of different design
characteristics this condition may result in
a pump back between them at no load; but,
when the load is light enough for this to be
objectionable, all but one or two of the con-
verters may be shut down so no trouble will
be experienced from this cause. It is quite
important to note that for non-reversible
13S February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 2
operation with converters of older design
it is usually not advisable that the new
converters be designed to divide the load
proportionately to their rated capacities. To
produce this condition may necessitate a
higher cost and result in other disadvantages
without any real gain in operating character-
istics.
Reversible Operation
There are a few cases where it is desirable
to operate the converters in parallel through-
out the entire range of load in both directions,
and then it is desirable that the load division
be proportional to the rated capacities
throughout this entire range. This con-
closing the switch when the synchronism
indicator shows that the two are in phase. If
the converter is used to supply power from the
generator without operating in parallel with
an}- other generator, it is not necessan,- to have
any synchronizing devices at all. Adjustment
of the stator frame need not be provided
where only one converter is to be used, but
it is usually advisable to have the stator
frame made adjustable as the future instal-
lation of other sets for parallel operation may
make this feature ver],- convenient.
It is evident from the preceding discussion
of load angles that when an unloaded con-
verter is to be paralleled with sets that are
carrying load the incoming generator will
Fig. 2. A Frequency Converter similar to the one in Fig. 1, except that the stator of the near unit can be shifted
by an auxiliary motor-driven device. This arrangeincnt readily permits of synchronizing with other
sets without disturbance and of adjusting the load while the set is in operation
dition requires that the full-load lag angle
of the converters be the same or very nearl\-
the same. The use of the motor-operated
phase-adjusting device described later will
usually be the most satisfactory method of
operation under these conditions.
Synchronizing
When only one converter is used between
the two systems, it is only necessary to staft
the motor and throw it on the line in the
ordinary way for starting synchronous motors.
Then the generator is synchronized with the
second system by changing the speed of one or
the other of the systems until the generator
runs at appro.\imatcl\- the same si^eed as the
system to which it is to be connected and then
be ahead of the bus voltage by the load angle
of the converters already operating. The
synchroscope needle will not rotate as the
converters already tied in will hold both
systems at a ratio of frequencies corTes])ond-
ing to the ratio of the number of pole.s. and
the incoming set will be held at exactly the
same speed as the others by the synchronous
motor. The angle between the neetlle and
the in-phase position will be the load angle
of the converters already operating when the
incoming converter is on the proper pole for
synchronizing. This angle may be from W
to 75 deg., (lepending on the load and which
unit is operating as a motor, being greater
for the higher frequency generators. As
the change in the angle indicated by the
SYNCHRONIZING OF FREQUENCY CONVERTERS
l.-ji)
synchroscope due to slipping a pole is rela-
tively large, it is very easy to determine the
proper position for synchronizing. The
synchroscope dial may be calibrated to
indicate this position for various values of
load on the sets already operating.
The operations necessary for synchronizing
a converter with sets already loaded are as
follows :
(1) Start the converter from the motor end
and operate from the line with full field.
(2) Adjust the voltage of the incoming
generator to approximately the bus voltage.
(3) Slip poles on the motor by reversing
the field until the generator is ahead of the
bus voltage by the load angle of the sets
already operating.
(4) Close the generator on the bus and
adjust the field until the proper reactive cur-
rent is taken by the incoming generator.
Methods of Phase Adjustment
The usual standard method of adjusting
the. phase angle of frequency converters to
secure proper load division is to mount the
stator of one unit in a cradle rigidly secured
to the base. The stator may be securely
clamped in this cradle when the required load
adjustment is obtained. With the frame
once set it is unnecessary to change this
adjustment. A converter equipped with this
type of cradle is shown in Fig. 1.
The synchronizing of converters equipped
with this type of adjustment requires that the
generator be put on the line when its voltage
is out of phase with the bus by an angle of 30
to 75 deg., which will cause some little dis-
turbance when the switch is closed and there
will be some swinging of the load between the
converters before they settle down to stable
operation. This may be objectionable espe-
cially if the converters are of large capacity.
and to eliminate the action a motor-operated
phase-shifting device may be used. This
device enables the stator frame to be shifted
while tmder load. When s>'nchronizing with
this device, the stator frame is shifted until
the generator voltage is in phase with the bus
voltage, and the switch closed. There will
then be no disturbance on closing the switch,
but the converter will not take any load until
the stator frame is shifted. This shift ma\- be
Fig. 3. Close-up View of a Motor-driven Phase-shifting
Device Similar to That Shown in Fig. 2
adjusted until the desired load is taken b\-
the incoming set. The converter may also
l3c unloaded before taking it off the line, and
there will be no disturbance on opening the
generator switch. It should be particularly
noted that this device is of use only where two
or more converters are to be operated in
parallel.
This shifting device is very useful where
reversible operation of sets of recent design
is required with sets of the older types.
The recent designs will very likely have
larger load angles than the older ones, unless
larger and more expensive machines are used ;
and they will therefore have a tendency to
take less load than they should unless a
shifting device is employed.
A converter equipped with this type of
adjustment is shown in Fig. 2, and a more
detailed view of the device itself is given in
Fig. 3.
140 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, x\o. 2
Motor-generator Sets
By G. H. Tappax
Alternating-current Engineering Department, General Electric Company
In the smaller motor-generator sets an induction motor is usually employed on the alternating-current
side, while in the larger units the synchronous motor is found more desirable. Sometimes the field of the
synchronous motor is excited from the direct-current generator and in other cases a separate exciter is provided.
With frequency converter sets either a single exciter may be used for the two field windings, or a separate
exciter for each, depending upon the service for which the set is intended. The author briefly describes the
basis on which motor-generator sets are rated. Provision for heax-j- overload is often required and a method of
compounding the exciter to automatically take care of such conditions is mentioned. — -Editor.
M'
G. H. Tappan
OTOR-GEN-
ERATOR sets
in which one or more
of the machines are
of the synchronous
type may in general
be divided into two
classes; those convert-
ing alternating cur-
rent to direct current,
and those converting
alternating current at
one frequency to alter-
nating-current at a
different frequency.
Machines of the first class are commonly
known as motor-generator sets; while those
of the latter, although properly motor-gen-
erator sets also, arc now known as frequency
converters.
Those which convert alternating current
to direct current arc the more numerous of
the two classes.
There are many conditions of
senice where the rotary or syn-
chronous converter is not quite
as suitable as the motor-gener-
ator set. Often a direct-current
supph- is desired where the volt-
age can be varied or controlled
over a considerable range. Also,
the synchronous motor may be
built for higher voltages than
the rotary converter, thus doing
away with the necessity of using
transformers. Such sets arc used
for electrolytic work, battery
charging, mining, railway, light-
ing and power sennce, etc.
Motor-generator sets which con-
vert alternating current to direct
current and which are of less than 100-kw. ca-
pacityare usually built with an inductionmotor
as the driver, while those above 100 kw. arc
driven by synchronous motors. Of the two
methods of driving, the synchronous motor
has some advantage. This motor runs at
constant speed and may be used for correct-
ing the power-factor of the system on which
it operates. The desirability of this latter
feature is being more and more emphasized
inasmuch as public ser\-ice companies are
beginning to place a premiiun and give a
bonus on this type of load.
The majority- of s>-nchronous motor-gen-
erator sets are built for operating from a
60-cycle circuit, although there are many
operating from 25-cycle circuits and some
at other frequencies such as 50, 42, 40,
30.
The present day tendency is to build
motor-generator sets to run at as high a
speed as practicable because of the smaller
space required for the same kilowatt output
and the lower cost.
There is also a tendency to enclose ma-
chines and obtain a greater output from them,
with the same heating, by the addition of
forced ventilation or improved natural ven-
Fig. I.
Motor-generator Set <900 r.p mA, Consisting of an 800-kv-a . 2300-
volt Synchronous Motor and a 600-kw., 2S0-volt
Direct -current Generator
tilation. Also, since enclosure reduces the
noise it is ])crmissible to design motors for
higher peripheral speeds than is practicable
with the o])cn type of machines.
Small size (id-cycle motor-generator sots,
such as .iOO-kw. or less, are not usually built
AlOTOR-OENERATOR SETS
141
for voltages higher than 400(1. Larger sets are
built for voltages up to lo.L'OO.
Where the voltage of the direct-current
generator is fairly constant and not over 275
volts, the field of the synchronous motor is
usually excited from the generator end of the
set; but where the voltage of the generator
is greater than 275 volts, or where the service
required of the generator necessitates a
greatly varying or fluctuating voltage,
separate excitation for the synchronous motor
is required. For this case a direct connected
exciter is generall)' used with the set.
The ratings given to the synchronous
motor-generator sets which have been stand-
ardized by the General Electric Co. are those
recognized by the A.I.E.E. These are the
"nominal" rating and "continuous" or 50-
degree rating.
Following the demand for the power-factor
correction feature, the synchronous motors
of the standard sets are designed for opera-
tion at a leading power-factor. In the case
of the nominal rated sets, the motors are
designed for 80 per cent power-factor; and
for the continuous or 50-degree rated sets,
the power-factor chosen usually is 85 per
cent.
There are frequently cases where a greater
corrective effect is desirable, which condi-
tions require a m.otor having a much lower
power-factor rating and approaching very
close in design to a synchronous condenser.
Quite a number of such sets have recently
been built with synchronous m.otors designed
for 70, 60 and 30 per cent power-factor.
The character of the average load and the
conditions of momentarv overloads are taken
Fig, 2. Motor-generator Set (720 r.p m.i. Consisting of a 1600-kv a.. 12.500-volt
Synchronous Motor, a 1000-kw.. 600-volt Direct-current
Generator and an Exciter
The nominal rating is based on a certain
normal load maintained continuously until
the temperatures have become constant, and
this followed by a 150 per cent overload for
two hours on the generator. Under these con-
ditions the m.aximum. temperature rise is
guaranteed not to exceed a specified value.
This rating is generally applied to the
motor-generator sets for mining and railway
service. The synchronous m.otors of the sets
in this case are generally designed to operate
at 80 per cent power-factor at normal load,
and the same field strength which is necessary
for normal load is maintained on all loads
above the normal full rated load.
The continuous or 50-degree rating is
the load which the set will carry continuously
without exceeding a tem.perature rise of 50
degrees. The momentary overload is the
only overload guaranteed in this case.
into consideration in designing the synchro-
nous motors. Thus the motors of the nominal
rated sets are designed to withstand a mo-
mentary overload of 100 per cent, while the
motors of continuous rated sets are designed
for a momentary overload of 50 per cent.
Sometimes the severity of the conditions,
with respect to the nature of the overload
peaks, requires even a greater margin of
safety. This is especially true in railway
service. For such cases the motors are
designed for an overload as great as 200 per
cent, i.e., capable of withstanding a momen-
tary load of three times normal without the
synchronous m.otor dropping out of step.
However, another method of taking care
of this condition is sometimes used where
the motor line conditions pennit; viz., that
of compounding the direct-connected exciter
which excites the field of the synchronous
142 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 2
motor. This compounding is done with the
direct-current generator line current or a
part thereof being passed through the series
field of the exciter. The adjustment can be
so made that, with a certain direct-current
generator load, the field current of the
synchronous motor will be of the proper
value to give the desired kilovolt-amperes
input and power-factor. It will also be so
adjusted that the field strength for a peak
load will be sufficient to keep the motor
from dropping out of step.
When a momentarj' heavy load occurs on
the generator, the increase in current in the
series field of the exciter will cause a corre-
sponding increase in the exciter voltage,
which in turn will result in a motor field
current of sufficient value to hold the motor
in step. This method usually increases the
the starting period, it is the common practice
on synchronous motors forming part of
alternating current to direct current sets, to
start with the field closed through a resistance.
The voltage required to start the motors
varies considerably with the design. As a
general rule, 25-cycle motors will start on
a lower voltage than GO-cycle motors. Thus,
the 25-cycle motors of motor-generator sets
require approximately 20 to 30 per cent
normal voltage, while the 60-cycle motors
will average approximately 35 to 45 per cent.
The other class of motor-generator sets,
properly known as frequency con\'erters,
is used to con\-ert power at one frequency to
power at another frequency. The most
usual application is in converting 25-cycle
power to 60-cycle power for tying two sys-
tems together.
Enclosed Frequency Converter of 10.000 Kw., 60 25 Cycles, 300 R P M and 13,200 11.000 Volu
size of the exciter due to the large scries
field, or due to the higher insulation required
on the series field, as the sets where this
method is used may have a generator voltage
much higher than that of the exciter. Also,
the design of the exciter for the heavy peak
duty requires a larger exciter than otherwise
would be necessary. On the other hand, it
allows the use of a smaller synchronous
motor, inasmuch as the momentary overload
can be taken care of and the size of the motor
be determined by the heating at the normal
or average speed.
Motor-generator sets are started from
either end, but the general practice is to
start from the synchronous motor end using
an auto transformer or low-voltage taps on
the transformers. Due to the high induced
voltage in the field coils of the motor during
Very little choice is possible in the matter of
the speed of these sets as the maximum speed
for which 25-c\xle to (10-cycle converters may
be designed is 300 r.p.m. , and this is low enough
for any capacity so far demanded.
The size of the standard line of frequency
converters ranges from 300 kw. to 3000 kw.,
the smaller sets being standard for 2300 volts
and the larger sets for 2300 to 13,200 volts.
Sets are built also for 4000 kw., 50(Ht kw. and
larger, there being a 10.000-kw. set now in
process of manufacture.
The usual power-factor for which the 2.">-
cyclc motor is designed is 00 per cent, anif
for the GO-cycle generator SO per cent.
When conditions of service permit of
raising the (iO-cycle frequency to (».2o
cycles (and this may ordinarily be done
if the prime movers of that system allow
SYNCHRONOUS CONDENSERS
143
25-cycle power to
operating reversed
power to 24-cycle
it and the system is not already tied to
the 25-cycle system by a 25-cycle to 60-
cycle converter), a speed of 750 r.p.m. may
be used, converting the
62.5-cycle power or, if
converting the 60-cycle
power. The higher speed allows the use of
smaller machines for the same capacities,
therefore smaller space is required and the
cost is less.
The bulk of the frequency con^■erters manu-
factured are for the two conversions of fre-
quency mentioned, although several other
conversions are common, such as 40 cycles
to 60 cycles, 42 cycles to 50 cycles, the latter
being frequently used in Europe.
The frequency converter is usually built
with one direct-connected exciter for excit-
ing the fields of both units, but when a
voltage regulator is used on the generator
it is usually customary to have a direct-
connected exciter for each unit, one being
placed on each end of the frequency con-
verter.
One of the units of the frequency converter
is built with adjustable feet, so that the
stator may be rotated to get the correct
phase relation for synchronizing. A complete
explanation of this arrangement is given in
the article, "Parallel Operation and Synchro-
nizing of Frequency Converters," by O. E.
Shirley, in this issue.
Synchronous Condensers
By E. B. Plenge
Alternating-current Engineering Department, General Electric Company
As a means of controlling voltage on high potential long distance transmission lines, the synchronous
condenser is an absolute necessity. It is also an economical investment for users of large quantities of power
at low power factor, as it is now common practice for central stations to make a charge for supplying wattless
current. Greater latitude in the electrical characteristics is permissible in the design of synchronous con-
densers that are not intended to carry mechanical loads. In most applications a synchronous condenser is
required to carry leading currents, but occasionally, as for example when the load is removed from a trans-
mission line, it may be necessary to operate the synchronous condenser with a lagging current. The require-
ments of machines for such service are shown vectorially. Starting characteristics, ventilation and prevention
of noise are also briefly referred to. — Editor.
T"
E. B. Plenge
'HAT the econo-
mies effected in
the operation of gener-
ators, transformers,
and transmission
lines by the use of
synchronous condens-
ers are now generally
appreciated is evident
by the rapidly in-
creasing number of
applications of these
machines for power-
factor correction.
The higher rates
charged for loads of low power-factor by many
power companies is undoubtedly partly
responsible for this increase. The use of
condensers on long transmission lines for
voltage control is a practical necessity'. This
article will be confined to the design and
characteristics of the condensers them-
selves, as other articles have dealt at length
with the determination of the capacity re-
quired to attain certain results.
In general, the design of a synchronous
condenser differs little from that of a syn-
chronous motor, but the fact that the con-
denser runs idle; i.e., with no mechanical load
and at approximately zero power-factor, per-
mits of certain modifications which result in a
less expensive machine than a motor or gener-
ator of the same kilovolt-ampere capacity.
Speed and Capacity
A reduction in cost is obtained by an
increase in speed, up to the point where the
greater mechanical stresses necessitate a
m.ore expensive type of construction or where
a change is required in the electrical loading
(such as lower armature reaction) to relieve
the duty on the field. As the speed of the
condenser is usually left to the choice of the
manufacturer, it is made as high as possible
with a salient pole type of rotor. This con-
struction has been found best suited for con-
densers and is now used almost exclusively.
The mechanical design is therefore the same
as for high-speed waterwheel-driven gen-
erators with the added problem of designing
an amortisseur winding to withstand the high
stresses. Fig. 8 in the article entitled, " Large
Horizontal Alternating-current Waterwheel-
driven Generators and Synchronous Condens-
ers," page 151, illustrates the rotor of a 12,500-
kv-a.. 500-r.p.m. condenser and is typical of the
construction used. Small capacity condensers
have been built with speeds as liigh as 1800
144 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 2
r.p.m. The most economical speed will, of
course, be lower as the capacity is increased,
so moderate size 60-cycle machines are built
to operate at speeds of from 900 to 720
r.p.m. and the largest sizes at 600 r.p.m.
The size of individual units has followed
the increased capacity of generating stations
and transmission lines. At present, there are
a considerable number of machines in opera-
tion with capacities ranging from 5000 to
15,000 kv-a. and a 50-cycle condenser of
30,000-kv-a. capacity at 600 r.p.m. is now
under construction.
Electrical Design
As a condenser runs with little or no
mechanical load, it can be designed without
consideration of the breakdown torque. The
armature reaction can therefore be made
higher and ttie no-load excitation lower than
in a synchronous motor. The reduction in
the length of air gap is limited principally by
the heating of the pole faces and the increase
in the leakage reactance.
A machine designed with a large ratio of
armature reaction to no-load excitation,
however, will have a very flat phase char-
acteristic or "V" cur\-e with the minimum
input point m near the origin, curve A, Fig.
1. With no excitation whatever it will
U
Fig. 2.
Vector Diagram of Current, Voltage, and m.m f.
Relations at Zero Power-factor Leading
120
5110
L.
olOO
I
S 90
i2 60
s 60
a
5 50
.3 40
o
S
< 30
a
I 10
0 10 20 30 4 1 50 60 70 80 90 100 UO
Amperes Field in Fcr Cent of Full Lo&d Excitation
Fig. 1. Comparative PhaHc Characteristic Curves for Two
Values of Armature Reaction and No-load Excitation
Taking advantage of these factors and also
the fact that the mechanical parts, such as
the base, shafts, standards, and bearings can
be made lighter than for a motor or generator,
results in a compact and ellicient machine
requiring comparatively little floor space antl
having excellent starting characteristics.
1 ■
1
//
\
—
—
--
. —
—
__,
--
—
7
(
\
/
/
\
/
/
\
^
//
/
\
/
/
\
\
/
,/
'b
1
\
\
/
/
/
1
1
\
\
/
/
1
1
\
/
V ,
/
1
1
V
Y
1
1
operate at a lagging power-factor at only 40
or 50 per cent of its rated capacity and will
require a range of approximately 30 to 125
volts across the collector rings from the
minimum input point m to full capacity
leading power-factor. This is not objection-
able when the condenser is used solely for
raising the power-factor and is not required
to operate at lagging power-factors, since
automatic regulators can be built for this
range in excitation voltage.
The reasons for this wide range in excita-
tion will probably be most readily under-
stood by a consideration of the vector
diagrams. Fig. 2 illustrates the current.
voltage, anil m.m.f. relations at zero power-
factor leading. E is the terminal voltage
and li\ the internal voltage. Ir is the resist-
ance dro]) at right angles to the voltage and
is of so little elTect that it can usually be
neglected. Ix is the reactance dro]) which
adds directly to E. F and Ft arc the ampere-
turns required to produce the fluxes cor-
responding to E and Ei respectively. .4 is
the armature reaction.
It will be noted that, in atUlition to the
ampere-turns on the field required for the
voltage E,. there must lie atiiled an amount
equal to .4 as the armature reaction is entireh'
demagnetizing at zero power-factor leading.
The total amperc-tums requia^d for excita-
tion is represented by Fo.
SYNCHRONOUS CONDENSERS
145
At the minimum input point m, F is
approximately the same as Fi since both Ix
and Ir are small under this condition. An
inspection of Fig. 2 will show that if the
armature reaction is increased and the no-
load excitation reduced, there will be a rapid
Fig. 3. Vector Diagram'of Current, Voltage, and m.m.f.
Relations at ZeroPower-factor Lagging
increase in the range of excitation required
from minimum input to full capacity leading
power-factor. This range is represented
approximately by the ratio -^
Fig. 3 shows the relations at zero power-
factor lagging. In this case Ix subtracts from
the tenninal voltage E, and as the armature
reaction A is magnetizing it also subtracts
from Fi. The resultant excitation Fn thus
becomes very small.
When a condenser is used at
the end of a long transmission
line having considerable capacity,
it often becomes necessary to
operate at lagging power-factors
when charging the line or during
periods of light load in order to
reduce the voltage at the receiving
end. In certain instances it is
necessary to hold the voltage
constant on the high-tension side
and the requirements on the con-
denser then become even more
severe as the reactance of the
transformers adds to Ei. Fig. 2,
at leading power-factors and sub-
tracts from El, Fig. 3, at lagging
power-factors.
It is evident from Fig. 3 that
a condenser will not operate at
full capacity lagging if the arma-
ture reaction .4 exceeds Fi, the
ampere-turns corresponding to
the internal voltage E\. Theo-
retically, the proper relation be-
tween A and Fi can be obtained by lengthen-
ing the air gap and thus increasing Fi but
this results in a proportionate increase in the
full-load excitation Fo. The reactance is
slightly less with the longer gap, but this is
offset by the increased leakage between
poles. It is seldom that the additional excita-
tion can be provided on the field on account
of the limitations as to space and heating.
The armature reaction must therefore be
reduced at the same time that Fi is in-
creased, in order to keep the full-load excita-
tion approximately the same. The effect
is that the angle of the phase characteristic
cur\'e is accentuated and the minimum input
point m is moved away from the origin, as
shown by cun^c B, Fig. 1. This results in a
larger and more expensive machine. For
example, a 2000-kv-a. condenser to operate
at leading power-factors only can be built
with a core length of say 20 inches. To
operate at full capacity at both leading and
lagging power-factors, and still keep the
excitation voltage within the range of a
voltage regulator, it is necessary to increase
the length to say 30 inches. The cost,
however, is not increased in the same pro-
portion.
If separate excitation is provided for the
voltage regulator, it will operate from prac-
tically 0 to 125 volts and a considerable
saving can be effected in the size of the
condenser. This excitation can be supplied by
i
Fig. 4.
A Synchronous Condenser withSDirect-connected Exciter for the
Main Field and Small Exciter for the Regulator
146 Februarv, 192n
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 2
storage batteries or any other fairly con-
stant source. Fig. 4 illustrates a condenser
furnished with a direct-connected exciter for
the main field and a small exciter for the
regulator.
It is realized of course that by reversing
the field of the condenser the lagging kilo-
volt-amperes can be increased, but behavior
under this condition is uncertain, and no
attempt has so far been made to operate in
this manner commercialh', although it is done
in the Testing Department of the Company.
If the negative excitation is increased bevond
turbance to the line, starting induction
motors have been applied as shown in Fig. 5.
Motor-driven pimips for supplying oil to the
bearings at a sufficient pressure to lift the
rotor and thus reduce the torque required to
break the machine from rest have also been
used.
Ventilation and Noise
The problem of ventilation is the same as
for generators with the exception that con-
densers are more often in rooms by them-
selves. The smaller sizes are entirely open.
Fig.
5. Synchronous Condenser Eqxipped with an Induction Starting Motor to Reduce
to the Minimum the Line Disturbance at Starting
a certain point the machine will slip a pole
and ^continue to operate, but the kilovolt-
amperes will drop to the amount correspond-
ing'to the same positive excitation.
Starting
These machines are provided with high-
resistance amortisseur windings and arc
usually started as induction motors by means
of compensators or low-voltage taps on the
transformers. In general, the starting kilo-
volt-amperes do not exceed 50 per cent of the
rated capacity and are usually much less.
In some cases where it is particularly im-
portant to start with the minimum dis-
tho moderate sizes scni-encloscd; i.e., these
draw air in around the shaft and discharge
it into the room throu.yh holes in the stator
irame.
enclosed
through
purpose
The largest sizes are completely
air being drawn in and discharged
duels especially jirovidcd for the
All the condensers are self ventilat-
ing, air being circulated in the larger machines
by means of fans attached to the rotor
Condensers are occasionally instalitd in
residential sections and it then becomes
necessary to take special precautions to make
the operation as quiet as possible. This is
accomplished by either jiartially or totally
enclosing the machine.
147
Large Horizontal Alternating-current Waterwheel-
driven Generators and Synchronous
Condensers
By M. C. Olson
Alternating-current Engineering Department, General Electric Company
In the November, 1919, issue of the Review the writer dealt with designs of hirge vertical waterwheel-
driven generators. In the article below some features of large horizontal waterwheel-driven generators and
synchronous condensers are considered. The importance of these subjects is emphasized by the fact that the
number of waterpower installations is constantly increasing and yet only 16 per cent of the latent waterpower
resources of the world has so far been utilized. More and more synchronous condensers are being used on
power circuits; the writer describes the largest condenser ever designed and also the highest voltage condenser
built. — Editor.
T'
'HE trend in the
design of water-
wheel-driven gener-
: 1 tors and synchronous
condensers is contin-
ually toward higher
speeds and larger
capacities. As a rule,
the higher the speed
the lower is the cost
of the waterwheel, as
well as the generator
or synchronous con-
denser.
There are cases,
however, where higher speed machines may
be more- expensive than those of a lower
speed, due to the special and more expensive
M. C. Olson
%
Fig. 1.
A High-speed. (630 r.p.m.) Totally-enclosed. Self-ventilating, 10,500-kv-a.
6600- volt. Three-phase, Waterwheel-driven Generator
design of rotor occasioned by the overspeed
requirement of the waterwheel. The poles
on the high-speed machines become large and
heavy, and consequently require special con-
struction for attaching them to the field rim.
Two or three supporting brackets between
poles are sometimes required to prevent the
heavy field winding from bulging out.
The rotors of high-speed machines are
balanced with extreine accuracy, so that
when running no vibration is transmitted to
the bearings.
Fig. 1 shows one of the highest speed gen-
erators. It is rated 10,500-kv-a., 0.75-p-f.,
6600-volt, three-phase, 42-cycle, 630-r.p.m.,
and two of them were built for the Breda
Power Company, Milan, Italy. Each unit
has a 40-kw., 250-volt, compound-wound,
direct-connected exciter.
These machines are located
at an altittide of 5610 feet
and are to operate at any volt-
age between GOOO and 6600
volts. The guarantees under
these conditions are 50 deg.
C. by thermometer on all
, parts.
The machines are totally-
enclosed and are of the self-
ventilating type, the air for
\'enti]ation being supplied
through ducts in the founda-
tion and being drawn into
the rotor at each side by the
poles and by fans vi^hich are
attached to the rotor rim.
After passing through the
ir achine the air leaves through
a duct located at the bottom
of the stator frame. The
amount of ventilating air
rcTuired is approximately
oO.COf) cu. ft. i3er min. Around
\
148 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 2
the periphery of the stator frame there are
several holes which are closed by small sheet -
iron covers that m.ay easily be detached if
the operator desires to allow the warm air
from the m_achine to escape into the room
in cold weather. The enclosed shield on
Fig. 2. Rotating Field of the Generator shown in Fig. 1.
This photograph was taken before the ventilating fans
were assembled on each side of the rotor rim
each side also has several covered openings
through which to inspect the inside of the
machine. In addition to obtaining a definite
flow of cooling air through the machine, these
enclosing features reduce to a minimum the
noise occasioned by the revolving parts.
The rotor is shown in Fig. 2. The field
spider consists of two cast-
steel wheels. The poles are
attached to the field rim by
four T-dovetails, as the rotor
must be capable of withstand-
ing an overs]Deed of SO per
cent above norm.al. The rotor
has fans at each side of field
rim, but these are not shown
in the illustration. The 11 7\-
of this rotor is 420,(KJ(), this
being the amount required
by the waterwheel makers
for proper speed regulation.
Six leads, that is, the ends
of each phase, are brought
out to the terminal board of
this generator for use in con-
nection with current trans-
formers and relays for ])rotec-
tive devices. The reactance
is approximately Ki per cent,
and the test efficiency at p^ 3 ^ Generator
10,500 kv-a., 1.0 p-f. is 97. :i Fig
per cent, and at 10,500 k\'-a., 0.75 p-f. is 96.3
per cent.
A generator of similar design and ventilation
of lower speed and larger diameter is shown in
Fig. 3. Its rated output is 7050 kv-a. at 0.85
p-f., 0600 volts and 375 r.p.m. Two of these
generators were built for the Tasmanian
Govemmient, Australia, and two more are now
being built. The rotor is designed to with-
stand a runaway speed of two times normal
without distortion of any of the parts.
Alachines of different capacities and speeds
require different methods of construction
and ventilation. Waterwheel-driven genera-
tors are, in m.ost cases, self contained; i.e.,
they are supplied with shaft, bearings, and
base or foundation caps. The flywheel eflfect
required is usually embodied in the rotor of
the generator. In som.e cases the waterwheel
is overhung on the generator shaft, which
m„akes it necessar\- to design the shaft and
bearings to take care of the weight of water-
wheel and water thrust, if any.
A somewhat unusual design is shown in
Fig. 4, a 10,000-kv-a., 5000-volt, three-phase,
50-cycle, 300-r.p.m. generator, six of which
were built for the Andhra Valley Power Sup-
ply Company, India. As it was required to
provide for moving the stator in the direction
of the shaft to facilitate repairing and to
utilize the space thrust required, the venti-
lating hoods were so designed as to take in
all the air from one side instead of from both
sides of th'^ iront^rator.
of Larger Diameter and Lower Speed Than That Shown in
1. but of Similar Design and Ventilation
LARGE ALTERNATING-CURRENT WATERWHEEL-DRIVEN GENERATORS 149
Fig. 4. A 10,000-kv-a. Generator of Somewhat Unusual Design, in That All the
Ventilating Air is Drawn in from One Side
This design has the further advantage that
it relieves the weight on the bearing, due to
the overhung Pelton waterwheel, which in this
case is opposite from the usual side. The
center line of the overhung waterwheel is
32 in. from the outside of the bearing housing.
The machine is entirely enclosed. All the
cooling air is taken in through a duct below
^/**
Fig. 5. An 8,750 kv-a., 500 r.p.m., 6000-volt. Waterwheel-driven Generator
of the Usual Construction, Wherein the Ventilating Air is Drawn in
Around the Shaft and Expelled Through Openings in the Stator
the generator on one side, the duct being 4 by
(i ft., and is discharged into another duct 6 by
12 ft. 4 in. through the bottom of the stator
frame. Air dampers, made of steel plate, are
located in the air inlet and outlet to regulate
the am.ount of air and to prevent air entering
when the machine is not in use. Underneath
the feet of the stator are rollers so arranged
that, after the ventilating hoods
have been removed, tlie amiature
can be easily moved along the
shaft for repairs.
The bearings of this machine
are cooled by water circulating
through copper coils imbedded in
the babbitt. On all horizontal
machines the bearing pedestals
are insulated from the base to
prevent circulating currents that
may be produced by unbal-
anced magnetic conditions, from
flowing through the shaft and
bearings.
The rotors of two of these
machines were run in a testing
pit at SO per cent above speed
for fifteen minutes without any
distortion of field coils or poles.
The temperature guarantees on
the generators are 60 deg. F. rise
bv thermometer for continuous
ojDeration at 10,000 kv-a., 0.8 p-f.
and SO deg. F. rise on a rating
150 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 2
^ Fig. 6. A 10.000-kv-a., 300-r.p.in., 6600-volt Generator of the
Same General Construction as Tiiat Shown in Fig. 5
of 12,000 kv-a., 0.8 p-f. for ten hours based
on a room temperature of 110 deg. F.
Figs. 5 and 6 show* the usual construction
of waterwheel-driven machines, with sheet-
iron enclosing shields. The air for ventilation
is drawn into the machine around the shaft
and expelled into the room through openings
in the stator spider. This last machine has
water-cooled bearings designed to carr\- one
half of the weight of a 15-ton fl>T\-heel at the
coupling end.
Synchronous condenser designs are very
similar to waterwheel-driven generators both
in mechanical arrangement and ventilation,
except the poles are equipped with a squirrel-
cage winding for stability of operation and
for self starting. The number of slots in the
stator and the size of the rotor bars and end
rings are so proportioned as to require the
lowest amount of kilovolt-amperes at starting.
A sectional view of the largest capacity
condenser under construction is shown in
Fig. 7. It is a 30,000-k-A--a., (iOOO-volt. 50-
cycle, 10-pole, GOO-r.p.m. machine, and will
be capable of operating at 20,000 kv-a. lagging.
The machine is arranged with hoods for
the intake of approximately s.'i.Ofiii cubic
I
Fig. 7. Sectional Drawins of a 30.000-kv-a., 6000-voIt, 600-r.p.m., Totally-enclosed Synchronou* Condenser
LARGE ALTERNATING-CURRENT WATERWHEEL-DRIVEN GENERATORS 151
Fig. 8. A 12,500-kv-a., 22.000-volt Synchronous Condenser
feet of air per minute and is designed to
exhaust the air vertically at the top. A special
double ventilating hood is provided for ad-
mitting this amount of cooling air which is
drawn into the rotor by the poles and fans.
On accovint of the ver}' long stacking of this
machine and the amount of air required for
cooling, the fans at each end of the rotor are
double " and have curved blades. The guar-
antees on this condenser for continuous opera-
tion are 50 deg. C. by thermometer and 60
deg. C. rise by temperature coil, except the
field which will be 80 deg. C. rise by ther-
mometer. Special attention has been directed
to the elimination of those harmonics in the
wave shape that would produce inductive
interference with communicating lines.
The rotor, instead of being built in the
usual way of steel castings, is built of steel
plates in four sections. Each section consists
of a number of H-in. plates and 2-in. plates
riveted together and shrunk on the shaft.
The rotor center is heated to approximately
SO deg. C. above the room temperature for
Figs. 9 and 10. Stator Coils of the Condenser shown in Fig. 8, showing the method used to support
them at the ends of the windings
152 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 2
Fig. 11. Rotor with Squirrel-cage Winding for the
Condenser shown in Fig. 8
assembling on the shaft. The rotor spider
and shaft are to be shipped assembled. The
bearings are arranged for water cooling and
oil pressure will be used when starting. The
direct-connected exciter is 1.5()-kw., 2oO-Yolt,
and is compound wound. The exciter arma-
ture has a stub shaft with forged coupling and
is bolted to the end of the condenser shaft.
The magnet frame of the exciter is supported
by the bearing housing.
The condenser is to be .started
by a compensator in connection
with a 50 per cent tap in the
transformer. The potential taps
a\-ai]able for starting will be .'50,
.3712, and 45 per cent of the
normal voltage.
The highest voltage condenser
built is shown in Fig. S, being
22,000 volts, 12,500 kv-a., 500
r.p.m., three-phase, 50 cycles,
two of which were constructed
for the Andhra Valley Power
Supply Co., Bombay, India.
The temperatures arc guaran-
teed not to exceed SO deg. F.
on the armature and 100 deg.
F. on the field when operating
continuously at 12,500 kv-a.
leading, based on a room temper-
ature of 110 deg. F. (433^
deg. C).
On account of the very high p^^ ,2 ^ 5500
voltage and great expense in- exciter at
volved in making up the armature coils, un-
usual precautions have been taken in making
the coils and assembling them in the stator.
The projecting ends of the armature coils are
laced to three steel binding bands firmlj- suji-
ported from the stator.
Figs. 9 and 10 show the stator coils and
their method of support at the ends of the
windings. These windings were giv^en a
high-potential test of 50,000 volts for one
minute, between phases and between phases
and frame.
The ventilation of this condenser is the
same as that for the m.achines shown in Figs.
1 and 3, the air being taken in at each side
of the ventilating hood and expelled at the
bottom of the stator frame through an
opening 3 ft. 9 in. by 10 ft. 5 in. In order
to protect the high-voltage coils from hand-
ling as much as possible in erection, the
stator without ventilating hoods, but with
punchings and windings, is to be shipped
completely assembled. The rotor complete
with its squirrel-cage winding is shown in
Fig. 11.
Fig. 12 shows a G500-kv-a., 50-cycle, 6300-
volt, 750-r.p.m. condenser, with direct-con-
nected exciter at one end and starting induc-
tion motor at the other end.
The design of the stator and rotor of water-
wheel-driven generators and synchronous con-
densers are so proportioned that the wave
form follows verv closelv a sine wave.
kv-a., 6300-volt. 7S0-r.p.ni. Condenser, having a direct -connected
the near end and a starting induction motor at the far end
Measurement of Losses and Efficiency by
Temperature Rise of Ventilating Air
By Wm. F. Dawson
Turbo-generator Department, Lynn Works, General Electric Company
The determination of the losses and efficiency of a machine by measuring the temperature rise of 'its
ventilating air possesses many advantages which will undoulitedly make this type of test popular. Two
methods are applicable; first, measure the average inlet and outlet temperatures and the volume of the
ventilating air, assuming the specific weight and heat of the air; second, measure the aveiage inlet and oa*:let
temperatures, pass the discharged air through an electric heater of known capacity and measure the average
temperature of the air fiom the heater. J3oth methods arc described and discussed in detail belowjand
several actual tests are included for the purpose of illustration. — Editor.
d;
Wm. F. Dawson
|URING the dis-
cussion of a group
of papers on ' ' The
Method of Determin-
ing Losses." before
the American Insti-
tute Electrical Engi-
neers in 1913, H. M.
Hobart made a strong
plea for this method of
determining losses.*
The author had previ-
ously made some ex-
periments and since
then has followed the
subject with considerable success. An im-
portant contribution to the subject is con-
tained in the paper by S. P. Barclay and
S. P. Smith, entitled, "Determination of the
Efficiency of the Turbo-alternator, "t
The A.I.E.E. papers referred to and their
discussion made it plain that, with the
methods available at that time, an accurate
determination of full-load efficiencies was
practically impossible without the great ex-
pense of making special input-output tests,
in conjunction with calibrated loss supply
apparatus.
Load losses cannot as a rule be measured
directly and can be determined only by
subtracting known losses from a reasonabh'
accurate determination of all the losses made
either by the input-output method or by the
method described herewith. Obviously, the
latter system can be used only where the
arrangements for ventilating the machine
under test are such that there is a distinct
path for the ventilating air, and where it is
possible to measure accurately the average
temperature of the inlet air and the average
temperature of the outlet air at such jjoints in
the air path that the temperature difference
* Transactions A. I. E. E.; Vol. 32. Pt. I, Page 645.
t Journal I. E. E. (London). Vol. 57, April, 1919. •
is affected wholly and only by the losses of the
machine. Such losses as those of the bearings
and those by convection must be determined
in a different manner and care must be
exercised to insure that extraneous heat such
as that from steam pipes, etc., is not added
to the discharged air before its temperature
is measured.
There are two methods of determining the
losses by the air method: The first is by
measuring the average inlet and outlet tem-
peratures, and the volume of air passing
through the machine; assuming the specific
weight of the air and its specific heat. The
second method, and apparently the more
satisfactory one, necessitates the measure-
ment of average inlet temperatures and
average outlet temperatures; the passing of
Plan showmqdischarqc
opening divided into
SQuares by stretching
thin cordsacross
1-Linttd with thin
s^ect metal.
// '~<^ Expanded or
// '-^'perforated
tX nietal screens
li
Fig, 1. Temporary Discharge Trunk to
Facilitate the Accurate Measurement
of Air Volume and Temperature t
the discharged air through a suitable tunnel
or duct in which is placed an electric heater
that supplies a known quantity of heat
during the test; and the measuring of the
average temperature of the air as it is dis-
charged from the heater.
154 February, 1920
GENERAL ELECTRIC REVIEW
Vol XXIII. Xo. 2
Measuring of the Air Volume
Barclay and Smith* describe the testing
and calibration of anemometers for this work.
They also describe the Pi tot tube, Ventur
tube, and electrical methods. Their choice
was the anemometer and this was applied
« 1500
ischarqe opcnrnq ^
Fig. 2. Curves Plotted from Pilot Tube
Readings. Taken Over the Opening of
a Temporary Discharge Trunk
to Show the Effect of Baffle
Plates on the Uniformity
of the Air Velocity •
opposite the end of a discharge trunk arranged
as shown in Fig. 1. It will be noted that this
discharge trunk was fitted with expanded or
perforated metal screens to level out differ-
ences in velocity othen\- ise due to the direction
of rotation of the machine. Fig. 2 shows the
variation in velocity across the opening,
curve A with the expanded metal baffler and
B with baffle plates omitted. Our own
earliest experiments were made without baf-
fle plates by dividing the cross-section of
the discharge opening into equal squares and
. 55" -^
Top
JI6
.l?T
,!'>('
.158
.145
.145
.uw
.142
.159
.146
.146
.154
J42
.120
.122
.122
.150
.158
.140
.I5Z
.159
.157
.128
.150
.150
.150
.IZ9
.125
.107
.092
.095
.114
.141
.152
.102
.095
.094
.106
.127
.125
.120
.090
.082
.095
.102
.122
.100
.095
.060
.055
jm
.091
:9i-
J06
.090
.070
.066
JDIA
.085
.110
.110
.085
.050
.050
.062
.085
.114
.090
.079
.084
.084
.100
.108
.144
xm
.064
.065
.079
.101
.109
.101
.098
.101
.109
.117
.121
.121
.117
.115
.098
.090
,096
i
Bottom
Fig- 3. Diagram showing the Variation of Hook Gauge. Read-
ings ( expressed in inches of water i Over the Area of a
Discharge Trunk Not Having BafBe Plates
measuring the \-elocity of each square by
means of a "hook" gauge. An example of
the variations over the discharge area is
indicated graphically bj- the readings shown
in Fig. 3 and taken on a I563-kv-a., 3600-r.p.m.
•Journal. I. E. E. (London); Vol. 57. April. 1919. Page 294.
turbo-alternator. It wiU be noted that the
hook gauge readings vary from a minimum of
0.05 in. of HiO (water) to a maximum of 0.154
in. The lower value represents a velocity of
15 ft. per sec. and the higher, 26.3 ft. per sec.
Other careful tests, where the average velocity
Fig. 4. Or-fice and Impact Tube
was about the same, actually show some
negative readings. These data are given to
indicate the extreme care necessar>- for this
method of testing. The mean of repeated
tests when carefully worked out check with
reasonable accuracy- . but the labor involved
in determining the square of the average of
the square roots is considerable.
T
A
Orifice
Diam
B
■ C
D
e
F
21-
42-
59h-
56-
12-
12"
28-
48"
59fe-
45-
\6K
16V
30'
48-
591i,-
42*
I6'
16*
32-
48-
59fc-
45-
15V
15tf-
Fig. 5. Dimensions of Some MeasuKng Orifices
Tests of this sort had been resorted to more
as a means of measuring the volume of venti-
lating air to check fan calculations than to
measure the losses of the machines. Even
for the former jjurpose, however, the measure-
ments were cumbcr.some and arrangements
MEASUREMENT OF LOvSSES AND EFFICIENCY
155
were therefore made to make future tests by
means of a calibrated orifice, such as is
shown in Fig. 4. This latter method has been
described in detail by Dr. S. A. Moss.*
These tests demonstrated not only the
accuracy and accessibility of this method,
but also the effect on the quantity of air by
restrictions in the outlet ventilating ducts,
as shown in Fig. 6. These curves show the
effect of external restriction on the volume
of ventilating air and the effect of volume on
static pressure in the generator casing, f The
apparatus required for this test consisted of
a wooden elbow, a wooden pipe connecting
the discharge under the generator base to an
IS-foot straight length of 42-in. pipe, and
standardized orifices of the general shape
shown in Fig. 4. These were respectively
of 21, 28, and 32-in. orifice diameter and pro-
duced restrictions at 3600 r.p.m. (see Fig. 5) of
0.844, 0.314, and 0.189 in. of H2O respectively.
Fig. 6 also demonstrates that with machines
fitted with fans, which give practically con-
stant pressure independent of the volume,
the external restriction may approach 10 per
cent of the static pressure without reducing
the air volume an objectionable amount.
The rotation losses indicated by the zero-
field test were 24,100 watts at 3600 r.p.m.
This method of measuring the volume of
ventilating air is much more satisfactory
than either of the methods described pre-
viously. Its only disadvantage is the neces-
sity of providing a sufficient length of
straight discharge pipe and the calibrated
0.2425
*j 0.2420
"2^
t?..
r.^
j~100
0 ^"^^'V." "'^ "r *-■*: ." ".'.'"'5000"" "'■.
A '■ ■■ShortC.rcuit.5600 - I1.2XKv-a,
nStaticPres5ure /Current
» 90
^.80
't
_\Eff«ct of External IfesisUncBjeOOUPUl
6 -70
\
^ 1
<.eo
-I
-
-
-
^
-
=
=^
^
J
Turb.Er
d_
=
L
S.,c
J
\
~
-
-
i
-
=
=
\
~
-
-
h
t!
;0C
PF
M
% .40
V
\
-
III.
_
\
^ ?0
V
'1 1 f ' <
\
'"
~
M
\
t 10
Resistance^
5000 RPM
A
S
V
V
i2 .10
n
J
L
...
Fig. 6,
8000 9000 ■
Quantity of "Standard Air"
Cubic Ft. per Minute
Curves of Static Pressure in Turbine Generator Casing
vs. Quantity of Standard Air
and 20,300 watts at 3000 r.p.m. or 1 per cent
of the kilowatt rating.
*"The Impact Tube." read before the American Society of
Mechanical Engineers. Dec. 5. 1916.
t The annular spaces which are occupied by the armature end
windings and into which the fans at the ends of the rotor dis-
charge.
} Most tables and text books give this value as 0.237. but the
values shown on the curve by W. F. S. Swann are considered
more reliable.
-0.2415
1^0.2410
0-2405^Qo 60° 80' 100° 120°
Temperature of air In degrees F-.
Fig. 7. Curve showing the Variation
of the Specific Heat of Dry Air
with Variation of Temper-
ature. (From values de-
termined by W. F.
G. Swann )
nozzles; but, when these are available, the
quantity of air can be measured to within
about 2 per cent.
All measurements of air volume require
correction to "standard air" which is defined
as air at 60 deg. F. (15.55 deg. C), having a
barometric pressure of 14.7 lb. per sq. in.
(29.92 in. mercury). The weight
of such air is 0.0764 lb. per cu.
ft. and its specific heat at con-
stant pressure is 0.2416, as shown
in Fig. 7.t Obviously the specific
weight varies directly with the
barometric pressure and inverse-
ly as the absolute temperature
(60 deg. F. equals 519.5 deg. F.
absolute), as shown by Fig. 8.
There is still another source
of error, this being due to the
moisture content as shown in
Fig. 9, but it will be noted that
even with 90 deg. F. inlet air,
saturated, the correction is less
than 1 per cent. Hence, this cor-
rection has not been included.
In the following formula for
calculating the velocity and
quantity of air through the dis-
charge orifice:
a = inches of water shown by the hook gauge
j3 = inches of mercury (barometer)
27,700 represents the height in feet of a
column of air having a pressure of 14.7
lb. per sq. in. and a uniform density of
0.0764 lb. per cu. ft.
U U
L U
00
L. ^ ^
in
156 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No 2
27.7 represents the height in inches of a
water coltimn for one lb. per sq. in.
459.5 deg. F. is the absolute temperature
of zero deg. F.
29.92 is the barometric reading correspond-
ing to 14.7 lb. per sq. in.
Velocity (Ft./Sec.) = Orifice CoefiV-Z gh
... |64.34X 27.700 X a
= °-^N 14.7X27.7
= 0.99V4375Xa = 65.6"v/a
The above is correct onlj- for 'standard
air.
0 085O
j: 0 0700
Q0650
^
W W 70* 90° 110° 150*
Temperature of air indegrcesF
' Fig. 8. Curve showinE the Weight of
a Cubic Foot of Dry Air at Vari-
ous Temperatures. iBarom.
29.92 inches of mercury.)*
Actual velocity of non-standard air
r-f ;^y 459^5 + 7^ 29.92
'459.5-1-60 ■
/3
= 65.6X0.241
|aX459.5-f Ts
^
= 15.72-.
a X 459.5 -I- r,,
Quantity (Cu. Ft./Min.) of "standard air"
from measurements of non-standard air
= 60 X 15.72 X
4
aXmi-y + T, 519.5 g
/3 ^ 459.5 -I- r, 29.92 ^^"l-^^
P
= 60xI5.72xI7.:j.SVaxv,-n - , -r X^q- Ft
= 16,3S0\/aX-Jp:7r|-r=FXSq- Ft.
\4o9.5+rj
•Journal I. E. E. (London). Vol. 57. April. 1919.
The heat carried off by one cubic foot of
standard air per minute for one deg. C. rise
equals :
Dee F
lb. per cu. ft. XSp. ht. X„ ^ ' XFt. lb. in one B.t.u.
Deg. C.
Ft. lb. in one Watt-minute
0.0764X0.2416X1.SX77S
33000
= 0.585
746
Second Method
It is obvious that if the inlet temperature
and outlet temperature of the ventilating air
lOIOtr
"^' 40* 50° 60° 70° «r
D«w point tcmDcrature of iirat inlet m D«^
Fig. 9. Correction in Respect of Variation of Air Density
and Specific Heat with Humidity. (This correctioa
may be ignored for commercial work and is not
taken into account in any of theformulK.)*
are accurately recorded after having reached
a steady value, and that if after the outlet
air has passed the thermometers, an electric
heater be introduced with sufficient energy to
raise the temperature an amount equal to
that caused by the losses of the machine, the
energy dissipated in the heater will equal the
losses. In practice, it is more convenient
to supply a heater of fixed resistance with a
constant input from a constant potential
supply: the input being as large as possible,
preferably equal to the full-load losses of the
m.achine. In this case the following simple
formula applies:
Loss in watts = H^ — '
wherein :
ri = Inlet temperature
rs = Outlet temperature
Ti = Temperature after passing the heater
H = Heater watts
This method has the advantage that cum-
bersome and expensive air flues and hook-
MEASUREMENT OF LOSSES AND EFFICIENCY
157
fjaugc readings with awkward fluctuations arc
avoided. Also, there can be no dispute as to
the specific heat of the air, the effect of
humidity, or the corrections for temperature
and barometer. The volume of standard air
volumes of the air were measured by a
discharge pipe with calibrated orifice and
"hook gauge." The tests indicate a consid-
erable variation in the quantity of ventilating
air. It is hard to explain why test No. 8
TABLE NO. I
Air Readings
5750 Kv-a.. 2300 Volt, 3600 R RM Turbo- A Iter nator
28 in. Orifice. - 4.2g Sq-.Ft.
Test
Number
HookCiauige
Inches
Water
Barometer
Inches
Mercurg
Air Temperature at
Nozzle
Ts in Degrees F.
Vc, Inciti) ( Ft.Der Sec.) reduced to
•standard Air">E73V"H20X'Jg^^^J±9pj^
1
0.309
50.15
9 4. 5
275X0.5S5X O.Z53 -35.3
2
0 288
30.39
9 5. S
273X 0.537 X 0.234-34,4
5
0,521
30.354
100.7
2 73X0.5S7 Xn7?i3-360S
4
0.515
30 078
102.0
273X 0.561 X 0.231 = 35.4.
5
0508
30.354
103 0
2 73 X 0.55 5 X O 2 32 -36.4
e
0 329
30.14
106.0
2 73 X 0 58 2 X O 2 31 = SS. 7
7
0.330
30.15
120.0
275X 0.575X 0.228 -55,8
8
0.Z39
30 584
122. 0
273X0489X 0.229 -30.6
9
0.505
29.97
112.0
273X 0 552 X 0.229 = 34.5
10
0.317
30.36
116.5
2 73 X 0 565 X 0.2 30 = 35.4
11
O30O
30 3S
124.0
2 73X0 548X0.228-34.1
TABLE NO.n
Losses and Efficiencies By Air Tests
5750 l<v-a..25D0 Volt.5600l?.P.M. Turbo-Alternator
Calibrated Orifice
Test
Number
Kv-a.
Load
Power
Factor
Volts
Arm.
Amp
Arm.
Volts
Field
Amp,
Field
Air DeareesC.
CuFt/Min
Stafid.Air
Kw.
Loss
firm Rise =icld Rise
DegC DegC.
,Tl
Tz
1
Zero
Field
2 6.7
5 5.7
9070
47,7
7.4
76
2
Zero Field
27.5
368
8840
48,2
10 0
7,7
3
100%
O.C,
2 300
38.5
8 5.0
25.3
4 0.0
9 2 50
7 7.2
19,2
16.6
4
1 1 0%
OC.
2 530
45.9
990
23.0
3 8.9
91C0
850
20-0
28.2
5
100%
S C
941
55 7
1115
24.5
4 1.0
9350
90 2
42,0
56.4
6
12 0%
S C.
1140
67.0
1536
23.2
4 3.0
9410
110.0
44,8
650
7
3770
100%
2270
960
73.0
141, a
29 0
5 3.3
9200
113 0
51 0
510
8 •
3760
80%
2260
960
100, 6
180,4
29 0
54.0
7850
115 0
44,3
73,5
9
3700
100%
2280
935
70,0
1400
2 8 0
48.0
8870
98.5
41.5
43.3
10
5900
100%
2 2 00
990
75,0
142,0
P7''^i
^^ S
9100
967
39.2
5T5
11
402O
82%
2 260
1055
104.5
192,0
??m
U;?
8750
121.3
<7 O
77.0
Losses
Test-
Numbci
ConvttctKMl
*
10
1.490
1.610
1.675
2 540
Bearing
rnction
10.00
10.00
10.00
10 00
10. 00
Windage
47.95
47 95
47 95
47.95
47 95
47.95
47.95
Core
Loss
2553
25 55
25,53
25 55
25 53
Ann,
17.30
T5TT^
Load
Loss
1897
51.90
1115
"X3T
17,00 -178
19, 10
20,90
6 82
I^'R
Field
5.27
4,54
5.48
10.37
18.20
10.65
20. 10
Rheo
Total
Loss
7.40
4,40
Percent Loss
rullL(ad|54Load|ll;Load
5914
8869
9661
12200
342
Kw Lossit
^roLobd
Normal
Voltajc
655
96.eo
* Convection Surface- 180X45= 8100s<).in.
can be quickly computed from these tests:
,, , Heater watts
Tables I and II give data on air tests of a
3750-kv-a., 2300-volt, 3600-r.p.m. turbo-alter-
nator. Mercurial thermometers were used
for measuring the temperatures of the inlet
air and outlet air in these tests,* and the
* Electric thermometers were inserted as a check in tests Nos.
10 and 11. but the results are not tabulated.
should show only 7850 cu. ft. of standard air
while tests No. 11 should show 8750 cu. ft.,
but in spite of this apparent discrepancy, the
load losses worked up as 5.32 kw. and 6.82 kw.
respectively — a discrepancy of only about
one per cent of the total losses.
One difficulty experienced was caused by
the shape of the discharge elbow creating a
centrifugal swirl which, impressed upon the
discharged air was sufficient to give decidedly
uneven readings across the calibrated orifice.
158 February, 1920
GEXER.\L ELECTRIC REVIEW
Vol. XXIII, Xo. 2
Under more favorable conditions, the read-
ings across the calibrated orifice var\- less
than one per cent from a single reading
taken at the center. A wooden cross of
boards, having an axial length equal to about
twice the diameter of the discharge pipe, was
placed at inlet of the pipe and eliminated the
difficulty.
Another source of perplexity was the fact
that on some readings, at least, there was a
greater variation between the maximum and
minimum readings of the inlet air than be-
tween the average of these readings and the
average discharge temperature.
An insurance against errors due to im-
properly averaged air readings seems to lie
in the use of electric resistance thermometers,
either as a substitute or as a check to the
mercurial thermometers. Ver\- satisfactory'
results were obtained bj- plotting the com-
plete cross section of both the inlet and outlet
air with standard ten-ohm resistance coils
arranged four in series and four in parallel,
thus giving the average of 16 readings with
what was in effect a ten-ohm coU. A spe-
cialized application of this principle, consisting
of wooden frames exactly fitting the air inlets
and air discharge pipes was made as follows:
Wooden pegs were fitted to peripheries of
these wooden frames and at approximately
equal intervals, Xo. 24 (0.020-in.) copper
wire was zigzagged around the wooden anchor
pins, vertically on one side and horizontally
on the other side of the wooden frame; in
sufficient quantity- to provide the standard
resistance. Ten ohms (25 deg. C.) were pro-
vided in the discharge pipe and five ohms in
each of the two inlet areas; the latter coils
being connected in series. Actual tempera-
tures, Ti, Ti and Ts, were read on a standard
temperature indicator wherein a standard-
ized ten-ohm manganin wire resistance forms
one arm of the Wheatstone bridge and the
heated ten-ohm copper wire, under obser-
vation, forms the other arm of the bridge. A
refinement of this method would be to sub-
stitute the ten-ohm coil T2 for the constant
resistance manganin coil of the instrument
and to read 7"i and T3 as differences instead
of as actual temperatures. Besides auto-
matically averaging the temperatures, the
electric thermometer has the advantage that
it can be placed sufficiently close to the
machine to insure against any appreciable
change in the air temperature between the
source of cnerg\' loss and the thermometer.
Also, the instrument for indicating the tem-
perature is alwaj's visible and is fully acces-
sible, while mercurial thermometers if prop-
erly placed are inaccessible.
It is natural to suggest that all of the
generator losses are not indicated by the
temperature rise of the ventilating air,
particularly in respect to the heat radiated
from the generator casing. Data vary as to
the convection loss which ensues from a
given area at a measured difference in tem-
perature between the exposed surface and the
room. Some data are available indicating
that there is only — - of a watt dissipated per
square inch for each degree centigrade dif-
ference in temperature. The writer, however,
has noted a loss of approximately — of a
watt for each degree centigrade difference per
square inch of black surface. This figure is
used by the author in his computations and
it appears that there is rather less than 5
per cent of the total loss radiated from the
shell of a 500-kw. 3600-r.p.m. turbo-alter-
nator and that the ^"alue drops to approxi-
mately two per cent in the case of a 30tiO-kw.
machine. Certainly a fairly close approxi-
mation can be made by computing the area
(in square inches) of the stator frame exposed
on the inside to the heated air from the core,
and on the outside to the room temperature,
and assuming a loss of 3- of a watt per square
inch per degree centigrade temperature dif-
ference. Other losses consist of those in the
bearings and in the field rheostat, where these
arc chargeable to the turbo-generator. Bear-
ing losses can be computed from the bearing
reactions and cur\xs of co-efficient of friction,
provided these factors are known. AATiere
these data are not available, it is quite
accurate to assume the loss equal to one third
of one per cent of the kilowatt rating; where
the generator is "maximum" rated, ."JCjOO
r.p.m. and O.S p-f.
The rheostat losses will be:
(Fid. amp.) X (Exciter volts — Collector
ring volts).
In respect to hand-wound electric resistance
thermometers, it is important to guard against
long spans of fine wire. There appears to be
good evidence that these may be stretched
sufficiently, by the pressure of rapidly moving
air, to increase their resistance and impair
their accuracy as much as one or two degax^s.
Where the span is great, it is best to wind the
resistance wires as helices on glass or wooden
MEASUREMENT OF LOSSES AND EFFICIENCY
159
rods, to mount these in the outside frame,
and to connect the helices together. The
supports of the helices must be thin enough
and be spaced sufficiently far apart to prevent
undue obstruction to the air flow.
Electric Heater
It seems advisable to utilize heaters spe-
cially built for the air tunnels employed, in
order that the air be heated uniformly across
the duct section. Excellent results were
obtained from 0.015 by 1.5-in. German-silver
to carry about 25 kw. and the frames were
set in the discharge duct about eight inches
apart. Due advantage was taken of the fact
that, when in use, these heaters would be
subject to a strong blast of air. A tempera-
ture rise of about 85 deg. C. was considered
satisfactory and the expected heating was
based on the formula:
Deg. C. rise for one watt per sq. in. =
896
VAir velocity, ft. per min.
TABLE NO. nr
Losses and Efficiencies ByAi'rTests
1250Kv-a.,Z300 Volt, 3600 R. P.M. Turbo-Alternator
Electric Heater Method
Test
Number
Kv-a.
Load
Power
Factor
Volts
Arm.
Amp.
Arm.
Volts
Field
Amp.
Field
AirDaqrees C.
Heater
T5-T2
Heater
Watts
Kw.
Tz-Ti.
Tl
T2
1
Zero Field
Hq29.4
El 31.5
33.2
36.0
15.5
14.5
48.670
12 0
15 1
Avq.»
13 55
2
110%
O.C.
2550
5L8
59.5
Hg277
£131.5
38.7
41.5
15.9
16 0
51 500
55.6
32.2
Avq =
53.9
3
120%
S.C-
377
68.5
72.3
Hg36.8
£140.7
46.8
52.0
16.4
14.1
50 400
40.3
30.7
Avq,=
35.5
4-
lOO^o
80%
2300
314
110
102
Hq331
E 1 572
472
52.8
16.3
14.4
50.030
45.2
54.2
Avq. =
48.7
Losses
Test
Number
Convection
*
Bearing
Friction
Windage
Core
Loss
12R
Arm.
Load
Loss
1=R
Field
Rheo.
Total
Loss
Percent Loss
Km Loss at
ZeroLoac
Normal
Voltage
Full Load
Uload
y2 Load
1
0.197
'/3%
3.533
12.00
13.55
15.10
—
15.53
17.08
18.65
2
0.504
3.333
12.00
13,55
15.10
20.51
17.25
14.00
3.09
39.43
3 2.73
36.03
3
0.508
3.333
12.00
13.55
1 5. 10
12.08
10.52
4.19
-2 15
4.95
44.10
3 9.35
3 4.54
4
0.720
3.333
12.00
13.55
15.10
16.90
14.20
11,50
8. 90
-6.05
-6.32
7.5
11.20
1.53
4 853
53.11
5 9.76
5.03
5.88
7 60
58.55
5 +
0.720
3333
12.00
13.55
15.10
16.90
14.20
11.50
8.90
10.00
1.40
52.2
4.00
4.68
6.2 7
38.55
JtConvection Surface= 5840 Sq in. +Computed at 1.00 Power- Factor, lOOyo Kv-a
ribbon having a resistance of about 150 ohms
per square mil-foot. The ribbon was cor-
rugated by being run through loosened gears
having about ^-^-in. pitch. This provided
sufficient resilience to prevent any slackness
when heated and also increased the radiating
surface for a given span. The strip was
zigzagged vertically on one side and hori-
zontally on the other side of a steel frame
provided with pins having porcelain insulator
supports and the ends were provided with
ample terminals. Each frame was arranged
A current of 139 amp. for each 0,015 by 1,5-in.
ribbon gave 3.57 watts per sq. in. and con-
sequently (with an air velocity of 1500 ft.
per sec) an expected rise of 82.5 deg. C.
The current density was 6170 amp. per sq. in.
Repeated heat runs of 33^ to 4 hours gave
wholly satisfactory results, but due to the
difficulty of placing thermometers no tem-
peratures were taken. That the temperatures
were consen^ative is shown by the fact that,
on the initial trj^out, the heater ribbon was
zigzagged around wire nails driven into a
160 February, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII. Xo. 2
TABLE IMO.H
Losses and Efficiencies By Air Tests
5750 Kv-a.,5000 Volt, 3000 R.P.M. Turbo-Alternator
E-lectric Heater
Test
Number
Kv-a.
Load
Power
Factor
Volts
Arm.
Amp.
Arm.
Volts
Field.
Amp.
Field
Air DcqracsC
Heater
T^-T2
Heatar
Watts
Tz-T,
flrmRisa
Deg.C
riekil?i«
Ti
T2
1
Zero Field
Hg29.7
E 1 30.7
34.5
36 8
11.3
12.7
49.5
21.10
25 80
Avq.=
22.4 5
11.5
12.2
2
llO'^
O.C.
3300
55.3
129
Hq280
E 1 51.5
44.5
50.0
15.0
110
49.7
54.€0
8 3.50
25.0
39.1
Avq-
69.05
3
A
120%
100%
S.C.
ao%
867
58.5-
96-0
129
200
Hq270
E 1 300
42.0
47.0
14.5
12.0
49.6
51.50
70.20
44.0
57.3
Avq.=
60^5
1
Test
Number
Comection
«
0.573
0.850
1.970
2.210
1.790
2.050
Z.500
Bearing
FriclKKi
lO.O
10.0
10.0
10.0
Losses
Windage
21.10
2 3.80
Avq.=
22.45
21.10
2 5 80
Avq.-
22.45
21.10
25.80
Avq. =
22.45
22.45
Core
Loss
26.35
52.55
Avq.=
39.44
32.60
Arm.
22.0
15.2
Load
Loss
0.65
16.85
Avq. =
8.81
e.l2
Field
7.15
7.55
^Convection Surface. 180X55«955O Sq-.in
19.20
Rheo
Ava =
53.16
S.a
Total
Loss
Fter Cent Loss
FullLoadBiLoadl^Load
31.67
34.65
66.57
95.71
Avq =
81.14
65.09
82.25
Avq.=
7 2.66
113.67
5.62
4.25
KlKLosSAt
Itntoati
Normal
Voltage
5.65
80-52
TABLE N0.2
Lo5Szs and Efficiencies By Air Tests
4575 Kv-a, 5600 Volt, 5600 R. P.M. Turbo-Alternator
Electric Heater
Test
Number
Kv-a.
Load
Pby»«r
Factor
Volts
Arm,
Amp.
Arm.
Volts
Field
Annp
Field
AirDeqrecsC
Heater
T5-T2
Heater
Watts
K«v
T2-T,
arm Rise
Deq^C
DegC.
Tl
T2
1
Zero
Field
Hq24i3
El 26^
55.0
35.5
9.e
10.0
49 3
46.2
44.4
Avq -
12.5
6.03
2
110%
O.C.
3960
56.2
151
Hg 2 1-5
El 240
41.5
46 0
11 0
100
501
91.0
HOC
29.5
S4.6
Ava-
96 0
5
12 0%
S.C.
867
584
129
Hq22.5
£1250
400
440
126
100
502
69 7
955
480
61.0
Avq-
826
A
100%
80%,
97.0
202
Losses
Test
Number
Con»edion
*
Bearing
Friction
Windage
Core
Loss
12R
Arm
Load
Loss
Field
Rheo
Total
Loss
PiirCent Loss
OUnsat
Ecrx> LOM
Normal
Voltl^
FullLoad
Xload
Moad
1
1.070
1,070
12.0
46 2
44.4
59 27
57.07
Avq -
58 37
Avq -
45.5
2
2.385
2 150
12 0
46.2
44.4
3S44
57. ?4
Avq- =
47.35
7.360
10558
12515
Avq «
45 3
Avq.-
11426
3
2.090
2.260
12.0
46.2
44.4
Avq-
45 5
22.0
-608
7 540
85 79
1?9 7^
Avq •
96l7
/I
2 500
12.0
4 5.3
39.00
IS. 2
5.4
I960
565
134.7
561
4 2S
5 3.^
1:0 3
•* Convection Surface- 180X55-9550 Sq. in.
MEASUREMENT OF LOSSES AND EFFICIENCY
161
frame of soft wood and there was no signs of
charring after the experiments.
A few examples of air m.easurements, illus-
trating the different m.ethods outlined in the
foregoing and the results obtained, are given
in Tables I, II, III, IV, and V.
The tests recorded in Table III were
made by using the electric heater in the air
discharge directly after the air had passed the
thermometer indicating outlet temperatures
(Ti). The air temperature was measured by
both mercurial thermometers {Hg) and elec-
tric thermometers (El). The results averaged
from the two methods are also indicated.
It is difficult to reconcile some of the dis-
crepancies observed, but the author favors the
results from the mercurial thermometers in
this particular test. He feels that the ac-
curacy of the electric thermometers was
impaired by the stretching and vibration of
the unsupported resistance wires. It is hoped
that this difficulty will be considerably re-
duced by winding the resistance wires as
helices upon suitable supports.
As indicated in Table IV, no tests were run
under energy load but efficiency at full load
of 3000 kw. and 80 per cent power-factor was
computed from the three tests made at zero
field, 110 per cent voltage open circuit, and at
120 per cent kv-a. current short circuit. These
tests were duplicated at 3600 r.p.m.. Table V.
The open-circuit heat run was here made at
the same flux as previously, the short-circuit
heat run at the same current and the efficienc\'
computed at a rating of 3600 kw., SO per cent
power-factor.
The tests were made by introducing an elec-
tric heater of approximately 50-kw. capacity
just outside the thermometers measuring the
discharge air. The tabulations indicate the Hg
readings averaged from the mercurial ther-
mometers, the El readings averaged from the
resistance thermometers, and the average of
these.
The efficiencies were computed from the
average of the two methods.
It is interesting to note that, in spite of
some apparent discrepancies, the full-load
losses at the two speeds check within 1/100
* It is important to observe that the load losses wnich exist
on short circuit are often very greatly reduced when the machine
is running at full voltage and rated power-factor, see Tables I
and II. The A. I. E, E. Standardization Rules. 1918. page 458,
provide that the load losses measured on short-circuit tests
shall be charged against the efficiency of the machine. Our air
tests have demonstrated that, in certain machines, this is a
reasonable rule; but there are sufficient tests on other machines
to show that load losses which are of considerable magnitude
when measured on short-circuit test are practically eliminated
at rated load and power-factor. Dr. S. F. Barclay pointed this
out in his paper on "Mechanical Design of Turbo-alternator
Rotor," page 482, Journal I. E. E. (London). Vol. 56. July, 1918.
of one per cent and that the losses at frac-
tional load also agree very closely.
This machine shows the remarkably high
efficiency of 96.4 at full load, SO per cent
power-factor.
As measured by mercurial thermometer,
the short-circuit test run at 3600 r.p.m. shows
a negative "load loss" of 6 kw. This is un-
doubtedly due to some error in the test.
The procedure in making such tests should
be as follows wherever possible :
(1) Zero-field heat run. This will indicate
the windage friction of the machine plus the
power required to drive the self-contained
rotor fans.
(2) Open-circuit heat run at normal volt-
age and also, if possible, at 110 per cent
normal voltage. This will indicate the core
loss by subtracting the windage loss and the
loss occurring in the fields.
(3) Short-circuit tests at normal kilovolt-
ampere current.
(4) Short-circuit tests at 120 per cent
kilovolt-ampere current.
The load-loss can be obtained from test
(3) by subtracting the windage loss, the PR
loss of the armature, and the field loss (volt-
amperes at the collector rings). The test at
120 per cent kv-a. current is used as a check.
This load loss will include the short-circuit
core loss, the loss due to the circulation of idle
currents in the armature conductors, and
whatever losses are induced in the pole pieces
and the retaining rings by induction from the
armature windings.*
(5) Full-load unity power-factor test.
(6) Full-load fractional power-factor
(usually 80 per cent) test.
(7) Tests at full kilovolt-amperes and zero
power-factor can sometimes be taken when
(5) or (6) are not available on account of in-
sufficient power. Such tests are valuable but
(5) and (6) are preferable.
It is obvious that, to be reliable and to be
in accordance with contract requirements,
tests of this class should be continued long
enough to insure constant temperature in the
various parts of the machine as well as in the
rise of the ventilating air. Our experience
in this line of apparatus indicates that con-
stant temperatures will be attained in from
three to four hours.
Acknowledgment is made to the Turbine
Department of the West Lynn factory (for
permission to publish the results of tests, and
also to the Turbine Research Department for
guidance and assistance in making the air
tests and checking the formulae.
162 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 2
Bearings and Lubrication for Vertical Shaft
Alternators
B3- T. W. Gordon
Alternating-current Engineering Department, General Electric Company
We have published several articles in the Review on the new type of spring thrust bearing that has been
developed for large vertical shaft machines. These articles dealt chiefly with the design of the bearing and
the success that it has attained in actual installation. The present contribution discusses methods that have
been adopted for lubricating and cooling this type of bearing. On some of the larger units a central station
oiling system connecting with all main bearings is found to be necessary-, and in such systems the arrange-
ment of piping may be greatly simplified by careful study. In the smaller units a self-contained oiling sys-
tem is often provided, which gives excellent results and requires very little attention. — Editor.
T'
^HE usual vertical
shaft alternator
IS equipped with a
I thrust bearing and
two guide or steady
bearings. The thrust
bearing supports the
weight of the rotor
and is of first im-
portance because of
the ver\- heavy loads
often imposed upon
it. The duty required
of the guide bearings
is less severe. Many
conventional designs of horizontal journal
bearings, with sttitable changes in the oil
grooves, mav be used for vertical shafts.
T. W. Gordon
turbine runner is suspended from a bridge
or bearing deck, over the generator, on which
the thrust bearing is located at the upper
end of the shaft. Fig. 1 shows a standard
type of vertical waterwheel-driven 2857-
kv-a. 55.5-r.p.m. generator. The thrust
bearing is located in the housing at the top
and carries the entire weight of the revolving
parts and the water thrust. The total load
or downward thrust is 340,000 lb.
Fig. 2 shows a spring thrust bearing built
for large vertical shaft hydro-electric gen-
erators. The rotating ring (standing on its
edge) has radial oil grooves in the rubbing
surface. This part is made of a special
grade iron, and the bearing surface is ground
and polished to a high degree. The stationary
bearing ring (which is raised to show the
Fig. 1. Modern Type of Large Vertical Waterwheel-driven Generator. The thrust bearing is located
at the top and supports the weight of the entire revolving clement, plus the water thrust
In the design of hydro-electric generating
stations, the early practice of supporting
the rotating parts of the generator and
waterwheel on a step bearing below the
generator has been discontinued. Now the
combined weight of the generator rotor and
springs and dowel pins) is made of steel with
a babbitted rubbing surface. It is a continu-
ous ring, but it has a saw ait in one oil
groo\e to prevent an>- tendency of the plate
to "dish" with a change in temjicrature.
The total thickness of the babbitt is small as
BEARINGS AND LUBRICATION FOR VERTICAL SHAFT ALTERNATORS 163
compared with the diameter. This flexible
ring rests on a multiplicity of steel springs,
which press the babbitted surface against
the rotating ring with approximately the
same intensity at all points.
Fig. 2. Spring Thrust Bearing with Babbitted Stationary
Ring Raised to Show Springs and Dowel Pins
Fig. 3. Plate with "Compressed Springs" Used in Thrust
Bearings for Machines Having? Small Clearances
Heretofore, the aim of bearing designers
has been to provide a very rigid support for
the rotating part. Such bearings must be
carefully fitted to the running surface in
order to have the load well distributed. With
an oil film of only about three ten-thou-
sandths of an inch in thickness between the
rubbing surfaces, it is quite evident that
Figs. 4 and 5. A comparison of these illustrations shows the extent to which oil piping may be simplified
164 February, 1921)
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 2
even with good macHining and erection ver\-
severe conditions ma}' exist in a rigidly sup-
ported thrust bearing. The spring supported
thrust bearing furnishes the runner with
a flexible support which will automatically
adjust itself, while in operation, to any
Fig. 6.
If^ft e-rfiC
Arrangement of oil and water pijTCs on large vertical-ihaft Kenerator&
when connected to a station oiling system
n ff
Z>sii''r
Fig. 7.
Arrangement of oil and water pipes for tclf-ciling vertical-shaft
generators having a combined thrust and upper bearing
tendency toward unequal distribution of the
load, caused by inaccuracies in workmanship
or alignment. This flexibiHty is particularly
advantageous in connection with large gen-
erators, which cannot be constructed as
accuratelv as small ones. With the ever
increasing size of hydro-electric units, the
thrust bearing with a flexible support will
be found superior to any of the rigid types.
Some of the smaller waterwheels have ver>-
little vertical clearance between the runner
and the stationar\- parts. In such cases.
it is sometimes desirable for con-
venience in erecting to precom-
press the springs to a position
corresponding to full load on the
bearing. This is accomplished by
the use of washers and clamping
screws. Fig. 3 shows a set of "com-
pressed springs." When a bearing
with compressed springs is installed
there is no further deflection of the
springs as a whole while the weights
of the generator and waterwheel
rotors are being placed on the
thrust bearing. If, however, there
are high local pressures on any part
of the babbitt surface, the springs
directly below will be further com-
pressed and the pressure at these
spots will be relieved before it
reaches a value that will cause
■'wiping" fo the babbitt.
Many power plants containing
vertical shaft alternators have
elaborate central station oiling
systems lor furnishing large quan-
tities of oil for cooling the thrust
bearings. These systems include
an extensive equipment of large
filters, tanks, pumps and pipe
lines, together with thermometers,
meters and any other devices which
might aid in insuring a continuous
supply of clean, cool oil to the gener-
ators. The installation of water
cooling coils in the thrust bearing
housing to remove a part or all of
the heat from the bearing will allow
of a considerable reduction in the
capacity and cost of the lubricat-
ing system. Fig. 4 shows a typical
installation of thermometers, inte-
grating flow meters, oil sights,
strainers, and large return pipes,
such as has often been used on
generators in the past. In addition
to these thermometers, there is a
recording thermometer on the thrust bearing
housing with its bulb in the oil bath near
the bearing. Thermometers in return pipes
as here shown are a]3t to be unreliable because
these pipes do not run full and the bulb of
the thermometer is often not covert^d bv the
3
BEARINGS AND LUBRICATION FOR VERTICAL SHAFT ALTERNATORS 165
oil. On account of the small stream of oil
passing through a guide bearing, there will be
a considerable drop in the temjierature of the
oil before it reaches the thermometer at the
outside of the machine. Oil sights in both
feed and return pipes are not necessary and
the records of the meters showing the amount
of oil pumped per day are of little value.
The oil piping seen on the generator in
Fig. 1 is sufficient to meet the requirements
for successful operation of all bearings. The
water pipes to the cooling coils in the thrust
bearing housing are seen at the left of the
photograph, and the valve connected to three
small pipes at the right controls the operation
of the brakes. The oil pipes are shown to
better advantage in Fig. 5. A comparison
of the equipments seen in Figs. 4 and 5
indicates the extent to which the oil piping
may be simplified and the reduction in the
size of pipes for the thrust bearing when
cooling coils are used. The drain pipes have
openings for observation of the oil flow. A
mercury actuated indicating dial thermom-
eter is mounted in a conspicuous place on
the bearing bracket arm and is connected
by a capillary tube to its bulb, which is in
the oil bath near the thrust bearing. Ther-
mometers for the guide bearing are not
considered necessary.
Fig. 8. A 1000-kv-a., 360-r.p.m. Waterwheel-driven Generator
connected Exciter. All bearings are self-oiling and
station oiling system is not required
Fig. 9. Pump in Oil Pan of Lower Guide Bearing on Ver-
tical Generators. One-half of pan removed
for inspection of pump
Fig. 0 is a sketch showing the arrangement
and size of pipes for a large high-speed gen-
erator connected to a station oiling system.
All thrust bearing pipes pass under the plat-
form and up to the oil space. With
this arrangement the platform is
free from obstructions. By using
cooling coils in the thrust bearing
housing, the oil required in one
case was reduced from sixty gallons
to five gallons per minute. Such a
reduction in the oil for each gener-
ator makes possible a correspond-
ing reduction in the size of the
station oiling system.
It is very desirable to make the
smaller units self-oiling and thus
eliminate a central station system.
Fig. S shows a 1000-kv-a. 3(30 r.p.m.
generator which is self-oiling. The
upper guide bearing is above the
thrust bearing and runs in the
sam.e oil bath, from which the
heat is removed by the water cool-
ing coils. Fig. 7 shows the arrange-
ment of bearings and oil piping on
this self-oiling unit. There are
radial grooves in the rotating sur-
face of the thrust bearing which
pump oil from the thrust bearing
up through the upper guide bear-
ing. The oil for the lower guide
with Direct
166 Februarv, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 2
bearing is pumped up to a convenient height
where the operator on the generator floor
can regulate and obser\-e the flow to this
bearing. In the right-hand view the small pipe
at the left comes from the ptunp and returns
to the lower bearing. The pipe between the
water coil connections is for filling and
draining the thrust bearing housing. An oil
pimip is located in the drain pan of the
lower guide and circulates the oil for this
bearing. The pump is geared to the generator
shaft and is mounted so that one half of the
oil pan may be removed without disturbing
either the pump or the discharge pipe leading
to the floor above. Fig. 9 shows this arrange-
ment. The oil in these generators is removed
occasionally and the housings refilled with
new or filtered oil.
Large generators, having the upper guide
bearing below the thrust bearing, as in Fig. 6,
are sometimes equipped with unit oiling
systems. This is done by the addition of a
pvunp located in the lower drain pan, as in
Fig. 7. The returns from all bearings are taken
to a small filter 30 by L5 by 30 inches deep
mounted on the wall of the pit below the
generator. The oil is pumped from the clean
oil compartment of this filter to the bearings.
A Unique Design of Waterwheel-driven Alternator
By A. E. Glass
Alternating-current Drafting Department, Gener.\l Electric Company
The waterwheel-driven alternator described in this article is one of three which are being built to operate
in a powerhouse hewed out of the solid rock of a mountain in Norway. The conditions which prevail in this
most unusual installation have necessitated the adoption of a unique design for the generating units. These
features of design are clearly expla-ned and illustrated by the author. — Editor.
A. E. Glass
nfHE increasing de-
-'■ mand for large
waterwheel-driven
alternators of high
speed has led to the
design of some rather
tmusual units, espe-
cially those of the 25-
cyclc type. These
alternators are novel,
not only in the
method of ventilation,
but in the construc-
tion of the revoh-ing
element or the "re-
volving field" as it is commonly called.
Alternators of high rotative speed necessarily
have high stresses. The unit is liable to an
overspeed of from SO to 100 per cent in case
the electrical load is suddenly removed and
the waterwhecl governing mechanism fails to
function. In order that no failures occur at
this overspeed, the rotors should be so
designed that the stresses at the overspeed
will not exceed one half the elastic limit of
any of the material. These considerations
lead to the use of as small a diameter as
is consistent with other factors in the design.
Therefore the core length of the alternator
must be neccssarilv long for units of lanrc
capacity; and it is at once apparent that,
with small diameter and long core length,
ventilation becomes one of the most difficult
problems. The location and lajout of the
power house is also a determining factor of
design with respect to ventilation. The
General Electric Co. has under construction
three 2.j-cycle, l-i.OOO-volt. 750-r.p.m. hori-
zontal shaft generators for the Aktieselkabet.
Saudefaldcne, Xorway. to be direct driven
by waterivhcels built by A. S. Myrens
Veakted. Christiana, Xorvvay.
Generating Unit
Each unit consists of a generator, flywheel,
direct-connected exciter, and waterwheel.
The electrical part of the set consists of an
alternator mounted between two bearings
and an exciter overhung at one end. The
mechanical part of the set consists of a fly-
wheel moimted between two bearings with
the waterwheel overhung. The generator,
flywheel, and exciter are mounted on a com-
mon base. The bearing which supports the
weight of the waterwheel runner must carr>"
in addition a force of 2000 lb. due to water
thrtist. Fig. 2 shows the general design of
the imit. It is unusual to cqui]> a unit of
this design with a separate fl\-whet>l; but as a
total UK- of Hli.OOO was required, and the
A UNIQUE DESIGN OF WATERWHEEL-DRIVEN ALTERNATOR
167
generator WR- was
but 84,000, the fly-
wheel was necessary
to make up the addi-
tional Il'i?^ of 62,000.
Construction Features
of the.Stator
The construction of
the stater core is not
novel, the usual ducts
are provided for the
passage of air through
the core to the stator
frame. The inside of
the stator frame at the
top is free of ribs to
facihtate the passage
of air through the
frame to the exit.
The stator coils are
of the usual barrel
tvpe, connected one
circuit Y for 14,000
volts, and are well
supported from the
stator frame to pre-
vent vibration or dis-
tortion due to short
circuits. The stator
frame is split into two
parts; and as there is
no provision made for
repairing the stator
or rotor coils by slid-
ing the stator along
the base, the top half
of the stator must be
unbolted and some of
the coils removed in
case repairs to the
stator winding are
necessary.
Construction Features of
the Rotor
Due to the small
diameter of the rotor,
and to the high peri-
pheral speed of the
poles, the usual def-
inite pole construc-
tion could not be used.
Instead of a laminat-
ed pole keyed directly
to a spider, the loose
or removable tip con-
struction was adopted
168 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 2
<
■a
in order that the rotor
coils could be wound on a
form and assembled sepa-
rately or disassembled
readily when making re-
pairs. Fig. 1 shows the
general design of the rotor.
The rotor body and poles
consist of a series of steel
plates machined to shape;
each plate is slotted across
the pole face at right angles
to the shaft axis to receive
the pole tip, which is a
separate steel bar machined
to the shape of a pole tip
as shown at A in Fig. 1 .
After machining and drill-
ing the individual plates,
they are bolted together in
two sections and these
sections are again bolted
together with through-
bolts, the whole forming
the revolving field without
coils. The rotor complete
is then shrunk onto the
shaft; the shrink fit is re-
quired so that therewill be
a tight fit between rotor
and shaft at the runaway
speed of the rotor.
Rotor Coils
The rotor coil is of the
usual ribbon type, wound
edgewise on a form, fur-
nished with top and bottom
insulating collars and a
metal retaining collar at
the bottom of the coil.
This retaining collar is slot-
ted to receive keys for
tightening the coil on the
pole as shown at B in
Fig. 1.
Rotor Coil Assembly
Each coil is well insu-
lated and mounted directly
on the pole body. The pole
tips arc inserted in the slots
in the pole body, as shown
at C" in Fig. 1. Long bolts
are then driven through
holes provided in the pole
body and polo tip to defi-
nitely lock the pole tips
into jjosition.
I
A UNIQUE DESIGN OF WATERWHEEL-DRIVEN ALTERNATOR
J 09
170 Februarv, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 2
The distance piece, as shown at D in Fig.
1, is placed on top of the coil end. This
distance piece is held and locked by a retain-
ing plate shown at E, Fig. 1, this plate being
rabbeted to the distance piece at F and to
the lower coil support at G previously built
up as part of the pole body. After the dis-
tance piece and retaining plates are assembled,
bolts are inserted through the distance piece
and plate at the top of the pole as shown at
H and screwed directly into the pole body,
and bolts are also inserted through the
retaining plate at the bottom of the coil and
screwed into the lower coil support as shown
at J. All bolts are then carefully locked to
pre\-ent backing off.
exceed 16,000 lb., hence the unit is split into
a larger number of sections than is usual.
Ventilation
A glance at Fig. 4, representing the location
and design of the power house, will show that
the ventilation problem of the generator was
a peculiar one. As all the air for the gen-
erators had to be taken into the station and
discharged from it, the incoming and out-
going air ducts were located on the "down
stream" side of the power house. The sheer
rock of the mountain forms the other three
walls of the power house; in other words, the
power house is gouged out of the solid
mountain rock.
Fig. 3 shows the general scheme of ventila-
tion. Air is carried through two ducts into
a pit directly tmdcr each generator. Each
duct is capable of delivering approximately
15,000 cu. ft. of air per minute to the gen-
erator at a velocity of 1000 feet per
minute. Fans are mounted on the re-
taining plates of the rotor spider and
these, added to the fan effect of
the poles, carr>- the air through
the generator and discharge it
Fig. 4. Sectional Elevation Showing the Unique Location of the Power House and Its Hydraulic Connections
Supporting brackets, well insulated from
the coils, are then placed between adjacent
coils and bolted directly to the pole body to
prevent distortion of the coils due to the side
strain produced by the large centrifugal
forces in the rotor coils.
Balancing
Drilled and tapped holes are provided in the
pole tip for the insertion of weights to balance
the rotor. Each rotor is given a static and
running balance after assembly, and a high-
speed or runaway-speed test is then made
after which all parts are carefully inspected.
Shipment
Due to the comparative inaccessibility of
the power house and the mountainous nature
of the country, the unit had to be so designed
that no one part boxed for -shipment would
into the stator frame, thence to an out-
going central duct directly under the stator
frame and expel it outside the station. The
generator is entirely enclosed above the floor
line to prevent the escape of the ventilating
air into the station. Ventilating hoods,
attached to the stator frame, guide the air
into the rotor.
Summary
The following figures give the relative
weights and size of one of these units exclusive
of the waterwhecls. The length given includes
the watenvhcel shaft extension and exciter.
Stator weight complete (two halves) .... 70,000 lb.
Rotor, exclusive of shaft (two halves). . . 31,000 lb.
Flywheel 13,500 lb.
Total weight 150,500 lb.
Length overall 24 ft. 7 >2 in.
171
Belted Alternating-current Generators
By A. L. Hadley
Engineering Department, Fort Wayne Works, General Electric Company
This article describes the principal features in the construction of a line of small alternating-current gen-
erators ranging from 37 J^ kv-a. to 300 kv-a. These generators are used principally on lighting circuits, but
also find a wide field of usefulness in supplying power to induction motors, heating devices, welding machines,
etc. The machines are built according to the best practice in alternator design and require the minimum
amount of attention in service. — Editor.
A. L, Hadley
'T'HE advent of the
-*- high-efficiency
tungsten and nitro-
gen-filled lamps has
created an unu.siial
demand for lighting
generators. A 30-kw.
generator using tung-
sten lamps will supply
as much light as a
90-kw.generatorusing
carbon lamps. Also,
on the same basis,
a 240-kw. rrachine
equals a 720-kw. m^a-
chine in lighting capacity. The substitution
of nitrogen lam.ps for carbon lamps makes a
240-kw. generator equal to a 1200-kw.
machine.
To meet this demand, a standard line of
small belted generators is being built, in sizes
ranging from 373^ to 300 kv-a., 0.8 p-f., .')()
dcg. rise (maximum rating) or rated in kilo-
watt sizes from 30 to 240 kw. These machines
are built standard both for 60 and 50 cycles
in the following voltages: 240/480 (same
windings used by reconnecting), 1150/2300
(same windings used by reconnecting) and
000 volts. They are equipped with either
three-phase or two-phase windings. Special
machines of higher voltage are frequenth'
built, up to (JOOO volts. For 6600 volts,
special stator punches and dies have been
developed to provide for the extra insulation
required. The kilowatt ratings, however, are
reduced 20 per cent.
These machines although used in a large
measure for lighting purposes are also used
for supplying power to induction or syn-
chronous motors, heating devices, welding
machines, etc. For lighting purposes the
generators may be belted to steam engines,
gas engines, counter-shafts, or used as parts
of motor-generator sets. Also they may be
belted or direct connected to waterwheels of
the vertical as well as the horizontal type.
Fig. 1 shows a vertical-shaft alternator for
direct coupling to a waterwheel; this unit
has a direct-connected exciter mounted above
the thrust and upper guide bearings.
These generators are frequently used for
butt welding and spot welding, in which case
the load is approximately 0.7 power-factor.
Standard 125-volt fields are used but the
excitation is taken from a 250-volt circuit
with a large resistance in series. A special
/ara^
li
"^*«^
Fig 1 Vertical Alternating-current Generator with Exciter
Arranged for Direct Coupling to a Waterwheel
control is provided for short circuiting part
of this resistance at the tim.c the weld is made,
thus rapidly boosting the field current and
maintaining the generator voltage.
Fig. 2 shows a standard belted generator
with belted exciter, and Fig. 3 a generator
172 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 2
^^.
c td
< s-
i
n
I
< «
BELTED ALTERNATINC-CURRENT C'tENERATORS
173
with direct-connected exciter. Fi^. 4 shows
a three-bearing ■240-kw. generator with sHde
rails.
These machines have been used largeh- in
motor-generator sets, transforming from direct
current to alternating current; also from
alternating current to alternating current for
special applications where special control of
the voltage was required, Figs, o and 1.3.
These machines are also used in motor-
generator sets, to transform from alternating-
current to direct-current, the alternating-
current machine being equipped with an
amortisseur winding and being operated as a
synchronous motor.
Stator
The stator frame consists of two castings
made from the same ]mttern, s])lit at right
uniform construction. The stator frame cast-
ings are made with openings so that air can
freely circulate through and around the core,
l^roviding exceptional means for ventilation.
End-Shield Bearing Brackets
The end shields are heavy castings each
provided with three arms for supporting the
bearing. They also have heavy flanges which
extend over the ends of the stator coils, to
serve as a protection against injury from the
outside. These end flanges are bolted into
recesses machined into the stator frame, and
form a very rigid construction. The top half
of the bearing housing consists of a cap which
can be removed readily, thus facilitating
inspection or removal of the split bearing
without dismantling the remainder of the
machine, Fig. 7. The bearing cap on the
Fig. 6. Rotor Shaft, Spider, Pole, and Staler Laminations
angles to the shaft, fitted together by an
accurately machined rabbet joint and fastened
together with heavy bolts extending through
openings in the outer edge of the laminations
outside the magnetic circuit; the two halves
of the stator frame serve also as clamping
rings for the stator laminations. This is
illustrated by the laminations shown in Fig. 6.
The stator coils are machine wound, using
rectangular wire, and are insulated and treated
with tape and varnish to make them thor-
oughly moisture-proof. The stator core is
made of thin sheets of electric steel punchings
built up with generous air ducts for ventila-
tion ; the air passes freely through and around
the ends of the core between the core and the
stator frame castings. The air ducts are
formed by I-beam separators spot welded to
the laminations; these I-beams extend one
against each tooth, thus making a very rigid,
collector end has two cast-on lugs for support-
ing the brush studs that carry the brushes for
the collector. Each bearing cap has two
hinged oil hole covers which provide a ready
means for inspecting the oil rings. The hous-
ing is made with a generous reservoir for
oil, the lower portion of the housing being
tapped with two holes, one on each side, so
that the sight feed overflow oil gauge may be
installed on either side. A pipe plug, in the
side opposite the oil gauge, can be readily
removed for draining the oil.
Bearings
The bearings are split horizontally and arc
liberally designed for a large diameter shaft.
They are of cast-iron, 1:)abbitt lined, and have
two oil rings. Fig. 7. The collector-end and
pulley-end bearings have the same diameter
for each size of machine, the ratio of length
174 Fel)ruarv. IIIL'O
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 2
Fig
7. End Shield showing the Ease with which
the Bearing Can be Dismantled
Fig. 8 Rotor with Amortisseur Winding for
Synchronous Motor Operation
to diameter of the eollector-end bearing beinj;
two to one, and that of the pulley-end bearing
being three to one. Space is thus provided
for the collector without detracting from the
symmetrical appearance of the machine.
Rotor
The rotor spider is built up of j^-in. thick
laminations riveted together. These lamina-
tions have punched dove-tail openings for
receiving the laminated pole pieces. Each
pole piece is held in i)lace by two small taper
wedges driven in from opposite sides, to
make a good mechanical and magnetic joint.
Pole Pieces
The pole pieces are made of i^e-'"- thick
laminations held together by heavy rivets
through cast steel flanged end plates, the
flanges being so made and located as to serv^e
also as a support for the outer \'enccr board
insulating washers. Each pole piece is care-
fuUv insulated, and together with the two
\-eneer board washers, one at each end, serv'cs
as a form o\"er which the field wire is wound
and anchored. All rotor coils of each machine
are alike, no rights and lefts, the leads being
brought out between the coils from the end
next the spider. This construction has several
distinct advantages: (1) the wire is wound
light to the pole piece, making a better
mechanical construction than if the coils
were wound separately and afterwards placed
t)n the pole, {'2) the coils being wound tight
to the poles, the heat is largely conducted
through the i^olc piece, thus reducing the
coil temperature, (3) the increased space
between coils permits of better ventilation,
4) the losses are less due to the shorter mean
length of turn, (5) all the coils being alike,
fewer spare coils and pole pieces need to be
carried in stock by the operator, and when
ordering new coils there is no question as to
what is wanted, and ((i) the making of all
the coil connections at the inner end of the
])oles next to spider, where the speed is less.
Fig 9 Alternating-current Generator Rotor with Exciter Armature
BELTED ALTERNATIN(',-CT:RR1-:XT GEXERAT( )RS
I ( .J
insures against vibration ami break-
age of leads, and throwing out into
the air gap. After the coils are
wound on the pole pieces thc>' arc
baked to ex]3cl moisture, and then
filled with varnish and baked lo
make them moisture-proof. All
pole pieces are built to permit of
the use of an amortisseur wind-
ing for the synchronous motors of
motor-generator sets, to facilitate
jjarallel operation with other gen-
erators, for generators driven by
gas engines, or for single-jihase
machines.
Collector Rings
The collector rings consist of two
heavy cast-iron rings insulated
from and mounted ui)on a cast-
Fig 10. A Direct-connected Exciter and Its Alternator
Fig U
Alternator and Direct-connected Counter-
electromotive-force Generator
iron spider pressed on the shaft, as shown
in Figs. S and 0. The rotor in Fig. S is
equipped with an amortisseur winding, that
in Fig. 9 has no such winding but has an
exciter mounted on the same shaft. Both
illustrations show the cast ventilating fans
which are part of. the end plates used for
carrying balance weights and for covering the
dove-tail joints between the poles and spider,
to give the rotor a finished appearance.
Brush Holders
The brush-holders and brushes are liberal
in size, there being two brushes for each ring,
mounted in one holder in tandem. The use
of two brushes per ring permits the removal
of either brush without interfering with the
operation of the m.achine. Furthermore,
there is less likelihood of sparking in case
Fig. 12. The Machine shown in Fig. 10 as Dismantled for Mule Back Transportation
ITfi February, 1020
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 2
dirt or some foreign particle becomes lodged on
the collector ring and lifts one brush at a time.
Shaft
The shaft is made of forged steel, is finished
to size by grinding, and is equipped with
heavy shoulders for taking the end thrust.
A key is sunk in the shaft for driving the
rotor spider which is pressed on by hydraulic
pressure, Fig. 6.
Exciters
The exciters may be belt-driven or direct-
driven. A belt-driven unit is shown in Fig.
2 and a direct-driven unit in Fig. 10. The
direct-connected exciter is the more com-
monly used. Its exciter frame is mounted on
the collector-end bearing housing, being held
by means of a heavy cast-iron ventilated
bracket which is fitted into a groove of and
bolted to the bearing housing. The same
shaft extension on the collector end may be
used either for carrying the driving pullej^ for
a belted exciter, or for carrj'ing the armature
of a direct-connected exciter. The direct-
connected exciter armature is overhung, and
does not have any outboard bearing. This
construction is ver\' simple, and for these
belted alternators of comparatively high
speed the size exciter is such that it costs ver>'
little if am- more than the ordinary- belted
exciter which has to be equipped with sliding
base, pulley, exciter belt, and exciter drive
pulley. This combination of alternator with
direct-connected exciter is usually used, one
of the principal advantages being that it
occupies less floor space.
Fig. 12 shows the various parts of the
machine dismantled for mule back trans-
portation, where the weight must be kept
down to a certain minimum.
Fig. 11 shows a generator for a special
application where the voltage control is
effected by means of a small counter-electro-
motive-force generator mounted on the col-
lector-end bearing.
I
Fir. 13. AUcrnatinR-currcnt to Direct current Motor-Bcncrator Set
it:
Sine Wave Testing Sets
By E. J. BlRNHA.M
Alterxating-cvrre.nt Encinekking Department, General Electric Company
There are many uses in the laboratory and in schools and colleges for a sine wave generator, and in
response to this demand the sets described in this article have been developed. Oscillograms and test data
show that the generator maintains practically a true sine wave of voltage from no load to full load unity
power factor and full load zero power factor. The generator is specially useful for testing the magnetic
properties of iron and for meter testing. — Editor.
B'
E. J Barnham
!ECAUSE the sine
wave is the ideal
or standard fonn of
voltage wave, a gen-
erator that will pro-
duce this wave under
different load condi-
tions is desirable in
many kinds of elec-
trical testing.
In iron testing, for
example, it is very
essential that the
variation of the flux
in the iron take the
form of a sine wave and this can be obtained
only bj- the use of a sine wave voltage.
A-Ieter calibration and testing are other ap-
plications which require a sine wave voltage.
Schools and universities also are often in
need of sine wave generators for use in their
testing and laboratory work.
Development and Description of Generator
After careful study and in\'estigation a
special generator has been developed to meet
these needs and any others that require an
expectionally good wave form. This gener-
ator is made small and compact, is three-
phase, four-pole, and of the revolving field
type. At ISOO r.p.m. or 60 cycles it has a
capacity of 5 kv-a. at 220 or 1 10 volts; and at
750 r.p.m. or 25 cycles it has a capacity of
2.1 kv-a. at 110. volts. The good wave form
is in part due to the use of a large number of
stator and rotor slots and to the use of a
Fig 2.
Full-load Unity Power-factor Voltage Oscillogram of
the sine Wave Testing Generator
cylindrical rotor on which the exciting coils
are displaced in phase. In addition, the rotor
is enclosed in a magnetic sheath which
maintains a sine-wave distribution of flux
under the most severe conditions.
n^ig 1
No-load Voltage Oscillogram of the Sine Wave Testing
Generator
Fig. 3. Full-load Zero Power-factor Heading"* Voltage Oscillo-
gram lower curve, of the Sine Wave Testing Generator
178 February, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 2
0 15 50 45 60 75 90 105 120 155 150 165 180 0 15 50 45 60 75 90 105 120 155 150 165 180
Degrees Degrees
15 50 45 60 75 90 105 120 155 150 165 IM
Degrees
Fig. 4. No-load and Full-load Normal Voltage Waves of the Sine Wave Machine Compared with an Equivalent Sine Wave
The full line is the Equivalent Sine Wave; the dots are points plotted from the Machine Voltage Wave
For a tabulur comparison see Table II
If a phase displacement between voltage
and current is desired, as in meter testing.
one of the generators can be furnished with a
special device for shifting the stator.
The generator is capable of operating under
a wide range of load conditions. Many
different voltages may be obtained by con-
necting the leads of the stator windings in Y
or delta, and for either one or two circuits
per phase. A total of 13 leads as shown in
Fig. 5 are brought out from the machine, one
of which is for a special connection to be used
in iron testing. The frequency may be varied
between 60 and 25 cycles by varying the speed
of the generator between 1800 and 750 r.p.m.
Mechanical Arrangement of Sets
In order to meet the different uses and
driving requirements, arrangements have
been made so that the generator may be used
alone as a single unit or with a driving motor
in either a two-unit or a three-unit set.
Single-unit Set
The single-unit set is mounted on a sliding
base and has a driving pulley.
Two-unit Set
Fig. 5 shows a standard two-unit set in which
the generator is direct connected to an8-h.p..
INOO 750-r.p.m. direct-current driving motor.
If alternating current only is available, an
induction motor may be substituted for the
direct-current driving motor. In such a case
speed variation cannot be obtained.
As it is often convenient to use the machine
as a single-phase generator. Table I is given to
show its capacity single-phase and three-phase
for both 60 and 25 cycles.
TABLE I
SINGLE-PHASE AND THREE PHASE. 60 AND
25-CYCLE CAPACITY OF THE SINE-
WAVE TESTING GENERATOR
Phases
Cycles
Kv-a.
3
60
5.0
;j
2.i
2.1
1
liO
3.5
1 1
25
1.47
A comparison of the no-load and full-load
voltage waves of the machine with an equiv-
alent sine wave is given in Fig. 4 and Table
II. The deviation from a sine wave, as given
TABLE II
NO-LOAD AND FULL-LOAD NORMAL VOLTAGE VALUES OF THE SINE WAVE TESTING
GENERATOR COMPARED TO THOSE OF AN EQUIVALENT SINE WAVE
FULL-LOAD UNITY POWER-FACTOR
I I LL-LOAD ZERO POWER-FACTOR
Degrees
Ordinates
Per Cent
Deviation
from
Sine Wave
Ordinates
Per Cent
Deviation
from
Sine Wave
Ord
Voltage
Wave
0
nates
Equivalent
Sine Wave
Per Cent
Voltage
Wave
Equivalent
Sine Wave
0
Voltage
Wave
Equivalent
Sine Wave
Deviation
from
Sine Wave
0
0
0.0 !
0
0
0.0
0
0.0
15
2.. 50
2.58
0.8
2.67
2..'i8
0.9
2.50
2.58
0.8
30
5.02
5.00
0.2
5.03
5.00
0.3 -
5.00
5.00
0.0
45
7.12
7.07
0.5
7.08
7.07
0.1
7.18
7.07
1.1
60
8.70
8.00
0.4
8.02
8.66
0.4
8.72
8.66
0.6
75
y.GC
(1.00
0.0
11.05
9.66
0.1
9.75
9.66
0.9
90
9.85
10.00
1.5
10.05
10.00
0.5
9.90
10,00
1.0
I
SINK WAVE GEXERATORS
179
in Table II, is found according to the A.I.E.E.
Standardization Rules; i.e., by placing the
equivalent sine wave on the actual wave so to
give the least difference between ordinates,
and then determining the deviation h\-
dividing the maximum difference between
Two-unit Sine Wave Testing Set; Sine Wave Generator at left, and
Direct-current Driving Motor at right
corresponding ordinates by the maximum
value of the sine wave.
It will be noticed in Table II that the
maximum deviation of the voltage wave
from the equivalent sine wave is only 1.5 per
cent. In fact, the deviation is so small that
when points from the voltage wave
are plotted, as in Fig. 4, they ap-
parently fall on the equivalent sine
wave.
Three-unit Set
The standard three-unit set is
shown in Figs. 6 and 7, in which
two of the sine-wave generators
are direct connected to a IG-h.p.,
I.SOO 75()-r.]j.m. direct-current
driving motor. The inner bearing
bracket has been omitted on one of
the generators so that the machines
may be assembled closer together,
thereby minimizing floor space. The
Fig. 6. Front View of the Three-unit Sine Wave Testing Set. Driven by the Direct-current Motor on the right.
The machine in the center is a High-voltage Low-current Stationary-frame Sine Wave Generator.
The machine at the left is similar to the one in the center except that it is wound for
Low-voltage High-current, and its Stator is adjustable
Fig.
7. Rear View of the Three-unit Sine Wave Testing Set Shown in Fig 6 The hanj wlieel on the right-
hand Generator is used to adjust the Stator in its cradle
180 February, 1»2()
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 2
base is strong and self-supporting, so the
assembled set can be easily lifted from one
place to another. Fig. 7 shows the phase-
shifting device, which consists of a hand
wheel, worm, and worm segment for shifting
the stator frame of one of the machines.
Fig. 8. Apparatus for Testing the Magnetic Properties of Sheet Iron
The worm segment is of sufficient length
for rotating the frame 50 mechanical or
100 electrical degrees in either direction
from the neutral position. The armature
windings of the two generators are different
because one of the generators is used for
current and the other for voltage, hence one
is designed for low voltage and high current
while the other is designed for low current
and high voltage. Fig. 6 shows the adjustable
frame 20-volt generator used as the high
current low-voltage machine. The four leads
from the stator windings, one of which is the
netural, are brought out near the top of the
machine as shown so they will not interfere
with the rotation of the frame.
Iron Testing
The sine-wave generator either as a single-
unit or as part of a two unit set is extensively
used for testing iron... For testing sheet iron
conveniently and accurately the use of the
(General Electric) Epstein Tester in con-
nection with the sine^wave generator is
recommended. Fig. $ .shows the essential
part of the Epstein teSter with sample sheet
iron held in ])lacc in;'^he coils by clamps.
Fig. 9 shows the diagram of connections to be
used in making the tests. Fig. 10 shows that
the form factor of a sine wave, namely, 1.11,
is ]3raclically maintained under the severe
distorting effect of the magnetizing current
in the testing of iron.
Meter Testing
The three-unit set described is \-ery useful
for calibrating and testing wattmeters and
also for many other kinds of work where it is
necessarv' to make tests at diflcrcnt power-
factors. In wattmeter testing, the 20-volt
To Garterqtor-
eratA
[U]5-
/«/-
c:
mvft—ee*'-
\
■•—Tb Motor Fitta
(\
> Seneroto*- /•'«/*
Fig 9. Diagram of Connections for Testing the Magnetir
Properties of Sheet Iron by the Use of the Sine Wave
Generator and the Apparatus Shown in Fig. 8
or high-current generator is used to excite
the current coil of the wattmeter and the
220/110-volt or high-voltage generator is used
to excite the voltage coil of the wattmeter.
The desired power-factor setting is then ob-
tained by mechanically shifting the stator of
the adiustahle frame machine.
Fig 10 VoltnKC ;tn.l Hxciling Current Whvc of the Sine W*»ve
Testing Gencratot While Testing Iron Saniplc by
Method Shown in Fig 9 Form Factor IMS
ISl
IN MEMORIAM
T S EDEN
Timothy Sharpe Eden, Enj^inccr in
Charge of the Generator and Syn-
chronous Motor Division of the Alter-
nating Current Department of the
General Electric Company, died on
Wednesday morning, October 1, 191!),
following a brief illness. To those
who knew Mr. Eden best, his passing
came with the force of a personal
grief.
Mr. Eden was born on the island
of Jamaica, W. I., where he received
his early education in a private high
school. On coming to the United
States in 1892, he entered Lehigh
University. After graduating from
Lehigh in LS90, he was employed
by the Bethlehem Steel Compan}^ at
Bethlehem, Pa., as an assistant met-
allurgical engineer. In September,
1.S97, he entered the Drafting Depart-
ment of the General Electric Com-
pany, and in January, 1900, was transferred to the Alternating Cur-
rent Engineering Department, where he remained until his death.
In the death of Mr. Eden, the General Electric Company has
lost a faithful and able engineer. He possessed, to an unusual degree,
the qualifications of the ideal designing engineer — a good sense of
proportion, a well trained mind of mathematical bent, great care in
the consideration of alternatives, patience in working out details, a
good memory, sound and sober judgment. He seemed to delight to
be in close touch with his machines as they came through the shops,
the test, and the installation, and after they were put into ser\-ice by
customers. He was respected by his fellow workers who accepted
his opinion in engineering matters with great confidence.
Mv. Eden was indeed much beloved b}' his friends and associates
for his amiable qualities of mind and heart, never sparing himself
when he could be of service to others. He won his way to their
hearts where he will long be held in esteem.
Besides his wife, Air. Eden is survived by his mother, a step-
father and three half sisters, Alice, Elsie and Winnie, all of Jamaica;
one brother, Alfred, of East Orange, N. J., two half brothers. Dr.
Arthur Henderson, of Montreal, and Brooke Henderson, of Xew
York Citv.
XVI! I
GtNEKAL ELECTRIC KE\IE\V
WHERE TO GET G-E SERVICE
'io insure correspondence against avoidable delay, all communications should be addr«ss^ to
the G-E sales office. G-E distributing jobber, or G-E foreign representative nearest the writer.
FOR BUSINESS IN
G-E Sales Oficc G-E OiuriboiinC Jobber
AUbima. Rirtnlncham M^thcwj EIcc i^uppl; Co.
Ark&osu. Little Rock
California. Lo« Anjelest - Paclflc SutM Elwtrif Co
California. Oaktand: Pacine Slates Elertric Co.
Calirorr^la. San rnincl»rott Parlllc Slates Etectrlr Co.
"-' — -•- " * , TlieMendrleJi BolUKtOMfS. A
Sup. Co.
Colorado. Denvert .
Conncrtlrut. Hartford
Connertlrul. New Haven ,. ...
ConniN-ticut, Wat«»buryt N'*w F.nctaod En* Co.
DUt-<rt of Columbia. Waahlos-
Ion National E1«:'I f^upply C*.
Florida, Ja^-kaoovUle >~lorida Eke. Supply Co.
Florida. Tampal . . .: Florida Elec Supply Co.
Ceorcia, Atlantajt Carter Etectrle Company.
neorcia. Savanoabt Carter Elrrtrlc Company.
IlUnoia. ChlcagoS I Central Electric Company.
Commonwealtb Edtson Co.
Indiana. Fort Wayne
Indiana. Indianapolis Indianapolis Flee Supply Co.
Indiana. South Bendt Soutli Itend ElM-trlc Co.
Iowa. r>*a Molno Mld-Wcat Electric Co
Kentucky. LoubvUle Belknap Hardware A Maau(M»
turlnj Co .Inc.
Louisiana. New Orleans! Gulf :'<ates Electric Co.. Inc.
Maryland. Balllmore Soint>*m Elertrtr Co.
Manacbuaetts. Doflonf Pett 'nccU-Andrewv Co.
Massachusetts, (^prlnRneU
Ma^uchuseits. Worw»t«f .... , . . .
Mlchlcan, DetroUI Frank C. Tnl Company
Mlchleao. Grand Rapids
Minnesota. Duluth Nonhwntem Elecwlc Equl^
ment Company.
Mlnnr^oia, MlnneapoUat Pwrhai Etectrfral Co
Mlnnrwta. SI. Paul Nortbweatem Elec. Equip. Co.
Mt-tsourl. Joplin •
Mbsoiirl. KamiL^ (*1iyt Tbc B-R Elertrlr Co
.Mhs-inri. St Loulat Wesco Supply Company.
.Motitoim. lluttc' Butte Electric .*!Mpplv Co
NcbrasWa. Omaha MM-WV-it Flwtrtc Co.
New Jf-mcy, Newark: T>-l-Olv ricctrlc Co.. Inc.
INoC^EOmrc
THE UNITED STATES
G-E Salea Oficc
New York. Albany:
New York. Buffalo
New York . Elmlra . . - - .
New York CHylf E B Latham A Company.
Royal EaMcm Elec'l f«ap. Co.
SIbley-Pitmaa Elec. C«rp-
G-E Distnbatial Jobber
Hairns Elrrtrtr ^ o . Itie.
RobrrtsoD-CaLaraci Elcr. Co.
. Wbeeler-Creen Elee'l S«p. C«.
'.Mohawk Elec'l Sup. Co.",
New York, Roch<
Mew York.Scbeocctady
New York. Syracuse..... . .
North Carolina. Charlotte
Ohio. ClDclanailt The F I» I^wrrnre Elee. Co.
Ohio. Cleveland Republic Flrctrtc Co
Ohio. Columbus The l-mcr A Hopkins Co.
Ohio. Da>-ton The Wra, MaH Ekrctric Co.
Ohio. Toledo W.G. Na«el Ekctfle Co.
Ohio. YoutUBitown . .
Oklahoma. OklabomaCltTt <*outhwe«i G-E Co.
Orccon. Portland t .PadAc States Electric Co.
Peoasyh-ania. Erie
ProaiylvaoU. P&lladelphU(l. . PhlladelpbU Ehietrte CempaBT
Supply Demnmeot
Prnttiyh'anta. Plttsburcbl Vnioo Electrtc Cocnpaay
Rhode Island . PrwrMeo'-e ....
Souih Carolina. ColumbUl, , .Prrrj-Mann Elee. Co . lac.
Tenne*iee. Challanooo Jamea Supoty Company
Tenn«»ce. Knojville .....
TennesMe, Memphis Elcctnc Supply Company
Tennessee. NatfivlDe .
Texas. Dallas ' 9outh*<« C-E Co.
Te&as. EI Paso* iUmthwesI C^E Co.
Te*as Hou«t..nt .... . AoatbwesK G-E Co.
t IA^ <.ili l..>f otyl Capital Eteetrte Caapaajr.
■'arise ttatat Elactfto C*.
I >.!.■ llawaiLa-i biutncn addrewt Cattoa. NslU * Coapaay, Ltd..
Honolulu
•Warehooar. (t'ertli-e Shop.
Diilhbulor* lor the General Fleetnc Company OuliiJc ofihc Uoiicd Stales
INTERNATIONAL GENERAL ELECTRIC COMPANY. INC
Geacral Sales Offices: 120 Brosd^av. New York, and Scheocctadr. N. Y.. U. S. A.
At<ois and Reprcacniaiivcs
Arcentlns: Cla General Electric Sudamcricaua. Inc
Au«iralln: Auiirallan General FlcrtHc Compam
Burt
J line J and
Melbourne
Belelum- ITHlnesCareliFrtfes. Druswll.
noUvla: Intemailonal Machinery Co.. La Pa« and Oruro.
Draill: CL\ General Flectric do llraill. Rio de J.inetro and
Sao Paulo -
Canada: Canadian Genera) Electric Company. Ltd. . Toronto. Onl,
Cblle; International Marblnery Company, Saoilaco ValparatM
Aiitolacasta.
China: Andcfjen. Meyer A Co. ltd.; IntcmaUonal Ceocnl
Electric Company. Inc , STtansha'.
Colombia: Wemclboeft A Poor BarranqullU and Bogota.
Cuba: General Electric Company of Cuba, Havana.
Dutch r Indies: Interaatlonal Geacral Electric Company. Inc..
Soerabala, Java.
General Oltico
Sch<nectad\':N.Y
'1 Snrirta Anonltna. Mtlaa.
r I: ! 3IJ-S I naltM«Tlt\( u ark». Tokyo; Tokyo EVrtrte Co.,
. hawa.*akl.
o: Mcilran General Elecirtc Company. Mealco Clir aad
■ilalaiars
'.'■ ii,.r.il T^»■ Nn'i-r,.ii 1 x-r-tri.-ii »-..! Fndneertnc Company
and A»fklarMl
-.••By. Manila
-irw CooipMir. lAd..
El© c tone
17°?)"^ y Sales Ollices in
, 11 il y all largo cities
I
Sav \ou saw it advertised in the General Electric ReView
TWO DOLLARS PER YEAR
TWENTY CENTS PER COPY
GENERAL ELECTRIC
REVIEW
VOL. XXIII, No. 3
Published by
General Electric Company's Publication Bureau.
Schenectady. N. Y.
MARCH, 1920
ELECTRIC STARTING MOTOR AS APPLIED TO A 12-CYLINDER LIBERTY AIRPLANE ENGINE
(.See article. "Electric Starting Systems for Automobiles," page 186)
For
Fractional H. P. Motors
^\T0 one objects to paying more for a
■^ ^ machine that is worth more. And
ability to stand up to its work longer, is
a factor in a machine's value which goes
far toward making its first cost to be for-
gotten. The time-tested service qualities
of 'tiORm^" Bearings under high-speed
conditions will make any machine a bet-
ter machine which is proved by the
service qualities of hundreds of thousands
of electrical machines in which HORma"
Bearings are standard.
See that your Motors
are "MORmfl" Equipped
Ball, Roller. Thrust and Combination Beai-inqs
General Electric Review
A MONTHLY MAGAZINE FOR ENGINEERS
Associate Editors, B. M. EOFF and E. C. SANDERS
Manager. M. P. RICE Editor, JOHN R. HEWETT j^ ^^^^^^ „f Advertising. B. M. EOFF
Subscription Rates: United States and Mexico. $2.00 per year; Canada. $2.25 per year; Foreign. $2.50 per year; payable in
advance. Library and Student Rates: United States and Mexico. $1.50 per year; Canada. $1.75 per year; Foreign. $2.00 per year;
payable in advance.
Remit by post-office or express money orders, bank checks, or drafts, made payable to the General Electric Review, Sche-
nectady. N. Y.
Entered as second-class matter. March 26. 1912. at the post office at Schenectady. N. Y.. under the Act of March, 1879.
Vol. XXIII, No. :j ,, co,^^a^?;;,^ cl/,.„. March. 1920
CONTENTS Page
Frontispiece 184
Editorial: Electricity on the Automobile 185
Electric Starting Systems for Automobiles 186
By F. C. Barton
The Outdoor Generating Station . . 194
By H. W. Buck
Methods for Alore Efficiently Utilizing Our Fuel Resources
Part XXXI. Petroleum 198
By C. G. Gilbert and J. E. Pogue
Professor Elihu Thomson's Early Experimental Discovery of the Maxwell Electro-Mag-
netic Waves 208
By Prof. Monroe B. Snyder
Effect of Color of Walls and Ceilings on Resultant Illumination 209
By A. L. Powell
Short-circuit Tests on a lO.OOO-kv-a. Turbine Alternator 214
By E. S. Henningsen
The Engineer Can Do More About It Than Pay and Grin ' 222
By Calvert Townley
Helium, the Substitute for Hydrogen in Balloons and Dirigibles 227
By W. S. Andrews
Silent Spokesmen in the Factory 229
By Roscoe Scott
A Biographical Sketch of the Late William Olney Wakefield 232
Question and Answer Section 234
GENERAL ELECTRIC
REVIEW
ELECTRICITY ON THE AUTOMOBILE
For a number of years the inability of the
internal combustion engine to be self-startins^
plainly militated against its use as the motive
power of the ultra popular self-propelled
vehicle. The woman who had the courage to
drive the stalling, balky automobile was a
rarity. The condition today is wholly changed
and we observe a goodly percentage of women
drivers, even in the densest city traffic. We
must attribute this rapid change to the de-
velopment and application of the electric
starting system.
When we review the state of development
of electric motors and storage batteries as
early as 190.T, we are caused to wonder wh>'
a satisfactory electric starting equipment did
not sooner present itself. The first starting
motors differed but little from existing series
motors for other applications; and the storage
batteries, while being subjected to unusually
severe service, resembled in all essential
respects storage batteries for electrically
propelled vehicles. It would seem that the
application of the electric motor to starting
the internal combustion engine awaited the
development of a satisfactory method of
connecting the two — an arrangement fool-
proof, simple, and sturdy. A form of shift
which automatically meshes the motor
pinion with the flywheel teeth immediately
the motor is connected to the battery, and
throws it out of mesh and holds it there as
soon as the engine begins to fire, is now almost
universally used on American built cars.
To one who has experienced the back-
breaking job of cranking a cold engine it is
incredible that a motor only 4^2 inches in
diameter can perform the work twice as fast
and for an indefinite period. All the credit,
however, does not belong to the creators of
this small motor; the battery manufacturers
have accomplished wonders in providing such
a bountiful reservoir of electric energy in so
small a space. Discharge rates on starting
and lighting batteries range from an amijere
or two for lighting to three or four hundred
amperes for starting in cold weather.
The principles of the electric starting system
for gasolene engines are described at length in
the article bv Mr. F. C. Barton in this issue.
Electricity was first used on the automobile
for igniting the gas in the cylinders, and in
this capacity it is indispensable to the opera-
tion of the gas engine. Other means of
ignition were attempted in the early stages of
gas. engine development, but they were crude
and woefully inadequate. True, the Diesel
engine dispenses with electric ignition, but
this high compression engine as at present
developed is out of the question for auto-
mobile propulsion.
With electric starting systems incorporated
as standard equipment on practically all
cars, electric lighting exists as a matter of
course. Now, instead of insufficient light
for safe driving, the injudicious use of the
brilliant miniature Mazda lamps has effected
the opposite extreme, and has provoked
restricting legislation in many states.
The electric generator-motor system of
transmission and speed reduction has met
with favor. With this arrangement the clutch
and transmission gears are replaced by a
generator -motor set, the armatures of which
are mounted on a common shaft intervening
between the engine crank shaft and the
driving shaft. Speed reduction is effected
by increasing or decreasing the magnetic pidl
between field and armature. The generator
also serves as a starting motor; and on long
down grades is made to act as an electro-
magnetic brake.
Further uses for the electric current are
found in the electro-magnetic gear shift; a
method of heating the mixture in the intake
manifold to facilitate starting in cold weather;
electrically heated hand grips for the steering
wheel; and cigar lighters. There has been
some application of electrically operated
brakes by means of a motor driving a drum
on which is wound a cable connected with the
brake bands. A motor-operated jack would
be a boon to the tourist.
A consummation devoutly to be wished is
an electrically propelled automobile having a
radius of operation comparable with that of
the gas engine driven car, and as readily
replenished. Its realization is contingent
only on the appearance of a suitable accumu-
lator for the magic "juice." — B.M.E.
186 March, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 3
Electric Starting Systems for Automobiles
By F. C. Bartox
Lighting Department, General Electric Company
After briefly describing the straight mechanical, compressed air, and acetylene methods which were used
to some extent years ago for starting internal combustion engines and detailing the reasons for abandoning
such methods, the author confines his description of starters to the various electric types. The design, char-
acteristics, and operation of the single-unit, two-unit, and combination-unit types of starting and lighting
equipment are discussed in great detail. Description and examples are also given of the methods employed
in selecting the size of units for assumed requirements. The conclusion of the article relates what has been
done by the Society of Automotive Engineers in standardizing the mountings of starting motors and lighting
generators. — Editor.
Ever since the early development of the
explosion or internal combustion engine, it
was realized that the inherent drawback to
the use of this type of prime mover lay in its
inability to be started by energi,- stored within
itself. The problem of starting the engine
with the least expenditure of human energy
has therefore occupied a large place in the
minds of designers, with the result that vari-
ous forms of starters were devised.
There were straight mechanical devices
emplo^-ing springs or their equivalent to
give the initial impulse; then, too, there were
devices in which the internal combustion
engine by the use of special distributor valves
was converted into a compressed air engine,
taking air under pressure from a storage
flask. This flask was in turn charged by some
form of pump connected to the engine and
driven by it during periods of normal opera-
tion. There were gas devices in which an
explosive charge of acetylene, or other gas,
was introduced through suitable distribution
valves directly into the cylinders and was
there exploded by the usual electric ignition.
Almost without exception these devices
lacked reliability. The springs did not store
enough energy to make second and third
attempts at starting in case the first failed.
The air starters developed leaks and pump
troubles which resulted in the slow discharge
of stored air with attendant loss of starting
ability. The gas starters were always
"touchy" and frequently the mixture intro-
duced for starting would not ignite when the
spark was applied, and, when it did ignite,
the resulting explosion was apt to be of
greater violence than is desirable from a
mechanical standpoint.
There were also electrical starters, which
took energy- from a storage battery to
drive an electric motor mechanically con-
nected to the engine, the battery being re-
charged bv an electric generator driven bv
the engine during normal operation. Other
things being equal, the type of starter using
electrical energy- acquired a tremendous
advantage over all others by reason of the
possibility of combining starting with the
most satisfactory form of lighting, viz.. that
employing Mazda electric lamps. Further-
more, it might be combined to furnish energy
for the now extensively used batten.' igni-
tion. Hence, electric systems always include
starting and lighting, and, frequently, start-
ing, lighting, and ignition.
It is not the purpose of this article to dis-
cuss cither lighting or ignition systems, but to
give a brief outline of the various ways in
which electric motors and generators are
employed in modern automobile design and
construction.
Electric starting and generating sets may be
divided into three general classes, as follows:
First : The single unit system in which the
same electrical machine acts as both motor
for starting and generator for charging the
batter>'.
Second : The two unit system in which the
motor is employed for starting only, and is not
in use for any jjurpose except during the start-
ing period. The generator is used only for
charging the battery, and is an entirely
separate unit driven independently by some
means from the engine during normal opera-
tion.
Third: A combination of the two systems
already mentioned. This system usually
includes a single field structure and an arma-
ture having two windings and two commuta-
tors, one being employed when the machine
is operating as a motor, and the other when
operating as a generator.
The single unit system requires an electri-
cal and mechanical compromise. The me-
chanical reduction ratio between the anna-
ture of the machine and the engine crank
shaft must be such that the speed of the
ELECTRIC STARTING SYSTEMS FOR AUTOMOBILES
187
armature will not be dangerous when the
engine is driven at speeds equalling maximum
car speeds. These engine speeds may be in
the neighborhood of 3000 r.p.m. or above.
It is therefore advantageous, from the gen-
erator standpoint, that the driving ratio be
as low as possible but, from the motor stand-
point, where a high torque is required at the
crank shaft, it is desirable to keep this ratio
as high as possible, as the lower the ratio the
larger must be the electrical machine to accom-
plish a given result. The electrical compromise
lies between the speed at which the machine
will crank as a motor and that at which it
will charge the battery as a generator.
The combination-unit system employs a
single field structure and a double wound
armature. In this system the armature shaft
is usually extended through both ends of
the machine, the rear end being connected
through suitable gearing to the engine fly-
wheel (upon the periphery of which gear
teeth are cut) during starting operations.
After starting, the mechanical connection to
the engine flywheel is disconnected, and the
armature of the machine is then driven by
means of the forward shaft extension from a
suitable power take-off on the engine arranged
to drive the armature as a generator at a
suitable speed.
To accomplish the change-over from motor
action to generator action, various automatic
or semi-automatic mechanical devices are
necessary. These usually consist of a manually
operated gear shifting device and switch, for
engaging the motor reducing gears with the
flywheel gear and completing the electric
circuit to the motor winding, and an over
running clutch on the generator drive which
permits the armature to rotate free from the
generator drive while it is running as a motor
cranking the engine, but which will cause the
armature to be driven by the engine when the
starting gears are disengaged and the motor
circuit broken. This arrangement permits
the motor ratio to be selected independent!}'
of the generator ratio.
The two-unit system employs a motor
and a generator, the generator being driven
through an ordinary coupling, or by chain
or gear by the engine, and the motor being
connected automatically, or by a manual
shift, to the flywheel gear ring during start-
ing operation.
The means employed for making the me-
chanical connection between the motor shaft
and the engine flywheel has been the subject
of a great deal of engineering development.
At this date, by far the greatest number of
devices make this connection and discon-
nection automatically. These automatic
"shifts" consist, in almost all cases, pri-
marily of a pinion connected by some means
to, or made part of, a nut which runs on a
screw thread mounted on, or cut in, an exten-
sion of the motor armature shaft. When the
motor circuit is closed, the armature starts
to rotate, but the pinion and nut, because
of their inertia, remain almost stationary.
This causes the lead screw on the motor
shaft to propel the pinion forward toward the
flywheel in a direction parallel to the axis of
the shaft until it encounters and engages
with the flywheel teeth. Contact between
the edges of the flywheel and pinion teeth
checks any tendency the pinion may have had
to acquire the rotative action of the armature,
thereby causing the lead screw to propel the
pinion positively to the Hmit of its travel.
It then can travel no farther axially and
must, therefore, either stop the armature or
rotate with it, and, being in mesh, if it rotates,
it must also rotate the flywheel and thereb)'
crank the engine. As soon as the engine com-
mences to run by its own power, its speed is
sufficiently great, with relation to that of the
motor, that the jjinion is driven by the engine
faster than the screw shaft is driven by the
motor. This causes the action of the lead
screw to be reversed, and the pinion is there-
fore propelled by the engine back along the
motor shaft to the out-of-mesh position. At
this point the motor circuit should be broken.
If it is not, it merely continues to accelerate
until free running speed is reached, but as the
pinion is then running at approximately the
same speed as the armature, there should be
little tendency on its part to re-enter the fly-
wheel gear.
The foregoing merely outlines the funda-
mental actions of engaging and disengaging.
A description of details, such as the method
of absorbing shock, and the prevention of
re-entry, and the obtaining a correct angle of
entrance follow.
Shifts
Generally speaking, there are two types
of automatic screw shifts in extensive use.
One transmits the torque developed by the
motor to the pinion through the medium of
a coil spring wound around the shaft. The
other delivers the motor torque to the pinion
through a self-tightening friction clutch.
The object of either the spring or the clutch
is to minimize the shock that would take place
ISS :\Iai-ch, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, Xo. 3
when the pinion reached the end of its travel
on the lead screw on the motor shaft, or the
point at which its axial motion is translated
into rotative motion. It must be remembered
that the rate of acceleration of the motor
armature is very high, and by the time it has
Motor shaft /tet/cti
tos/?/ftatt/i/s
point.
Engine
'f/i/iv/iee/
loose S/ccve
Fig. 1. Spring Drive Automatic Shift (Inboard Shift)
rotated the necessar\- one or two revolutions,
which carries the pinion into mesh, its angular
velocity is great enough to damage the gear
teeth or armature shaft if the shock at the
instant of starting to crank is not cushioned
in some way.
These devices are also designed to minimize
the liabilit}' of encountering what is known
as a "butt." This means a condition where
the fl\'wheel teeth and the pinion teeth are
not so lined up that they can slide directly into
mesh. In other words, a pinion tooth may
strike end on against a flywheel tooth, and,
without some flexibility in the drive, will not
be able to enter, and the two will then lock
tight in what is commonly known as a "jam."
To reduce further the possibility of a "jam,"
the front end of the pinion teeth are cham-
fered to produce the smallest frontal area,
and still maintain a liberal
mechanical margin of safety
against breakage. This
chamfering is very similar
to that used on transmission
gear teeth, where it is done
for the same jmrpose.
The flexibility of drive also
provides against another con-
dition known as "hunting."
This condition is particu-
larly in evidence \\'ith four
cylinder engines, and is a
result of the reaction of
gases compressed in the com-
bustion chambers of the cylinders by the
pistons on their compression strokes. The
expansion of each compressed charge, on
what would be the working stroke of the
cycle if the charge were fired, causes the
engine to tend to over-run the starting motor,
which in turn tends to run the motor pinion
out of mesh. The two factors which prevent
this from actually taking place are the
flexibiUty of the drive and the high rate of
acceleration of the motor, which enables it
to keep up with quite violent changes in
angular velocity. The tendency to "hunt"
decreases as the number of cylinders is in-
creased, until, with a twelve-cylinder engine,
the torque required h\ the engine for start-
ing is, due to overlapping of power impulses,
almost uniform throughout a revolution.
When the engine fires, causing its sudden
acceleration from cranking speed to running,
the motor pinion, as previously explained, is run
back along the lead screw to the out-of-mesh
position. This throw-out is frequently quite
violent; therefore, some form of cushion stop
or detent is provided at the out end of the
screw to prevent the possibility of a rebound
of the pinion, which might bring it into con-
tact with the flywheel again, and, due to the
relatively high speed of the latter, might cause
serious damage to the gear teeth.
Another point, which is given consideration
in "shift" design, is "angle of entrance."
Normally, the pinion is approximately Y%
of an inch away from the flywheel when out
of mesh. While travelling this '^% of an inch
along the lead screw, and being restrained
from turning only by inertia, a certain
amount of rotative movement is acquired.
Experiment has demonstrated that a definite
amount of rotati\-e movement is desirable,
and reduces the liability of "butt," and that
this amount is usually in excess of that which
would be normally acquired; therefore, some
leoc/ Screnn
Dri'v/ng C/utch
Thronout Stop
Clutch Drive Automatic Shift pinboard Mrsh
form of friction clutch, or loading device, is
provided to give "initial" friction between
pinion and lead screw to give the desired
number of degrees of rotation.
Two forms of each type of automatic
shift are used; that in which the pinion is
ELECTRIC STARTING SYSTEMS FOR AUTOMOBILES
189
propelled rearward away from the starting
motor when going to mesh, this being known
as "outboard" mesh, and that in which the
pinion in normal position is to the rear of the
flywheel gear, and is therefore propelled for-
ward toward the starting motor into mesh.
This latter is known as "inboard" mesh.
demand is high. Under these conditions the
current necessary to turn over a stiff engine
mav be three or four hundred amperes, which
means only 3.5 to 4 volts at the motor ter-
minals. This voltage is used up in two ways :
first, in overcoming brush and brush contact
drops and winding resistance, and, second, in
o
o
I IIP
i r
I i!f,l
Fig. 3. Starting Motor for Outboard Shift
Car builders who manufacture their own
engines and clutch housings usually ]:)ro^•ide
for inboard shift, as such changes as are
necessar}- are purely internal matters with
them, and can be easily provided for. But
manufacturers of assembled cars purchasing
engines and gear sets, which usually include
clutch housings, almost invariably use the
outboard form of shift.
The shift description has been carried out
to some_length, as it is a very vital part of the
whole system, and, while fundamentally
simple, has undergone much re-design and
development to bring it to the present posi-
tion of reliabilitv and sturdiness.
Fig. 4. Starting Motor Flange Mount (Inboard Shift)
Motors
Starting motors are always straight series
wound and of very low internal resistance,
both as to windings and brushes. This is nec-
essary to meet cold weather conditions when
the batter V voltage is low and the current
the production of useful work. So whatever
fraction is saved from the former is available
for the latter, thereby improving the per-
formance of the motor.
The conditions outlined in the preceding
paragraph will be found only in extremely
cold weather, but they must be met if the
starting is to be successful at all times.
Fig. 5 gives characteristic horse power,
speed, and torque curves of a 4i^-inch
diameter Bijur motor. With this curve as a
base, the most desirable ratio of pinion to fly-
wheel to give the most satisfactory cranking
can be determined. It is of course to be
desired that when conditions are adverse,
viz., when the engine is cold, the motor speed
shall be' such that it will operate as nearly as
possible at the peak of its horse power curve,
that being the point at which it will do the
most useful work.
Take, for example, a six cylinder engine of
a size suitable for the moderate sized car.
This engine will have a displacement of 303
cubic inches or cylinders 3}/2 ^y 5M inches,
and a flywheel having 126 teeth. We know
that this engine will require about 30 lb. -ft.
torque at the crank shafc to crank when hot,
and that it will need three to four times that
torque to crank at zero or below, and that
under this severe condition the cranking speed
must not fall below 50 r.p.m.
By a cut-and-try method it will be found
that a nine-tooth pinion will be suitable.
190 March, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No.
The ratio will be 120 :9 or 14: 1, or at 50 r.p^m.
crank shaft will give a motor speed of 700
r.p.m. 700 r. p.m. = 8.0 Ib-ft. torque or 120
lb. -ft. at engine crank shaft will take 400 amp.
and deliver 1.2 h. p.. which is near the peak of
the horse-power cur\'e, and, therefore, at the
2e 13 too
i4 /-2 SJO
22
20
\i It
I.
10
I.I 500
1.0 4S0
09 400
OS •p 350
0.i
OS
04
0.3
02
0.1
■ 250
200
ISO
100
50
O
_J ^2
10 12 M
Torque (Lb. Ft.)
Fig- 5. Starting Motor Characteristic Curves
most desirable point. This is satisfactory for
cold performance.
To find what will happen with a warm
engine requiring only 30 Ib-ft. at the crank
shaft or 2.12 Ib-ft. at the motor shaft, read
straight up from the 2.1 torque point. It
equals 13.3 amp., 20S0 r.p.m. or 147 r.p.m.
crank shaft and 0.S2 h.p. which is satisfactory.
Generators
The principal factor in determining the
size of generator suitable for a given car is the
ratio between the driven sjjeed of the gen-
erator and the miles per hour of the car.
This factor is usually given in terms of revolu-
tions per minute of the generator per mile per
hour of the car. This is affected by;
First ; the road wheel diameter.
Second: the rear axle ratio.
Third: the ratio of the generator drive to
the crank shaft.
This last ratio is usually determined by the
number of engine cylinders, as the generator
drive is in almost every instance made to run
at a speed suitable for magneto drive. This
would be 1 :1 for four cylinders and 1.5:1 for
six cvlinders.
For example: A four cylinder car having
33-inch wheels and a 4:1 rear axle ratio and
a 1:1 generator to crank shaft ratio would
have a generator speed of 41 r.p.m. at 1
m.p.hr. If this happened to be a six-cylinder
car and the generator to crank shaft ratio was
1.5:1, the generator speed woxild be 61.5 r.p.m.
at 1 m.p.hr.
Experience has shown that a generator to
meet average conditions should deliver 10
amp. at a car speed not much in excess of
14 m.p.hr. and should give maximiun output
at some speed between 20 and 25 m.p.hr.
The choice of a generator, therefore, is
merely a matter of selecting a standard ma-
chine which will fulfill the current output con-
ditions outlined above at the speed available
at 14 m.p.hr.
Take the six-cylinder example for illustra-
tion. A generator having an output like the
cur\-e Fig. 6 would be satisfactory'. This
machine delivers 10 amp. at almost exactly
S60 r.p.m., which equals 61.5X14, or 14
m.p.hr. car speed. It reaches a maximum of
between 16 and 17 amp., at 1400 r.p.m.,
equaling 23 m.p.hr.
After the maximum output has been
reached a further increase in speed causes
the current rate to fall off. This falling off
of the charging rate at high speeds is a most
desirable feature of a generator employing
the third brush type of regulation, as it means
that the a\-erage city dri\-er, who operates at
^T
^"^ ""=s^
n / = «.
i "^ -H^
s" ^ It ^
\,<i T5
K ^tZ
^1-4
^ t
6 -.'
1 jZ
* t a-
, J
t X
n -J 1
*x> 40C /aoe hoc ttco ttoe moc tooe ztoo t40O itcc ztmf
Generator /f.fM.
Fig. 6. Battery Charging Curve, Third Bruth Generator
low speeds but who uses the greatest amount
of current for lighting and starting, gets the
highest charging rate, whereas the tourist or
country driver, operating over longer periods
of time and at higher average speeds than the
city man, gets a lower rate of charge which
ELECTRIC STARTING SYSTEMS FOR AUTOMOBILES
191
saves his battery from heating and loss of
electrolyte due to the decomposition of the
water when gasing.
The foregoing remarks on charging rates
relate to the current regulated or third brush
type of machine. This is the type
most extensively used on moderate
and low-priced cars. One other
system of regulation is in fairly ex-
tensive use, especially among the
higher-priced cars. It is the system
employing voltage control. The
feature of this method of control is
that it supplies a high current when
the battery is low, and a low current
when it is high. It approximates
what is known as a "taper" charge
or one in which the generator if con-
nected to a discharged battery will
deliver a high rate at the start of the
charge, but as time progresses the
rate will gradually fall until, at the
end of the charge, it is down almost
to zero.
This system usually includes a
straight shunt-wound generator which
builds up to a voltage equal to that
necessary for the maximum charging
rate at comparatively low speed and
some form of vibrating voltage regulator
whose function is to hold constant generator
voltage. This is done by alternately cutting
an external resistance in and out of the shunt
field circuit. Its rate and period of vibration
depend upon the speed at which generator
is being driven and the battery current
requirements.
240
220
200
V
o
-u 120
160
140
too
I
80
60
40
20
\
-
Characteristic Batterq Charqinq Curve
Voitaqe Requ fated Si/stcm
-
V
\
\
\
\
s.
V
s.
s
s
S
■
s
*»
^
'^
___
0123456789/,
Time Hours
Fi^. 7. Characteristic Battery Charging Curve, Voltage
Regulated System
Fig. 8. Starting Motor and Voltage Regulated Generator
Taper Charging
The voltage regulated or constant poten-
tial systera of battery charging, which gives a
ta]jering charge. Fig. 7, is based on the fact
that the counter electromotive force or oppos-
ing voltage of a battery is lower when the
battery is discharged than when it is charged.
The difference will be in the order of 0.6 volts
per cell, or for a 3-cell 6-volt battery will be
1.8 volts. Therefore, if the generator is
set to hold 7.S volts, equalling a fully charged
battery, it will have, with a discharged battery
having a counter electromotive force of
only 6 volts, l.S volts available for
forcing the charging current through the
battery; consequently, the charging rate
will be high. Leaving the regulator and
generator characteristics out of considera-
tion, the high current rate will be determined
by Ohm's law, where E or voltage is the dif-
ference between generator voltage and battery
counter electromotive force and R or resist-
ance equals the sum of the battery and
external circuit resistances. If £=1.8 and
R = 0.06, then the charging rate to the batterv
will be 1 .8 -H 0.06 = 30 amp. at the start. The
rate will taper to zero when the charge is com-
plete, at which point the battery counter elec-
tromotive force equals the generator voltage.
192 March, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 3
In actual practice this condition is only
approximated, that is, the regulator or gen-
erator, or both, may be so designed that
the initial rate will be lower than the rate
indicated by the foregoing formulas, and the
final rate will not be zero, but rather some-
Fig. 9. Starting Motor for Sleeve Mount (Outboard Shift)
thing of the order of 5 or C amp. This is
done to prevent the generator from being
excessively overloaded during the first part of
the charge, and to insure the battery receiving
a low rate overcharge after completion of the
regular charge.
Reverse Current
Cut Out
the cable by means of a clamping band
v.'hich encircles the insulation at a point
beyond the bared portion to which the solder
is applied. Above all, terminals should be
tight on connection boards, as loose ter-
minals mean extra resistance, and extra resis-
tance in the lamp or ignition circuits means
decreased brilliancy of lights or unreliable
ignition. In the generator circuit of a third
brush machine, extra resistance means in-
creased generator voltage with attendant
heating of the generator; and in the generator
circuit of a voltage regulated machine means
decreased current output to the battery.
Society of Automotive Engineers
The work of the Societ\- of Automotive
Engineers toward the standardization of all
parts of the automobile has been of great
value in simplifying and standardizing the
mounting of electrical apparatus. The
Society through the medium of its standards
committees has recommended for adoption
by manufacturers: three methods of mount-
Shaft Extention
f"or Ignition Driver
JIL
^.
Q
Flange
OriveEnd
Fig. 10. Generator for Flange Mount
Terminals
Car builders in many instances do not give
the subject of connections and terminals the
consideration thac it should have. Terminals
should be rugged to withstand vibration, and
should so hold the cable that the effects of
vibration at the point they are attached will
also be minimized. Terminals should always
be soldered to cables, but the solder should
never extend beyond the last point of sup-
port of the cable; in fact, it is preferable that
the terminal be so designed as to support
ing starting motors ; two methods of mounting
generators; one form of pinion and gear tooth.
The three mounts are :
First: for inboard flange mount with three
sizes of flange.
Second: for outboard flange mount, with
three flange sizes.
Third : for outboard sleeve mount. This
in only one size.
In each of these, all dimensions wliich
affect both motor and engine manufacturers
ELECTRIC STARTING SYSTEMS FOR AUTOMOBILES
193
are given. Roughly, these are: flange bolt
drilling and location of holes; diameter of
pilot; distance from flange face or dowel
screw to flywheel teeth; and height of fly-
wheel teeth above flywheel proper.
The gear and pinion tooth selected is of
standard S-10 pitch 20 deg. pressure angle.
' The generator mounts are :
First : flange with two sizes to accommodate
large or small machines.
Second: bracket with but one size laid out
to accommodate the largest generator that
may reasonably be encountered.
The flange method of mounting is employed
when the generator is driven direct by a gear
or sprocket running in the engine timing gear
case. The engine half of the flange mount is
then machined on the rear face of the gear
case. The bracket mount is used where a
separate shaft is brought out of the timing
gear case for driving the water pump, igni-
tion apparatus, or generator, or sometimes
two or all three of them. In this case, the
generator is mounted on the engine bracket
and driven by means of a flexible coupling.
In these layouts, as in the motor layouts,
all common dimensions are given, including
shaft and sizes, coupling fits, and height of
shaft above bracket, and in the case of the
flange mount, shaft end sizes for gears or
sprockets and drilling and shape of flange.
Fig. 11. Generator for Bracket Mount
Fig. 12. Inboard Mesh Starting Motor
194 Marcli, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 3
The Outdoor Generating Station
By H. W. Buck
Vice-President, Viele, Blackwell and Bvck, Engineers
That a power plant must be sheltered by a "house" has become a habit of mind in engineering design.
In the opinion of the author it is now time to analyze the situation and to determine whether there is really
any need for the expensive buildings that have always been erected for housing hydro-electric generating
equipment. In the final analysis it would seem that the function of such a structure is to house the station
operators and switchboard panels and to provide favorable conditions for initial installation work and subse-
quent repairs. Drawings are shown of an outdoor generating station which was designed and submitted to
the War Department for a development at Muscle Shoals, Alabama. It is shown that a plant of this kind is
entirely feasible and offers decided advantages from the standpoint of economy in construction. — Editor.
operating expense and thermodynamic effi-
ciency. In this respect there is still hope for
improved economy and lower generating
cost. In the hydro-electric plant, however.
there is little to be expected in the way of
lower operating expense, which is already
very low. Water turbines have now reached
an efficiency of about 94 per cent, so that there
is small hope for improvement in this respect.
The large item of annual cost of power from
the water power plant is the interest charge,
and this can only be reduced by reducing the
ca])ital cost of construction. The most hope-
ful field for saving in this part of the account
lies in simplifying the power house con-
struction, particularly by eliminating the
costly super-structure now universally built
to house the hydro-electric generators.
The modem vertical shaft internal revolv-
ing field generator is essentially a waterproof
stnicture. The vital parts of the machine are
all on the inside protected by a massive casing
of cast-iron, and the openings in the upper
spider are usually jjlated with steel for pur-
poses of ventilation so that the top is naturally
])rotccted. With slight modifications in its
design, the standard vertical shaft alternator
can be made absolutely proof against all
stresses of weather. The waterwhecls them-
selves naturally need no housing, since they arc
imbedded in the concrete substructure of the
power house and are designed to run in water.
Thrust bearings are housed in heavy steel
or iron casing and need little protection from
weather. The various auxiliaries, including
governors, connccteil with an hydro-electric
I)lant can easily and adxantageously bo
installed under the main generator floor, in
the various compartments naturally existing
in the substructure of such a plant and there
protected from the weather. It is therefore
interesting to inquire why millions of dollars
are expended to house machinery which
really does not need housing at all.
In the last analysis it ajipears that the
function of the hydro-electric superstructure
The installation of electrical apparatus in
the open air without the protection of a hous-
ing structure is of increasing importance, due
to the constantly increasing cost of building,
and also to the increasing cost of construction
funds.
The installation of small transformers out
of doors is as old as the electric lighting busi-
ness, since such apparatus has always been
considered intrinsically weather-proof. In
recent years other types of apparatus, such as
oil switches, disconnecting switches, light-
ning arresters, busbars and other substation
ajjparatus have been forced out of doors for
economic reasons on account of their increas-
ing size with higher voltages, and the high cost
of housing.
Thus far, however, little progress has been
made on the out-door installation of generat-
ing and other rotating electrical machinery,
for such apparatus has been regarded as more
or less perishable. There have been some
isolated cases of outdoor installation of
generators, such as some mining power plants
in the arid regions of Arizona and Mexico.
where boilers, engines and generators have
been installed in the open. There is also a
modern installation of a waterwheel-driven
generator on one of the power systems in
Utah. The question, however, has been
under discussion by engineers for a number of
years, but prejudice has worked against it.
That a jiower house must have a "house"
in order to be a workable combination has
become a habit of mind in engineering design.
The time has now come, in the opinion of
the writer, to analyze the situation and to
determine whether there is a basis of justifi-
cation for the large investment required in the
construction of the modern and expensive
power house superstructure in connection
with hydro-electric plants. The economic
situation at the present time forces this
question i)romincntly to the front.
In the steam dri\-en generating ])lani the
annual cost of power is largely a matter of
THE OUTDOOR GENERATING STATION
195
lv;-.-;;'j.,:vt,:.-^
%;yigg'i^.iy^A:^?S'";"g::'^''^^>!g^«''yii^jjN'^?^
^^-X^
Fig I. Cross Section of a Generating Unit in the Outdoor Station Proposed for the Muscle Shoals Development in Alabama
196 .March, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 3
is to house two or three station operators, to
house the switchboard panels, and to produce
conditions favorable to original installation
work, and for repair work required from time
to time thereafter.
In regard to the switchboard panels, there
is apparently no reason why they should not
be housed in a small pilot house of a size not
over five per cent of the size of the total power
house superstructure required.
The drawings. Figs. 1 and 2, show a general
design of the Muscle Shoals Development
prepared for the War Department by Viele,
Blackwell & Buck in August 1918. The
design, as shown, comprises the outdoor
installation of fifteen 20,000-kw. generating
units. The situation at Muscle Shoals is ver\-
plant auxiliaries would also be housed. All of
the vital parts of the alternators could be
inspected from below the machines. Only the
occasional inspection of thrust bearings would
have to be made in the open. In this
connection it should be remembered that
sailors are not housed in the work which they
are obliged to do, and there is no reason why
some of the operators of a power plant should
not perform some of their duties in the open air.
In order to perform, under protection from
the weather, the work of installation and
repair on the generating units of this outdoor
^luscle Shoals plant, it was planned to install
a housed-in traveling gantry crane of suffi-
cient length to cover a single unit, which
would tra^-cl on rails the total length of the
Fig. 2. General Design of the Outdoor Generati
favorable for such an outdoor generating
station. The climate in Alabama is mild,
and the number of generating units in the
plant is large, which gives a maximum of
saving from eliminating the superstructure.
It was proposed to install within the generat-
ing structure itself only the water turbines,
generators, exciters, governors and the various
pumps. The switchboard panels were to be
located in a relatively small "pilot house" on
the bluff overlooking the generating structure,
which would also contain the low voltage oil
switches and busbars. All of the transformers
and high tension equipment would be installed
in the open, adjacent to the pilot house. In
this arrangement, the switchboard operators
would be housed and the generator attend-
ants also protected in the substructure. The
generator structure. In this way it would be
available for any one of the units. The crane
could be placed over any desired unit. and.
with the various openings in the ends and
sides of the gan t ry closed wi t h ad justable panels,
the unit could be completely protected from
the weather for handling and repair work.
In the case of this Muscle Shoals plan tthe
saving in construction cost by omitting the
superstructure was very large. Approxi-
mately 3000 tons of steel for superstructure
columns and roof trusses would be saved, and
after paying for the "pilot house" and gantr>-
crane, the net saving would amount to over
STOo.oon.
It will be noted from the cross section of the
plant. Fig. 1, that the generator assembly
is of a somewhat novel type. The usual iron
I
THE OUTDOOR GENERATING STATION
197
frame of the stationary armature is omitted,
and the armature laminations and coils are
attached to the surrounding concrete. The
only function of the armature frame of a
standard generator is to provide general
stability, and this can be provided by the
surrounding concrete, which must be there
in any case for supporting the weight of the
machine. By this arrangement approxi-
mately 70,000 lbs. of irori per generator can
be saved in the plan shown. It is also possible
by this design to save a large amount of boring
mill work on the armature frame at the
factory, which is apt to be the limiting
element in production in most machine shops.
This method of construction was first sug-
gested, in the knowledge of the writer, by
and it might be necessary to use a specially
light oil in the upper bearings in cold weather
on an outdoor machine.
The saving in the construction cost of a
power house superstructure is not the only
saving resulting from the outdoor construc-
tion. The annual maintenance of such a
building is a considerable item, which would
be entirely eliminated.
The advantage in cost of eliminating the
])ower house superstructure will increase, of
course, in proportion to the number of
generating units in the plant. In a single
unit plant, for instance, it might be a fact
that a housed crane for handling the unit
would prove almost as expensive as an
enclosed fixed superstructure; but with, say
Dn Proposed for the Muscle Shoals Development
H. G. Reist. It is mentioned here because it
is admirably adapted for the outdoor instal-
lation of generators.
It may be argued that an outdoor generat-
ing station, which would be successful in the
mild climate of Alabama, would not be
practicable in an installation where severe
winters are experienced. There does not
appear, however, to be much weight- in this
argument. An outdoor generator can be
made snow-proof as well as rain-proof. The
operator would be normally housed, in any
case, in the pilot house or between decks in
the substructure, and would be required to go
on deck at occasional intervals only.
In cold climates it would proljably be neces-
sarj' to make some special provision against
freezing, where water-cooled bearings are used.
three or more units installed, there should be
no doubt of a large saving in cost of plant
construction. In such plants as Muscle
Shoals, Keokuk, Cedars and Niagara Falls,
where there are large numbers of units
installed, the possible saving is very large.
Whether the outdoor plan can be applied
to steam turbine driven plants must be
decided by future development. Some com-
plications might be encountered from freezing
in idle steam pipes, valves, water pipes, etc.
It does not appear impossible, however, to
install turbo-generator units, and also the
boilers, etc., in the open, if special protection
can be worked out for certain parts of the
equipment. Such an installation would
afford an opporttmity for a very large saving
in the construction cost of such plants.
198 March, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 3
Methods for More Efficiently Utilizing Our
Fuel Resources
PART XXXI. PETROLEUM*
By Chester G. Gilbert and Joseph E. Pogue
Division of Mineral Technology, United States National Museum
In this series of articles we are at present reviewing the fuel resources of the Western Hemisphere. Previous
installments have described the fuel resources of Canada and Alaska; also the coal and natural gas resources
of the United States. The present installment is the first of a group which will treat of the petroleum resources
of the United States. It is introductory in character and treats of the nature and occurrence of petroleum and
the essential features of the petroleum industry, including production, transportation and refining, as well as
the distribution of the products. The next installment will treat of the petroleum reser\'e and its limitations.
Then the conservation of petroleum will be taken up. — Editor.
Petroletim is of peculiar value to society
because it is the sole source of gasolene, the
dominant motor fuel; provides kerosene, the
most important illuminant outside of cities
and yields lubricating oil, upon which the
wheels of industry' revolve. In addition, it
has come to be an essential fuel in the South-
west and on the Pacific coast, where coal is
lacking; is requisite to the operations of an
oil-burning navy ; and forms the starting point
for an oil by-products industry, a branch of
chemical manufacture still in its infancy and
offering unlimited possibilities of development.
The liquidity of the crude product makes
petnoletun unique among mineral raw materi-
als, contributing wide commercial availability
through the ease with which the substance
may be mined and handled; while the magni-
tude of the resource has given confidence for
the extensive mechanical developments essen-
tial to its use. As the petroleum deposits of
the United States have been drawn upon with
extraordinary rajjidity and the supplies have
already suffered serious depletion, the matter
of their approaching exhaustion assumes the
light of immediate importance. The com-
fortable assertion that such considerations
may be safely left to future generations does
not apply to petroleum.
Nature
Crude jjetroleum, as the raw or unrefined
product is often termed, is an oih" liquid
varying considerably in appearance according
to the locality from which it comes. It is an
extremely complex mixture of organic com-
pounds, chiefly hydrocarbons, but substances
containing sulphur, oxygen, and nitrogen are
also present in small amounts.
If exposed to the air, it gradually thickens
until a solid residue is left. The first product
♦Extract from Bulletin 102. Part 6. U.S. National Museur
"Petroleum: A Resource Interpretation."
given off is natural gas; then liquid com-
ponents evaporate in the order of their light-
ness; and the final residue is composed largely
of either paraffin wax or asphalt. Petroleum
is thus seen to be a mixture of different liquids
dissolved in one another and holding in solu-
tion also natural gas and solid substances.
This conception correlates natural gas as a by-
product of petroleum and affords a simple
epitome of the changes more rapidly induced
when petroleum is subjected to refining.
The asphalt lake of Trinidad and the ozokerite
deposits of Galicia and Utah represent natural
residues from the prolonged evaporation or
natural distillation of petroleum.
While petroleums vary considerably in
character, they fall chiefly into two classes
according to whether the residue yielded is
predominantly paraffin wax or asphalt. The
first are said to have a paraffin base; the
second, an asphallic base, or called merely
asphaltic petroleums. There are also inter-
mediate oils with almost equal proportions of
])arafiin and asphalt. This broad distinction
is of great economic significance, because the
paraffin petroleums, occurring chiefly in the
eastern part of the countn.-, came first into
use and therefore determined the current
refining practice and the existing demand for
petroleum products; while the asphaltic
petroleums, exploited later in the Gulf region
and California, found their immediate com-
mercial outlet in the fonn of fuel. The
higher gasolene content of paraflin oils,
coupled with the distance of coal from the
Califoniian region, gave free scope to the
economic differentiation of the two types.
Occurrence
Because of its liquiility. i)etroleum differs
markedly in geological occurrence from all
other minerals. It appears on the surface in
some localities in the fomi of oil seeps, hut
METHODS FOR MORE EFFICIENTLY UTILIZING OUR FUEL RESOURCES 199
commercial quantities of petroleum are found
only at depth inclosed within the rocks of the
earth's crust. Its occurrence is very similar
to that of artesian water, with which, indeed,
it is frequently associated. It saturates
certain areas of porous rocks, such as beds of
sand or sandstone, tending to accumulate
where such strata occur beneath denser,
impervious layers. Occurring in this way
under the pressure that obtains at depth,
carrying immense quantities of natural gas in
solution, and almost invariably associated
with water, petroleum is capable of movement
and in general migrates upward until it
encounters a layer of impervious rock so
disposed in structure as to impede further
progress and impound the oil into a "reser-
voir" or "pool," Fig. 1.
the world, therefore, oil production can now
be made as definitely an engineering project
as the mining of a clay bank.
The migratory character of petroleum,
coupled with the general tendency of stratified
rocks to occur in broadly imdulating folds and
shallow domes, gives peculiar significance to
the underground disposition of the oil deposit.
Thus the process of winning the oil consists in
puncturing the structural feature that holds
it so as to give free scope to a movement
upward to the surface. Accordingly the
position of the oil grows highly unstable as
soon as the deposit comes under exploitation
and this affects the entire geological unit or
pool. In consequence the joint ownership or
joint exploitation of a single pool results in
the inability to apportion the product on any
Fig. 1. View of the Occurrence and Mining of Oil and Gas
The geology of petroleum, therefore, is the
geology of rock structures, and the skillful
mapping of the surface disposition of rock
formations gives the means for determining
the structure at depth and hence the position
of structural features favorable to the accu-
mulation of oil. When this information is
supplemented by careful records of the rock
layers encomitered as wells are drilled, a three-
dimensional knowledge of the earth's crust is
obtained, remarkable for its detail and accu-
racy. Thus by the aid of geological methods
the development of petroleum fields may be
changed from a gambling venture to an exact
science, and, if the scale of operations be
sufficiently large, it may be- figured rather
closely how much oil can be obtained from a
given expenditure of money. Instead of
representing the most uncertain venture in
arbitrary basis of vertical boundary planes,
and the oil, therefore, is practically no man's
property until it is got above ground. This
circumstance is almost invariable and the
customary method of exploiting the single oil
pool by a series of small, independent holdings
has cost an inordinate toll of waste and loss.
The economics of oil production is out of
adjustment with the geological occurrence of
Origin
Few questions in geologic theory have met
with more discussion than the origin of
petroleum. It is reasonably certain, however,
that petroleum in the main is of organic
origin and represents the natural distillation
products of plants and animals buried in the
muds and oozes of ancient swam|_)s and seas.
200 March, 1U20
GENERAL ELECTRIC REVIEW
VoL XXIII. Xo. 3
Vast rock formations, indeed, are known
which are nothing more than the accumulated
debris of innumerable organisms, compressed,
hardened, and changed into rock. Fossili-
ferous limestones, phosphate rock, and coal
seams are familiar examples which underlie
thousands of square miles of the earth's sur-
face. It would be strange, in fact, if in the
Unitad states
Meat CO
Dutch £asl Indies
^oumonto
fnd/o
Oof/cto
/I// Others
.- 1 I ! I I
:
1 ' I I I I
100 lio 200 ;so
Millions of Barra/S
~^
I
.J
Fig. 2. World's Production of Petroleum in 1916
process of formation oils were not produced,
when organic products today, subjected to
heat and pressure, yield oily substances not
unlike petroleum. Sediments carrying organic
remains are sufficiently abundant and wide
spread to account for all the petroleum that
the oil fields of the world give promise of
producing.
Distribution
While petroleum is of ver\- common occur-
rence in traces, areas underlain by commercial
quantities are somewhat restricted and fields
of great importance are few. Thus in spite
of an intensive search for new oil regions and
vigorous campaigns of development carried
on in all parts of the world, the entire supply
comes largely from three countries, as shown
in Fig. 2.
In the United States the output is derived
from a number of widely scattered regions
known as "fields." In a broad way, these
fields fall into two groups — those of the east-
em half of the United States, bound into a
single unit by an extensive system of pipe
lines, and those of California, connected with
the rest of the countn.- by railroad transporta-
tion onl}-. The intermediate fields of Wyo-
ming do not come within this rough geographic
classification, but with further development
they will presumably be joined \>\ pipe lines
with the group of the eastern half of the
countr>'. The Kansas-Oklahoma field of the
eastern group and the California field are
about equal in production and dominate the
petroleum output of this country, together
contributing over two thirds of the total
supply.
The development of petroleum production
in the United States from 1881 to 1917. is
indicated graphically by Fig. 3. From the
situation there depicted, two features of
particular significance stand out — the slow
increase in domestic production up to 1900,
less marked than the increase in the corre-
sponding foreign production, and the rapid
domestic growth between 1900 and 1917, con-
trasted with a nearly constant production for
foreign countries during that period. This
emphasizes the fact that since the beginning
45O.0OO.OO0
/
400.000.000
/
/
550.000.000
/
I
y
/
500.000.000
n ^
/
'
A*°V
h ^
Jo ZSO.OOOflOO
o***" / -
X/
J'
/ >/
/
Ar-
-{/ >
X
40'J^
,.c^' y
150.000.000
^J^
' y
y
r*'" >
J
, hf "S
1 00.000.000
^c,r y
.*^"''"
y
..Aor^'"
y
■^...uoo^"
^ ' ^
.l,c r'
Importtc
ifiT>mMnt>c
0
0
00"'*'^
yr^
/S85
J890
/S95
1905
1910
I9'5 /9I7
1900
Years
Fig. 3. Chart Showing Petroleum Used in the United State] and the Rest of the World from 1880 to 1917
METHOES FOR IMORE EFFICIENTLY UTILIZING OUR FUEL RESOURCES 201
of the twentieth centur}-, the rapidly increas-
ing use of petroleum throughout the world
has been met largely through the intensive
exploitation of American deposits. Thus the
United States has assumed a dominant posi-
tion in respect to this commodity, producing
now two thirds of the world's supply.
THE INDUSTRY
The activities concerned with the produc-
tion, transportation, refining, and distribution
of petroleum constitute the petroleum indus-
try'. In quantity, value, and importance of
production, this industrial field stands among
the foremost in the country. It is notable,
especially, for the scope of its operations, which
embrace diverse activities usually the function
of separate industries — a characteristic arising
from the peculiar nature of petroleum. In
most other industries, to cite the most striking
distinction, transportation over alien lines
separates the producing activity from the
manufacturing activity, creating a break
between continuity of operations ; in the case of
petroleum, however, the liquidity of the crude
product adapts it to specialized transportation
through pipe lines, themselves a part of the
resource development. In consequence, the
petroleum industry in its ideal form represents
a type of industrial activity more highly
coordinated than other industries of the
present day, affording, therefore, an important
object lesson for constructive consideration.
The petroleum industry, in point of fact,
however, is not coordinated throughout, but
at present breaks into two portions, by no
means in complete adjustment — the produc-
tion of petroleum and the handling of pe-
troleum with its threefold aspect of trans-
portation, refining, and distribution. The
conditions of producing crude petroleum are
wholly different from those involved in its
treatment after it is above ground. This is
reflected in the circumstance that over 15,000
individual companies are engaged in the min-
ing of petroleum, while the organizations con-
cerned with the handling of the product are
numbered by a few hundred. About SO per
cent of the crude production appears above
ground through the efforts of a great many
small operators, while the bulk of the trans-
portation, refining, and distribution is taken
care of by a very few large organizations.
Production
Petroleum is won in commercial quantities
through wells drilled to varying depths into
the crust of the earth. The drilling is com-
monly done by means of a heavy string of
tools suspended at the end of a cable and
given a churning motion by a walking beam
rocked by a steam engine. This method is
known as the standard or percussion system of
drilling. The steel tools, falling tmder their
own weight, pulverize the solid rock encoun-
tered and literally punch their way to the depth
desired. To prevent the caving in of the hole,
but especially to avoid the inflow of water
from water-bearing formations, the well is
lined or "cased" wholly or in part with iron
piping, which is inserted in screw-joint sec-
tions at intervals during the drilling and
forced down to positions needful of such
protection. The well does not taper, but if
deep changes to successively smaller bores at
several points, resembling in section a great
telescope.
Another method of drilling, known as the
rotar\- system, is also in common use, being
particularly adapted to regions where the
sides of the well tend to cave badly, as in
California and some other localities. This
system requires more elaborate machinery
than the standard, as the drilling and insertion
of the casing is simultaneous. The iron cas-
ing, indeed, is tipped with a steel bit and
rotated so as to bore its way downward like a
great auger.
The oil well is marked by a tall wooden
framework called a derrick, which permits the
string of tools and the casing to be inserted
or withdrawn when necessary. It is the
presence of derricks that gives the character-
istic appearance to an oil field landscape. Oil
wells vary froin a few hundred feet or less in
depth, requiring a few weeks only to drill, to
those thousands of feet deep and demanding
months of continuous labor before production
starts. The deepest wells are slightly over
7000 feet, but such depths are exceptional.
The cost of drilling, before the war, ran from
$1 up to $15 and more a foot, while the rate of
progress, except for shallow wells, ranges from
about 60 down to 10 feet a day, slowing, of
course, with depth. It is apparent, then,
that oil-well drilling is a slow and costly proc-
ess and makes a heavy draft upon the iron
and steel industry, consuming, indeed, about
one twelfth of its output in ordinary times.
A well favorably located eventually pene-
trates an oil-bearing bed, and the petroleum
maj' spurt forth in a lavish stream under the
influence of the natural gas held in solution
under pressure. Such wells are called gushers
and some pour forth prodigious quantities
of oil. Other wells flow with less violence,
202 March, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. .3
and many, lacking in notable quantities of
natural gas, yield only under the inducement
of pumping. All wells, however, soon reach
a maximum production, after which they pass
into a period of decline, and eventually
become' extinct. So inexorable is this pro-
cedure that a cun'e may be plotted in advance
depicting the future behavior of a given
group of wells. Wells during decadence are
spurred into temporary renewals of activity
by the explosion of charges of nitroglycerine
at their bottoms. The life of an oil well varies
from a few months to twenty years or more.
The average life of Penns\dvania wells is
estimated to be seven vears.
When a gusher is struck, adequate facilities
are often lacking for catching and storing the
product, so that veritable lakes of oil gather
between quickly thrown-up earthen embank-
ments. Quantities, in such instances are dis-
sipated through seepage and evaporation,
while disastrous fires of spectacular nature
are not uncornmon. With more careful
development, however, field storage tanks
shaped like huge cheese boxes are in readi-
ness to receive the oil and prevent the glaring
waste inherent in more hasty operations.
Turning attention from the single well to
the oil field, we obser\'e that in petroleum
mining sustained production depends upon an
Fig. 4. Model of an Idealized Petroleum Refinery
When an oil well becomes extinct, its
nonproductiveness does not signify that all the
oil is exhausted. On the contrary, current
practice in general leaves over half of the oil
underground still clinging to the pores and
capillary spaces in the rock. To obtain a
greater yield from productive ground con-
stitutes a problem of the first magnitude, and
promising results have been obtained by forc-
ing compressed air into some of the exhausted
wells of a group, with the result that the
laggard oil is swept to the neighborhood of
other wells from which it may be pumped.
' As an oil field ages, new wells yield less than the inili.il
yields of the earlier wells, hence a growinR number of active
wells is necessary to maintain production.
unl)n)ken campaign of drilling operations.
Thus the producers must not only draw oil
from existing wells, but at the same time
must persist in the drilling of an incrcasinf:'
number of new wells and in locating promising
territory in advance of drilling. Any factor
that retards any one of these three related
activities quickly reacts to cause a falling off
in ])roduction.
Output, development, and exi)loration.
therefore, must go hand in hand. In a general
way, this threefold activity of production is
carried on either as a large scale engineering
|)rocedure or as a composite of small, in-
dividual operations. Large o\\ companies
engaged in production naturally adopt what
METHODS FOR MORE EFFICIENTLY UTILIZING OUR FUEL RESOURCES 203
might be called the engineering procedure,
while small companies and individual opera-
tors tend more to follow what is picturesquely
termed "wildcat"- operations. Thus the
production of oil is in part dependent upon
stable conditions, but in larger part is still
a type which operates in considerable
measure as a gambling venture. This is wh>'
oil mining is generally looked upon and
commonly described as hazardous from a
financial standpoint. The hazard is inherent
only in small-scale operations.
The engineering type of production makes
use of skilled geological knowdedge in its
campaign of oil production. The modern oil
company employs a large geologic staff, which
determines by detailed field sur\'eys the most
promising spots for drilling. The growth of
oil geology has been rapid and while, of course,
geologic science can not strike oil with e^•er\•
drill, it does multiply by many times the
chances of each drilling operation. It has
been stated that "the operator who plays
geology has a fifty times better chance of
striking oil than he w^ho does not."
But in spite of numerous highly organized
production activities, the fact remains that
the petroleum production of the United
States is in considerable measure dependent
upon a hit-or-miss plan of exploitation. Were
it not for the wildcatter, who stakes his all
(sometimes borrowed) on the chance that a
random hole drilled in the general vicinity
of productive territory will yield the hoped-
for return, the output of petroleum in a
country which produces two thirds of the
world's supply would fall to an utterly in-
adequate figure. The gambling instinct is
still the prime motive power that lifts most
of the oil produced in this countrv*.
It is not intended, of course, to throw oil
production into an unfavorable light by thus
focussing attention upon its gambling aspect;
to exert onerous effort (such as oil field
development demands) under the incentive of
rich possibilities of reward is a straightforward
and legitimate business activity. It is fre-
quently questioned whether oil development
could be sustained without prospect of large
pecuniary- gain. The point is merely made
that under present circumstances petroleum
production is dependent upon this psycho-
logical aspect, acutely developed, which is
both subtle and intangible, yet profoundly
important in conditioning the output; this
factor must be reckoned with in contemplat-
ing the course of the resource development.
Production and consumption, of course,
can not coincide in amount; hence, of neces-
400,000,000
350,000,000
500.000.000
250.000.000
§ ZOO.OOOfiOO
150.000.000
100.000,000
5O.0O0.000
2 In strict oil-field parlance, to 'wildcat is to drill a well
^here oil has been proven not to exist, as opposed to drilling a
well in the midst of producing wells. Thus both large com-
panies and small may alike engage in wildcatting, although, as
a matter of course, most of the yjitdcalling is done by small
operating units.
/90a 1909
Years
P"ig. 5. Chart Showing the Relative Values of the Principal
Petroleum Products Manufactured in the United States
from 1899 to 1914
sity, there are reserves of petroleum above
ground which serve as an expansion and con-
traction joint, so to speak, between supply
and demand. When there is an overproduc-
tion in respect to current needs, the reser\-es
or, as commonly termed, the stocks increase
conversely, with industrial expansion or les-
sened output, drafts are made upon the stocks,
which then decrease. The condition of the
stocks, therefore, is a sort of pulse to the crude-
oil market, since prices, under the influence
of the same factor of supply and demand,
fluctuate in like manner. The stocks, under
conditions of unorganized production, have
come to be unusually great during the past
few years, representing roughly in 1916 a
six-months' supply. Under war conditions,
the stocks were rapidly depleted to meet a
consumptive demand which was greater than
the productive capacity of the country.
The price of crude petroleum at the well
varies considerably according to quality,
distance from market, and other factors. The
paraffin oils of light gravity, such as those
produced in Pennsylvania, are the most valu-
204 March, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 3
able because they yield the largest percentage
of products in demand, while the asphaltic
oils of heavy gravity, such as those of Califor-
nia and part of the Gulf region, command a
price roughh^ a fourth of that which the best
oil enjoys. Thus the Pennsylvania crude
commenced 19L5 with a price of about $1.50 a
barrel and ended 1917 at about $3.75, while
during the same period California crude
climbed from about 35 cents to practically $1.
These two types of oil represent the extremes
of quality, with the factor of distance from
markets nearly the same in the two instances.
Between these limits range the prices of all the
other oils of the country, the quotation at
any given time and location being a complex
of quality and of balance between supply and
demand, with all the qualifications that the
latter expression involves. The wide range
in prices for a single raw material, with the
utmost concession to differences in location
and composition, suggests an undue dis-
crepancy to be credited against the conditions
under which oil is produced.
The dependence of sustained production
upon an unbroken campaign of drilling
exploration, and the extent to which such a
campaign is carried on by "wildcat" opera-
tions on the part of small companies and
individuals, lead to many perplexing legal
and economic difficulties. Land, of course, is
rarely owned by the operator, so that he must
ordinarily either purchase or lease the oil
(and gas) right. The laws connected with oil
lands have not been modernized, but are
confusing and in part conflicting, so that the
operator is put to undue trouble and expense
in meeting the legal requirements of his hold-
ings. Moreover, the method of leasing under
small unit operations leads to a wasteful com-
petition between neighboring wells in their
race to secure a maximima production within
the period of the lease — haste, with waste,
being an economic necessity in such instances.
In regard to lands owned by the Govermiient,
the legal regulations are so ill-adapted to prog-
ress that R. n. Johnson and L. G. Huntley
in their "Principles of Oil and Gas Produc-
tion," remark: "Most of the public lands
which seem promising for oil and gas have
been withdrawn, since there is universal
agreement by both Government and pro-
ducers that the present law, by which oil and
gas lands are taken as placer claims, is utterly
unadaptcd to the industry. The develop-
ment of the lands which are not withdrawn
would best be postponed until a new oil and
gas prospecting permit and leasing law is
passed, and the oil placer claim law revoked
except where work is already started."
Transportation
One of the remarkable and impressive
features of the petroleum industr\- is the fact
that the crude product is transported through
a system of pipe-lines that connect the points
of production with refineries, markets, and
seaports. This method of handling is natural
and inevitable with a liquid product con-
sumed in bulk, as e\-idenced by a somewhat
analogous method of transportation adopted
for the municipal water supply. While pe-
troleum shares with coal the main responsi-
bility for energizing the mechanical acitivities
of the country, it is interesting to note that
crude oil, unlike raw coal, imposes normally
no appreciable burden upon the railroads.
The pipe lines of the United States, com-
prising those of the subsidian,- companies of
the Standard Oil and a number of independ-
ent companies, aggregate thousands of miles
in length and form a network spread over
much of the country. They consist of trunk
lines, the longest of which connects Oklahoma
with the Atlantic seaboard by way of Illinois,
and gathering lines leading into the main
channels. The approximate mileage of the
principal lines of the United States amounts to
28,995 miles. The total length of all the pipe
lines is much greater.
The pipes varj- in diameter from 2 to 1 2 inches,
but G to 10 inches represent the common sizes.
The piping is made of iron plate and is ordi-
narih- placed below the surface of the ground.
At inter\-als of from 15 to 30 miles, according
to the viscosity of the oil, there are pumping
stations. In the case of heavy, viscous oils,
such as those of California, it becomes neces-
sary to heat the product at each pumping
station to facilitate its progress. Unlike a
railroad, the pipe-lines, in general, follow a
direct course, uphill and down. An S-inch
pipe weighs 2S pounds per foot, and its cubic
capacity is about 32S barrels of oil a mile.
This means that millions of barrels of oil are
required merely to keep the pipe lines of the
country active. The pipe-line facilities of the
country are ample to handle the normal dis-
tribution of the current production.
The significance of the pipe line in the
development of the petroleum industry has
been great. It has made crude ]ietrolcum inde-
pendent of the railroads and through cheap-
ness of operation has lowered the cost of petro-
leum products; it has freed the refineries from M
geographic allegiance to areas of production "
METHODS FOR MORE EFFICIENTLY UTILIZING OUR FUEL RESOURCES 205
and ])ermitted their establishment at strategic
points in respect to consumption of products;
it has pennitted and induced integration of
activities, with marked advantage to the
consuming public, but not unaccompanied
by hardships and abuses falling upon small
units of the industry itself; and by stretching
out to meet a growing area of exploitation it
has unified widely separated fields and enabled
production to grow to its present imposing
size. The pipe line has woven the- scattered
strands of adventurous exploration into a
steady flow of bulk raw material.
Some crude petroleum is transported in
tank cars, but most of the 60,000 tank cars
in operation in this country are engaged in
moving petroleum products — gasolene, kero-
sene, and fuel oil chiefly. For transportation
by sea, steel tankers and towing barges, fitted
with non-communicating compartments, are
employed for both crude petroleum and its
bulk products. The development of the tank
steamer has been an important factor in
building up an important foreign trade in
petroleum products, is responsible for a con-
siderable coastwise movement of crude and
fuel oil, and has opened the oil fields of Mexico
to the United States and other markets.
Refining
Crude petroleum may be burned as fuel
and nearly a fifth of the domestic consi-mip-
tion is utilized in this way. But most of the
petroleum is manufactured into a series of
products which have wider usefulness and
higher value than the crude oil, and it is
upon this dominant part that the petroleum
refining industry depends.
At the present time petroleum yields, when
completely refined, four main products —
gasolene, kerosene, fuel oil, and lubricating
oil — and a large number of by-products, of
which benzine, vaseline, paraffin, road oil,
asphalt, and petroleum coke are well-known
examples. These are commercial terms and
therefore carry no exact meaning in a chemical
sense. Since the products nierge one into the
other, there can naturally be between them
only an arbitrary line of demarcation.
Gasolene, as here used, covers those products
of crude oil which are more volatile than
kerosene; the term therefore embraces some
benzine and naphtha. Kerosene, as here
used, is the coinmon type of illuminating oil
representing the distillate heavier than gaso-
lene, but lighter than fuel oil. Fuel oil is used
2 See Part VIII of this series. Dec. 1917. [Ed.]
as a broad term, including all distillates
heavier than illuminating oils and lighter than
lubricating oils; it includes so-called gas oil —
a high-grade fuel oil used in the manufacture
of gas — as well as fuel oil proper, used largely
for steam raising. The term lubricating oil
includes a variety of heavy oils used for
lubricating purposes. Most of these products
in turn may be broken up into other sub-
stances, each the starting point of further
refinements. Under present practice petro-
leum yields only a few hundred substances of
commercial value, but the mind can set
absolutely no limit to the number of useful
materials that chemical research ma}- still
wrest from this raw material.
While refinery practice is a highly technical
matter and varies both according to the
chemical nature of the oil and the local
demand for products, we may, for the sake of
simplicity, ignore all details' and note merely
that there are three main types of refineries.
The first of these is called a "skimming" or
"topping" plant, because the light oils,
gasolene and kerosene, are removed from the
rest of the products, which are left behind as
a residual oil and sold in this semicrude
condition for fuel purposes. The "skim-
ming" plant, as its name implies, make an
incomplete recovery of products, suppying
only those in greatest demand or easiest to
make ; most of the plants of this kind are
situated west of the Mississippi River.
The second type of refinery may be termed
the "straight-run" plant; this produces all
four of the main products — gasolene, kero-
sene, fuel oil, and lubricating oil — together
with by-products, the process separating the
crude oil into its natural components with the
minimum of chemical change. The " straight-
nm" refinery lacks flexibility, because it has
no power of producing, for example, more
gasolene than the crude oil naturally contains.
Such plants are situated in the East and other
parts of the country where the demand,
especially for lubricants, justifies the expense
of the practice.
The third type of refinery is of recent birth,
but has made rapid strides toward a great
future; it employs the so-called "cracking"
process, which yields, like the "straight-run"
plant, a full set of products, but a greater
percentage of gasolene than the crude oil gives
upon ordinary distillation. This is accom-
plished at the expense of the heavier com-
ponent oils, whose molecules are broken or
"cracked" into lighter molecules, which
constitute just so much additional gasolene.
20G Alarch, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 3
It is obvious that cracking has developed in
response to a growing demand for gasolene ; its
significance is apparent in the fact that it per-
mits the production of a more valuable product
from one less valuable. With an increasing
call for gasolene and a decreasing supply of
petroleum, cracking may be called the hope
of the future as regards refinery advance.
If we pause for a moment to contemplate
the consumption of petroleum in the crude
condition, and then the three types of refin-
ing— skimming, straight-run, and cracking —
it becomes evident that each treatment
represents a step in advance over the preced-
ing, and that, while all four prevail toda}', the
cracking refinery is in line with true progress
and will eventualh' dominate the situation.
Refineries, whatever the type, employ the
principle of distillation in their operations.
The petroleum is heated in stills and the prod-
ucts vaporize, pass off, and are condensed in
fractions, representing roughly the materials
in demand. These products are then purified
by chemical treatment or transformed by
chemical means into a series of secondary-
products. The production of the various
kinds of lubricating oils needed for diverse
uses represents an intricate, yet single, part of
petroletun refining; and is merely one aspect of
the many ramifications found in refinery
technique. The refining of petroleum makes
heavy drafts upon other chemical industries —
for example, in normal times, about one tenth
of the sulphuric acid produced in the United
States goes into petroleum refining — but the
refiner},' in turn contributes many essential
products to other chemical manufacturing
activities. These industrial interrelation-
ships, ofttimes overlooked, are of the utmost
significance — a fact strikingly brought out
when one activity is called upon to expand
more rapidly than some other activity with
which it is geared.
The refining of petroleum, requiring elab-
orate plants, is by nature a large-scale
enterprise; hence such activities in the main
have naturally come under the control of a
few large organizations. While several hun-
dred individual refineries are in operation, the
bulk of the output is due to the efforts of less
than 10 companies. The refining of petro-
leum, therefore, is largely an integrated
activity, in close alliance with transportation
of crude, on the one hand, and distribution of-
refined products on the other. It has already
been pointed out that the development of
pipe-line transportation has permitted the
establishment of refineries at points distant
from oil fields, but convenient to centers of
consumption and to seaports.
With the broad outlines of refinen- tech-
nique in mind, it will be of interest to obser\-e
the shifting focus of development that has
characterized the production of petroleum
products in America. When the famous
Drake well struck oil on Oil Creek, Pa., in
1S59, an illuminating oil distilled from coal
and called "coal oil" was in general use
throughout the countn.-. Petroleum, there-
fore, found a market already established for
its illuminating constituent, which it usurped
at once, quickly supplanting the coal-oil
industr>- with a production of kerosene.
Although other products were also produced,
and lubricating oils made from petroleum
found quick favor in connection with a grow-
ing application of mechanical energy, kerosene
became the chief petroleum product and for
over 40 years its use expanded until this
illimiinant penetrated literally to the utter-
most comers of the globe. It would be
difficult, indeed, to estimate the value to the
world at large of this cheap and convenient
source of light, which has been aptly termed
"one of the greatest of all modem agents of
civilization." During this period there was
little demand for the light products of distil-
lation, the liquids now sold under the com-
merical name of gasolene, which were, there-
fore, largely waste products in an economic
sense, and even in some instances physically
destroyed for want of any adequate demand
for their utilization. Gasolene for a long time,
then, was a by-product of little value turned
out in the manufacture of kerosene.
Toward the close of the nmeteenth centun,-,
however, the commerical application of the
incandescent mantle in gas lighting and the
development of the electric light introduced
types of illumination so superior to the kero-
sene lamp in convenience that the use of the
latter was gradually relegated, in large part.
to the small town, the countr>-, and foreign
regions, where gas and electricity had not
been introduced. Accordingly, in spite of a
most aggressi\-c campaign for foreign trade on
the part of the petroleum industry, the
refinen,- faced the restrictions of a slowing
demand for kerosene which presaged a limit
to the output of the whole set of petroleum
products. But the menace of this limiting
circumstance was destroyed, before it became
effective, by the introduction and rapid
advance of the intcmal-combustion engine.
The phenomenal growth in the use of the
automobile built up such a heavy demand for
METHODS FOR MORE EFFICIENTLY UTILIZING OUR FUEL RESOURCES 207
gasolene that this product came into the lead
and took up the burden of justifying the
increasing refinery consumption of crude
petroleum — a burden which kerosene, even
with the aid of a growing market for fuel oil,
lubricants, and other oil products, was
scarcely longer able to sustain. Gasolene, now,
is the main prop to the whole cost structure
of petroleum refining.
With the industrial quickening due to the
entrance of the United States into the world
war, the demand for fuel oil became so insistent
that the complexion of the oil situation again
changed and the emphasis fell upon fuel oil.
And as the production of crude petroleum was
not able to keep pace with the attempted con-
sumption of fuel oil, a serious shortage of
this product resulted ; even while the supplies
of gasolene were ample to maintain the
activities of war, business, and pleasure.
If the course of development, as indicated
by this broad survey of refinery evolution, be
projected into the future, we may foresee a
time when the petroleum industry will yield
a range of fuels for the internal combustion
engine only ; illuminating kerosene in quantity
narrowing to that desirable for country use
and export trade; lubricating oils adjusted
to the growing demands of mechanical power
and an ever-widening range of chemical prod-
ucts supporting a great oil by-products
industry, rivalling if not exceeding the coal-
products industry in importance. In respect
to the last, it should be emphasized that the
United States today faces an opportunity
similar to that which 20 years ago confronted
both Germany and the LTnited States as
regards the manufacture of dyestuft's, explo-
sives, fertilizers, drugs, and other chemicals
from the non-fuel components of coal.
Distribution
Many industries terminate their activities
with the manufacture of commercial prod-
ucts, turning these over to independent
agencies for distribution. With the petroleum
industr^^ however, distribution forms an
integral division of the industrial activity, a
carefully planned out construction of markets
as part of the resource development being
substituted for a demand ordinarily left to
natural growth or maintained by costly
advertising. Thus, once the oil is produced,
it passes through the various stages of trans-
portation, refining, and distribution under the
influence of a highh' organized economic
machine, a coordinated industrial unit, en-
gaged not merely in adapting a crude material
to diverse uses, but also in shaping and
developing latent needs the world over into a
demand which will sustain a balanced output
of products.
We have already seen how the pipe line
and to a less extent the coastwise tanker,
brings the crude petroleum to the refineries
which are favorably located in respect to
distribution. From the refineries the gasolene,
kerosene, fuel oil, lubricating oil, and other
petroleum products are sent forth to supply
the needs of surrounding territory, while
refineries near seaboard furnish heavy con-
tributions to foreign trade. As distribution
is a diverging process, and, moreover, the
crude petroleum is broken into numerous
products requiring separate handling, the
pipe line is not broadly adapted to this diverse
haulage. Railroad tank cars, and barges (where
water transportation is advantageously avail-
able), therefore, receive the bulkier products
and carry them to distributing depots, where
storage tanks release the railroad carriers and
supply tank wagons that radiate to fill the
local needs. In this way the entire country is
covered by a network of specialized trans-
portation, each step employing a bulk carrier
best adapted to its particular purpose both
as to size and mechanical facility, the whole
involving the maximtun of expedition and
simplicity. Without this highly organized
system, with its far-reaching ramifications, the
present widespread use of gasolene and kero-
sene would not be possible. From the oil
field to the consumer, the handling of petro-
leum is remarkably efficient.
The arrangements whereby a foreign trade
has been built up and sustained are no less
elaborate. Fleets of tank steamers and
freighters carry the products in bulk or in
suitable containers to all parts of the world.
Fuel oil, gasolene, and lubricants go in greater
measure to industrial countries, but kerosene
penetrates to every corner of the globe, a
system of depots and distributing lines adapt-
ing the product to the needs of the most out-
of-the-way regions. The care that has been
bestowed upon the extensions of the market
for kerosene, against every conceivable obsta-
cle of climate, topography, and racial prejvi-
diee, is a striking example of industrial fore-
sight; yet without this policy, the whole oil
industry would have been unable to expand to
its present proportions.
208 March, 1920
GENER,\L ELECTRIC REVIEW
Vol. XXIII, No. 3
Professor Elihu Thomson's Early Experimental
Discovery of the Maxwell Electro-
Magnetic Waves
By Prof. Monroe B. Snyder
Philadelphi.\ Observ.\tory
It must be borne in mind that this very early and very remarkable investigatioa of the electro-magnetic
waves by Professor EHhu Thomson was quite incidental to an investigation intended to set aside a claim then
made by a famous inventor for the existence of an alleged Wetheric force. It is not possible here to reproduce
the relentless logic of description of special experiments made to prove the fallacy of the claim. But it is very
clear how Professor Thomson was led by his mode of testing for "induction" effects to the far more extensive
testing for the aether waves produced, and thus to the tests which Professor Snyder has so definitely described
as performed by his former colleague. It was indeed a misfortune for American science, as Professor Snyder
indicates, that Professor Thomson could not then continue his investigations in the incidental field. And this
very clearly appears from the ingenious use then made in the Thomson experiments, as described in the Frank-
lin Insiitute Journal cited, of "balanced circuits" and of other devices thoughtfully dealing with the Kther
waves concerned. It is indeed gratifying that these notable tests of the electro-magnetic waves by Professor
Elihu Thomson in 1875 have now been so specifically and reliably placed on record by one appreciating their
significance. The story was originally published in the Central High School Mirror, Philadelphia. — Editor.
In the annual lectures to my classes in
astronomy on the vast electro-maj^netic
spectrum of radiation of more than 50 gamuts,
and in which the interesting visible light
occupies but a single gamut, I have again
and again been reminded of the fact that my
former colleague, Professor Elihu Thomson,
had already in 1875 experimentally dis-
covered the long electro-magnetic waves first
announced in the mathematical theory of Clerk
Maxwell in 1873, and later concretely revealed
by the experimental work of Hertz in 18S7.
One day in 1875, while busily engaged in
some work in the old Central High School
Obsen-atory, at an elevatorless height that
usually obviated intrusion, I was surprised
by a bustling visit from my associate. Pro-
fessor Elihu Thomson. Ho was bent, as I
soon found, on testing whether the aether
disturbance, which he was exciting by means
of a Ruhmkorff coil in the Physical Room of
the first floor of the building, could be ob-
served in the obscrvaton.- hallway on the
sixth floor. Applying the sharpened point
of a short lead pencil near the brass knob of
the obser\-atory library door, Thomson called
attention to the delicate sparks that were
passing between the pencil point and the
door knob. With due elation over the success
of the test, he then told me that he had
similarly traced the a?ther disturbance all
through the building; in the Lecture Hall on
the first floor at a distance of about (50 feet;
at the room of the professor of mathematics
on the third floor at a distance of about SO
feet; and now at the door knob of the observ-
atory library, distant perhaps over 100 feet
from the experimental apparatus.
It is interesting to know what odd electric
radiating system was at that time kept in
action in the Physics Room. In an effort
to magnify the electrical oscillations, then
being studied for another purpose, Thomson
had connected one terminal of that famous
induction coil to the water pipe and the other
terminal to a large metallic still, which stood
at hand, and which he duly insulated by
placing it on a glass jar. \'igorous sparks of
a few inches were then passed, and the unique
radiating system produced results that soon
induced the professor to widen the area of
his obsen'ations in the manner mentioned.
The invisible long electro-magnetic waves
were thus definitely and repeatedly traced
by Professor Elihu Thomson in 1875 to the
distances and cft'ects stated, and through five
intervening floors, by a means much simpler
than the detector of Hertz, and yet 12 years
prior to Hertz's celebrated verification of the
Maxwell theory. The insight and the accu-
racy of the conception of Professor Thomson
as to what was really happening, so clearly
reflected in an article on the same experi-
ments undertaken for the purpose of correct-
ing a misconception of Edison's {Journal
of Franklin Institute. April, lS7b'), show how
unfortunate it was that Professor Thomson
was then diverted from a continuance of the
study of those aether waNcs, observed so
many years before their elucidation by Hertz.
Elihu Thomson's demonstration of 1875
seems, be_\-ond doubt, to have been the very
first discovery, by means of repeated tests
at large and varying distances, of the trans-
mission of the invisible electro-magnetic
waves through the aether, the first experi-
mental discovery of what are generally known
as the Hertz- Maxwell waves, now so widely
and triumphantly used in wireless telegraphy
and wireless telephony.
209
Effect of Color of Walls and Ceilings on
Resultant Illumination
By A. L. Powell
Edison Lamp Works, General Electric Company
The color of walls and ceilings plays a very important part in the illumination of interiors, and architects
and others who are responsible for systems of interior illumination should make a special study of the reflect-
ing powers of different colored walls and ceilings with special reference to the qualities of paints and other
pigments. The author outlines briefly some color schemes for walls, ceilings and fixtures that have been
found to give good results in industrial plants, offices, schools, stores, and residences. Some valuable infor-
mation is given on the reflecting power of several different kinds of paint, both when new and after ageing
one year. Some helpful suggestions as to the best methods of applying paint to secure satisfactory reflection
are also ofifered, and a method is described for determining the coefficient of reflection of any surface. — Editor.
No matter how carefully designed a lighting
system may be with respect to lamps, reflec-
tors, spacing, height, etc., if the surroundings
are not adapted to reflecting such light as
strikes them an inefhcient system may result.
The ]jroper painting of walls and ceilings is
tlierefore of great importance.
The ceiling and wall surfaces in a room are
secondary sources of light — receiving and
reflecting light from the lamps. ]\Ierely
increasing the reflection coefficient of the
ceiling a slight amount may greatly increase
the effective ilhunination.
It is therefore very important to see that
the ceilings are as light in color as possible.
Pure white is usually to be preferred, although
if a tint is demanded for artistic eft'ects it
should be a light cream rather than gray or
some similar tone. Not only is the color of
the ceiling important, but the actual finish
must also be considered. A glossy surface
reflects images of the lamp filament and intro-
duces glare causing eye strain. A flat or
matt finish is therefore essential. A thin
coating of white paint through which a dark
surface may be seen has the same eff'ect as a
thin coating of enamel on a reflector. In
other words, utiless this surface is thick the
light gets through the surface and becomes
absorbed.
It is safe to say that well painted white
ceilings give an increase of between 20 and
30 per ceiit in illumination over ordinary
light buff or similar colored ceilings where
semi-indirect or similar lighting systems are
in use. This is really a conservative figure.
Industrial Plants
Efficiency of utilizatioii of light is highly
important in the industrial plant, and pillars,
walls and ceilings should be pure white. Any
light striking these surfaces is reflected in a
degree depending upon the color. If dark
brown or smoke covered, possible' only 5
per cent will be reflected; if pure white the
reflection coeflJicient may be as high as 70
per cent. Even the floors should be kept as
light as possible, for a portion of the flux
which strikes this is reflected to the ceiling
and then back to the work.
A recent test in a new factory building
with white ceiling, light wood floor and light
colored side walls showed more units of light
reaching the working planes than were gen-
erated by the lamps themselves. This para-
dox is explained by a consideration of the
multiple reflection. Of course the extremely
high ^•alue would not have been obtained if
machinery were installed.
The lower part of the side walls is of less
importance in reflecting light, and for pur-
poses of appearance it is often desirable to
have a dado of dark green or some neutral
color, as fingermarks and other disfigurements
are not so noticeable. This treatment of the
walls also reduces the brightness of the back-
ground in the field of ^-iew — a desirable
feature.
In many instances the painting of certain
parts of a machine white or a lighter color
will materially soften the shadows and
improve working conditions, eg., on a large
\-ertical slotter the surface which faces the
table, or on a lathe the area surrounding the
work.
High grade oil painting, as discussed later,
is most desirable, but where whitewash is
absolutely necessary frequent cleanings will
speed production and keep the lighting bills
at a minimum. A clean, bright shop has a
decided effect in itnproving the morale of the
workmen.
Offices and Schools
Because of the justh' wide-spread use of
the indirect lighting systems and the likeli-
hood that they will be installed at any time,
the ceilings should always be light in color.
210 March, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 3
With most systems of lighting a consider-
able portion of the flux strikes the upper part
of the walls; for these surfaces, a soft pale
olive green with a light blue cast in north
rooms and a yellow cast in south rooms is
recommended. However, this question of
wall tint is largely a matter of personal
preference. Some individuals prefer a green-
ish tint which is soft and restful, while others,
for artistic reasons, prefer a light buff or
cream. It is recognized that it is often worth
while to sacrifice lighting economy for artistic
effect. The lower surfaces can well be of a
darker neutral color to provide space on which
the eye can rest in comfort.
A light-colored room is decidedly more
cheerful than one finished in dark colors. In
many cases dark surroundings have given the
impression of bad lighting, while in reality
there was a sufficiently high intensity on
the desks. The psychological effect of gloomy
interiors is well known, and it is, of course,
desirable to keep the clerks or pupils buoyant
and cheerful.
In general, light surroundings reduce the
conditions of glare. An artificial light source
viewed against a bright ceiling is less annoy-
ing than in another jjosition. Light-colored
walls diffuse the light back toward the window
sides of the room and thus lessen the contrast
between the bright sky and adjacent walls.
As has been mentioned before, glossy wall
surfaces should not be used, and even the
furniture and trim should not be highly
varnished. In this connection close cooper-
ation between the builder and lighting engi-
neer is essential.
Light buff window shades are desirable,
and if these are drawn at night they materially
assist in reflecting the light rather than allow-
ing it to escape to the street. If these shades
are slightly translucent they are very useful
in the daytime in cutting down the direct
sunlight, diffusing the light which ])asses
through them and ijreventing a sharp line
of shadow demarkation which may result
if opaque shades are used. The Code Light-
ing School Building issued by the Illuminat-
ing Engineering Society gives some interesting
data on this subject as well as the question
of design of blackboards.
Stores
A pure white finish throughout is most
universally applicable to stores, not only for
its effect on the amount of light utilized and
general bright appearance desired, but for
the color result secured.
White light striking a colored surface will
have some of its rays absorbed and be re-
flected as colored rather than white light.
(This property is that which makes the sur-
face colored.) Hence, if an illuminant
approximating daylight is used, and all of
the reflected light is tinted, the resulting
light will be of a different color from that
given out by the lamp. A practical illustra-
tion of this undesirable condition will be
seen where daylight lamps are used in semi-
indirect units in a room with a yellow ceiling.
W^hite surroundings do not modify the color of
the reflected light as do colored surroundings.
Residences
It is true that efficiency of light utilization
is not at all important in the home, yet the
color of walls and ceilings, particularly the
former, has a remarkable bearing on the
pleasing appearance of the room.
If these are of such colors that they do not
reflect the light satisfactorily, no matter how
much light is supplied the room will never
appear bright and cheerful. Dark green wall
paper, for exami)le, reflects very little light,
and a room finished in this way frequently is
dull. A room finished in dec]) brown wood-
work and side walls is often uncomfortable
when lighted by ordinary' methods of illumi-
nation. With general lighting systems no
matter how much ])recaution is taken to
shield and diffuse the light, the lamp and its
accessories show up in contrast to the dark
background and become annoying bright
spots. The only satisfactor\- method of
lighting such an interior is by the use of table
or floor lam])s giving spots of fairly bright
illumination, around which the occupants are
grouped and the attention concentrated,
allowing the room as a whole to be comjiara-
tivcly dark or in shadow.
Light-colored wall paper and paint are
therefore generally to be desired if the room
is to be cheerful at night. An object or an
interior looks cheerful and bright in pro-
portion to the amount of light it reflects back
to the eye. although jmre white finishes are
not to be desired from an artistic standpoint.
The study of the ixsychological effect of the
different colors is most interesting, but space
does not permit a discussion of this phase of
the subject.
Permanency of Various Wall Finishes
The freshly scraped surface of a block of
magnesium carbonate reflects more light
than any other object, SS per cent of the light
I
COLOR OF WALLS AND CEILINGS ON RESULTANT ILLUMINATION 211
falling on it being sent back. We cannot
expect as good results from ordinary painted
surfaces, because the usual mediums, includ-
ing eA-en zinc white, are quite gray compared
with magnesium carbonate. A paint made
with magnesium, carbonate as a pigment
more nearly approaches this value and is
desirable from a standpoint of light reflection.
As comparative average values for properly
prepared and freshly mixed samples, the
following figures apply :
Coefficient of Reflection
Paint
New
After Ageing
One Year
White lead and oil
Lithopone . .
0.85
0.77
0.67
0.72
Calcimine type
Flat enamel (magnesia
0.74
0.76
0.75
0.67
0.73
Gloss enamel
0.75
There is comparatively little choice between
any good white paints when fresh. The story
is different, however, after being exposed to
normal daylight conditions.
It is seen from this table that the enamels
have held their own very well. The lithopone
paint has fallen off by 6 per cent of its initial
value. Calcimine and white lead have fallen
off about 10 per cent. The falling oft" of
calcimine is due largely to its porous nature,
which permits it to absorb dirt readily. The
falling off in white lead and calcimine is
progressive and does not decrease in rate.
Numerous observations on lead and oil paint
in use for two years indicate a falling off of
about 20 per cent. The slight falling off of
flat enamel occurred in the first month, no
further decrease being observed. The co-
efficient of reflection of the gloss enamel was
constant throughout the test.
These tests were all made under constant
laboratory conditions and must serve only
as a guide for judgment. They form a start-
ing ]3oint for observation and practice.
Method of Applying Paint
Now as to the actual painting itself:
What conclusions do we draw from the data
presented? It is obvious that a gloss enamel
will not ftilfill one of our initial conditions,
due to its high value of specular or image
reflection. We are therefore reduced to the
use of some form of what we have called flat
enamel. This paint must contain no lead
and probably no linseed oil. It must be
composed of chemically inert white sub-
stances ground exceedingly fine (to produce
density) and mixed in an inert vehicle which
is impervious and non-porous when dry.
It must dry flat and be washable.
The most permanent and highest practical
coefficient of reflection and diffusion can be
obtained with plaster svirfaces treated as
follows :
First coat. — Good impervious surface.
Second coat.- -Straight lithopone paint.
Third coat. — Gloss enamel and lithopone mixed
equal parts.
Fourth coat. — Flat enamel (magnesium bearing
flowed on).
For metal surfaces, after the iistial prepara-
tion apply a first coat of red lead thinned with
raw linseed oil drier and "turps" to give an
eggshell finish. Over this a coat of lithopone
paint, mixed one gallon to one quart of good
varnish; then the second, third and fourth
coats as applied to plaster.
From an illuminating standpoint the walls
of a room are not as important as the ceilings,
and they should be less bright. A simple
ijainting formula will apply. It is in brief:
First coat, good imj^ervious surfacer mixed
with equal part lithopone paint; second and
third coats, straight lithopone paint, the last
tinted with japan tint thinned with "turps."
If it is necessary for any reason to use a
gray tint it should never be obtained by mix-
ing lamp black in the paint. This is the
substance having the lowest known coeffi-
cient of reflection. To obtain the gray it is
desirable to mix vermilion and emerald green
to get black and then thin out with white.
This produces what is known as a warm gray
and has a reasonably high coefiftcient of
reflection.
In any painting, the surface on which it is
applied should be properly prepared and non-
porous so that it will not absorb any of the
vehicle of the final coat. It must also be
chemically inert with respect to this final coat.
A number of paint manufacturers have
investigated the subject of "painting for
light" and have produced pigments which
give results comparable with those specified
above. Any of the prominent paint manu-
facturers will gladly furnish detailed infor-
mation on their jirodtict upon request.
Economics of Situation
It is true that it is somewhat more expen-
sive to- paint the surroundings correctly than
to apply calcimine, mill white, or some other
paint which depreciates quite rapidly, yet
the economies of the situation well warrant
212 :\Iarch, l'J2U
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 3
this expenditure. For example, the following
calculation applies:
If we consider a room in which the ceiling
is painted with white lead and oil, as de-
scribed in the test quoted above, we may
expect at the end of two years that the
illiunination efficiency has decreased not
less than 15 to 20 per cent, due alone to the
reduction in coefficient of reflection of the
ceiling, other conditions being constant. But
in actual work other conditions are not con-
stant; for one thing the coefficient of reflec-
tion of the paint on the wall also undergoes
a decrease. It is probably safe to say that
as Affecting Illumination," for many of the
figures presented in connection with painting.
Measurement of Reflection Factor
There are a number of laboratory- methods
of obtaining this value, some of which
employ elaborate apparatus and which take
into account with a high degree of accuracy
the direction of the incident light, color of
incident light, and similar features. A com-
plete description of these methods will be
found in the technical press, references to
which are given in the bibliography which
follows the article.
Fig. 1. Determining by Means of Portable Photometer the Reflection Factor of a Wall Surface
the use of load and oil (or calcimine) as an
interior paint entails a progressive loss of
light amounting to lo per cent at the end
of the first year. Thus a room of 4U() sq. ft.
floor area, so painted and initially lighted by
four 100-watt lamps, will require an addi-
tional 100-watt lamp at the end of two years
to bring the illumination back to what it
was in the beginning — an increase of 25 per
cent in energy consumed and lamp renewals.
Surely this figure is striking enough to
warrant the necessar}' expenditure. The
writer is indebted to an article by Mr.
Bassett Jones. Consulting Engineer, on "The
Characteristics of Interior Building Finishes
The practical determination of the co-
efficient of reflection of a wall or ceiling
(diffuse reflection) is quite simple indeed, and
can be made by anyone familiar with the
operation of a jjortable photometer employ-
ing a detached test plate. The standard on
which reflection factors are ba.sed is a freshly
scraped block of pure magnesium carbonate.
One of these standards can be s<.»cured at
any drug store. A block approximately four
inches square and two inches thick can be
purchased for a few cents. The first step in
the determination is to scrape the surface
of this and place the block in any convenient
position relative to an artificial light source.
COLOR OF WALLS AND CEILINGS ON RESULTANT ILLUMINATION 21;:
The ]3hotometer is then pointed at the block
from some angle not too far from the normal
and a reading taken and recorded. A second-
ary standard, or working standard, such as
a sheet of blotting paper, is next calibrated.
This is substituted for the magnesium block,
and with the same illumination incident on
it as with the previous reading, a second read-
ing is taken and recorded. We then have the
following proportion ap]3lying: Reading ,4
is to 88 per cent as reading B is to the co-
efficient of reflection of the blotting paper.
Taking care that the blotting paper or
secondary standard does not become dirty,
it is placed at a convenient position on the
wall or ceiling the reflecting factor of which
is desired, and a reading taken of the blotting
paper with the nonnal illumination received
on the wall incident on the paper. The paper
is now removed and a reading taken of the
wall surface. We have already determined
the coefficient of reflection of the blotting
paper, and the following proportion applies:
Reading on blotting paper is to the coeffi-
cient of reflection of blotting paper as reading
on wall is to coefficient of reflection of wall.
If the surface to be tested is polished or has
a considerable element of specular reflection,
the detemiination of the coefficient is more
complex, and several readings at different
angles should be taken to insure a fair aver-
age value.
No difficult mathematical equations are
involved in this determination. Simple
readings of the photometer and a proportion
are all that is necessary. Example calibra-
tion: Magnesium carbonate block, apparent
foot-candles 10.5; white blotting paper, appar-
ent foot-candles 9.1. Then.
1CK5_!2_1
O.SS~ X
Coefficient of reflection of blotting paper is
7G per cent.
Test of wall surface: Apparent foot-candles,
white blotting paper in place, 3.7; apparent
foot-candles of wall, blotting paper removed,
2.6. Then
3J^2X)
.76 X
Coefficient of reflection of wall is therefore
53 per cent.
Coefficient of Reflection (Reflection Factors)
The following table indicates in general the
amount of light reflected by the different
colors. It will be noted that there is a con-
siderable variation in percentage for any
particular color. This is necessary, as we do
not have any means of specifying the exact
shade or tint of the various colors. The
figures presented are the result of a consider-
able number of tests by different authorities
and are representative average values. We
believe them to be fairly typical within
reasonable limits.
Color
White — new
White— old
Cream
Buff
Ivory
Gray
Light green
Dark green
Light blue
Pink
Dark red
Yellow
Dark tan
Natural wood brown stain
Light wood varnish
Percentage of
Light Reflected
74 to 80
67 to 76
56 to 72
44 to 59
66 to 70
15 to 57*
4:5 to 67
10 to 22
31 to 55
32 to 55
12 to 27
55 to 67
27 to 41
15 to 26
38 to 44
* Grays vary remarkably, depending on the way they are
prepared. A gray made by mixing lamp black with white paint
has a low co-efficient of reflection. A gray made by mixing red
and green paint with white base has a relatively high co-efii-
cicnt of reflection. It is known as a warm gray.
BIBLIOGRAPHY
Effects of Reflection from Floors. J. R. Cravath, Electrical
World, October 28. 1911.
Reflection Co-efficients. Paul Bouder, I. E. S. Transactions,
Vol. 6. p. 85.
Determining the Reflecting Power of Opaque Bodies. P. G.
Nutting, I. E. S. Transactions. Vol. 7, p. 412.
Reflecting Properties of Painted Interior Walls. C. W.
Jordan, I. E. S. Transactions, Vol. 7. p. 529.
Influence of Colored Surroundings on the Color of Useful
Light. M. L. Luckiesh, I. E. S. Transactions. Vol. 8, p. 62.
Reflection from Painted Surfaces. Louis Bell, Electrical
World, Jan. 22, 1915.
Air Shaft Illumination as Studied by Models. C. H. Sharp,
I. E. S. Transactions. Vol. 9, p. 598.
Report of Committee on Glare — Diffusing Media VI Interior
Furnishings, I. E. S. Transactions, Vol. 10, p. 397.
The Light Reflecting Values of White and Colored Paints.
W. S. Gardner, Journal Franklin Institute, January, 1916.
Effect of Interior Colors and Finishes Upon the Lighting of
Rooms. S. G. Hibben, Electric Journal. July. 1916.
Measurement of Reflection Factors. M. Luckiesh. Electrical
World. May 19, 1917.
Effect of Wall Colors on Lighting Requirements. Electrical
World, Aug. 16. 1919.
214 March, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 3
Short-circuit Tests on a 10,000-kv-a.
Turbine Alternator
By E. S. Hexningse.x
Alternating-current Engineering Department, General Electric Company
The series of tests described in this article were made for the purpose of rounding out the data that has
been compiled on the behavior of alternators under short circuit. The performance of definite pole machines
under short circuit had previously been analyzed by means of extensive tests, but the characteristics of the
alternator with smooth core rotor were not so well known. The tests were carefully conducted under numerous
conditions of short circuit, with various arrangements of reactors in circuit and with no reactors in circuit. Special
precaution was taken to eliminate errors and meter readings were taken as a check on oscillograph records. The
results of tests are shown in tabular form and by means of curves plotted from the table values. The deduc-
tions that may be drawn from this series of tests are stated in a series of concluding paragraphs. — Editor.
The question of predicting the amount of
current that will flow when an alternator is
short circuited under various conditions has
been widely discussed, partictilarly in the
past few years. Although not a new problem
at all, its importance has increased ■udth the
increase in size of central stations and trans-
mission systems. While the theory under-
h'ing the phenomena of short circuits is now
well established, the weight to be given
certain factors in the theory can be deter-
mined only by actual tests on a variety of
types of machines under different conditions
of load, voltage, etc. As there seems to be
phase synchronous impedance 87.5 per cent.
For the tests with external reactance in the
circuit, standard current I'miting reactors
were used, Fig. 1. These were wound with
270 turns of copper wire in eighteen layers
of fifteen turns per layer. The resistance of
each reactor was 0.292 ohms and the ohmic
impedance as obtained from an average of
some fifty volt-ampere readings was 4.96
ohms, which corresponds to 49.6 per cent
reactance three-phase and 28.7 per cent
single-phase on the basis of the generator.
Three oscillographs were used in the test:
one to record the three currents, another the
\B^^^^,.3 ^.#^%A5_<
t
Fig. 1. Group of Current Limiting Reactors Arranged for Three-phase Operation
less accurate data on large turbine alternators
than on definite pole machines, an elaborate
series of short-circuit tests were made some
time ago on a 10,00()-kv-a., lO.OOO-volt,
2400-r.p.m. turbine-driven altcrnalor installed
in the power house of the Schenectady Works
of the General Electric Company. The
results of this test are briefly described in
this article.
The armature leakage reactance of this
generator as calculated from saturation and
synchronous impedance curves was 12..) per
cent. The three-phase synchronous imped-
ance from test was 130.2 per cent and single-
three voltages and the third the field current
and voltage and the voltage across one ex-
ternal reactor. The voltages were read from
the secondaries of potential transformers,
but the currents were read by means of direct-
current shunts instead of current trans-
formers, so that no distortion or inductance
due to transformers would be recorded on
the films.
Saturation and synchronous impedance
tests were made and are recorded in Fig. 2.
The short -circuit tests were made as follows:
Starting with initial conditions of 5000
volts, 10,000 volts and 12.000 volts oiwn
SHORT-CIRCUIT TESTS ON A 10,000-KV-A. TURBINE ALTERNATOR 215
circuit, the generator was short-circuited
three-phase with, first, one reactor in each
leg; second, two reactors in parallel in each
leg; third, three reactors in parallel in each
leg, and fourth, four reactors in parallel in
each leg. Fig. 3 shows the connection
employed for one reactor in each leg.
These same tests were then repeated for
single-phase short circuits at the same volt-
ages and various numbers of reactors in the
line. Fig. 4 shows the connections for the
one reactor test.
An automatic voltage regulator was then
wired in and the 10. 000- volt condition for
all the foregoing tests was repeated. These
tests represent as nearly as possible the actual
operating conditions when the generator is
being regulated. On these tests records
were also obtained showing how rapidly the
regulator opened and closed and how soon
after the short circuit came on the regulator
contacts closed to put full field on the exciter.
Short circuits were made under 10,000 and
12,000 volts full load zero power-factor, the
load being obtained by means of reactors in
series. These tests were made with and
without external reactance. Figs. 5 and 6.
The generator was dead short-circuited at
the three voltages with zero reactance in the
MOOO
13000
12000
11000
10000
^ 9000
■^^ sooo
7000
<n 6000
§ 5000
4000
3000
ZOOO
1000
/
^
'
/
y
1
1 ' '
I'l ■
/
1
U
V*..! .
/
/
/
1 ,
'A'
' X !■■
' AirCnn^
A^n
1
1
1
'7
f
1
ji' 7^-
n'
A
H
?
/
n^
.■1
U
c
/ ,
,1
h
^1
/
'/
1 .'"
#'
J
/
1
r
■I-
,\f
.■^
/
//
/
'
l~^
^
<■/
/
'/
/
trw
.o"/
V/^
/
J/>
'//
/'
/
■^/-^
7
/
/
w
y//
/
^ /
/
'. /
/\
1
//
//
///>
/
V
^f
j/A
i
\
^
1
1000
900
800
700 \
600\
500 I
400 ;
300 I
s
^oc^'
lOO
O 100 ZOO 300 4O0
Arriperea Field
Fig. 2. Saturation and Synchronous Impedance Curve on
10,000-kv-a., 10-000-volt Alternating Current Generator,
with and without current limiting reactor in circuit
line and also dead short circuits were thrown
on when the machine was carrying practi-
cally full load unity power-factor. Figs. 5
and 6.
The oscillograph films were run at a speed
that would allow a record of the phenomena
from approximately one second before the
short circuit came on to about eight seconds
afterward. Meter readings were also taken
as a check on the oscillograph records.
More than 1.50 oscillograms were taken
and therefore it is not practical to include
Reactor ,
vwv — '
30 5C
With Reactors
Fig. 3. Three-phase Short
Circuit on A-C. Gener-
ators with Reactors
in Each Phase
Reactor ,
^/W — '
10SC-1 Reactor
Fig. 4. Single-phase Short
Circuit with One Re-
actor in Circuit
copies of the films in this article. Figs. 7
and 8, however, are typical of the current
records. Fig. S is particularly interesting in
that it shows the operation of the automatic
voltage regulator contacts controlling the
current in the exciter field. One vibrator of
the oscillograph was connected in series with
the regulator contact for the purpose of
observing how quickly, after the short cir-
cuit came on, the contacts would close. As
shown, the action is very rapid.
It was not considered necessary to take
into account the resistance of the armature
^
To Zero or Un/'ti/
Power- Factor Load
3(SSC- Under Load
Fig. 5. Three-phase'Short Circuit, No Reactors
- To Zero Unity
Power-Factor L oad
J0 5(: {With Reactors)
Under Full Load
Fig. 6. Three-phase Short Circuit with Reactor in Each Phase
circuits in calculating the results because
from test the reactance and impedance were
practically identical. Likewise the spacing
of the reactors (approximately four feet
between centers) was considered sufficient
to disregard any eftect of mutual inductance
between them.
Measurements were made of the current
films on the basis of a symmetrical current.
Lines were drawn through the tops and
bottoms of the current waves and the instan-
216 March, 1920
GENER.\L ELECTRIC REVIEW
Vol. XXIII. No. -.i
taneous values obtained by measuring the
distance between these lines, and dividing
by two times the square root of two to obtain
the effective symmetrical current.* Readings
were taken at the following points: the
instant the short circuit occurred, which
Fig. 7. OsciUogram of 3-phase Short Circuit.
Reactor in each phase
assumes that the current could rise instantly
and is the value used as the instantaneous
short circuit current; the first peak; second
peak; fourth, eighth, twentieth, fortieth and
eightieth cycles; the point where the current
waves become symmetrical; and the sus-
tained value. The duration of the armature
and field transients, and the time from zero
to the first peak were also recorded. The
duration of the armature transient was also
Particular attention was given to the
elimination of errors and all precautions
were taken to obtain as consistent results as
possible. Meter readings were taken as a
check on the oscillograms just before the
short circuit was thrown on, and as soon after
as conditions again were steady. The meter
readings are a little more consistent than the
values obtained from the oscillograms; and
while many of the oscillograph records are
within one to two per cent yet, in the aggre-
gate, it is believed that the error in the values
at the instant of short circuit are of the nature
of five per cent. However, for practical
considerations this is negligible. In reading
sustained values from the films, the amplitude
of the waves is so small that these readings
are not closer than ten to twenty per cent;
but for all sustained conditions, synchronous
impedance curves were taken using standard
meters.
Table I gives a summary of results of a
few of the oscillograms taken. The values of
voltage and current given are effective values.
The meter readings recorded were taken just
before the short circuit came on, and after
sustained conditions were reached, as a
check on the values obtained from the
oscillograms. The calculated values of react-
ance do not include the field leakage reactance
which must be included to obtain the true
value of the transient reactance limiting the
Fig, 8. Single-phase Short Circuit, Alternator Controlled by Automatic Regulator. Lower
curve shows the rapidity with which regulator contacts close to increase excitation
found from the field current film, using the
arbitrary rule of taking the number of cycles
from the instant of short-circuit until the
ripples in the field current show an amplitude
equal to twice (api)n)ximately) the thickness
of the light line as the duration of the arma-
ture transient.
•"Analysis of Short-circuit Oscillograms." by O. E. Shirley.
General Electric Rkview. Feb.. 1917. page 121.
current at short circuit. However, the field
leakage reactance on this machine is very
low.
The values of transient reactance obtained
from these tests are tabulated in TaVilo II.
Where duplicate tests showed varying values
of reactance, the several values obtained arc
given .so that the accuracy of the results may
be judged.
SHORT-CIRCUIT TESTS ON A 10,000-KV-A. TURBINE ALTERNATOR
211
TABLE I
Per Cent
External Reactance
VOLTS AT TIME OF
SHORT CIRCUIT
FIELD
AMPS.
. at
U V ■
<B(0
c c
PS
Duration Arm.
Transient, Cycles
Duration Field
Transient, Cycles
SUSTAINED
.AMPERES
Test
j3
0.
oo;
So!
1
|«
o.E
■11
o<2
^ C
|S
SoS
c,
a?
o.S
" ft)
o«
1
0
10040
10200
180.6
180
4320
12.5
13.4
21
135
440
466
2
12.4
10000
10380
180.6
171.5
2160
24.9
26.8
17
230
410
435
3
16.5
10020
10150
181.8
182.7
1820
29.0
31.8
18
220
412
404
4
24.8
10000
10000
183.6
181
1390
37.3
41.6
16
202
428
448
5
49.6
10000
10000
182.4
183.8
842
62.1
68.5
11
310
339
336
6
0
10000
9920
281
351
3780
15.3
23
81
772
820
7
0
10030
9900
315
430
4760
11.1
12.1
24
110
1060
1180
8
49.6
10000
9970
92.8
90
980
61.20
59
13
612
570
9
12.4
10000
9910
183.3
173
2040
24.9
28.4
18
952
1060
10
49.6
10000
9960
154.5
826
62.1
70
11
928
883
11
0
10000
10000
312.5
337.4
4420
13.1
27
155
575
776
12
0
9360
9260
291.5
293
3630
15.9
24
110
760
13
0
12000
12100
246.3
227
5280
10.4
10.95
21
107
616
683
14
41.4
12000
12040
222.3
250
965
51.8
59.8
12
200
466
451
15
0
11680
11680
261.6
5760
9.3
10.0
22
90
1480
1442
16
41.4
11980
11890
1118
51.0
51.9
9
205
820
800
17
0
5020
5090
80.8
3510
25
38.2
24
135
262
18
99.2
5000
4980
83.7
83
412
124.2
140
12
340
152
152
19
7.2
10000
lOIOO
181.8
178.8
2370
19.7
24.4
15
216
652
661
20
9.6
10000
9940
183.6
182
2300
22.1
25.1
16
221
620
656
21
14.35
10030
10050
179.7
169
1760
26.85
32.8
205
600
614
22
28.7
10000
10110
188.4
1230
41.2
47
9
250
552
552
23
23.9
12020
12000
225
247
1460
34.3
39.6
222
784
704
24
28.7
5000
5100
84
77.4
895
53 7
64.6
11
230
260
307
Tests 1, 2, 3. 4, 5, 13, 14, 17, 18. 3<t> short circuit, no load, no voltage regulator.
Tests 9, 10. 3<^ short circuit, no load, with voltage regulator.
Tests 6, 11, 12. 3<^ short circuit, 1.0 p-f. full load, no voltage regulator.
Tests 7, 8. 15, 16. 3<^ short, circuit 0. p-f. full load, no voltage regulator.
Tests 18, 19, 20. 21, 22, 23, 24. 10 short circuit, no load, no voltage regulator.
TABLE II
PER CENT REACTANCE OF GENERATOR FROM SHORT-CIRCUIT TEST
THREE-PHASE
SINGLE-PHASE
Ohms
External
No-load
Full-load
No-load
React-
ance
10000
Volts
12000 .5000
Volts Volts
OP-F.
12000
Volts
OP-F.
10000
Volts
1.0 P-F.
10000
Volts
10000 12000
Volts Volts
.5000
Volts
0
1.24
1.65
2.48
4.96
13.4
16 0
16.0
14.4
17.4
15.3
16.5
14.8
16.0
16.8
18.9
20.4
10.25
10.95
13.2
12.7
12.9
13.3
14.0
18.4
38.2
29.4
38
39.4
39.8
10
10.5
12.1
14.5
9.4
15.3
15.9
17.0
17.2
15.4
17.5
18.3
18 4
18.3
18.3
15.7
15.7
16.5
35.9
42,5
45.5
21S March, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 3
Values of currents at ^"arious interA-als of
time from the instant of short circuit were
scaled off and are plotted in Fig. 9 without
the automatic voltage regulator and in Fig.
10 with the automatic voltage regulator con-
nected in.
DISCUSSION OF RESULTS OBTAINED
FROM THE TESTS
Sustained Short-circuit Current
In speaking of instantaneous short-circuit
current it is common practice to refer to
per cent leakage reactance as it is generally
understood what is meant. There is, how-
ever, more or less confusion when speaking
is the field current required to force normal
current through the synchronous impedance
of the machine, and Fe the field current to
give normal voltage, assuming the saturation
ciu-ve to be a straight Hne, then the per cent
synchronous impedance is Fj-t-Fe. The
value of Fe must also be given when specify-
ing per cent sustained reactance in order that
correction may be made for whatever field
current is being held on short circuit, since
the sustained current will be directly pro-
portional to the field current. By using this
\alue of per cent S}-nchronous reactance,
e.Kternal reactance expressed in per cent may
be added directlv to the generator reactance
7.6
^'•^
^6.8
C
I 5.6
I
A J. 6
'■->
1 ; :
1
! '
1
; :
1
t"
X "^
-u 1 -
1
i 1
IT
I 1
pr
~r
1
1
' T ^ ._
1 ;
i
■
1 ;
T"""-!- ■
1
rr "^
t ' ' "
\
t 1 '
i-L-i-
i ^
1
~7 T
i '
t
1 1
3
r
*
V* -•
'
\\ I r — oon
■ ' -f-^"
m&i
■ Ohr
'75 External Reuctanct
.
X^. Aii
7"___ _
>NS, r **-'
^x^it -
^.- -T
[
S^'^l^
^
— 1
~~^
s«= g-_^ 1
1 ,
~^ ', 1 r—
X Zl
1 1
X-X ^- . -
x±x
i 1
± ±
0 I z 3 4 s e 7
seconds fromJnstontofSfwrt Circwt
Fig. 9. Three-phase Short Circuit Test at 10,000 Volts, no
Regulator. Curves show values of armature current from
the instant of short circuit until sustained values arc
reached
I:
1
'
!
1
'
^ [
1
1
1
'
--
9:
9 '
1
r-
'* 1
\ 1
1
1
'
\b
1^
'^
'
1
1
_L
.^
\^,tri Rctfula
tor
l-
.
^
.
>*^
=— ^
■
-^ rp
■=»H
,-,
t.^
"iVpT
—
- —
—
—
-
-•
' ^iV
■-^
-— -1
1
'
' 1^
\l"-i
-r-
•^
!
T^
\
!
s^
^
1
^
^
^
^
>—
--
t7ih
cui Rcqutator^^
-
^-
1— J
=^
•
i
1
1 Z 3 * S
Seconds tromln:tant of snort Circuit
Fig. 10. Tests Made at lO.OOOvolt. Three-phase, with External
Reactance, with and Without Automatic Voltage Regulators.
Curves show values of armature ciirrent from instant of
short circuit until sustained values are reached
of the synchronous reactance that limits the
sustained value of short-circuit current. In
order to have as logical a means of expressing
synchronous reactance as there is for dealing
with leakage reactance, Mr. R. E. Doherty
has proposed that the voltage consumed by
the synchronous reactance be exjjressed as a
percentage of the normal open-circuit voltage.
Since the actual flux in the machine under
normal sustained short-circuit current is
only from five to forty per cent of the flux
at normal voltage, saturation need not be
considered. In other words, if /'"/ in Fig. 1 1
and the sustained short-circuit current cal-
culated. For example, in Fig. 2 the field
current for normal voltage no load with
straight line saturation is 175 amperes; that
required to give normal current on syn-
chronous impedance test is 228 am(>eres.
Hence the per cent synchronous or sustained
reactance is (22S-^ 17."i) X 100= 130.2 per cent
at 175 amperes field. The per cent reactance
of one reactor is 4Jl.(i. Therefore with one
reactor in each phase the total sustained
reactance is 13().2-|-40.(i = 17!t.S per ont at
175 amperes field, and with this vaule of
SHORT-CIRCUIT TESTS ON A 10,nnO-KV-A. TURBINE ALTERNATOR
219
field current the sustained short-circuit cur-
rent would be (100-Hl7it.S)X.57S (normal
current of the generator is o7S) =321 amperes.
Reference to the test results in Fig. 2 shows
that exactly this value was obtained. At
any other value of field current (until sat-
F/e/(^/lmperes
uration is approached) say ¥c. amperes, the
sustained value would be (F^-H 175) X321
amperes. Table III shows how closely the
test values check the amperes calculated by
this percentage method.
The prediction therefore of the sustained
short-circuit current of a generator is a simple
matter when all phases are short-circuited.
It can be determined of course from the syn-
chronous imepdance curve, or can be cal-
culated on a new design on which no test data
are available, with the same degree of accuracy
that the saturation curve is predicted. How-
ever, the sustained current that results when
a polyphase machine is short-circuited single-
phase is not so easily calculated. The single-
])hase armature reaction, pulsating from zero
to s/'UNI, where / is the armature current and
A' the armature series turns per phase, is op-
posed by the field current and also by the eddy
currents induced in the pole-pieces, damper
windings, etc., which are most difficult of
calculation. In most cases it matters little
whether we can calculate this, since a single-
phase synchronous impedance test will give
the necessary data and is easily made.
Instantaneous Short-circuit Current
.4; Normal Voltage No Load Zero External
Reactance
The almost universal method of calculating
the initial short-circuit current of a generator
is as follows: Calculate, estimate, or obtain
from the saturation and synchronous im-
pedance tests the per cent leakage reactance
of the generator. The instantaneous sym-
metrical short-circuit current will then be
approximately 100 divided by the per cent
reactance, times the normal current. The
total current with the wave completely ofi^set
may be two times the above value. This of
course neglects the field reactance, but so far
as present tests have detennined it does not
seem necessar\' to add the field leakage
reactance of modern turbine generators to
the armature leakage reactance in the deter-
mination of the transient reactance limiting
the current on instantaneous short circuit.
This is not the case with definite pole machines
where the field leakage reactance may be
from 30 to 60 per cent of the value of the
armature leakage reactance.*
.4/ Other Voltages
It is of course well known that at voltages
higher than normal the per cent reactance
limiting the instantaneous short-circuit cur-
TABLE III
PER CENT REACTANCE
SUSTAINED
SHORT-CIRCUIT CURRENT
Generator
External
Total
Calculated
Test
3 phase, 1 reactor
130.2
130.2
130.2
130.2
87.5
87.5
87.5
87.5
49.6
24.8
16.5
12.4
28.7
14.3
9.6
7.2
179.8
155.0
146.7
142.6
116.2
101.8
97.1
94.7
321
372
394
405
497
568
594
611
321
3-phase, 2 reactors
370
390
3-phase, 4 reactors
403
1 -phase, 1 reactor
1 -phase, 2 reactors . . .
504
568
1 -phase 3 reactors
593
625
■Reactance of Synchronous Machines and Its Applications," by R. E. Doherty and O. E. Shirley, Proc, A.I.E.E.. June. 1918.
220 March, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 3
rent decreases, and that it increases at lower
than normal voltage. Furthermore, as a
general rule, the reactance at half voltage is
higher in proportion to the reactance at
normal voltage on turbine generators than on
definite pole machines. However, basing
conclusions on all the data available, it does
not appear that any great error will be intro-
duced if between 50 and 120 per cent of
normal voltage the per cent reactance at
normal voltage is assumed to vary in inverse
proportion to the terminal voltage. Prac-
tically, of course, machines are seldom
operated at other than normal voltage, con-
sequently approximations are sufficient for
these conditions.
With External Reactance
External reactance has the same effect as
partial voltage in increasing the apparent
reactance. The reason for this is that since
part of the voltage is consumed outside of the
generator, the reactive voltage within the
machine must counterbalance only a part of
the total voltage. Due to the decrease in
saturation accompanying the low voltage,
proportionately less current through the in-
ternal reactance is required to generate a
reactive voltage equal to the partial voltage.
Placing an external reactance of value equal
to the internal reactance of the machine in
circuit when short circuiting has the effect
of increasing the generator reactance from
ten to thirty per cent or increasing the total
reactance from five to fifteen per cent. Since
in practice the external reactance is seldom if
ever more than the internal reactance of the
generator, it does not seem that any correction
for this increase in reactance is necessary.
Under Load Conditions
When a generator is o])erating under full
load, the flux in the machine is that necessary
to maintain normal voltage plus the reactive
flux due to normal current flowing in the
armature windings added in the proper
phase relation. This reactive flux is .v per
cent of the normal flux where x is the
armature leakage reactance of the armature.
At zero power-factor, these two fluxes arc in
phase and at unity power-factor at right
angles to each other. Since on short circuit
there must be enough current in the armature
to maintain the total flux that existed in the
machine prior to short circuit, and since
normal current in the armature will cause
X per cent of normal flux to flow in the leakage
paths, it follows that on sudden short circuit
under full load zero power-factor the initial
current will be I — + 1 1 times normal and
on full load unity power-factory,
m
times normal current . The tests made on open
circuit and on full load zero power factor con-
firm the above theory. It is not known why
the tests at unity power factor full load do not
check unless it is due to an error in test.
On Single Phase
The single-phase tests agree with those
made on other machines showing that the
reactance limiting a single-phase short circuit
is somewhat higher than the corresponding
three-phase reactance. The difference, how-
ever, is small and may just as well be due
to inaccuracies in the results as to any actual
difference. There seems to be no reason why
it should be different except that there may
be some slight effect of mutual reactance
between phase belts in the one case that is
not present to the same extent in the other.
Effect of Automatic Voltage Regulator
In this particular case, the automatic
voltage regulator had no effect on the short-
circuit current until one quarter of a second
after the short-circuiting switch was closed.
This time was the same for various amounts
of external reactance also. Oscillograph
records were taken which show the action of
the contacts of the regulator; Fig. S shows
one of these records. Full field was put on
the exciter in approximately one and three
quarter seconds and the generator current
reached its sustained value in from three and
one half to five and one half seconds.
CONCLUSIONS
While there is not yet enough data at hand
to decide finally the various points brought
out in these tests, yet the following con-
clusions are apparently well established.
I. Within all i)ractical accuracy the effec-
tive symmetrical initial short-circuit current
/ (effective amperes) of a turbine generator
equals :
(a) 1 = X/n on no-load normal voltage
with no external reactance.
(6) / = — ; Xin on no-load normal volt-
X, X -r,
age with external reactance.
(c) / = I h 1 I/k on full-load zero power-
<V-)'
factor and with no external reactance.
SHORT-CIRCUIT TESTS ON A 10.000-KV-A. TURBINE ALTERNATOR
221
(d) I = ( '"" +lV" on full-load
zero
power-factor and with external reactance.
w /=
1/1 GOV
-|-(1)- on full-load unity
power-factor with no external reactance
where
£ = normal voltage per phase of the gen-
erator
/„ = normal current per pliase of the gen-
erator
A'(. = per cent external reactance
_ohms external reactance X /nXlOO
~ E
A'j = per cent transient reactance of the
generator.
II. Percentage synchronous reactance A'^
of the generator equals:
As= =;-X 100 per cent at Fr amperes,
Fe
where
/^/ = field current required to give normal
current on synchronous impedance tests, and
F£ = field current no-load normal voltage
assuming a straiglu line saturation (i.e., air
gap ainperes). Then the sustained short-
circuit current with external reactance and
Fv; amperes field is /.s
Is — — i — y(.i ny^ 7^~
Xs + Xt rE
That is, external reactance is added directly to
the internal reactance of the generator in the
same way as for the instantaneous values.
III. It is close enough for practical
purposes to consider that the per cent leak-
age reactance varies inversely with the
voltage.
IV. As far as the first quarter second
after short circuit is concerned, it makes no
difiference whether the generator is regulated
1)V hand or by means of an automatic voltage
regulator.
V. Adding external reactance in a gen-
erator circuit equal to the internal reactance
of the machine reduces to one half the instan-
taneous short-circuit current. At the end of
one second, however, the current is prac-
tically the same whether the external react-
ance is connected in or not, if the leakage
reactance of the machine is small compared
with the synchronous reactance.
VI. All of the foregoing conclusions apply
to definite pole generators as well as turbine
generators except that in definite pole ma-
chines it is very essential that the field leakage
reactance be included in the transient react-
ance, while on this machine, neglecting the
field reactance did not introduce any appre-
ciable error.
222 March, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 3
The Engineer Can Do More About It Than
Pay and Grin
By Calvert Townley
President, American Institute of Electrical Engineers
In this address, which was prepared for presentation to the Schenectady Section A.I.E.E., Mr. Townley
discusses the pressing question of why prices are high. The analysis discloses that prices are high because
costs are high, and costs are high largely because wages have increased. The whole structure of costs and
wages has shaped itself according to the law of supply and demand. An increasing scarcity of materials and
labor since the beginning of the war, first in Europe and then in this country, has acted to bring about the
vicious circle of mounting prices. The problem of greatest moment is to foresee what will be the ultimate
outcome of the situation. Will prices continue at their present level? Will they rise still higher or descend
to their pre-war basis? History gives us a precedent, and it is almost certain that supply wiU eventually
catch up with demand, and if prices have not by then adjusted themselves gradually and orderlj' they will
come down with a thump which will be heard around the world. The engineer can do much to prevent the
disaster and hardship that would result from such a course of affairs. — Editor.
The purchase prices of most essentials —
not to mention luxtiries of life — are now
abnormally high. Can the engineer do any-
thing abotit it except pay and grin? George
F. Swain of Boston, says that the engineer is
the antithesis of the idealist and that the
idealist is a most dangerous individual. The
engineer approaches a problem with an open
mind; first obtains all the available facts and
then reaches his conclusion and bases his
action on those facts. The idealist, on the
contrary, first pictures the ideal result which
he would like to obtain and then proceeds
to make his facts fit; if they do not, so much
the worse for the facts.
In discussing higher prices, let me see if I
can qualify under Professor Swain's definition
of an engineer. The first cjuestion that arises
is, "Why are prices high?" And the answer
to this question is almost, if not quite, obvious.
Prices are high because costs are high and
costs are high because wages have gone up.
Of course material as well as labor goes into
cost, but material in the last analysis is very
largely labor, because coal, iron, copper,
lumber and other raw materials forming the
bulk of those used are governed as to their
cost by the wages paid to produce them. It
is also alleged, and with reason, that prices of
some commodities are higher than the in-
creased costs justify because their distributors
have taken advantage of existing conditions
to reap abnormal profits. This no doubt is
so to a limited extent. It can hardly be
claimed to be true in general and certainly
not to such an extent as to disprove the state-
ment that prices are high because costs are
high.
Following the analysis, if prices are high
because costs are high and costs are high
because wages have increased, the next
question is, "Why have wages increased and
in what way, up or down, may wages be ex-
pected to change in the future?" We have all
heard much about the "awakening" of labor
and its determination to hereafter demand and
obtain a greater "share in the reward of its
products." Our recent histor>- is not lacking
in examples of efforts on the part of workmen
to benefit through organization and collective
bargaining, efforts crowned with no mean
measure of success; but of catch phrases and
slogans it perhaps may be said that they lack
a sufficiently definite meaning to be inter-
preted alike by all. A catch phrase can
frequently be made to mean whatever its
user wants it to mean over wide limits. Let
us therefore adhere to a tcnninology of which
the meaning is understood by all and which
is always the same.
First of all, what do we mean by "labor"
and "capital'" Perhaps we all ought to
understand what these words mean, but do
we? If we say that "labor" is the perform-
ance of manual work and " capital ' ' is accimiu-
lated money, it can be pointed out that many
are classed with labor who do no manual
work, while a large number have accumulated
money who are not capitalists. Possibly the
walking delegate's definition would be that a
laborer is one who works all the time for pay
but never has any money and a capitalist
is one who has money all the time but never
does any work. Both of these are manifestly
incorrect definitions. For the purpose of
avoiding misleading terms perhaps we can
dodge the issue by not using them in the
l^resent discussion and instead of referring
to capital and labor speak of the "employer"
and the "employee" although even then it
becomes necessary to expand the term
"employer" to include not only him who
THE ENGINEER CAN DO MORE ABOUT IT THAN PAY AND GRIN 223
pays for the ser\-ices of others with his own
money but also him who directs the work of
others while himself employed, and also to
limit the term "employee" to those who do
not direct the work of others.
Is there any good reason to believe that a
"new order" of things has been created that
the working man or employee will hereafter
"demand" and what is more to the point
obtain a greater share of the reward of his
labor, and that therefore wages and conse-
quently the cost of e\-er\-thing into which
labor enters will stay up and may even go
higher? Has there been anything which
may properly be called an "awakening" of
labor? The only evidence that I can find to
support such an idea is the undeniable fact
that beginning in 1914 the employee has
demanded and has obtained a greatly in-
creased wage, and of course we know it is
human nature to get all we are able and to
keep it if we can. But these plain facts do
not prove the reasons why. The working
man like every other man has in the past
always wanted all he could get and human
nature today has neither gained nor lost
cupidity. There is no indication of a "new
order" in these facts.
It is conservative to look for ordinary and
natural causes before evolving new theories,
so before wondering whether there is or is
not any "new order" of things, suppose we
examine the old order and see what would be
naturally expected under it. Before the war
the country was prosperous, general business
was good, the employees if not contented
and happy were at least much less discon-
tented and unhappy than they are today,
and with wages very much lower than they
now receive. Then came the war. The men
of both sides threw down their tools and took
up arms. The productive capacity of every
warring nation was at once greatly reduced,
but their needs were not — they were greatly
increased — and naturally the United States
was called upon to help supply them. We
were by no means the only, but we were
certainly by far the largest, source of supply
in the world and otir productive capacity
was immediately speeded up to meet the new
and unusual demand made upon it. All this
was natural, ordinary, and logical and its
analysis so simple as to seem very obvious.
Then what happened? The business men
of this country — the employers — those in
command of industry saw their chance. They
had, if not exactly a monopoly or corner in
the supply market, at least something very
like it and they promptly took advantage of
the situation and boosted their prices.
Higher prices for export to Europe soon
reacted to cause higher prices at home and
the complaint of profiteering started and
spread. It reached such proportions as to
influence the government, and action was
taken to curb the business man's cupidity
and make him loosen up. To a considerable
extend he did it; sometimes with not very
good grace, but nevertheless as a class he
recognized the logic of events and acquiesced.
Then we got into the war ourselves and we
likewise took several million men out of our
shops to fight and in turn we increased our
demand for manufactured goods and de-
creased our productive capacity. You will
note that I say "productive capacity" not
our output. That our actual output was
greatly increased in spite of the reduction in
capacity is history but the reasons for it
were impro\-ed efficiency, concentrated effort,
etc., and do not contradict nor weaken the
preceding statement. Well — what happened
then? Why the workman — the employee —
waked up. He began to do what the business
man — the employer — did when the war
began and which caused the hue and cry
against profiteering. He saw a diminishing
supply of men and an ever increasing demand
for work and he cornered his market. He put
up the price of his services and having got
the new price easily he put it up again and
so on. You know the rest. Now all this
seems to be natural, simple, logical, and so
obvious as not to need argument, but it is all
based not on any "new thought" or on the
"awakening of the proletariat" or any other
new ideas or theories but on the plain old-
fashioned, simple law of supply and demand,
a law as old as the hills and just as immu-
table.
I do not lose sight of the fact that the
organization of labor played a conspicuous
part in bringing about increased wages.
Many people no doubt honestly believe that
organization did it all. Well, organization
did a great deal of course. The organization
was the machine tool or weapon which the
employees used to get quicker and greater
results. It is pertinent to remember, however,
that the organization of workmen is not new.
They have been organized for j^ears. There
is no essential difference between the way in
which nor the extent to which they are
organized now, and what they have been for
many years past. And ever since employees
began to organize they have been trying to
224 March, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 3
get higher wages by identically the same
methods they've been using during the war
period. But organization never before ac-
complished anything like the results which
latterly have been brought about, and it seems
clear therefore that we must seek the cause
for the great wage advances not in some old
condition, like organization which existed long
before the war, but in some new condition that
has been created since the war began. That
new condition is obviously a change in the
relation of supply and demand and we get
back to our first conclusion again as to why
wages have so greatly increased.
Another contributory cause is the decreased
efficiency of the workmen. Employees — as
a class — do less work, produce smaller results
per day's work than formerly. There seems
to be abundant evidence of this condition,
enough to warrant our accepting it as a fact.
One reason, of course, is that the employers
had to use "seconds." Just as a builder in
times of stress will put into a house lumber
that he would ordinarily reject, because he
cannot get enough first class lumber and he
must have the house, so employers were
forced to hire men in war time that ordinarily
they wouldn't have about the place; there
weren't enough others. Then of course there
is the matter of fewer hours per day and a
reluctance to work continuously through the
week. These features materially affect the
increase in cost. They affect it tremendously,
but while in fact they may be important they
nevertheless are not causes at all but merely
results incidental to the working out of the
law of supply and demand. They are super-
imposed upon and do not undcrly the existing
condition which we are analyzing.
Some say that a change in the value of the
dollar has caused the increased cost. On all
sides we hear and read the statement that
the dollar is only worth .50 cents, or some
other small fraction of its face, and that
prices and wages are really no higher than
they used to be because the dollar is now
worth much less. Well, suppose we briefly
examine that statement. What is a dollar
worth anyhow — -by itself? Why a dollar
isn't worth anything — by itself. You can't
do anything with it — by itself. A dollar is
simply and solely a convenient medium of
exchange. It has become valuable just to
the extent that men want it and will give up
something which they have to get it. When
we set out to place a definite value on the
dollar in commodities cr labor, it isn't suffi-
cient to take into account the conditions in
one place only; in Schenectady for example
or in a dozen or in 50 places, or even in any
one entire country. The dollar has value all
over the world, but when our economists
insist that the value of the dollar has fallen
permanently they are evidently thinking in
terms of conditions in the United States only.
Abroad the situation is ver\- different. A
dollar will buy about six shillings in London,
one and one half times its old rate; 1.5 francs
in France, three times its old rate; 100 marks
in Germany, 24 times its old rate, and the
end is not yet. Travellers returning from
Europe bring back specific information of
how much more than formerly can now be
bought with a dollar. Not only is foreign
money cheaper but things are as well. For
example, in November last the rate at the
best hotel in Vienna. Hotel Bristol, for a big
room with three beds and a bath, occupied by
three, was the equivalent of (i2 cents a day
American money. Does that look like a
depreciated dollar"' And if it be said that
this comparison is misleading because Euro-
pean exchange is only down for awhile and
the condition is therefore temporar>-, the
answer is how do we know, how does anybody
know that the values in the United States
are any more stable or permanent ?
Instead of theorizing, why not look to
history- for real information? The best
indication of what will happen in the future
is what has happened in the past. We
have had wars before, not so big, not so many
men were killed, but those that were killed
were just as dead, and a ver>' similar condition
as to the supply of and demand for labor was
created. This same condition of inflated
values that confronts us now existed right
after our Civil War and it seems reasonable
to attribute it to the same causes. In any
event the high cost of living after the Civil
War was not due to the activities of labor
organizations because they didn't exist then.
The more we examine the cause for the
abnormal ad\ance in wages from different
angles the more it seems evident that the
fundamental underlying and controlling cause
has been an increased demand and a decreaseil
sui>ply. Now if the law of supply and demand
has controlled the wages of workmen in the
l)ast and through them the cost and therefore
the price of commodities, this same law is
ver^• likely to exercise this same control over
the same conditions in the future. In other
words, prices will stay up. go higher, or fall
according as the supply of labor is equal to,
less than, or greater than the demand.
THE ENGINEER CAN DO MORE ABOUT IT THAN PAY AND GRIN
22=
You will have noted, I hope, that I have
tried to discuss facts and conditions and have
not referred to the so-called "rights" of the
interested parties, the emijloyer, the employee
and the public. That is another part of the
story, but it is my conception that the eco-
nomics of industry will continue to be
governed by economic laws which are just
as immutable as is the law of the attraction
of gravitation and other physical laws, albeit
not so generally understood and acknowledged ;
and that any so-called "rights" of the dif-
ferent elements of society, no matter how
skilfully or persistently asserted, must give
way absolutely before the inexorable opera-
ation of economic laws. The question of
"rights" is further one into which opinion
enters more than demonstrable facts and
no opinions of any party to such a discussion
have received particular credence, much less
acceptance from any opposing ]jarty. It
would be jjrofitless therefore in the present
discussion to diverge into any attempted
examination of this phase of the subject.
If now jjrices were raised to their present
level because the war required greatly in-
creased jjroduction and at the same time
withdrew from productive occupations so
large a number of workers, how will ]jrices
be affected by the operation of economic laws
in the future? The war is over. War
material is no longer demanded. Its jjro-
duction has ceased. The men who fovight
have been released to pursue again their
pre-war vocations. Yet prices are still up
where thei.- were. If my reasoning is correct,
why has the law of supply and demand not
reversed its effect and ojjerated to restore
pre-war conditions? That question has
to be asked to follow the analysis logically
but its answer seems fairly obvious. Pre-
war conditions have not been restored be-
cause there hasn't been time.
In this country during the war we couldn't
and didn't i:)roduce what was needed. We
Ijroduced all we could, selecting the things
most essential — war material. Industrial
needs had to wait. Now the country is
catching up. Our stimulated jjroductive
capacity has been diverted from war to peace
channels and a booming business still strug-
gles to meet the demand made upon it. In
Europe the conditions are similar but more
acute. Industry was put out of joint there
worse than here. It will take Eurojje longer
to recover and readjust, meanwhile some
of their immediate needs must be supplied
from this side and this requirement ]3uts an
added demand on us and still further defers
our return to a pre-war normal status. New
nations have been created, financial and many
other problems demand solution, all taking
time and yet more time and keeping the
United States still working extra hours.
Although the supply of workmen has been
greatly augmented by demobilization, the
peace demands have absorbed this supply
and as yet there is no surplus.
Many people say there never will be a
surplus; that the United States has taken
up a new place in the world's industries and
henceforward will be expected to produce and
will supply a so much larger share of the total
world's commerce than ever before that our
industries will be kept going at top speed and
every workman be busy. This is certainly
an optimistic picture. It is worth examining.
Europe owes a lot of money. Some of their
men have been killed. In the war zone the
country was laid waste. These are the only
real differences between then and now.
Their abilities are no less, their natural
resources are intact, their morale is good.
Europe competed with us before the war and
although we set up a tariff wall around our
own country and successfully protected our
domestic business she captured most of the
world's trade against our best efforts. Of
course we now have certain advantages we
did not possess before. We are a creditor
nation and can exert the influence attaching
to that position. We have made tremendous
inroads into Europe's foreign commerce while
she was down and out commercially and we
have thereby established relationships and
gained an entree previously denied to us and
which if judiciously followed up should pro-
duce results of great value. But the European
nations are not out of the running perm-anently.
They must recover a large volume of trade or
go bankrupt. They are diligent and by nature
and training thrifty. They will work under
the spur of necessity. We are naturally
si^endthrifts and are inclined to overcon-
fidence. When the world's productive ca-
pacity shall have caught up with the indus-
trial shortage occasioned by the great war and
the demands of commerce shall have again
become normal, and when Europe's industries
are once more functioning properly, it seems
reasonably certain that there will be an excess
of production over consumption and some
nations will lack a market. Translated to
workmen this means that the supply will
be greater than the demand and some must
go hungry. It will be then that the test will
226 March, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 3
come. If the United States shall have used
the prosperous period to prepare for it by
reducing the costs of production and by
teaching its people economy and thrift, the
readjustment may come perhaps without
any serious disturbances and we may save
a good share of what we had gained by our
running start. But if we sail serenely on
ignoring the possibility of a coming storm, if
we continue the policy of working less and
less and of paying more and more while
Europe buckles to, the resulting depression
with its period of unemployment, suffering,
and possible panics is appalling to con-
template. Then prices and wages will come
down suddenly and with a thump, such a
thump as the country has never known. The
inexorable law of supply and demand is no
respecter of people or of nations and its deadly
work will be deadly indeed.
We now come back to the question I asked
in the first place : ' ' Can the engineer do any-
thing about it except pay and grin"'" I
think he can. An engineer is by training
taught to think straight and to speak clearly.
Further it is a fixed tenet of his faith to tell
the truth and fear none. People know that
and believe him. If I have been fortunate
enough to have made my views clear to you
and if you agree with me, say so and keep on
saying it as you go about your daily tasks.
Do what you can to show up the idiocy of
dwelling in a fool's paradise and preach the
gospel that industrial prepaerdness is as
essential to commercial safety as military
preparedness is to national safety. Dispel
the hogy of class control. Brains always
have ruled the world and brains always will.
Show that we are dealing with a perfectly
normal problem which must be solved in
conformity to well known natural laws and
not with any mysterious unknown or novel
principles or with newly discovered rules of
life. No organization of a minority created
for the avowed purpose to taking from the
majority some of its property or just rights
can long prevail. Witness organized Ger-
many's eflort to subjugate the world. The
laws of supply and demand will ultimately
just as surely bring down the cost of living
and the wages paid employees as it put
these items up. The only uncertain features
are the time when the changes are to occur
and whether these costs and wages shall be
brought down in an orderly and gradual
manner so that the readjustment shall be
made without disturbance and with lasting
benefit to all, or whether they shall come down
with a thimip. heard around the world, amid
disaster and distress.
This is one of the most important problems
confronting our great nation today. As citi-
zens, it concerns us all. As members of the
A. I. E. E. who have enjoyed the privilege
of special training and of valued association
with our fellows, we have each a duty to
perform. It is perhaps as well stated as
may be on the old familiar railroad crossing
sign,
"STOP, LOOK, LISTEN."
227
Helium, the Substitute for Hydrogen in
Balloons and Dirigibles
By W. S. Andrews
CoNsiLTiNc. Engineering Department, General Electric Company
The successful transatlantic round trip made by the dirigible R-34 attracted attention again to the capa-
bilities of heavier-than-air machines for freight and passenger air traffic. One drawback to this mode of
travel however has been the fire danger inherent in the use of hydrogen as the buoyant gas. It has long been
known that this danger would be eliminated if the light and incombustible gas, helium, could be discovered
in sufficient quantity for use in place of hydrogen. Only recently has this welcome discovery been made.
Mr. Andrews, who has made a considerable study of this remarkable gas, describes below its historical features
and physical properties, one of which is that its'lifting power is 92.6 per cent that of hydrogen. — Editor.
Introduction
A considerable amount of public interest
has been recently focussed on the rare gas
helitim on account of its proposed use in
dirigible and observation balloons instead of
hydrogen, thus absohitely eliminating the
constant danger of fire that is connected with
the use of the latter gas. It is true that helium
is a little heavier than hydrogen, but both of
these gases are so light in comparison with
air that there is actually not much difference
in their voliunes for equal lifting power; and
the total elimination of the constant fire
hazard connected with hydrogen far over-
balances the disadvantage. Moreover, on
account of helium being a little heavier than
hydrogen, it does not diffuse so readily
through the thin covering of a balloon and is
therefore less subject to waste.
Helium is one of the so-called noble gases ;
it makes no chemical combination with any
other element, and therefore is absolutely
inert and incombustible under all conditions.
The great and hitherto insurmountable draw-
back to its use in balloons has been its
extreme scarcit}-, its only known source until
recently being the atmosphere and certain
rare minerals and mineral waters, from which
it has been extracted only in small quantities
by refined methods involving much time and
expense.
Within the past year or two, however, it
has been discovered that the natural gas
found in Kansas and elsewhere contains
sometimes as much as 2.5 per cent of helium,
and that the latter can be extracted and
purified in large quantities at a comparativeh*
small expense.
If, therefore, the present promise for the
cheap production of helium holds good, it is
reasonable to hope that its application to
aeronautics may almost revolutionize this
important branch of scientific and com-
mercial development.
Historical
It is stated that a brilliant yellow line was
first seen by Janssen in the spectrum of the
sun's photosphere in 1S6S. This remarkable
line has a wave length of .5876.6 Angstroms;
and, being up to that time unknown, it was
considered good evidence of a new element
existing in the stm but foreign to the earth.
Frankland and Lockyear therefore named
this new element "Helium," from the Greek
word "Helios," the sun.
Great interest was excited in the scientific
world by this discovery, and numerous
investigators began to search diligently for
further evidence of helium. As time went on,
the yellow spectral line of helium was dis-
covered in the spectra of some stars and in
1SS2 it was observed by Palmer in the
spectnun of flames issuing from Mount
Vesuvius, thus proving the actual presence
of the element on the earth.
Later on in 1895, Ramsay while investigat-
ing the properties of gas obtained from certain
rare minerals such as clevite, tiranite, etc.,
submitted it to spectrtun analysis, and once
more the bright yellow line of hehimi became
apparent, thus showing the new element to
be at length actually within oiu- grasp.
A practical, though very limited sotirce of this
new gas being thus discovered, it was soon pro-
duced in sufficient quantity for the examina-
tion of its physical and chemical properties.
Up to quite recently, however, the term
rare was applied to it with good reason, for
only a few years ago .the price of piure helium
was quoted at from $1700 and upwards per
cubic foot, whereas, according to recent
estimates, it can now be produced from the
Kansas natural gas, before referred to, for
about 10 cents per cubic foot.
Physical Properties
One of the most interesting physical
properties of helium is that it is the most
228 March, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. .3
difficult of all gases to liquify and for a long
time resisted all efforts. At length, however,
in 1908, Onnes succeeded in condensing it to
a liquid, by means of liquid hydrogen and
pressure, and by its rapid evaporation he
obtained a temperature within two degrees of
absolute zero.
The boiling point of liquid heliimi is — 268.7
deg. C. Its specific gra\dty in liquid state is
0.15 and it is therefore the lightest of all
known liquids.
Helium gas is generally believed to be
monatomic, although some phj-sicists con-
sider the possibility that its molecules consist
of two atoms so firmly bound together that
they have never yet been separated. It is
present in the air at ordinary levels to the
Therefore
1.2933-0.178 ^^^ ^ ,.r-
1.2933-0.089 = ^2-^ = '^^ ^^'^"^ ^^^' °^
helium in air as compared with that of
hydrogen taken at 100; or, in other words, a
given volume of heliimi will lift 7.4 per cent
less weight than the same volume of hydrogen
under similar conditions.
According to a statement in the L'. S. Gov-
ernment Bulletin 178 C (p. 76), issued by the
Bureau of Mines, about 15 per cent of hydro-
gen may be mixed with So per cent of helium
without imparting any dangerous inflammable
feature. This mixture would have 93.4 per
cent of the lifting power of hydrogen and its
use is proposed for dirigible balloons.
Red Yellow Green Blue Violet
The Spectrum Lines of Helium
extent of two to three parts by volume in a
million but at higher altitudes it is probably
more abundant.
As previously stated, helium is heavier than
hydrogen, which latter is the lightest of all
known gases its molecular weight being
2.016. The molecular weight of helium is
3.99 or practically twdce that of hydrogen,
but, owing to the wide difference between
the weights of air and hydrogen, helium has
only about 7.5 to 8 per cent less lifting power
than hydrogen, and this may be deemed an
insignificant feature in consideration of the
perfect safety from fire hazard which results
from its use.
Ordinary' dry air at zero deg. C. and at
760 m.m. pressure weighs 1.293 grams per
litre, but this weight is naturally subject to
variation under changing conditions of
humidity and altitude. However, assuming
the above figures to be correct for all practical
purposes, the difference in lifting power
between hydrogen and helium may be figured
thus:
Dry air weighs about 1.2933 grams per litre.
Hydrogen weighs about 0.089 grams per
litre.
Helium weighs about 0.1 7S grams per litre.
Helium shows a remarkable and beautiful
spectrum when excited by electricity, con-
sisting principally of eight lines which
include the colors red, yellow, green, blue,
and violet, the yellow line, as before referred
to, being especially brilliant.
The most remarkable feature connected
with helium is that although it is unquestion-
ably an element yet it is one of the dis-
integration products of another element as
discovered by Ramsay and Sody in 1903.
When the life cycle of a radium atom is
completed it breaks up into two elementary
atoms, one of which, the helium atom, is
permanent, while the other, the niton atom,
proceeds through a cycle of changes, until it is
believed to assume finally a stable form in
the shape of lead. Other atoms of helium are
also ejected at the instant of some of these
changes.
These elemental changes are due to intra-
atomic action for they can be neither hastened
or retarded by any known external means.
The forces which maintain the integrity of
the atoms of radium and all other elements
appear to be almost inconceivably more
powerful than any disintegrating force that
we are able to apply from the outside.
229
Silent Spokesmen in the Factory
By Roscoe Scott
National Lamp Works of General Electric Company
The problem of labor turnover is of great importance to the employer. Not only is it expensive to break
in many new men each month in proportion to the average number of employees on the payroll, but the manu-
factured product is naturally not up to standard where a great share of the work is performed by inexperienced
help. Many employees will make a change without giving a single thought to the matter of living conditions.
Good wages, good hours, good treatment and good working conditions would be supposed to create a feeling of
contentment, but it has been found that unless some special means are taken to focus employees' attention on
these advantages they are prone to make frequent changes in employment. This article shows that a judicious
use of " Silent Spokesmen" — placards and cards posted in conspicuous places in the factory— will serve to make
the employee appreciative of the advantages that have been provided for him by the management, hence con-
tented with his lot and reluctant to make a change without mature thought. — Editor.
to fill
In one of the General Electric Company's
factories in Providence, R. I., there is an
employee who has been with that particular
plant, and its predecessor, for forty-five years.
If every industrial- worker were of his type, the
much-discussed problem of labor-turnover
would not exist — unless it existed as a problem
of getting ' ' new blood ' ' into the works without
bidding farewell to a band of white-haired but
devoted employees of half a century's standing.
CP 6000
c
(0
(0 I
5000
"*- in
o (];4000
0 C13000
4JUJ
8 > 2000
U q;
3^2
j->.b
C
o
5
1000
O 10 20 30
Length of Stay of Average
Employee (^Months) -*■
Expense of Training New Employees
It is hardly likeh' that a condition of general
labor stagnation, such as that just suggested,
will ever be found in the electrical industry.
The person who stays with his or her job only
forty-five weeks is met a thousand times more
often than the one who sojourns forty-five years.
Accordingly, we find a widespread anxiety
among superintendents and employment man-
agers that the percentage vearly turnover,
200 ja+c-b)
a+b
may be brought as low as possible, so that the
continual expense of training new employees
the places of those who leave may be
brought within reasonable limits. (In the
above expression, a stands for the number on
the payroll at the beginning of the 3'ear, b the
number at the end, and c the number hired
during the year.)
Many employers have had the notion that
good wages, good hotirs, good treatment and
good working conditions — important as they
unquestionably are — would solve every diffi-
culty with labor, and have been disappointed
to find that all of these helps combined did not
produce the immediate results expected.
They did not create a "company feeling," nor
did the>- deter employees from throwing up
their jobs with disturbing frequency.
The Company may spend $20,000 on a
special ventilating and air-conditioning
system, to eliminate unpleasant working con-
ditions in certain departments, but does Sally
Smith give the fact a moment's thought before
taking it into her head to "lay off a few
weeks'" Perhaps she does; perhaps she con-
siders the efforts that have been made in her be-
half and decides that she, in turn, should do the
square thing by her employer. But if she does
thus take the company's efforts into account,
it is probably because she has been told about
them, and from time to time reminded of them.
Only a constant presentation of the facts in
their proper light can keep even the most
intelligent of us from occasionally viewing our
work through colored spectacles that distort
the good and make mountains out of mole-
hills. In the case of an operator at the bench
or machine, who has formed the mental habit
of looking at his work as meaningless and at
the quitting-time whistle as his best friend, we
should not blame him for the hiunan tendency
to overlook shower-baths, bonuses and other
benefits, but should, rather, blame ourselves for
not having tried to direct his thoughts along the
proper lines. How, then, shall this be done "1
First and foremost, the personal contact of
the superintendent with his foremen, and
230 March, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 3
through them with his people, can do wonders
towards promoting an atmosphere of har-
mony and co-operation, if systematic effort
along this line is made. But there are limits
to what can be done by personal contact.
Some foremen are ill adapted to explain the
company's policies to employees; others are
too busy, or take no interest in that sort of
thing. Right here is where the ' ' Silent Spokes-
men in the Factory," referred to in the title of
this article, play their part.
The "spokesmen" in question consist of
printed messages, systematically posted
throughout the plant, reminding the people
of facts they would otherwise overlook, and in
this way molding their attitude towards the
institution with which they are connected.
In the incandescent-lamp factories and
lamp-parts factories of the Edison and of the
National Lamp Works of the General Electric
Companj', "silent spokesmen" have been put
to work to build up goodwill among some
15,000 employees. A three-fold system of
placarding was started over two years ago,
comprising :
(1) Permanent signs, 15 in. by 11 in.,
protected by glass in green wooden frames.
(2) Smaller placards 11 in. by S^ i"-. pro-
tected by celluloid in slotted frames. These
placards are changed at six-week intervals, the
plan being to put up new copy before the old
copy has entirely lost its interest.
(3) 5j4-in. by-33^ in. cards (postcard size),
distributed in large numbers in suitable
holders of three specially-designed types,
throughout the factory. These cards also are
changed at six-week inter^-als.
The purpose of the large permanent signs
is to advertise certain definite working con-
ditions, or beneficial features of the plant, to
those who benefit by them. One sign, for
example, reads:
"HEALTH = $$$$$$
"Your health is priceless.
"You value it — we value it, too.
"We value it so much that we have put in a
special ventilating plant, costing many thou-
sand dollars, to change the air in here com-
pleteh- ever}- few minutes and prei^are it for
your lungs. The air is purified, then mois-
tened so as not to be harmfully dry, heated or
cooled (as needed), and forced into this room
by a powerful fan.
"No poor air conditions for our people — ■
not if the management can help itl"
Among other points similarly featured are
pure drinking water, large windows, fireproof
doors, "panic" doors, oily-waste cans, cloth-
ing lockers, protective conduits for electric
wires, toilet-room facilities, fire extinguish-
ers, sand and water pails, and guards on ma-
chinery. The simple facts are stated, and the
reader left to draw his own conclusions.
The 11-in. by 8j^-in. changeable placards
earn,' goodwill messages suited to special sea-
sons of the year, or deal with special conditions
that may arise, such as an epidemic of tar-
diness (these frames, by the way, are located
near the time-clocks and building entrances).
It is felt that the requirements of promptness
and regularity can be presented in such a way
as to create goodwill. Among the seasonable
placards posted are those relating to Christ-
mas, vacations, lectures at the factor}-, pre-
cautions against influenza, etc.
The small cards (5* o in. by 3^2 in) are more
general in tone. The goodwill messages which
they bear relate not only to the employee's
work, but to the spirit of the institution;
sometimes epigrams and quotations regarding
qualities that foster or hinder success are
given, for example:
" The Most Contagious Disease in the World
is Xot the Grippe — it's the Grouch."
"A Slap on the Back Beats Two in the Face."
" There's one sure way of getting more money
— and that's to do more and better work each
day"
Most of the cards are attractively illus-
trated in colors. Some are designed simply to
provoke a smile that will relieve the tension
of prolonged application to an exacting
operation. Others point out the usefulness
and importance of the work in supplying the
world's needs, or dwell on the greater earnings
that the employee can make by guarding
against breakage and shrinkage. Good sug-
gestions are ver\- frequently furnished by the
factor}- executives and foremen.
A census taken in one of the lamp factories
showed that out of 4 1 7 people engaged in mak-
ing electric lamps in this particular plant, only
1S9, or 45 per cent, had electric light in their
homes. This fact ser\-ed as a basis for a series
of cards of which the following is a sample :
"For home lighting, electricity costs much
less than many people think. You can bum a
15-watt Mazda lamp seven minutes for less
tnoncy than it costs to light a match. And think
how much safer, cleaner and more convenient !
" If renters would only insist on living in
wired houses, landlords would quickly wire
them."
" Do they pay? " is a question that naturally
suggests itself regarding the "silent spokes-
SILENT SPOKESMEN IN THE FACTORY
231
men in the factory." The answer, viewing the
matter from either the theoretical or the prac-
tical standpoint, is "Yes." The accompany-
ing diagram, indicating the great saving of
expense in "breaking in" new employees,
through lengthening the average term of
em]:)loyment, shows one reason why it pays.
A numerical example will show e\-en more
clearly what is meant.
Assume that the average cost of "breaking
in" a new employee is $100, which is a con-
servati^'e figure in many industries when all
factors are taken into consideration. In a
shop containing 300 employees, whose average
length of stay is 10 months (which is a typical
condition in industries where female operators
are in the majority), the monthly cost of
"breaking in" new operators will be .13000. A
little figuring will show that on any plan
which will increase the average period of
service by even 2 per cent, the management
will be justified in spending $58 per month.
The cost of the placarding system, when con-
ducted on a large scale and spread over a large
number of factories, is much less than this.
A test of this advertising service was made
in a certain department employing over lUO
male (union) operatives, with the object of
reducing, if possible, the number of un-
explained absences from work. Placards
were featured in which the subject was tact-
fully handled, with the idea of goodwill
uppermost. A check-up of time-cards for the
month preceding and the month following the
installation of the advertising showed a drop
of over 25 per cent in the percentage of
absences.
Placards and posters are not the only
mediums of internal information that can be
used to good purpose. The fact that they are
kept in sight of the operator more continu-
ously than other mediums, such as pay-
envelope enclosures, is one strong point in
their favor. The employees' house-organ,
particularly if it be carried into the homes,
can be made very effective.
While we engineers are improving the effi-
ciency of machinery and mechanical equip-
ment of all kinds, we shall do well to recognize
the importance of promoting mutual under-
standing and goodwill with the people in our
plants — the human element on whom we
depend for the translation of engineering into
commercial output.
232 March, lii2()
GENERAL ELECTRIC REVIEW
Vol XXIII, \o. 3
A Biographical Sketch of the Late
William Olney Wakefield
William Olney Wakefield was born on the
2nd of January, IS41 , in the town of Gardiner,
Maine. In his youth he had a varied experi-
ence in different trades, from shoeing horses
(his father being a blacksmith) to working
in the local paper mills, and the building of
boats, which was then an important industry
on the Kennebec River.
William Olney Wakefield
He finally was apprenticed as a millwright,
which in those days embraced nearly e\-er>-
phase of the mechanical trade from pattern-
maker to tinsmith.
After serving his apprenticeship he entered
business in Boston as a hydraulic engineer
and also operated a machine shop for several
years. It was there in 1S7() that he invented,
patented and manufactured the first water
motor, which was installed to operate a large
coffee-mill in the window of Cobb, Bates &
Yerx's Boston store. This red coffee mill
was for years a landmark at the corner of
Kneeland and Washington streets. He also
designed a hydraulic engine on which he
secured a i)atent in 1877.
A further de\-elopmcnt of his water motor
resulted in its application to blowers for
church organs, many being installed by
Mr. Wakefield throughout New England;
notably those in Trinity Church, and the
Holy Cross Cathedral, Boston.
It was while engaged in this business that
he became acquainted with the late Henry
A. Pevear, which acquaintance resulted in
his going to New Britain, Conn., in 1882,
where in the pioneer da}-s of electrical de\elop-
ment he worked with Professor Elihu Thom-
son and Mr. E. W. Rice, Jr. Thus he became
the first draftsman employed by the Thom-
son-Houston Electric Comjjany now known as
the "General Electric Company." In the
year 1883 the Thomson-Houston Electric
Company moved to Lynn, Mass.. with Mr.
Wakefield as the chief and only draftsman.
It is interesting to note that he personally
made Drawing No. 1, which is still in the
files of the Company.
Mr. Wakefield was a m.an of keen percep-
tion and always ahead of the times. He fore-
saw with rem.arkable accurac\' the great
future before the infant electrical industr\-.
and laid the foundation upon which was
built the Company's present system of
making, mmibering and cataloging drawings.
Through his persistence a standard nomen-
clature was adopted for the parts of machines
as well as the machines themselves. The
whole system, a revolution from general
practice, has since marked in a peculiar way
the Company's drawings, resulting in a
standard known the world over as "General
Electric."
In 185)4 Mr. Wakefield came to the Sche-
nectady Works and although relieved of
many of the onerous duties of Chief Engineer
of the Drafting Department, retained the
position and title of Chief Draftsman until
his death. In recent years he had his own
machine sho]i in Building No. 4, where every
op])ortunity was afforded him to develop his
mechanical ideas. In this work he seemed to
specialize in machines for the Blueprint
Department, which department under his
suj^erAMsion grew rapidh- from the day he
made the first blueprint to the day of his
death, November 4, 101(1, on which day 10.S3.">
])rints were produced.
Mr. Wakefield was an American patriot, a
"down East Yankee" of the old-fashioned
ty]ie. As an abolitionist, he enlisted in the
EDISON'S BIRTHDAY COMMENTS ON WORK
233
War of the Rebellion and became a private
in the 16th Maine Infantry. He was severely
injured on the field of battle and crippled for
some years. He was reticent, however, about
relating his war experiences and very few
except his closest friends knew of the result-
ant physical infirmities, which deprived him
of the sight of one eye and left one side of his
body nearly paralyzed.
Not only was Mr. Wakefield a mechanic
and a soldier, but he was also a deep student
of political economy. Of a literary turn of
mind, he nimibered among his acquaintances
such men as Sylvester Baxter, the American
poet and author; Robert Creelman, of news-
paper fame, and Arthur Brisbane, the great
editorial writer. He also enjoyed the per-
sonal friendship of Edward Bellamy.
Mr Wakefield for a number of years was a
member of the one-time famous "Cold Cut
Club" of Boston, which numbered among its
membership some of New England's most
noted men of science, art and literature.
Mr. Wakefield was active in politics,
invariably leaning to the progressive and even
the radical theories.
His religion was that of the "Brotherhood
of Man and the Fatherhood of God." He
took a deep interest in the personal welfare
of his associates and was often a last resort
in counsel. He had little patience with
orthodoxy, although numbering many of
the clergy among his circle of friends. His
intimates recognized in him a character of
rare achievements, a philosopher, a man
wonderfully versatile, in a word "unique."
He was a member of the Bay State Lodge,
Independent Order of Odd Fellows of Lynn,
Mass., and the General Electric Quarter
Century Club. His remains are interred in
the beautiful Forest Hills Cemetery, a suburb
of his beloved citv, Boston.
Edison's Birthday Comments on Work
In celebrating his 73rd birthday, February
1 1th, Thomas A. Edison made some comments
on the value of work that all of us could
cogitate to advantage. Mr. Edison's capacity
for work has been the subject of wonder the
world over; and while he is not opposed to
the eight-hour day for his fellow workers,
imagine how he would have chafed had the
working of such a ruling restricted his activ-
ities in his younger days! He does not be-
lieve that a young man should tie his hands
by limiting his efforts by the time clock. Mr.
Edison's birthday comments follow in part:
"I'm glad that the eight-hour day had not
been invented when I was a young man. On
my birthdays I like to turn for a moment and
look backward over the road I have traveled.
Today I am wondering what would have
happened to me by now if fift}' years ago some
fluent talker had converted me to the theory
of the eight-hour day and convinced me that
it was not fair to my fellows to put forth m\-
best efforts in my work.
"This country would not amount to as much
as it does if the young men of fifty j'ears ago
had been afraid that they might earn more
than they were paid. There were some shirk-
ers in those days, to be sure, but they didn't
boast of it. The shirker tried to conceal or
excuse his shiftlessness and lack of ambition.
"I am not against the eight-hour day or
any other thing that protects labor from
exploitation at the hands of ruthless em-
ployers, but it makes me sad to see young
Americans shackle their abilities by blindlv
conforming to rules which force the indus-
trious man to keep in step with the shirker.
If these rules are carried to their logical con-
clusion, it would seem that they are likely
to establish a rigid system of vocational classes
which will make it difficult for the working-
man to improve his condition and station in
life b\- his own efforts.
"Of course, I realize that the leaders of
union labor have their political problems and
that the}^ must appeal to the collective
intelligence of their followers, which is lower
than the average individual intelligence of
the same men, but there ought to be some
labor leader strong enough and wise enough
to make trades unions a means of fitting their
members for better jobs and greater responsi-
bilities. I wonder if the time will ever come
when the unions, generally, will teach their
members how to be better workmen, and
train the ablest and the most ambitious to
become bosses and employers. If that time
ever does arrive trade unionism will be one
of the world's greatest forces in social prog-
ress, and I think there will be a much better
understanding between capital and labor.
"I hope I may have enough birthdays to
enable me to witness something of that kind.
I feel like it now. Inasmuch as the pro-
hibitionists have buried Johnny Walker under
the Eighteenth Amendment and he has no
further use for his trade mark in this country,
I'll borrow it and say 'I'm still going strong.'"
234 March, 1920 GENERAL ELECTRIC REVIEW Vol. XXIII, Xo. 3
Question and Answer Section
Beginning with the May issue, we will resume the
Question and Answer Section of the GENERAL
ELECTRIC REVIEW, which was discontinued in 1917,
at the commencement of war.
This section provides a valuable service to our readers
in making available to them the consulting service of a
large corps of engineering experts, and in publishing
Questions and Answers of general interest and educa-
tional worth.
Address your inquiries to Editor, Question and Answer
Section, GENERAL ELECTRIC REVIEW, Schenectady,
N. Y.
TWO DOLLARS PER YEAR
TWENTY CENTS PER COPY
GENERAL ELECTFIC
REVIEW
VOL. XXIII, No. 4
Published by
General Electric Company's Publication Bureau,
Schenectady, N. Y.
APRIL, 1920
* ^■^ '■^'4^vi'*c! /^^syxi^
^^»ifc'^
Mallet Type Freight Locomotive Formerly Used on the Mountain Divisions of the C. M. 86 St. P. R. R.
and the New Electric Passenger Locomotive for the Cascade Division o." ^he same Railroad
A SPECIAL ISSUE ON
ELECTRIC TRACTION
For
Fractional H. P. Motors
TF the price that is a little higher, for
-*■ the machine that is a little better, buys
a stand-up-ability that adds a year to its
useful life is there any ground for objec-
tion to the higher price? "noRfr\a" Bear-
ings are standard equipment in hundreds
of thousands of electrical machines that
are leaders in their class their leadership
being based upon their ability to do more,
last longer, earn more. Are your ma-
chines leaders in their class?
Sie that your Motors
are "NORmfl" Equipped
THE m^mm/^ c^mf/^my
ILs'iM] Iglaifii^ €i^
Ball, Rollei-, Thrust and Combination Bearings
General Electric Review
A MONTHLY MAGAZINE FOR ENGINEERS
Manager, M. P. RICE
Editor. JOHN' R. HEWETT
Associate Editors. B. M. EOFF and E. C. SANDERS
In Charge of Advertising. B. M. EOFF
Subscription Rates: United States and Mexico, $2.00 per year; Canada. $2.25 per year; Foreign, $2.50 per year; payable in
advance. Library and Student Rates: United States and Mexico. $1.50 per year; Canada. $1.75 per year; Foreign, $2.00 per year;
payable in advance.
Remit by post-of&ce or express money orders, bank checks, or drafts, made payable to the General Electric Review, Sche-
nectady. N. Y.
Entered as second-class matter, March 26, 1912, at the post office at Schenectady, N. Y., under the Act of March, 1879.
Vol. XXIII, No. 4
Copyright. 1020
by General Electric Company
April, 1920
CONTENTS
Frontispiece : The Speed King of the Rails
Editorial : Electricity Opens Wide the Door to Advancement
Summary of French Mission's Report on Railway Electrification
By A. Mauduit
Railway Electrification in the Super-power Zone
By W. B. Potter
The Last Stand of the Reciprocating Steam Engine
By A. H. Armstrong
Electrification of the Coast and Cascade Divisions of the C, ]\I. & St. P. R
By E. S. Johnson
Passenger Locomotives for Chicago, Milwaukee & St. Paul Railway'
By A. F. Batchelder and S. T. Dodd
Control Equipment of the New Locomotives for the C, M. & St. P. R
By F. E. Case
New Type of High-speed Circuit Breaker
By J. F. Tritle
Power-limiting and Indicating System of the C, M. & St. P. Ry.
By J. J. LiNEBAUGH
Electrification of the Hershev Cuban Railway ....
Bv F. W. Peters
Summary of High-voltage Direct-current Railways .
By W. D. Bearce
Control for 1200 and 1500-volt Car Equipments
By R. S. Beers and C. J. Axtell
The Public Trusteeship of the Boston Elevated Railway .
By Edward Dana
Operating Costs of Various Types of City Cars ....
By J. C. Thirlwall
Motor Busses or Trackless Trollevs
Bv H. L. Andrews
Improvements in the Design and Construction of Railway Motors
By E. D. Priest
Importance of Simplicitj' in Locomotive Design . . . .
By A. F. Batchelder
Modern Devices and Control for Automatic Railway Substations .
Bv Cassius M. Davis
Page
236
237
239
246
249
263
272
278
286
292
307
313
314
319
326
331
335
340
342
s
u
0
2
a
u
c.
y:
HI
X
■
3
a.
c
z
EML ELECTODC
ELECTRICITY OPENS WIDE THE DOOR TO ADVANCEMENT
Electrification is the open sesame to better
]ierformance and to achievements otherwise
unattainable. Upon its ability to open the
door to these accomplishments, we wish to
la>' particular stress in introducing this
special electric railway issue of the Review.
Electricity for traction purposes made its
debut when it displaced the horse car and
the cable car. Even in those early days,
when comparatively little was known of its
workings and the equipment was crude,
electricity rendered a better performance
than did the motive powers it replaced.
Its promise of a brilliant future attracted
engineering talent which, in the years that
followed, has so developed our city trollev
system that its service is no longer capable
of duplication by any non-electrical method.
When the increasingly heavy traffic in our
larger cities outgrew the capacity of surface
lines, on account of the limited space avail-
able in the streets, the elevated railway came
into being as a parallel transportation system.
In the earlier of these installations, steam
locomotives were employed. This method of
operation was comparatively short lived,
however, because rapid progress in the devel-
opment of electric traction equipment made
evident the advantages of electrification in
this new field. The resulting electric oper-
ation produced not only an improved service
but one incapable of accomplishment by
steam. The same characteristics are true of
electrification in subways.
The interurban trolley system was designed
to furnish a service where none existed before.
Its electrification was therefore a feature of
installation — not one of substitution for prior
motive power. The development of this type
of transportation has been so successful as to
discourage any ambition of the steam engi-
neer to produce an equivalent service in this
field.
The initiation of electrification in the steam
traction field took place at railroad tunnels
and terminals. Increasing traffic had resulted
in the steam locomotive smoke becoming so
dense as to limit operation in tunnels and to
be declared a public nuisance at terminals.
In both these locations, the limitation of
trackage had resulted in such congested
traffic conditions as could be relieved only by
a radical change in the motive power em-
ployed. Electrification was the logical solu-
tion to these difficulties ; and the record made
by all the resulting installations has unequiv-
ocally established the fact that electricity is
the only tractive power capable of meeting
all the requirements.
The experience gained in the operation of
this type of electric traction paved the way
for the next progressive step — the electrifi-
cation of main line divisions. When congested
traffic conditions demanded the actual taking
of this step, mountain divisions were the first
to be electrified because the long grades and
sharp curves in these sections had already
taxed steam locomotion about to the limit
of its capacity. The operation of the equip-
ments developed for this service has conclu-
sively demonstrated that the performance of
electric traction is better than that of steam
under these exacting haulage requirements.
In fact, the operating data of the later installa-
tions indicate that the art of electrification has
already entered the stage wherein its applica-
tions produce results impossible of attainment
by steam motive power.
Now that the substitution of electricity
for steam has surmounted the greatest
difficulties experienced in heavy traction,
namely, the handling of traffic at terminals
and the haulage through tunnels and over
mountains, there no longer remains any
engineering obstacle to the electrification of
entire steam railwav svstems. E. C. S.
u
<
239
Summary of French Mission's Report on
Railway Electrification*
By A. Mauduit
Secretary of the Mission and Professor of the Faculty of Science of the
University of Nancy (France)
Destruction of French coal mines by the invading Germans and intensified production during the war so
depleted the coal resources of France as to force the Government to take active steps toward providing for
future requirements. Because electrification of the steam railroads would relieve the situation immensely, a
commission of experts was sent to America for the purpose of studying our systems of railroad electrification,
comparing them with those employed in Evirope, and making recommendation as to the best system to install
on French railroads. The commission has completed its work; and M. Mauduit, Secretary of the Mission,
has prepared the following summary of its activities. It is of particular interest to note that he states that
lie "does not hesitate to formally conclude in favor of the adoption of this [high-voltage direct-current) system,
and he believes it to be actually the only system suitable for the electrification of heavy traffic lines." — Editor.
The Minister of Public Works (France)
formed, by the resolution of November 14,
19 IS, in accordance with the upper chamber
of Public Works, a commission of sttidents
charged to examine the propositions stibmitted
by the railway systems of the Paris-Lyons-
Mediterranean, the Orleans, and the Midi
for the electrification of approximately lU.UUU
kilometers of the lines of their systems.
This committee, composed of the most
qualified technical men of the administration
and of the railway systems, believed that
it was necessary to propose to the A^linister to
send to the United States a commission of
engineer specialists, instructed to obtain all
information relative to the recent progress
of electric traction.
Organization and Composition of the Mission
The mission was comprised of thirteen
members as follows:
Major D'Anglards, and Professor A. Mauduit of
the faculty of Sciences of the University of
Nancy, attached to the Administration of
Railways, delegates of the Ministry of Public
Works and Transports.
M. Pomey, Chief Engineer of the Post and Tele-
graph, and M. Lecorbeiller, Engineer, deleg-
ated by the Administration of the Post and
Telegraph.
M. Debray, Chief Inspector, and M. Barillot,
Inspector, delegates of the State Railways.
M. Sabouret, Chief Engineer, attached to the
Administration, M. Balling, Principal Main-
tenance Engineer of the line, and M. Parodi,
Chief Engineer, all three delegated by the
Orleans Railway.
M. Japiot, Chief Engineer of material, and M.
Ferrand, Chief Engineer of the central
maintenance service, both delegated by the
Paris- Lyons- Mediterranean Company.
M. Bachellery, Chief Engineer attached to the
Administration, and M. Leboucher, Principal
Engineer of Motor Power, delegates from the
Midi Railway Company.
* Translated from the Journal Ojficiel De La Retyublic Francais,
August 13. 1919.
The greater part of the members of the
Commission left Paris the 15th of April for
America and returned to Paris the 22nd of
July, 1919.
Itinerary and Work of the Mission
Arriving at New York on April 2.5th, we got
in touch with the representatives of different
construction companies, manufacturing com-
panies, and railway companies, and visited
the following electric railways :
New York Central; direct current 000 volts with
third rail.
New York, New Haven & Hartford; single-phase,
at 11,000 volts, 25 cycles.
Pennsylvania Railroad and Long Island; direct
current, 600 volts, with third rail.
Suburban Lines; that carry a considerable freight
traffic.
We also visited a certain number of steam
central stations for electric power; the Inter-
borough Rapid Transit Company and the
New York Edison Company, together with
the Hydro-electric Central Station of Niagara
and Steam Central Station at Buffalo.
From May 8th to 10th we made a visit to
the works of the General Electric Company
at Schenectady, New York, and discussed
with the principal engineers of that Company
questions concerning railway electrification
in general and particularly the electrification
with high-tension direct current (.3000 volts)
of the Chicago, Milwaukee & St. Paul (710
kilometers in operation) installed by the
General Electric Company.
From the 11th to the 25th of May, we
visited the following installations ;
Electrification of the Norfolk and Western Rail-
way; single-phase, 11,000 volts, 25 cycles, from Blue-
field to V'ivian, Virginia.
Electrification of the Pennsylvania Railroad; sin-
gle-phase, 11,000 volts, 25.cycles, from Philadelphia
to Paoli.
240 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
Washington, Baltimore and Annapolis Electric
Ry.; direct current, 1200 volts (Interurban).
The Baldwin Locomotive Works at Philadelphia.
The repair shops and factory of the Pennsylvania
R. R. at Altoona.
From May 25th to 2Sth we visited the
factory of the Westinghouse Mfg. Co., at
Pittsburgh, and discussed with their engineers
the subject of electrification in general, and
particularly the single-phase and single-to-
three-phase systems installed by that Com-
pany and also the new direct-current 3000-
volt locomotives for the extension of the
electrified portion of the Chicago, Milwaukee
& St. Paul Rw-]*-.
From May 29th to June 4th different visits
were made to the electric locomotive factory
of the General Electric Company at Erie, Pa.,
to automatic railway substations for 600-volt
direct current, both of Westinghouse and
General Electric designs, and to the Chicago,
Lake Shore, and South Bend Electric Railway
single-phase 6600 volts 25 cycles.
From June 5th to June 14th a complete
study of the Chicago, Milwaukee & St. Paul
Railway was made, including the sections in
operation from Harlo^\i:on to Avery (710-
kilometers. Rocky Mountain and Missoula
Divisions), the section in the course of con-
struction from Othello to Tacoma and
Seattle (360-kilometer Cascade and Coast
Divisions) all direct current at 3000 volts, and
the repair shops and supply depot at Deer
Lodge, Montana.
A visit was made to the three hydro-
electric stations of the Montana Power Com-
pany, which furnish the three-phase 100,000-
volt 60-cycle current to the electric railroad.
The stations were the Rainbow (35,000 kw.),
the Great Falls (48,000 kw.), the Holter
(48,000 kw.) all three on the Missouri River.
The following installations were also studied :
Central California Traction Company; the line
from Stockton to Sacramento (72 kilometers) in
California, equipped with the inverted third-rail
for 1200 volts direct-current, a unique American
example of the application of the third rail to rather
high voltage.
The Pacific Electric Railway system; suburban
and interurban lines around Los Angeles, from 600
to 1200 volts direct-current.
The hydro electric station' of the Puget Sound
Light & Power Company on the White River,
near Seattle, Washington; 48,000 kw., 55,000 volts
with a head of 130 meters.
The principal incoming substation of the Utah
Power & Light Company at Salt Lake City, Utah;
outdoor substation at 120,000 volts 25,000 kw. with
a regulation by synchronous condensers (located
inside a small special building).
The Great Western Power Company of San
Francisco, California; their hydro-electric plant
at Los-Plumas, California, on the Feather River
with a head of 138 meters, 65,000 kw., 115,000
volts and a double line on unique poles for 246
kilometers to the incoming substation at Oakland.
The oil burning steam central station of the
Pacific Gas and Electric Companv at San Francisco;
57,000 kw.
The Southern California Edison Company of
Los Angeles; the two hydro-electric plants nearby
at Big Creek in the Sierra- Xevadas, each of 25,000
kw., 600 meter head, 150,000 to 160,000 volts and
the two 400-kilometer lines made of aluminum and
steel on separate towers; the incoming substation at
Eagle Rock, near Los Angeles, 150,000 volts, with
regulation by synchronous condensers.
Apart from the general duty of the Mission,
consisting of collecting all useftil documents
on the electrification of railways and the
distribution of electric energy- at high tension,
the principal duty was to find out. on sum-
ming up all the information gained by the
study of the Swiss and Italian Electric Rail-
ways on one side and the American on the
other, if a system of electric traction existed
for large systems distinctly superior to all
others and able to be adopted to the exclu-
sion of all others by all the different compa-
anies interested for the projected electrifica-
tion in the center and the south of France.
From the fovu^ systems of electric traction
actually in operation on great lines of the
world, that is, the single-phase, three-phase,
single-to-three-phase, and high-tension di-
rect-current, the three-phase has already
been studied in detail in Italy, where it is
largely used, while it is not used to any
appreciable extent in any other country-, and
the single-phase has been equally studied in
operation in France on the ^Iidi Railway and
in Switzerland on the Loctchberg Lines and
in construction on the Swiss Federal Rail-
ways which have adopted this system for the
gradual electrification of all their systems,
the electrification actually intended and even
in the course of construction for the Gothard
Railway.
The single-to-three-phase, and the high-
tension direct-current systems are used only
in America, and so became the principal
object of the work of the mission. At the
same time, the examination of American
single-phase installations (25-cycle, while
the analogous French installations are 16-
cycle.) allowed the completion of the study
of monophase installation.
The total information of all kinds gathered
in America forms the subject of a detailed
report by M Mauduit. This report was
submitted at the October, 1919, meeting of
the Technical Sub-Commission in order to
SUMMARY OF FRENCH MISSION'S REPORT ON RAILWAY ELECTRIFICATION 2-11
serve as a basis for the discussion of a pro-
position tending to make a choice of a traction
system, different for the individual com-
panies but following a general formula
established by this Sub-Commission with the
approval of the whole committee.
The purpose of this summary of the report
is to give only the most important results and
the principal impressions obtained from the
American experience, together with the per-
sonal conclusions of the writer. The docu-
ments have been gathered by all members of
the commission, perhaps together and perhaps
separately, but the opinions expressed in this
article, while they are in general the concensus
of the general impressions of the Mission, are
personal opinions and only bind the writer,
since they have not been approved by the tech-
nical sub-commission in the presence of all the
members of this commission, called before this
commission to complete and discuss them.
Monophase Electrification
The principal lines equipped with mono-
phase current are the New York, New Haven
& Hartford Railroad and the Pennsylvania
Railroad, from Philadelphia to Paoli.
x'\lthough these lines are suburban lines,
they are interesting to study since the system
of traction employed is applicable to larger
lines, and the same as that of the French
Midi road save that the frequency is 25
cycles instead of 16.
New York, New Haven & Hartford Railroad
The electrification of this system was
decided upon in accordance with the order
of New York State; it has a total of 102
kilometers electrified and takes in a part of
the direct-current inverted third-rail system
in the common terminal with the New York
Central Railroad, when leaving New York.
The outlying part is 11,0UU volts single
phase, with an overhead trolley wire. The
necessity of operating partly on (JOO ^'olts
direct-current and partly on 11,000 volts
single-phase greatly complicates the equip-
ment of the locomotives which must run into
the city of New York.
The traffic is important and the technical
operation adequate after many difficulties of
the first years were surmounted. These
difficulties mainly consisted in struggling
against accidents due to the frequent short
circuits on the trolley wire, or on the power
feeders, and against the interference set up
in the telegraph and telephone lines adjoining
and belonging to different companies.
The solution of these problems has been
found, but at the price of complicated
organization, delicate and costly to install
and maintain. The telephone lines have been
put underground in lead covered cables: the
distribution of power has been made at
22,000 volts by means of 30 compensating
auto-transformers, spread over the 102 kilo-
meters of the road to lessen the height of the
voltage surges in the line, and to reduce the
interference on the telegraph and telephone
lines. This installation is in place of the
transformers for this work on the Midi road
with the additional advantage of the reduc-
tion of the voltage.
The equipment includes 103 locomotives
and 26 motor cars; the cost of maintenance is
comparatively high and the personnel of the
repair shops quite numerous. The single-
phase motors are very delicate and require
ver^' careful watching of the commutator.
Pennsylvania Railroad
The lines from Philadelphia to Paoli are 32
kilometers with four tracks and from north
Philadelphia to Chestnut Hill are 20 kilo-
meters with two tracks.
The equipment includes only motor cars,
not locomotives, and the service varies from
suburban t^^pe to heavy traffic. The techni-
cal operation is good, the motors are not
required to operate on both direct current and
monophase current, are of a more modern type,
and possess better commutation.
Special precautions have been taken to
prevent short circuits, and the struggle
against interference on the telegraph and
telephone lines has been solved after a
fashion: (1) by placing these lines in lead
covered cables underground; (2) by the use
of frequent feeder transformers (5 for the
52 kilometers of roadj ; and (3) by the use of
track transformers placed along the track at
very short distances, approximately one kilo-
meter.
Under normal conditions, the operation
of the signal lines is adequate, but short
circuits although rare produce important
disturbances. A very interesting preparatory
register connected on an extra wire, placed
in a cable, permits the control at any moment
of the interference voltage induced in the
telegraph and telephone lines.
The American monophase traction instal-
lations, especially on account of the high
frequency adopted (25 cycles instead of the
16 cycles in Europe) a frequency which was
imposed by the local conditions in order that
242 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
the numerous distribution systems at this
frequency might be directly utilized, and the
employment of motors often not quite so good
as those which have been found on the Midi
and in Switzerland, showed an installation
less perfected than the similar installations
in Europe.
At the same time the struggle against
interference with telephone and telegraph
lines has been carried to a considerable per-
fection and there will certainly be a con-
siderable discussion in the large report of
this system of traction in France, if it is
adopted. On the other hand, the trolley lines
with catenary suspension are remarkably well
made.
If we assemble now the experience of
France, Switzerland and America, we are
forced to conclude that the monophase system
is still far from the point of presenting the
solution to a ntunber of problems insuffi-
ciently solved in actual practice, notably the
production of a motor capable of exerting a
heavy torque for a considerable time without
rotating, in order to be able to start heavy
trains on the important grades, and of
regenerative braking.
Furthermore, this system leads to com-
plicated equipment for the protection of the
neighboring telephone circuits, which con-
siderably augments the cost of installation.
Without this consideration the cost would be
distinctly less than similar costs with the
three-phase and high-tension direct-current
systems.
The expenses of maintenance of the rolling
stock are always higher than in the latter
two systems and the motors are less rugged
and capable of less overload.
Single-to-three-phase Electrification
In the single-to-three-phase system which
the American calls split-phase, the power
is furnished to a single contact wire as
in the monophase with the return by the
rails as in the single-])hase form, but it is
transformed in the locomotives by means of
a special converter to three-phase power and
the motors used with this last locomotive are
three-phase induction motors. The aim of
this installation is to jjrofit from the single-
contact wire of the monophase system (while
the Italian three-phase requires two trolley
wires in addition to the rail .serving as a
return) and from the three-phase induction
motor, rugged and economical and capable
of exerting heavy torque for several minutes
without rotation and of pulling the heaviest
trains which, up to now, has not been ob-
tained with the ordinarv- commutator mono-
phase motor.
There onh' exists at this time one line
operating with this system. It is the line
from Bluefield to Vivian of the Norfolk &
Western Railway, in the Appalachian Moun-
tains in Virginia and West Virginia, for a
length of 48 kilometers with two or three
tracks, and numerous cur\-es and grades reach-
ing 20 millimeters per meter.
These locomotives are flexible and robust.
but their operation brings out different
mechanical and electrical faults which have
not been corrected up to this time in an
adequate fashion and, on account of which.
this installation may be considered to be as
yet only in the test period, and the mainte-
nance expense of the rolling stock is greater
than that of the other systems.
From the mechanical point of view the
transmission of motion from the motors to
the axles, which is made by "jack shafts" and
horizontal cranks, occasions rapid wear of
the bearings, and even a dislocation of the
frame or the breaking of the cranks, on
account of enormous forces de\-eloped at the
time of the vertical displacement of the frame.
From an electrical point of \-iew the
principal inconveniences are the following:
The three-phase power produced by the con-
verter actually is not perfectly s>-mmetrical.
and the phases do not have equal currents.
Furthermore, the rotors of the motors are
connected to different liquid rheostats, and
the loads are not always equally divided
between the different motors, very often with
considerable differences. A regulation of
loads by the engineer has been pro\-ided but
the latter, very busy, can only make sure of
a very imperfect adjustment, and the motors
consequently often deteriorate rapidly. The
power-factor is \-ery low. on account of the
presence of the induction converter which
adds its magnetizing losses to those of the
motors.
To remedy these defects, with the exception
of the distribution of the load between the
motors, the manufacturer is taking up at this
moment the use of a synchronous converter
to give a good ])ower-factor and to make the
three-]3hase current more s\Tnetrical: but no
practical application of this new apparatus
has been made yet, and it must be feared
that there will be very great instability on
the occasion of breaks in the trolley wire.
On account of the numerous repairs in
progress and of the lack of electric loco-
SUMMARY OF FRENCH MISSION'S REPORT ON RAILWAY ELECTRIFICATION 243
motives due to the war, the operation of this
portion from Bluefield to Vivian still requires
many steam locomotives. The Pennsylvania
Railroad is taking up on its own account a
single-to-three-phase application on the four
track line from Altoona to Johnstown on
the road from Philadelphia to Chicago. A
test locomotive is in the course of being tried
out, but no permanent installation has been
started on the road.
In conclusion, the single-to-three-phase
system, in which the principle at first glance
seems very interesting, and which supplies
an effectual assistance to the monophase
system by the emplo\-ment of locomotives
or motor cars with monophase only for the
express trains or light trains, and of loco-
motives single-to-three-phase for the heavy
and slow trains, all these locomotives being
supplied by the same trolley wire with mono-
phase current, is found to present in practice
numerous faults which have not yet been cor-
rected, and on account of which this system
has not come up to the hopes with which it
was regarded when started.
High-tension Direct-current Electrification
Already the 600-volt direct-current system
has been utilized for a long time, in a stan-
dard method for city and suburban electric
railways, either with a trolley wire for the
tramways, or with a third rail for the subur-
ban railways (the line of the Invalides to
\'ersaille and from Paris to Juvisy, of the
Metropolitan) .
In the United States, the greater part of
the interurban lines operate at 1200 volts
direct current with an overhead trolley wire.
A considerable number of these lines are
really railroads with both passenger and
freight traffic, and attain speeds of 60 to 80
kilometers per hour. Many of them were
originally equipped with single-phase current,
at voltage varying from 3000 to 6600 volts,
but have been made over for direct cturent
at 1200 and 1500 volts. The equipment for
this latter voltage is now as standard as that
for tramways at 600 volts.
Encouraged by the excellent operation of
these installations the Americans have tried,
with like success, to raise the direct-current
voltage to 2400 volts, and have equipped in this
manner the mining line from Butte to Anacon-
* This second section of the C. M. & St. P. Ry. will have
been placed in electric operation before this article is printed.- —
Ed.
t These breakers are described in the article "High-speed Cir-
cuit Breakers for the C. M. & St. P. Electrification," by C. H.
HiU. General Electric Review. September, 1918. and also in
the article. "A New Type of High-speed Circuit Breaker," by
J. F. Tritle. in this issue. — Ed.
da of the Butte, Anaconda & Pacific Railway
(Montana), 53 kilometers of main track.
Following this they have executed, at 3000
volts, the electrification of the world, from
Harlowton to Avery, 710 kilometers, main
track across the Rocky Mountains and the
Missoula region on the Chicago, Milwaukee
& St. Paul Railway.
The electrification of the second section of
360 kilometers, between Othello and Tacoma,
Seattle, as far as the Pacific, is in the course
of construction*, and that of the portion
comprised between Avery and Othello, about
the same distance, has already been decided
upon.
We studied with particular care this
installation of the Chicago, Milwaukee and
St. Paul, and all the members were unanimous
in considering that this electrification, by far
the most important in the world, was at the
same time greatly superior to all the others
on account of the excellence of its technical
operation from all points of view.
The electric power is furnished by the
Montana Power Company and is three-phase
100,000 volts. It is transformed to direct
current in rotary substations consisting of
motor generator sets, which are composed of
a synchronous motor and two generators for
direct current, mounted on the same shaft
and coupled electrically in series in such a
manner that each produces only 1500 volts
on the commutator.
These substations are the most delicate and
most expensive part of this traction system,
but they are only to the niunber of 14 for the
710 kilometers (about 1 every 50 kilometers)
and operate very excellently. They require
only a personnel of three men each, a chief
and two aides for continuous operation with
a power from 4000 to 6000 kw. By the use
of flash barriers on the comnuitators, and of
extra fast circuit breakersf in the main line,
accidents resulting from the most redoubtable
phenomenon of direct-current circuits, nameh\
the flash of fire on the commutators (flash
over) in the case of short circuit, have been
eliminated.
The excellence of the installation of these
substations counts for a great deal in the
success obtained by the high-tension direct-
current project.
At the relatively low tension, 3000 volts
on the trolley wire (in place of 11,000 to
15,000 volts for single phase) gives a cor-
respondingly great voliune of cturent to
obtain the pull on hea\n,- trains. Experience
has shown that with a double trollev wire
244 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
and a pantograph trolley with a double shoe
and quadruple contact a current of 1500 to
2000 amperes is easily obtained at a speed
of 80 to 96 kilometers per hour, and 4000
amperes at a speed of 25 kilometers per hour,
which is more than sufficient for the heaviest
trains and the greatest powers.
The locomotives are very easy to run and
operate perfectly, the series direct-current
motor being of all others the ideal motor
for traction work as has long been shown by
the experience of tramways and suburban
railways. They are capable of regenerative
braking, marvellously regulated, which assures
the most flexible progress on down grades
and occasions an important economy of
the power, the tires of the wheels, and the
brake shoes. A single armature winder -Rath
an assistant assures the operation of the
336 motors of the 42 locomotives in the
service, while the former storage of steam
locomotives at Deer Lodge, corresponding to
360 kilometers of line, is sufficient for the
installation of the storage of the electric
locomotives and the repair shops for the total
distance electrified, which is 710 kilometers.
A single locomotive is sufficient to pull
passenger trains of 900 to 1000 tons American,
even on grades of 20 millimeters per meter.
Freight trains of 2800 American tons are
pulled by a single locomotive on grades of
10 millimeters (the tractive effort is then 32.8
metric tons) and by two locomotives on
greater slopes. The average weight carried by
freight trains is about 1900 American tons. In
trains pulled by two locomotives, the second
machine is placed in the middle of the train
and not at the end. It must be said further-
more that the break-up of a train is not
feared in America as all the freight trains, like
the passenger trains, are provided with an
automatic air brake on every car.
A considerable advantage of the direct-
current system is that it does not seem to
have any but the slightest interference with
the telegraph and telephone lines, in fact
insignificant. We are well able to report that
one may telephone very easily on the service
lines of the railroad placed all along the tracks
on an aerial wire without any protection.
A multiplex printing apj)aratus for the
telegraph service, worked between Spokane
and Helena with an earth return, was diverted
especially for us in such a fashion as to use a
wire placed on the poles of the electric rail-
road for a distance of 270 kilometers. This
operated perfectly during eight days without
even being troubled by three short circuits
made very complete intentionally between
the trolley wire and the rail in the course of
the telegraph wire.
In spite of the loss of energy due to trans-
formation of three-phase current to direct
current in rotary substations, operating con-
tinuously, and although the load is compara-
tively light, that is, two passenger trains and
three to four freight trains in each direction
per day, the efficiency of the system is good;
27 watthours per metric ton kilometer which
corresponds to an over-all efficiency of 50 per
cent from the point of purchase from the
producer up to the point of consumption.
Conclusions Relative to the Choice of an Electric
Traction System
Oil account of the remarkable results obtained
by the Chicago, MilwOiUkee & St. Paul Ruy. with
3000-volts direct current, the writer does not hesi-
tate to formally conclude in favor of the adoption
of this system, and he belirces it to be actually the
only system suitable for the electrification of heavy
traction lines.
It is possible that with the single-phase
system, which at the first glance shows the
advantage of lending itself to a great variety
of combinations, satisfactor]*- operation ma\-
some day be obtained, but it is, without
doubt, the fact that the actual practice is far
from being desirable at this time.
Direct current presents the incon\-enience
of being a little more expensive in first cost,
on account of the rotar\- substations required
to transform the 50-cycle three-phase current
generally produced in the (French) central
stations. Nevertheless, it must be said. that
to obtain economy in this regard with the
monophase installation, it is necessary- to
generate directly the single-phase current
at a low frequency (16 cycles) by means of
special electric generating groups, so that if
it is wished to utilize the current produced
normally by the central station (three-phase
at 50 cycles) it is necessary to go back to the
rotary transformation, the same with the
single phase as with the direct current. From
this point of view, the direct current offers
the advantage of being able to use the current
of any station under the same conditions.
So far as the expense of operation is con-
cerned, the complete and exact calculations
compiled by the engineering ser\-ices of the dif-
ferent companies could only show the compari-
son between the different systems; the writer,
nevertheless, estimates that the dilTerenco
would not be great, and would not come into
consideration in the choice of the svstem.
SUMMARY OF FRENCH xVIISSION'S REPORT ON RAILWAY ELECTRIFICATION 245
The almost complete absence of inter-
ference on the telephone and telegraph lines
constitutes for direct current a very con-
siderable superiorit}^ over the other systems.
We have not spoken of the three-phase
system, which, in America, has only an insigni-
ficant local application. Dispite certain
advantages obtained by the Italians, we
are of the opinion that it should be rejected
especially in consequence of the complexity
and of the high price of installation and
maintenance of the two trolley wires.
Economic Considerations of Electric Traction
From the economic point of view, the
papers which we brought back from America
are much less complete and less accurate than
the technical information.
On the other hand it is necessary, in judg-
ing from the American experience the future
economy of European electric traction, to
make considerable modifications in the figures
in the case of two principle items, which
differ in the American installation from the
European installation.
(1) In America, the coupling employed has
a strength against breaking of about
135 tons, and the tract i'.; efl^orts are
allowable up to 40 tons. In Europe,
the draw bars are of two models of
which the strengths are respectively
35 and 55 tons, and tractive efforts
are limited to ten tons (except 12 to
15 tons in Switzerland).
(2) In America, all the passenger and
freight cars are equipped with a com-
pressed air brake.
The result is that in America locomotives
two or three times more powerful may be
employed, with freight trains two or three
times longer and heavier than in Europe, and
that the personnel on these trains is relatively
much smaller which completely changes the
expense of operation.
The accurate calculations made by the
Companies, and above all the results of the
first electrifications installed and the con-
sideration of the exact prices of coal, can alone
show under what conditions electric traction
will be more economical than steam traction.
It is known, however, from another source,
that the economy will be mostly felt on the
lines having steep grades and heaw traffic;
and it is probable that for many lines differing
too greatly from these conditions, electric
traction will be more expensive than steam
traction.
Nevertheless the necessity, more and more
important, of economizing coal, and the great
advantages, which, it is well known, are
linked with electrification, render it neces-
sary that the most rapid construction of the
first works be carried out in view of the
gradual electrification of the most interesting
lines of the systems of the Paris-Orleans
Railway, the Paris-Lyons-Mediterranean Rail-
wav, and the Midi Railwav.
La Gare de Lyon, Paris
246 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 4
Rail^vay Electrification in the Super-power Zone
By W. B. Potter
Engineer Railway and Traction Department, General Electric Company
The proposal that a Super-power Zone be created in the section of the Atlantic seaboard lying between
Boston and Washington, and extending 100 to 150 miles inland, calls for the electrification of the railroads and
industrials in this congested district and for the installation of a comparati\-ely few exceptionally large electric
generating stations of the most modern type to furnish the necessary power. At the Midwinter Convention
of the A. I. E. E., New York, February 19, 1920, Messers. W. L. R. Emmet, J. F. Johnson, H. G. Reist, F. D.
Newbury, W. B. Potter, P. Torchio, P. H. Thomas, W. D. A. Peaslee, and A. O. Austin, all experts in their
fields, contributed to a symposium on the subject of the Super-power Zone. Under the leadership of Mr.
W. S. Murray, Chairman of the Traction and Transportation Committee, truly astonishing facts were brought
out. For example, the electrification of this district would raise the load-factor from 15 to 50 or possibly 60 per
cent, would enable one ton of coal to do the work of two, and would produce a 25 per cent saving annually on
the Zone investment cost. The following article is Mr. W. B. Potter's contribution to the symposium. His
figures for this district and those of Mr. Armstrong for the entire United States clearly demonstrate the advis-
ability of the Government, railroad operator, and electric locomotive manufacturer taking immediate steps
toward the general electrification of our railroads. — Editor.
T'
^HE suggested eco-
nomic system of
interconnected power
generation and distri-
bution throughout the
proposed Super-power
Zone, Fig. 1, should
adequately and ad-
vantageously provide
for the electric opera-
tion of the railways
within this zone, as
well as for the power
required for indus-
trial and other pur-
poses. The electrification of these railways
would insure not only a substantial reduction
in the amount of coal othenvise consiuned by
the steam locomotives but also a material re-
duction in the cost of maintaining the mo-
tive power units. Electrification would also
provide a more reliable service for all classes
of traffic and would be a welcome improve-
W. B. Potter
ment to the traveler as passenger trains would
be less frequently late, especially during the
winter. The colder the weather the greater
is the reser\-e power of the electric locomotive,
which is a much better characteristic than
that of the steam locomotive whose power
under similar conditions is correspondingly
diminished.
There are ntmierous illustrations of elec-
tric operation which are comparable to the
ser\'icc within the zone under consideration,
as well as man>- other examples of railwav
electrification throughout the country and
abroad, which afford conclusive evidence as
to the successful operation of railways with
electric power. In fact, a large number of
railwa>- electrifications are already embraced
within the limits of the proposed zone, and
while they do not represent a large propor-
tion of the total mileage, their traffic statistics
are available and can readily be studied as a
basis for determining the demands of the
whole area. A tabulation of these electrifica-
TABLE I
STEAM RAILROAD ELECTRIFICATION IN THE SUPER POWER ZONE
Railroads
Dateof
Electri-
fication
Route
Miles
Total
Mileage
of Track
No. of
Locoa.
Near
Motor
Cars
Baltimore & Ohio R. R. . .
1895
1905
1906
1906
1907
1910
1911
1912
1915
3.6
88.63
54.00
74.60
81.63
18.73
7.97
18.23
30.5
8.
218.
268.
150.26
527.49
97.49
21.50
54.41
116.3
9
0
73
0
106
33
7
1
0
0
Long Island R. R
N. Y. C. & H. R. R. R
\V. J. & Seashore R. R
477
221
109
N. Y., N. H. & H. R. R
Penn. R. R. (New York)
27
8
Bo-Jton & Maine R. R
0
N. Y., West & Boston
40
Penn. R. R. (Phila.)
115
Totals
377.89
1461.45
229
997
RAILWAY ELECTRIFICATION IN THE SUPER-POWER ZONE
247
tions shows that in this area there are already
three hundred and eighty (380) miles of
electric route, embracing 1450 miles of single
track and operating 230 electric locomotives
and about 1000 motor cars for multiple unit
suburban service. Table I shows the data
of the various roads embraced in this state-
ment.
In order to obtain a general picture of the
railroad traffic which would be affected by
the power supply of the Super-power Zone,
a study was made of the traffic conditions of
would be embraced in the Zone. For the
purpose of the investigation, the mileage of
those divisions of each road which would
presumably be included in the proposed
Super-power Zone has been tabulated, thus
determining the percentage of the total
mileage of each road lying within the Zone.
This percentage, or ratio, has been applied
to all other data of the road in order to deter-
mine the traffic within the zone. This factor,
therefore, determines the number of loco-
motives, the amount of traffic which would be
Fig. 1. Map Showing the Proposed Super-power Zone
the territory' covered by the Zone. In making
this study, data were taken from the operat-
ing reports of the United States Railroad
Administration, extending over the months
of 1919 for which comparable figures are
available.
The reports of the Railroad Administration
do not give separate traffic statistics of the
various divisions of the roads which are
embraced in their report, and there is neces-
sarily some uncertainty in estimating the
portion of each road and the traffic which
handled electrically instead of by steam, and
the tonnage of coal which would be replaced
by electric power. In view of these assump-
tions as to the probable area of the Super-
power Zone and the amount of included
traffic, the estimate as given can only be
an approximation.
As the detailed figures obtained from the
operating reports do not apply to switching
service, 20 per cent has been added to the
mileage and tonnage to cover this service;
and as the power requirements per ton mile
248 April, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, No. 4
for switching are approximately double those
for main line senace, 40 per cent has been
added to cover the coal consumed in switching.
On this basis it is estimated that the rail-
road traffic in the region covered by the zone
can be approximately represented by Table II.
TABLE II
RAILROAD TRAFFIC IN THE SUPER-
POWER ZONE
(Passenger, Freight and Switching)
Miles of route 12,000
Miles of single track 30,000
Locomotives in service 8,100
Locomotive miles annually 185,000,000
Gross ton miles annually, including
main line and switching move-
ments of passenger trains, freight
trains and locomotives 170,000,000,000
Tons of coal consumed annually. . . . 21,000,000
Considering railway electrification broadly
throughout the whole countrv' and including
only those lines which handle freight and pas-
senger service with electric locomotives, there
are found to be about 700 electric locomotives
operating over 5000 miles of route.
There has been some data published on
the results of heavy electrification.
From data available,* it would appear
that the ton-miles moved by 6V2 lb. of coal
in a steam locomotive is approximately equal
to that which can be moved by one kilowatt-
hour delivered from the power station.
Applying this ratio to the last item in
Table II, the electric energy which would
be required to handle the traffic now han-
dled by the 21,000,000 tons of coal in steam
locomotives is approximately 6,500,000,000
kw-hr.
If we assume 40 watt-hours per ton mile
at the power station, which checks fairly
* Particular reference was made to the following papers:
" Electrification Analyzed and Its Practical Application to
Trunk Line Roads." by W. S. Murray. A. I. E. E.. Vol. XXX,
1911.
"Electrical Operation of the Butte, Anaconda & Pacific
Railway." by J. B. Cox. A. I. E. E.. Vol. XXXIII. Sept., 1914.
"Operating Results from the Electrification of the Trunk
Line of the C, M. & St. P. Ry," by R. Bccuwkcs. New York
Railroad Club, March 16, 1917.
well with the records of a mixed service of
main line and switching, the total energv-
for moving the assimied traffic of 170,000,-
000,000 ton miles would be approximatelv
6,800,000,000 kw-hr.
The actual requirements would, however,
be something less. It has been estimated that
of all the tonnage moving over the railroad,
approximately 12 per cent is taken up with
the movement of railroad coal to points of
distribution, including a second movement
of the same coal in the locomotive tenders.
Making an allowance for railroad coal that
would still be required, a reduction of Id
per cent would seem a fair estimate. This
would correspondingh' reduce the yearly
power requirements to about 6,000.o6o,0ob
kw-hr. On the basis of probable load-factor,
this load would call for about 1,250,000 kw.
of power station equipment.
The conclusions to which this analysis
points may be siunmarized as follows:
(a) Of the whole mileage included in the
Zone, not a very large proportion has been
electrified, but main hne electrifications now
in operation are of sufficient extent and carry
tonnage of such character to present data
which can be applied to the traffic of the
whole district.
(b) The traffic within the Zone now handled
by steam locomotives, if handled electrically,
would require an average output of less than
750,000 kw. and if produced entirely by coal-
burning electric j)owcr stations would reduce
the coal requirement for transportation pur-
poses from 21 to 7 million tons annually.
(c) As a certain proportion of the electric
power will be produced from hydraulic power
stations, this coal requirement will be reduced
in proportion as advantage is taken of
hydraulic operation.
(d) The reduction in cost of maintaining
the motive j^ower units would be a large
amount which estimated from the locomotive
mileage would be in the order of $15,000,000
or more, annuallv.
249
The Last Stand of the Reciprocating
Steam Engine
By A. H. Ar.mstroxg
Chairman Electrification Committee, General Electric Company
The reciprocating steam engine is gradually disappearing from the industrial field, and indications
point to a similar movement in the propulsion of ships. The author foresees an era of electrification in the
steam road field, as the trend of real progress. The remarkable success of the C. M.&St. P. electrifi-
cation warrants the belief that this method of haulage could be used to advantage, beginning with terminals,
mountain grades, and congested districts. Although treated in a broad way, the recommendations may well
be applied to specific cases, where roads are confronted w'ith heavy expenditures for improvement of existing
facilities. The article was originallv presented as a paper before the Schenectadv Section of the A.I.E.E.,
Feb. 20, 1920.— Editor.
D"
A. H. Armstrong
kURING the year
1920 the people
of the United States
will pay otit for auto-
mobiles, not commer-
cial trucks or farm
tractors, but pleasure
vehicles, a sum of
money considerably
greater than the esti-
mated requirements
of our steam railways
for that year. The
railways, however
may find it very diffi-
cult and perhaps impossible to secure the large
sums needed without government aid, not-
withstanding the fact that the continued
operation and expansion of our roads is of
vital necessity to the welfare and prosperity
of the country and all its industries. The
will of the American public has always been
constructive and undoubtedly, in due time, its
voice will be heard and properly interpreted
by its representatives in Washington with
the resulting enactment of such laws as will
permit our railways again to offer an attrac-
tive field for the investment of private capital.
The purpose of this article is not to discuss
the politics of the situation nor any necessary
increase in freight rates that may be required
to make our roads self-sustaining, but rather
to offer certain suggestions as to the best
manner of spending the sums that must
ultimately be provided for new construction
and replacements.
During the war period many lessons were
most clearly brought home to us and not
the least of these is that there is some-
thing inherently wrong with our steam
railroads. During the three generations of
its development, we have become accustomed
to look upon the steam engine as properly
belonging to the railway picture and have
given little thought to its wastefulness and
limitations. It is around the steam locomotive
that railway practice of today has gradually
crystallized.
During the winter of 1917- IS our railways
fell down badly when the need for them was the
greatest in their history. It is true that the
cold weather conditions were unprecedented
and the volume of traffic abnormal, but the
weaknesses of steam engine haulage were
disclosed in a most startling and disastrous
manner. Delayed passenger trains in cold
weather can be endured by the traveling
public in suffering silence or voluble expres-
sion, according to temperament; but the
blocking of our tracks with frozen engines
and trains, resulting in a serious reduction
of tonnage in cold weather and a prohibitive
delay in transportation of freight in times of
great stress, is quite another thing and
plainly indicates the inability of the steam
engine to meet overloads and adverse climatic
conditions.
In marked contrast to the adjoining steam
engine divisions, the 44()-mile electrified
section of the Chicago, Milwaukee and
St. Paul Railway continued to do business as
usual all through that trying winter of
1917-18. The electric locomotives brought
both freight and passenger trains over the
electrified tracks in schedule time or better;
in fact, it was quite customary to make up
on the 440-mile electric run fully two hours
of the time lost by passenger trains on
adjoining steam engine divisions. While the
results obtained upon the Chicago, Milwaukee
& St. Paul were perhaps more spectacular
due to the greater mileage electrically
equipped, other electrified roads contributed
similarly attractive records. The reliability
and permanenc}^ of the comparison between
steam and electric locomotive haulage is
sufficiently guaranteed, therefore, by the
results of several 3-ears' operation, to justify
drawing certain conclusions regarding the
merits of the two types of motive power.
The following analysis of the railway situation
is therefore offered for the purpose of exposing
250 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 4
the fact that railroading today is in reality
steam engine railroading and the general
introduction of the electric locomotive will
permit fundamental and far reaching changes
being made in the method and cost of hauling
freight and passenger trains.
The writer is not proposing the immediate
electrification of all the railways in the
equal to twice the estimated capacit}-
required for the electrical operation of ever\-
mile of our tracks today.
The tonnage passing over the tracks of
our railways may be subdivided in a most
interesting manner as shown in Table I.
The first four items, representing 85.56 per
cent of the total ton-miles made during the
TABLE I
TOTAL TON-MILE MOVEMENT
All Railways in United States — Year 1918
u
Per Cent
1 — Miscellaneous freight cars and contents.
2 — Revenue coal cars and contents
3 — Locomotive revenue, driver weight only
4 — Passenger cars, all classes
42.3
515,000,000,000
16.23
197,000,000,000
10.90
132,300,000,000
16.13
196,000,000,000
Total revenue, freight and passenger .
5 — Railway coal
6 — Tenders, all classes
7 — Locomotive railway coal
8 — Locomotive, non-driving weight
Total non-revenue
GRAND TOTAL (All classes) .
85.56
5.00
6.50
0.39
2.55
14.44
Ton Miles
1,040,300,000,000
60,600,000,000
78,800,000.000
4,700,000,000
31,000,000,000
175,100,000.000
100
1,215,400,000,000
aooo
;
Operat
\' —
^
i Operilcd — '
-
5
\
;
s 1
A
Ij.
1
['
^ Va
J
V
A 1
; *
1
\l \
i*
1\
«
i ^
^(
\
\
c
m
? 1
1 1 J^
\
'^3000
1
f\ ''
l/^'
■/;i
V ^
1 V
5
f •-■^v--^
2000
A
i *'
<!" jv ^
'£
1
1
5 1
JuV
1000
ri
''
J
, '
■-^'
: :
1
\/
, '
;
]/x
<aiuja
■
i
;
r
b
;
<
' I
\
§
z
i
i
s 1
I 1
l!
5
D
s
I
1
'
\
J
-
0
N
-*
W
N
0
u
«
7
»
fit
A
■<
c
B
<^
Miles from Seattle
Fig. 1. Profile of Electrified Line* of the Chicago, Milwaukee and St. Paul Railroad
United States, as many roads of lean tonnage
would render no adequate return upon the
large capital investment required, but is
offering the following table of total operating
statistics simply as a measure of the magni-
tude of the problem confronting us in the
future. In this countn,^ it should be noted,
however, that we have during the past
thirty years installed electric power stations
year 191S, may be regarded as fundamentally
common to both steam and electric operation.
By introducing the electric locomotive, how-
ever, the last four items are reduced to the
extent of completely eliminating items ((i) and
(7). reducing item (5) by possibly SO per cent
and item (S) by one-half. Of the total of
14.44 per cent affected, therefore, it may be
assumed for purposes of comparison that
THE LAST STAND OF THE RECIPROCATING STEAM ENGINE
251
approximately 12 per cent or 146,000,000,000
ton-miles at present hauled by steam engines
over our roads will be totally eliminated with
electric locomotive haulage. This ton-mileage
eliminated is equal to over 20 per cent of items
(1) and (2) representing the revenue produc-
ing freight traffic on our railways. In other
words, if all our railways were completely
electrified they could carry one fifth more
revenue producing freight tonnage with no
change in present operating expenses or track
congestion.
It is evident that the greater part of the
tonnage reduction effected by electrification
is included in items (5) and (6), representing
the railway coal movement in cars and engine
tenders. The steam engine tender will of
course entirely disappear, while the railway
coal haulage will be largely curtailed by
utilization of water as a source of power and
the establishment of steam power houses as
near the coal mines as an abundant supply of
good condensing water and load demand will
permit. While water power should be
utilized to the fullest economical extent, the
greater portion of the railway power must
undoubtedly be supplied by coal, due to the
unequal geographical distribution of water
power available.
Even with coal as the source of power, it
may not be fully appreciated just how
enormous is the saving made by burning
fuel in large modern power stations under
the most efficient conditions possible, instead
of under the boilers of 63,000 engines which
by necessity must be designed and operated
for service rather than for fuel economy.
During the year 19 IS the fuel used by railways
is reported to be as shown in Table II.
TABLE II
RAILWAY FUEL 1918
Total coal production (all grades) . .678,211,000 tons
Used by steam railways 163,000,000 tons
Percentage of total 24 per cent
Total oil marketed in U.S 355,927,000 bbl.
Used by steam railways 45,700,000 bbl.
Percentage of total 5.8 per cent
Coal equivalent of oil at 33^ bbl. . . 13,000,000 tons
Total equivalent railway coal 176,000,000 tons
A quarter of all the coal mined in the
United States is consumed on our railways
and the following analysis will point out
some features of this extreme wastefulness
which are inseparable from steam engine
operation.
During the year 1910, exhaustive tests
were made upon the Rocky Mountain
Division of the C, M. & St. P. Ry. to
determine the relation existing between the
horse-power-hours work done in moving
trains and the coal and water consumed on
the steam engines in service. Table III gives
the results of these tests:
TABLE III
C, M. 86 ST. P. RY.; ROCKY MOUNTAIN
DIVISION
Coal and Water Used
Water
per
H.p.-hr.
Water
per
Lb. Coal
Coal
per
H.p.-Hr.
Three Forks-Piedmont. . .
Piedmont- Donald
Deer Lodge-Butte
Butte- Donald
39.6
35.4
39.7
40.4
38.0
44.2
41.4
40.2
5.08
4.70
4.85
4.86
4.09
4.65
6.51
5.63
7.75
7.54
8.31
8.74
Harlowton-Janny
Janny-Summit. .
8.90
9.48
Three Forks-Piedmont. . .
Piedmont- Donald
6.37
5.78
Average of eight tests. .
39.86
5.04
7.86
The records were obtained during the
portion of the runs that the engines were
doing useful work in overcoming train and
grade resistance, that is, all standby losses
were excluded. The through run, however,
included such losses in the magnitude shown
in Table IV :
TABLE IV
STANDBY LOSSES
Coal per hour
Fire banked in roundhouse 150 lb.
Cleaning fires for starting 800 lb.
Coasting down grade 950 lb.
Standing on passing track 500 lb.
Adding standby losses to the average of
7.86 lb. per h.p.-hr. obtained in the preceding
eight tests, the total actual coal consumed
under the engine boiler in twenty-four hours
divided by the actual work performed by the
engine is found to be 10.18 lb. per h.p.-hr.
at the driver rims.
As the result of this particular series of
tests it was determined that the coal con-
sruned while doing useful work was raised
30 per cent by standby losses. It should be
appreciated in this connection moreover
that this value was obtained on through runs
with no yard switching service or adverse
climatic conditions. It may be concluded,
therefore, that under all conditions of service
fully one third the coal burned on our steam
engines today is absolutely wasted in standby
losses of the general nature indicated above.
Supplementing these tests, a 30-day record
was kept of all coal used on the entire Rocky
Mountain Division and the total engine,
tender, and train movement reduced to
horsepower-hours, resulting in a value of
252 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
10.53 lb. coal consumed per horsepower-hour
at the driver rims. Both the above values
were based upon constants of 6 lb. per ton
train resistance at all speeds and 0.7 lb. per
ton per degree of curvature as determined
in part by dynamometer car tests and
representative of general railway operation.
Reducing the average coal values of the test
runs and the 30-day record per horsepower-
hour to electrical constants, we arrive at the
data shown in Table V:
TABLE V
COAL EQUIVALENT PER KW-HR.;
STEAM OPERATION
Coal per h.p.-hr. at driver rims 10.27 lb.
Coal per kw-hr. at driver rims 13.75 lb.
Coal per kw-hr. at power supply on basis
55 per cent efBciency 7.56 lb.
It is this last figure of 7.5G lb. of coal burned
on steam engines to get the equivalent
tonnage movement of one kilowatt-hour
delivered from an electric power station that
is of special interest to this discussion.
Comparing coal and electrical records on the
Butte, Anaconda & Pacific Railway before
and after electrification results' in arriving at
a value of 7.17 lb. of coal previously burned
on the steam engines to equal the same
service now performed by one kilowatt-hour
input at the substations, a figure comparing
favorably with 7.56 lb. above arrived at by an
entirely different method.
TABLE VI
ANALYSIS OF ROUNDUP COAL USED
Fixed carbon 49.20 per cent
Volatile carbon 38.12 per cent
Ash 7.74 per cent
Moisture 4.88 per cent
B.t.u 11,899
Making due allowance for the fact that
roundup coal is somewhat low in heat units,
it is nevertheless within the limits of reason-
able accuracy to assume that the steam
engines operating over all our railways are
consuming coal at a rate closely approximat-
ing 12.75 lb. per kilowatt-hour of useful work
done, as measured at the driver rims or 7 lb.
per kilowatt-hour as measured at a power
station and including for convenience of
comparison the transmission and conversion
losses inherent to electrical operation.
An electric kilowatt can be produced for so
much less than 7 lb. of coal that we are now
in position to finally forecast the approximate
extent of the coal economy that would result
from electrification.
All power values in Tabic VII are given at
the point of supply from the Montana Power
Company at 100,000 volts and include
deductions made for the return of power due
to regenerative braking of the electric
locomotives on down grades, amounting to
approximately 14 per cent of the total.
Owing to the excessive rise and fall of the
profile of the electrified zone of the C, M. &
St. P. Ry., its operation is materially benefited
by regenerative electric braking and the
value of 33.2 watthours per ton mile for
combined and passenger movement should
possibly be raised to the round figure of 40 to
make it apply more nearly to conditions
universally obtaining on more regular profiles.
Hence referring again to the ton-mile values
of Tabid:
Total ton-miles, 1918 1,215.400,000,000
Watthours ton mile 40
Kw-hr. total movement 48,700,000.(K)0
Coal required at 7 lb. per kw-hr 170,000,000 ton
TABLE VII
RELATION BETWEEN KW-HR. AND TON-MILES
CHICAGO, MILWAUKEE & ST. PAUL RAILWAY
Avery-Harlowton — Year 1918
Passenger
Freight
Average weight locomotive
300 ton
651,000
195,000,000
434,406,000
629,406,000
24,890,000
39.0
31 per cent
284 ton
Locomotive miles,
1 431 ,500
Locomotive ton-miles
407.000,000
2,903,09'.1,000
3,310.049,000
105 ''87 000
Trailing ton-miles
Total ton-miles ...
Kilowatt-hours
Watthours per ton-mile
31 9
Ratio locomotive to total
12.3 per cent
Watthours per ton-mile combined movement , . ....
33 2
Ratio locomotive to total combined movement
^
THE LAST STAND OF THE RECIPROCATING STEAM ENGINE
253
The actual equivalent coal consumed on
our steam railways for the year 1918 is given
as 176,000,000 tons, closely approximating
the figure of 170,000,000 tons estimated
from the operating results obtained on the
C, M. & St. P. electrified zone. These several
values check so closely as to justify the com-
pletion of the fuel analysis of the railways
as shown in Table YIII.
TABLE VIII
COAL SAVING BY ELECTRIFICATION
Total ton-miles steam 1,215,400,000,000
Reduction by electrification 146,000,000,000
Total ton-miles electric 1,069,400,000,000
Kw-hr. electric at 40 watts 42,776,000,000
Coal on basis 23^ lb. per kw-hr.. . 53,500,000 tons
Equivalent railway coal 1918 176,000,000 tons
Saving by electrification 122,500,000 tons
developed to produce power more cheaply
than by coal in many favored localities.
Perhaps no nation can be justly criticised for
lavishly using the natural resources with which
it may be abundantly provided. In striking
contrast with the picture of fuel waste on
the railways in this country however is the
situation presented in Europe at this writing.
Faced with a staggering war debt, with two
millions of its best men gone and an unde-
termined number incapacitated for hard
labor, and with so much reconstruction work
to do, France has to contend also with the
destruction of half its coal producing capacity.
Before the war, France imported twenty-three
million of the sixty-five million tons of coal
consumed. It is estimated that the full
-^r^.
A
i
\
1
i
m
y
f^-
'■^^B
m
' ^' '^;
/
/ ""'
]
Mr '
'i
■'■'J
W '
v^
■ /
Fig. 2. bUUU-ton, lUU-car Freight Train on the Missoula Uivision. Chicago, Miiwaukee 66 St Paul Railway
The startling conclusion arrived at is that
approximately 122,500,000 tons of coal, or
more than two thirds the coal now burned in
our 03,000 steam engines, would have been
saved during the year 1918 had the railways
of the United States been completely electri-
fied along lines fully tried out and proved
successful today. This vast amount of coal
is 50 per cent greater than the pre-war
exports of England, and twice the total
amount consumed in France for all its
railways and industries. Moreover, the
estimate is probably too conservative as no
allowance has been made for the extensive
utilization of water power which can be
restoration of the coal mines in the Lens
region will take ten years to accomplish,
which nieans materially increasing the coal
imported into France if pre-war consumption
is to be reached, as the relief rendered from
the Saar District will not compensate for the
loss in productivity of the mines destroyed
by the Germans This situation is being
promptly met in part by France in the
appointment of a Commission* to study the
feasibility of the general electrification of
all its railways with special reference to
immediate construction in districts adjacent
to its three large water-power groups, the
•See article by M. Mauduit in this is.sue.
254 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 4
Alps, the Pyrenes, and the Dordoyne or
Central plateau region. It is proposed to
electrify 5200 miles of its total of 26,000
miles of railways during a period covering
twenty years. If this work is accomplished
at a uniform rate of 260 miles a year, it is
a most modest program, considering the
extreme necessity for the improvement.
In even worse plight is Italy with practi-
cally no coal of its own and compelled to
import its total supply of 9,000,000 tons.
The war has brought home to these countries
what it means to be dependent upon imported
fuel for their very existence and both Italy
and Switzerland are also proceeding with
extensive plans for railway electrification.
Contrary to general understanding, the mines
of Belgium are not destroyed, but the need of
fuel economv is ver\- acute and this countn^
From figures given, the conclusions in
Table IX are arrived at in the matter of power
station capacity required for complete electri-
fication of the railways in the United States.
TABLE IX
RAILWAY POWER REQUIRED
Kw-hr. electric operation,
1918 42,776,000,000 kw-hr.
Average load, 100 per cent
load-factor 4,875,000 kw.
Power station capacity at 50
per cent load-factor 9,750,000 kw.
It appears therefore that approximately
10,000,000 kw. power station capacity would
have been stiificient to run all the railroads
for the year 1918, or one half the station
capacity which has been constructed during
the past thirty years.
~
~^
~
~
~
1
\
v
1
,
\
\
\
\
-c
'
\
e
rric
/•wi
r-'
MM
on
^st
1
\
1
—
f
\
\
^
,
.
\
\
1 —
1
\
^
\
^
\
\
\
'M
\,
1
\
\
^
^
\
ta
\
<'
■^
'~^
X
kjj
=1-,
w
^
£3.
-«i
^
^^^^S_
^^^^.1
-'
^is
— ,
—
.^
s^-
~~1 TH"^
- 1 I —
^
3
IT"
'
7
7
a
?
s
jk
7
>
5
■*,
7
■*
•■
f
«
0
■&
Fig. 3. Characteristic Curves of Gearless Passenger Locomotive
also has broad plans for railway electrification
with immediate construction in view.
Recognizing the many advantages of elec-
tric operation of its railways, Europe further-
more considers this a most opportune time
to start the change rather than to spend its
limited funds in replacing worn out and
obsolete steam equipment in kind. Also in
marked contrast to the American attitude
is the sympathetic interest and constructive
assistance rendered by the Governments
abroad in regard to the vital matter of
rehabilitation of its railway systems. It
would not be without precedent if the next
decade witnessed England and the Continent
outstripping this countr\- in the exploitation
of another industry which, while possibly not
conceived here, has certainly been more
fully developed and perfected in America than
elsewhere.
TABLE X
ESTIMATED POWER STATION CAPACITY
UNITED STATES— YEAR 1918
Central stations. . . 9,00(1, (l(Xt kw.
Electric railways .3,0(K),lHK1 kw.
Isolated plants 8,000,000 kw.
Total 20,000,000 kw.
In the order of magnitude, therefore, it
is not such a formidable problem to consider
the matter of power supply for our electrified
railways and it becomes evident also that the
railway power demand will be secondar\- to
industrial and miscellaneous requirements.
Such being the case, the question of
frequency of electric power supply becomes
of great importance, if full benefit is to be
obtained from extensive interconnected gener-
ating and transmission systems covering the
THE LAST STAND OF THE RECIPROCATING STEAM ENGINE
255
entire country. Indeed with the full develop-
ment of interconnected power systems supply-
ing both railway and industrial load from the
same transmission wires, the above assump-
tion of 50 per cent load-factor for the railway
load can be materially bettered.
In this connection a method of limiting the
troublesome peak load hitherto considered
inherent to railway power supply has been in
successful operation on the electrified C, M.
& St. P. zone for the past year. With unre-
strained peaks, the load-factor was approxi-
mately 40 per cent, but this low value has
been raised to nearly 60 per cent by the
installation of an inexpensive and most
II
10
31
.♦J
to
a.
5«
>
J? 7
D
E
5 ^
5
4
5
2
I
0
/
^ /
il
^>
il
/
r'
i-
/
/
t
1
4
t>^
f
^
fo^
^
l^J
.^
!>,-
SS^
^J
^
--^
'
Mountain Division supplied by seven sub-
stations controlled as a unit. A load-factor of
nearly 60 per cent brings the electric railway
within the list of desirable customers and
makes it possible for power companies to
quote attractively low rates for power.
Returning again to the question of power
supply, it is instructive to note the general
trend toward a higher frequency as evidenced
by the turbine and transformer sales of the
General Electric Company during the past
decade.
It is quite evident that 60 cycles is rapidly
becoming the standard frequency in America ;
and many instances are on record where it has
12
11
10
a
09 '10 11
Fig. 4.
'V, '14 15 16 17 18 '19
Years
/
/
/
./
/
c^
//
P
/
/
/
9
/
/
r
if
/
/
f
/
f'-
f
/
^^
— •
/
y
,J^'
f^
^'
^l-
-'•'
'08 09 '10 11 12 15
14 15 16 17 18 l4
Years
Comparative Sales of 25 and 60-cycle Transformers and Steam Turbines
satisfactory device known as the
limiting and indicating apparatus.
power
TABLE XI
LOAD-FACTOR RECORDS
C, M. & ST. P. RY.
1919
April
May
June
July
August. . . .
September
Per Cent
Duration
of Peak
6.4
4.6
1.6
0.7
4.1
9.5
Per Cent
Load-Factor
59.3
56.1
56.5
55.6
54.7
58.8
The readings in Table XI cover the
performance on the 220 miles of the Rocky
replaced lower frequencies, principally 25
cycles. This fact in no manner handicaps the
future development of electric railways, as
entirely satisfactory power can be obtained
from 60-cycle transmission lines through
rotary converters or synchronous motor-
generator sets, depending upon the direct-
current trolley voltage desired. Indeed a
growing appreciation of the declining impor-
tance of 25-cycle power generation in this
country contributed largely to the demise of
the single-phase system, as its chief claim for
recognition is wiped out with the introduction
of the motor-generator substations required
with 60-cycle supply.
While America apparently has adopted
60 cycles as its standard frequency and can
look forward to unlimited interconnection of
256 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. i
its large power systems, European practice is
evidently crystallizing on 50 cycles. The
situation abroad is as yet, however, not
clearly defined. In such a small compact
country as Switzerland for instance, where so
much electrical development is taking place,
there is much conflict of frequencies. Appar-
ently there is little appreciation of the
advantages resulting from interconnected
power stations; in fact the Loetschberg
Railway is supplied with power from 15-cycle
waterwheel-driven generators placed in the
same power station with 42-cycle units
supplying industrial load while in the same
average load demand for the country- as a
whole.
A good example of the necessity for
improvement in power distribution conditions
in Switzerland is provided in the supply of
power to the Loetschberg Railway as illus-
trated in Table XII:
TABLE XII
POWER SUPPLY TO THE LOETSCH-
BERG RAILWAY
March, 1919
Total for month 540,180 kw-hr.
Average of six 15 min. peaks 3,489 kw.
Load-factor, basis 24 hours 20.8 per cent
Fig. 5. General Views of Grand Central Terminal Area from 50th Street Before and After Electrification
immediate district there is a 50-cycle trans-
mission line and no tie-in frequency changer
sets as yet installed to interconnect any two
frequencies. The power company, power
consumer, and electrical manufacturer pay
heavily for the complication imposed by
maintaining three frequencies where only
one is needed, and growing appreciation of
this fact may lead to the standardization of
50 cycles in Switzerland and thus swing that
country- in line with its neighbors and
ultimately bring about a more economical
ratio of installed generator capacity to
As the railway was operating for only seven-
teen hours per day, the load-factor during
actual operation is somewhat better than
20.8 per cent. On the other hand, the actual
momentary peak load greatly exceeded 34S9
kw.; and this very fluctuating railway load
furnishes a good illustration of the need of
combining it with other diversifitxl loads, in
order to keep down the fixed investment of
power station equipment now set aside for
this isolated railwav load. For example, the
00 per cent load-factor of the C, M. & St. P.
power demand is the ratio of average to
THE LAST STAND OF THE RECIPROCATING STEAM ENGINE
257
momentary peak while the Loetschberg
Railway peak load is determined by six
15-min. peaks with momentary peaks greatly
in excess of this figure.
Apparently the adoption of a standard
frequency of 50 cycles would meet all general
requirements in vSwitzerland, but would
necessitate the installation of frequency
changing substations to meet the demands
for 15-cycle single-phase railway power. If
the electrified railways are to benefit, therefore,
from the establishment of a common generat-
ing and transmission system in Switzerland,
the choice of the single-phase railway system
might possibly be considered unfortunate,
viewed in the light of modern development in
power economics and the successful adapta-
from one transmission system, the average
combined load-factor is raised to nearly
(iO per cent, a figure which could even be
surpassed on roads of more regular profile.
Furthermore, when the railway load is
merged with the lighting and industrial
power of the district and the whole diversified
load supplied from the same 60-cycle trans-
mission and generating system, it is quite
evident that all the conditions are most
favorable for the efficient production of
power. In this country such an achievement
will probably be governed by the laws of
economic return upon the capital required
because our vast natural fuel resources are
popularly regarded as inexhaustible, but in
Europe there is the compelling spur of stern
Fig. 6. High Speed Gearles5 Passenger Locomntivc fcr the C, M. & St. P. Rwy. 3000-volt Direct-current Electrification
tion of the less expensive and more flexible
direct-current motor to high trolley voltages.
From the power station standpoint, the
electrification of our railways admits but one
conclusion. We have some 63,000 engines
now in operation and their average combined
load amounts to approximately four million
horse power at the driver rims, or only an
insignificant total of 65 h.p. for each engine
owned. It is true that, owing to shopping
and for one cause or another, a large pro-
portion of these engines are not in active
service at all times, still the average twenty-
four hour output of each engine is less than
ten per cent of its rating. In the case of the
C. M. & St. P. electrification, the average
load of each individual electric locomotive is
only 15 per cent of its continuous rating, but
by supplying power to 45 electric locomotives
necessity behind the movement to utilize
economically the water powers they possess
in place of the coal they cannot get.
While the much discussed subject of power
generation and transmission is a very vital
part of the railway electrification project,
chief interest centers in the electric locomotive
itself. Few realize what a truly wonderful
development has taken place in this con-
nection in a comparatively few years and how
peculiarly fitted this type of motive power
is to meet the requirements of rail trans-
portation. Free from the limitations of the
steam boiler, and possessing in the electric
motor the most efficient and flexible known
means of transmitting power to the driving
axles, the electric locomotive gives promise of
revolutionizing present steam railway practice
when its capabilities become fully recognized.
258 April, 1920
GEXER.\L ELECTRIC REVIEW
Vol. XXIII, No. 4
The only limits placed upon the speed and
hauling capacity of a single locomotive are
those imposed by track alignment and
standard draft rigging Only questions of
cost and expediency control the size of the
locomotive that can be built and operated
by one man, as there are no mechanical or
electrical limitations that have not been
brushed aside by careful development. Just
what this means in advancing the art of
railroading is as yet but faintly grasped, any
more than the boldest prophet of twenty
years ago could have fully pictured the
change that has taken place at the Grand
Central Terminal as the result of replacing
steam by electricity.
Progress in utilizing the capabilities of the
electric locomotive has been slow. It is hard
to break away from life-long railway tradi-
tions established by costly experience in
many cases. In consequence the electric
locomotive has thus far simply replaced the
steam engine in nearly similar operation.
Even under such conditions of only partial
fulfillment of its possibilities, the electric
locomotive has scored such a signal operating
success as to justify giving it the fullest
consideration in future railway improvement
plans.
On the C, M. & St. P. Ry. 42 electric
locomotives have replaced 112 steam engines
and are hauling a greater tonnage with
reser\'e capacity for still more. On this and
other roads, electrification has set a new
standard for reliability and low cost of
operation. In fact, although no official
figures have yet been published, it is an open
secret that the reduction in previous steam
operating expenses on the C, AI. & St. P. Ry.
are sufficient to show an attractive return
upon the twelve and a half millions expended
for the 440 miles of electrification, without
deducting the value of the 112 steam engines
released for sen,dce elsewhere. As the electric
locomotive is destined to leave its deep
impression upon the development histor%- of
our railways, it is fitting that the remainder
of this paper should be devoted to its consider-
ation.
Our steam engine construction is unsym-
metrical in wheel arrangement, must run
single ended, and is further handicapped with
the addition of a tender to carry its fuel and
water supply. The result has been much
congestion at terminals; and the necessary
roundhouses, always with the inevitable
turn tables, ash pits, and coal and water
facilities, have occupied much valuable land;
and in addition steam operation has greatly
depreciated the value of neighboring real
estate. The contrast offered by the two
large electric terminals in New York City is
too apparent to need more than passing
comment, and similar results may be expected
on the fulfillment of plans for electrifying the
Chicago terminals.
While it has been a simple matter to design
electric locomotives to run double ended at
the moderate speeds required in freight
service, the problem of higher speed attain-
ment, exceeding 60 miles per hour, has
presented greater difficulties. The electric
motor is however so adaptable to the needs
of running gear design that electric loco-
motives are now in operation which can meet
all the requirements of high-speed passenger
train running. These results, also, arc
obtained with less than 40,000 lb. total weight,
and 9500 lb. non-spring borne or "dead"
weight on each driving axle, and finally, but
not least, with both front and rear trucks
riding equally well, a success never before
achieved in locomoti\-es of such large ca-
pacity.
In connection with the riding qualities of
electric locomotives, it is of interest to note
the conclusions that the Committee of the
American Railwa\- Engineering Association,
F. E. Tumeaure, Chairman, reached in their
report of 1917:
"From the results of the tests on the electrified
section of the Chicago, Milwaukee & St. Paul
Railway, the tests made in 1916 on the Norfolk
and Western, and the few tests made in 1909 at
Schenectady, N. Y., it would appear to be fairly
well established that the impact effect from electric
locomotives is very much less than from steam
locomotives of the usual type. Comparing results
obtained in these tests with the results from steam
locomotives, it would appear that the impact from
electric locomotives on structures exceeding, say.
25 ft. span length, is not more than one third of the
impact producted by steam locomotives."
There is as yet no general acceptance of a
standard design of electric locomotive. Geared
side-rod construction for heavy freight service
and twin motors geared to a quill for passenger
locomotives appear to find favor with the
Westinghousc-Baldwin engineers, while the
General Electric Company goes in for the
simple arrangement of geared axle motors for
freight and gearless motors for passenger
locomotives. In both Switzerland ami Italy
the side-rod locomotive enjoys an almost
THE LAST STAND OF THE RECIPROCATING STEAM ENGINE
259
exclusive field. How much of this preference
for side-rod construction is due to the re-
strictions imposed by the use of alternating-
current motors is hard to determine, but the
facts available indicate both in this country
and abroad the uniformly higher cost of
Armature and Wheels of 3000-volt Direct-
current Gearless Locomotive
repairs of this more complicated form of
mechanical drive.
The electric railway situation in Italy is
further complicated by the employ-
ment of three-phase induction motors
with all the attendant handicaps of kx-
double overhead trolleys, low power-
factor, constant speeds, and overheating ^'
of motors resulting from operation on ^
ruling gradients with motors in cascade
connection. In many respects the non- ^ ^c
flexible three-phase induction motor is §
poorly adapted to meet the varied re- ^ ^
quirements of universal electrification; t; ^
and in consequence Italian engineers ^
are still struggling with the vexing ques- ;; ■«
tion of a system, which may, however, iS?
be in fair way of settlement through ^
the adoption of a standard of 50 cycles ^
as the frequency of a nation-wide in-
terconnected power supply, thus throw- lo
ing the preponderance of advantages
to high-voltage direct current. °
The extreme simplicity of the gearless
motor locomotive appeals to many as
does its enviable record of low mainte-
nance cost, reliability, and high operating
efficiency, as exemplified by its unvarying
performance in the electrified zone of the
New York Central for the past twelve years.
Table XIII shows that the high cost of living
did not appear to have reached this favored
tvpe of locomotive until the year 1918.
' The records on the C, M. & St. P. loco-
motive are equally remarkable when con-
sidering their greater weight and more
severe character of the ser\'ice.
TABLE XIII
MAINTENANCE COSTS
NEW YORK CENTRAL
Number locomotives
owned
Average weight, tons
Cost repairs per loco-
motive mile
1913 1914
48 62
118 ,118
4.32 4.03
1915
63
118
4.45
1916
63
118
3.78
1917
73
118
4.01
1918
73
118
6.26
TABLE XIV
LOCOMOTIVE MAINTENANCE COSTS
CHICAGO, MILWAUKEE 8d ST.
PAUL RAILWAY
1916
1917
1918
Number locomotives owned
Average weight, tons
Cost repairs per loco. mile.
20
290
8.21
44
290
9.62
45
290
10.87
eear'ess
/^
■^~
-_^
-^
Pr
^
^^
■
/O /S £0 Z5 30 35 40 as SO S5 60 &S^
<5p^ed - Mi/e.-^ per /iour
Fig. 8. Comparative Efficiencies of Original Geared and New Gearless
C. M. & St. P. Rwy, Passenger Locomotives
In both these instances the cost of repairs
approaches closely to three cents per 100
tons of locomotive weight. Giving due
credit to the excellent repair shop service
rendered in each case, it is instructive to
note that three cents per 100 tons main-
tenance cost of these direct-current loco-
motives is less than half the figures
given for any of the alternating-current
locomotives operating in the United States
or in Europe.
260 April, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, No. 4
Compared with the cost of repairs for
equivalent steam engines, the foregoing figures
for electric locomotives are so verj' favorable
as to justify the general statement that
electric motive power can be maintained for
approximately one third the cost of that of
steam engines for the same train tonnage
handled. As locomotive maintenance is a
measure of reliability in ser\'ice and in a way
expresses the number of engine failures, it is
quite in keeping with the records available to
state also that the electric locomotive has
introduced a new standard of reliability that
effects material savings in engine and train
crew expense as well.
formerly took to handle the lesser tonnage by
steam engines. This means a material
increase in capacity of this single-track line
which may be conser\"ativeh- estimated in
the order of at least 50 per cent and probably
more. In other words, on this particular road,
electrification has effected economies which
sufficiently justify the capital expenditure
incurred and furthermore has postponed for
an indefinite period an>- necessity for con-
structing '' second track through this difficult
mountainous country.
A careful study of the seriously congested
tracks of the Baltimore and Ohio Railroad
between Grafton and Cumberland disclosed
Fig. 9. Latest Type Gearless Passenger Locomotive in Service in the New York Central Electric Zone
While the first cost of electrification is
admittedly high, it may in certain instances
be the cheapest way to increase the tonnage
carn,'ing capacity of a single track especially
in mountain districts where construction is
most expensive and steam engine operation
is most severely handicapped. In this
connection a comparison of steam and electric
operation on the C., M. & St. P. Rwy. may be
summarized as follows :
For the same freight tonnage handled over
the Rocky Mountain Division, electric opera-
tion has effected a reduction of 223-2 pt"r cent
in the number of trains, 24..") jjer cent in the
average time per train, and has improved the
operating conditions so that nearly 'M per
cent more tonnage can be handled by electric
operation in about SO per cent of the time it
vitally interesting facts. Company coal
movement in coal cars and engine tenders
constituted over 1 1 per cent of the total
ton-miles passing over the tracks. In other
words, due to the very broken profile of this
division, the equivalent of one train in every
nine is required to haul the coal burned on the
engines. Taking advantage of this fact and
the higher speed and hauling capacity of the
electric locomotive and its freedom from
delays due to taking on water and fuel, it is
estimated that the three tracks now badly
congested with present steam engine tonnage
could carry NO per cent more freight with
electric locomotive ojieration. The coal
output of the Fainnont District is largely
restricted bv the congestion of this division
of the B. &'0. R. R. and it is probable that
THE LAST STAND OF THE RECIPROCATING STEAM ENGINE
261
^'-<^y(A~^A<^\^
ly-e-
-33=
'4-5f-5-IO-— 7-8'-~5-0' ir-2i'
WEIGhfT-LOCOMOTIVE&TENDER 414500 lb.
WEIGHT OF TENDER 154000 -
WEIGKTON DRIVERS 201.000 "
CYLINDERS 24-»30-
STEAM PRESSURE... _ 200 lb.
HEATING SURFACE 3614- sq.ft
GRATE AREA _ _ _..488" ■■
TRACTIVE EFFORT .46.630 lb.
-5ni--6Kr-
•as -43*. -ST.
ffg'-^ffg- — irey "^s-er -^ tg-c -— icr-e-
- iS^o- -
WEIGHT- LOCOMOTIVE ^TENDER 555,700 lb
WEIGHT OF TENDER.. ..... 165700 •
WEIGKTON DRIVERS ...323500 ■
CYLINDERS _.:- 23i'&3r«30-
STEAM PRESSURE...
HEATING SURFACE
GRATE AREA
TRACTIVE EFFORT.
.35
S'-O- - e-7i'2
200 lb.
.6554 6 SOFT
724 - ■
,: .76.200 lb.
-36r- ~5Z^
-ir-o- - ig6'
ItZ'-O"
-iwr-
WEIGHT OF MECH EQUIPMENT 328000 lb.
WEIGHT OF ELEC, EQUIPMENT 24.8PO0 •
WEIGHT- TOTAI .576(000 •
WEIGHT ON DRIVERS .4SQ000 •
-KW-
5P-
-IHT-
-5ZV -36-
— — KW— +-7-IF— -e'-c-^e*-
MOTORS B
TYPE OF M0T0R....GE 253. 1500/3000 VOLTS
GEARING 82-18 RATIO 455
TRACTIVE EFFORT I hour eS.OOOIbi
•36- -44- I
.i4-^U4-io'"4'-7" 6-11" -4.'-7"-^'-7"4'-7"S'-S-3^S^-7"-4-7^4!-7— G-ll" -4-7'-4'-1(f-4'-6r-'
13-9"
. '- T-=rr 67-0 ,g,g.. - ^
WEIGHTOFMECH EQUIPMENT E95.000lb MOTORS 12
WEIGHT OF ELEC EQUIPMENT 235,000- TYPE OF MOTOR GElOO,100O/3000 VOLTS
WEIGHT- TOTAL 530,000" GEARING GEARLESS
WEIGHTON DRIVERS 458,000" TRACTIVE EFFORT 1 hour 46,0001b.
Fig. 10. Principal Types of Steam and Electric Locomotives on the Chicago. Milwaukee and St. Paul Railway, Puget Sound Lines
262 April. 1920
GEXER.\L ELECTRIC REVIEW
Vol. XXIII. Xo. 4
equal relief with continued steam engine
operation could not be secured without the
expenditure of a much larger sum for addi-
tional track facilities than would be needed
to put electric locomotives upon the present
tracks.
Further instances could be cited where the
benefits of electrification are badly needed
and many of these are coal carrving roads
among which the Virginian Railway stands
out conspicuously as a good opportunity to
make both a necessarv- improvement and a
sound investment.
securing increased track capacity and im-
proved ser\'ice than by laying more rails
and continuing the operation of still more
steam engines in the same old wasteful way.
To conclude the startling picture of our
present railway inefficiency, we are today
wasting enough fuel on our steam engines to
pay interest charges on the cost of completely
electrifying all the railways in the United
States, — fuel that Europe stands in sad need
of and which England and Germany, the
pre-war coal exporting coimtries, cannot now
supply. With operating expenses mounting
fl
>J
3000
1^
^
1
i^
%
^
>
s5
k-
■>.
jg
i£
K
W^
U^
I.
ii
jn'
/
\i
ZOOO
%
^
—
ft
^p"
^
.t4
?
'^
^^
. 1?
/
\
r*
1
5
.
a
1
V
^
-
y
^.
b
^
1
£i-
^
J-
*<,
1
fi
%
^
/ooo
^
<
^
^
1
^
[455
fi
U
Oo,
if
_^
^ ,
^~
^=
^z
:~
^;
^
r
~~
^PT'
~"
"
n
i
i
SO
«
Fig. 11. Map and Profile of West End Cumberland Division, Baltimore ti Ohio Railroad
Reviewing the progress made in a short
twenty->-ear period, we have seen the steam
turbine and electric generator drive the
reciprocating engine from the stationary
power field. The same replacement is now
taking place on our ships, big and small,
notwithstanding the fact that the marine
reciprocating engine is a very good engine
indeed and operates under the ideal condi-
tions of steady load and constant speed. And
now the steam locomotive must in turn give
way to the electric motor for the same good
reason that the reciprocating steam loco-
motive has become obsolete and fails to
respond to our advancing needs. Electrifi-
cation affords a cheaper and better means of
to 82 per cent of revenue, inadequate equip-
ment and congestion of tracks, what we need,
in addition to constructive legislation and
real co-operation on the part of the Govern-
ment in the matter of rates and safeguarding
invested capital, is wise direction in the
expenditure of the large sums that must
speedily be found and used to bring our
railways abreast of the times. Accord full
honor to the reciprocating steam engine for
the great part it has played in the develop-
ment of our railways and industries, but
complete the work by replacing it with the
electric motor and enter upon a new era of
real railroading, not restricted steam engine
railroading.
263
Electrification of the Coast and Cascade Divisions
of the C, M. CS, St. P. Ry.
By E. S. Johnson
Railway and Traction Engineering Department, General Electric Company
A convincing testinional as to the excellence of the system of electrification employed on the C. M. & St.
P. Ry., and of the performance of the apparatus, is disclosed by the fact that the equipment for the newly
electrified Cascade and Coast Divisions differs but little from that which for four years has been operating on the
Rocky Mountain and Missoula Divisions. The principal feature of difference lie: in the new passenger loco-
motives. These are designed especially for passenger service, while the original passenger locomotives were
in reality freight locomotives temporarily geared for higher speed until such a time as they would be put in
freight service and replaced by genuine passenger locomotives. The adoption of bi-polar gearless design for
the new passenger locomotives marks a distinct step towards simplicity and low maintenance cost of motive
power equipment. Advancement in the art of railroading has been very much speeded up by the successful
application of the many new features of this electrification, summarized in the conclusion of this article. — Editor.
capacity is practically doubled, which is of
very great advantage at certain times of the
year. Furthermore, during the past four
years the entire amount of coal that would
have been used for steam operation has been
saved and made available for other purposes,
which has helped in a small way to relieve
the coal shortage that has been so serious, as
the total electric power used has been obtained
from the waterpower plants of the Montana
Power Co. The comfort in traveling the
mountainous regions resulting from the elim-
ination of cinders, smoke, grinding and jarring
due to air brakes, and the saving in running
time have been very mtich appreciated by the
travelling public as shown by the increase in
passenger traffic over these lines.
The reliability of electric operation under
the very severe weather conditions such as
prevail in this part of the country has demon-
strated the fitness of the electric locomotive
to meet the most severe service requirements.
It is the opinion of engineers who have studied
electrification for the past several years that
it will supplant steam operation in the very
near future where it is necessary to increase
the tonnage capacity and where economical
operation and conservation of the world's
fuel resources are the watchword.
The service on this section is very similar
to that on the sections previously electri-
fied, consisting principally of the two all-steel
elegantly equipped trans-continental passenger
trains Olympian and Columbian in each
direction per day, a local passenger service
between Cle Elum, and Seattle and Tacoma,
and four to six through freight trains each way
per day. The freight trains are made up of
all types of cars varying in weight from 25
to 70 tons loaded, and thus the importance
of careful handling can be fully apjureciated
onl}- by the train crews having charge of this
E. S. Johnson
npHE original 3000-
-■- volt electrification
of the Chicago, Mil-
waukee & St. Patil
Ry., from Harlowton,
Mont., to Avery,
Idaho, a distance of
about 4-iO miles, com-
prising the Rocky
Motmtain and Miss-
oula Divisions was
described in the
General Electric
Review, November,
191G. Other articles
calling attention to the phenomenal success
of this great undertaking have appeared from
time to time in the technical press. The
section from Seattle and Tacoma to Othello,
Wash., a distance of 208 miles, including the
Coast and Cascade Divisions, is now being
placed in operation.
This electrification in the main is a duplica-
tion of the original undertaking, differing
only in various minor details which will be
pointed out later.
The electrification of the C. M. & St. P.
Ry. differs from practically all other work
of this kind in that it was undertaken for
reasons of economy and for the purpose of
increasing the tonnage capacity rather than
the elimination of smoke in tunnels or at
terminals, the taking care of suburban
traffic or other local conditions. The results
of operation for the past four years have
thoroughly demonstrated the reliability of
electric operation as compared with steam,
the increase available in the tonnage capa-
city, and the reduction possible in operating
costs. Preliminary figures indicate that the
economies effected more than justify the
capital expenditure and that the tonnage
264 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
lk_^
iX.J^.
u >
S o
E E
" 0
> S
il
C CU
in i
. o
^1
ELECTRIFICATION OF THE COAST AND CASCADE DIVISIONS
265
work. Trains of 3000 tons are now handled
with far greater ease at approximately twice
the speed on grades, and with less damage to
rolling stock, especially the pulling of draw
bars, than 2()00-ton trains were handled with
steam operation.
■-,»f-
Fig. 5. High-tension Room, Showing 2500-kv-a.. 100,000-volt
Transformers and Oil Circuit Breakers,
Tacoma Substation
The motive power equipment assigned to
these divisions consists of five bi-polar gear-
less passenger, twelve geared freight, and
two switching locomotives, all of which were
furnished by the General Electric Co. The
geared freight and switching locomotives were
described in the Review, October, IQKi.
The two switching engines are duplicates of
those previously furnished and the geared
freight locomotives are those originally used
in passenger service on the Harlowton-Avery
electrification, the gear ratio having been
changed to make the locomotives suitable
for freight service. (The original passenger
and freight locomotives were duplicates,
except as to their gear ratio.)
The bi-polar gearless locomotives are de-
signed especially for high-speed passenger
service and tests have demonstrated that their
riding and tracking qualities are superior to
any double-ended locomotives built. The
running gears consist of two three-axle trucks,
one at each end, and two central four-axle
**"■ MKffasnojAJTTai^ ,Qa9f'
266 April, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, Xo. 4
trucks, all of which are articulated together
and equalized in such a manner as to give
approximately the same weight per driving
axle and to insure proper tracking at high
speed. These important features received
much favorable comment from prominent rail-
vray engineers on the occasion of the exhibi-
tion tests at Erie, Pa., November 7, 1919.
The leading wheels on each of the three-axle
trucks are not equipped with motors, and by
a special arrangement of the journal boxes are
free to move axially a certain amount without
movement of the entire truck, thus assisting
in a more gentle turning of the trucks on
curves and a reduction of flange wear.
Fig. 6 gives the general dimensions and
Fig. 4 is a photograph of one of the loco-
motives, which was described in the Review,
December, 1919. They are now in service
on the Rocky Mountain and Missoula
Divisions, while the geared passenger loco-
The general arrangement and installation
of the electrical control apparatus is the same
as that used in the construction of the geared
locomotive. The contactors and other parts
are assembled on supports when built and
the work of installing them in the locomotives
consisted only in bolting the supports in place
and connecting on the proper cables. Back
view of the rheostatic contactor group is shown
in Fig. 7. With this type of construction the
apparatus can be more systematically arranged
and provision made for easy inspection and
maintenance. These equipment groups, to-
gether with the major portion of the other
electrical apparatus, are installed in the two
end cabs, with an aisle through the center,
and hatches are provided on either side of the
locomotive so that all parts can be readily in-
spected without the necessity of removing any
apparatus. Fig. 8 is of a cross section through
the locomotive showing the arrangement.
Fig. 7. Rear View of 3000-volt Contactor Group
motives are having their gearing changed for
freight service preparator\' to being trans-
ferred to the new electrification.
Each of these locomotives is equipped with
a high-speed circuit breaker * to prevent
damage to the electrical apparatus due to
short circuits. Repeated tests under the
most severe conditions and actual operation
have shown this feature to be a distinct
advance. These breakers are duplicates of
those installed in the substations for the
protection of the generators of the motor-
generator sets. With the protection thus
afforded to both locomotives and substation
apparatus, all damage due to short circuits is
eliminated and thus the dream of engineers
for decades has been realized.
*A view of this circuit breaker is shown in Fig. 1 of the
article by Mr. J. F. Tritle in this issue.
The profile of the entire electrification ex-
tending from Harlowton. Mont., to Seattle and
Tacoma, together with the 212 miles remain-
ing under steam operation, is shown in Fig. 9.
This illustration also shows the location of the
22 substations which have a total installed
capacitv of 91,300 kw. This equipment con-
sists of 39 2()00-kw. and 9 loOO-kw. synchro-
nous motor-generator sets with transformers
and switching equipment. The substation
spacing averages approximately 30 miles.
The maximum grades on the new section
are the 17 mile 2.2 per cent grade from the
Columbia River west and the 19 mile 1.7
per cent grade from Cedar Falls east to the
summit of the Cascades.
The 3000-volt power for the operation of
this section is supplied from eight substations
located as shown in Fit:. 10. This illustration
ELECTRIFICATION OF THE COAST AND CASCADE DIVISIONS
267
Fig. 8. Apparatus Compartment 3000volt Direct-current Gearless Locomotive
also shows the present and ultimate capacity
of each of these substations, the sizes of the
feeders and the tonnages of the freight trains
on which the feeder sizes were determined.
The lay-out arrangement of the substations
is the same as that of the original substations,
except that the synchronous motor starting
and running oil circuit breakers are installed
in cells in the basement instead of in the
dividing wall between the motor-generator
and transformer rooms, in order to afford
greater space in the motor-generator room
and greater reliability, and the direct-current
feeder disconnecting switches are installed on
framework outside the building instead of on
the walls inside the station. The installation
of these switches outside the buildings was
desirable, as practical operation had demon-
strated that it was necessary that they be of a
design that could be opened under load at
times of emergency. Their mechanisms are
so arranged that they are operated from
within the buildings. They have been
thoroughly tested under the conditions that
will exist in actual operation and they
meet every requirement, being capable of
opening a current from 7000 to 8000 amp.
successfully.
EL-ECTRIFIEC
DIVISIONS -
PROFILE
- CHICAGO MIUWAUKEE &
ST. PAUL RY
\
1
•— Elc
Iricallig
V —
— El«
ctncal
i Open
ted —
TlWI
opcrat^-u ^
1
1
5
\
!
ig 1
n
s
1
[ ^
,
! S
Q L
^\A
^ „
V
\i
1 i:
1 ^
;
T /^
1 1
f ^
>"
;;Aj
I
1
V\
W,!
/
N
V
iA
1 1
i
>
1
\^
3000
1 1
^< \
I 1
I
A^
;
a4»i^
V
/V
?
i)
/Ai
a
A
1 '"
L^
■it'
!
1
\' "
2000
^A
-/
1
* ■
0% li
y
^'
1
^ t
'
•^
w
^•'
1
M
\
\X
0
b
1
coAUJcr
1
1 s
\ I
\
3
>
S
a I
1 3
\
2
!
i
Miles from Scoltli
Fig. 9. Profile, Harlowton, Montana to Seattle and Tacoma, Washington
?6S April, 1920
GENER.\L ELECTRIC REVIEW
Vol. XXIII, Xo. 4
1 ' ' , . '. 1 1 i '1 1 1
! ? i -^
. « i^ ' T ±4^
i 5 .*! sj.p J I
• ^ «:l*llllll^ Ip
i'l 1 --fl lil ^^~^ -i -l^« -^ -ti '-^
j^ V "" t' ' ' ' ■ '"^ 'i
i trri--- SS - ; "?»M«!>-w\
Ti" ?■ i*=^"" :^'^'^ ^^ ■ it T 1 ^ \
■ ^' H — rpi" ' i _u V
'■ t' " ip ' ^ ^
"■ 1 5 ""xlstTI ' ■ ' .
^ ? \^ \ I r 1 1 «
i n ' ' "*
1
; ]i s, _j I.M 1 1 ■ 1 ; .
*..-j-J.4 ,-ja - .| w^afj.-^'--".^
- a-'sUjL III ' 1 i ij :.-'_!_
^ !:| t [r - J-^ ^ ' "/*» e^^ ?; -'
1 M 1' 1 1 '■'' \
J. 1 TIs^Iu ' ^^ 4-
S '8 ■ 1 : i=^ .»
tiZ "^ -l-i.J-^ X ^v ^-
\i > ; 0 „ 1 V
i -^ i'^^k 4-Am I ' '- -fi-H-
— ^ ... j iL-s-H Ou^^tfOfj
ITT "f q"" iK 1 L _(_
5^ ' '» "- i: I 4-i
d^ I R 1 - - -1-
r-, ,., __ ._^ . .... J
^^ 4 — \ -i- 4-
!? J _; I.J.I .§ .s_jL .§ ., s_^ ,§ _^..>
TE S 3 fi 6 iX s s-i-i 4- ^
'•^ '^+rHHr »!> «-- .T--<s •.'««>/)-(-4-f fs . rH" rH" "'^
°^, l" 7 _,_ 1^
'~^\ — r- X' i
ii^ '1^+ ' ' 1 ? yX d 1 1
, 5 1 t: 1 IS ^ „/^K->yJ 1 - .1
1 ! r ^'l . i 1 \ V \
',] k \ \ \ \ I / "> 1
i 5 ' : 1 > , ' c - ......
» 1 ' '^'H J '
g ! WTr , t, K s ^ 1 1
> :|ii% U^^.'^i, -r '
71 '- * li t -iT , s. "
^f--' T! «^ 1 ~«
1 1 , ^ 1 IT ^^ j:
~r-' CI ^. tl
-j-j^ — ^T- :: g ± is^X
■ Ji 3-L .- ^. , . -L -- ^vt T
4t li! ^ -uSI ^^ x
II ^U-C. - -r 4t » ^ ^.l>ec//i>j.,Gov y.
' 1 ,¥^\ lZi,x% ^_ -V ^^
i 1 ,2 inii''~ X -"X
i ilXil"" 4 X :i X
J. __.5__1_ _Sa-
itrx"'?tj"" §' ~" ' _ X' i\x
l» ,»A»,««=»-V 1
X-UU 1 ^ xs^
— tr-^TT-^ l ^ *= si,
Hi f.r jU. * > " *"* --o*^ ■
_^ **• 1 1 r -M*- OP"^foe73r~'m^»fSOKr ■
■ " i", 1 1 _ 1 1 1 MM ,
■ -■ -h- ; 1 , ■ ! t
' ] fM -.4 uj€se'»':j<'»S
■ " »; - ~ -.«.-^,->j-y.
|-| I -1 ; 5 : ! 1 1 1 1
■ iXi^l'X^ - ■" _ _. J 1 M M "
'n^r^f < J-' ■ ^•^''«•p~r^^
4LiX-sl X\\-^ -A- t -L A
1 , ,k^ •St ' ss^_-i_ _ X-<^JJ,l- '^.3
^^■^ -^'■^l = ''IjSufJ S* ^ 5 1'^ '
^ xxl iti ill- " " X 1
tz. i 1 ^ fls^ftl- - -f
1 '■J 1 1 2iX \
I 1 1 1 * ^
It is of particular interest to note
that all of the substation apparatus
for the Cle Elum, Hyak, Cedar Falls.
Renton and Tacoma substations is a
duplicate of that in the substations
on the Rocky Mountain and Missoula
Divisions except for the high-speed
circuit breakers and the direct-current
feeder disconnecting switches. One
high-speed circuit breaker is used per
motor-generator set instead of one
per station, which somewhat simplifies
the arrangement of connections. The
high-speed circuit breakers are of a
more simple design than those origi-
nally furnished and are interchange-
able with those used on the locomo-
tives. The contacts are held closed
magnetically, eliminating the use of
any latches or toggles. The opera-
tion is effected by a shifting of the
flux upon an increase in the main
current, and the operating arm when
released is moved quickly by a hea\->-
coiled spring. They will thoroughly
protect the direct-current generators
from damage due to a short circuit
by preventing the current from ex-
ceeding GOOO amps. This avoids
excessive strains in the armature
windings, which eventually might
weaken the insulation and result in
the burning out of an armature.
The motor-generator sets are of
2(100 kw. capacity each, consisting of
two 1000-kw. compound-wound l.)t((t-
volt direct current generators con-
nected in series for 15000-volts and flat
compounded from no load to loO ])er
cent load, driven by one 250(.VkA--a.
2:UX)-\olt three-phase ()0-c\-cle syn-
chronous motor, and two direct-con-
nected direct-current exciters, one of
12 kw. capacity for exciting the fields
of the two generators and the other of
;{() kw. capacity at 125 volts for ex-
citing the fields of the synchronous
motor. The synchronous motor ex-
citer is comi)Ounded by the line cur-
rent of the generators in order to pro-
vide the most economical excitation
for the s\-nchronous motor over the
wide variation in the load. This com-
])ounding also helps in the regulation
of the alternating-current line, to com-
pensate for drop in voltage due to
load, as it is arranged so that the
motor operates at a lagging power-
ELECTRIFICATION OF THE COAST AND CASCADE DIVISIONS
2G9
factor on light loads and at a leading power-
factor on heavy loads.
The sets are cooled by external automatic
blower equipments which are not started until
the load reaches a value sufficient to produce a
predetermined heating. The blowers are again
shut down as soon as the load is reduced to an
amount that will jjroduce a temperature below
this value. This arrangement materialh' in-
creases the all-day efficiency, as the average
load will probably be slightly below that neces-
sary for the blower equipments to operate.
tension winding is divided into 44 sections per
leg insuring a low voltage between sections and
thorough ventilation of the coil stack. Taps
on the low-tension winding give the desired
range of voltage in the high-tension winding
from !t2,40U volts Y to 102,000 volts Y.
The transformer tanks use external tubes for
circulating the cooling oil, which is the same
construction as was used on the transformers
previously supplied. The high-tension bush-
ings are of the oil filled type and the low-
tension bushings of the solid type. Both
Figs. 11 and 12. Front and Back Views 3600-volt Direct-current 1500-ampere Circuit Breaker
These sets are designed to carry 300 per
cent load for five minutes when operating
either as straight synchronous motor-genera-
tor sets or inverted. The operation of some
twenty sets of exactly the same design on the
Harlowton-Avery electrification for the past
four years has been very successful.
The transformers are of 2.500 kv-a. capacity
each, oil insulated, self-cooled, wound for
102,000 volts Y primary and 2400 volts delta
secondary with one half voltage starting
taps. They are of the circular-disc core-type
design with the windings mounted on three
vertical members of the core. The high-
bushings have their ground sleeves extended
from the cover beneath the oil level to obtain
an electrically neutral atmosphere in the
chamber above the oil. This construction
eliminates the possibility of an explosion due
to static discharge.
The main circuit breakers have com-
bined series and shunt blowout coils with a
large magnetic circuit suitably proportioned
and an improved narrow arc chute which
insures the circuit being opened under all
conditions of operation. At the saine time
a gradual reduction of the current is effected
so as to keep the potential strains of the
270 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
various parts at a comparatively low value.
The design of the breaker is very rugged and
great care was exercised in proportioning the
various parts in order to insure obtaining
the desired operating characteristics.
Protection from lightning and surges on
the transmission line is taken care of by one
aluminum-cell lightning arrester per sub-
station, connected to the high-tension bus
with choke coils installed in the high-tension
leads of each transformer. The horn gaps in
the case of the flat roof substations, which are
used where there is ver\^ little snow, are
installed on the roof; and in the case of the
hip-roof stations used in the snow belts
the}' are installed inside the station. The
protection afforded by those in operation
for the past four years has been remarkable,
as ver\' little if any trouble has been experi-
enced from lightning.
Great care was taken in the selection of
the protecting scheme of the high-tension
transmission system to insure continuity of
W
Fig. 13. Main Line, showing Overhead Construction
Looking East from Cedar Falls Substation
seirice, in order that when trouble occurs
on the high-tension line the power would
be cut off from only the section in trouble.
In order to meet the selective protection
necessan,' under the various conditions of
operation, three main types of relay are used;
viz., an induction relay, an induction three-
phase reverse-power relay, and an inverse
time-limit relay. The first two types of
relays have been in use for several years and
do not need further comment. The inverse
time-limit relay, however, is a new device
developed specially to meet the requirements
of selective protection for this particular
system. It has a truly inverse time-limit
cur\e and its construction is such as to insure
that the adjustment will remain permanent
for a long time. The working elements
consist of two parts, one of which is to all
intents and purposes an ammeter element,
and the other a definite time element.
The overhead construction is of the modi-
fied flexible catenary type using two-4/0 copper
trolley wires flexibly suspended side by side
from the same steel messenger by independent
hangers alternately connected to each wire.
Forty-foot wooden poles suitably guyed and
spaced are used except in crossing the Col-
umbia River and on other special work where
steel construction is used. Bracket construc-
tion is used wherever the track alignment
will permit, and cross-span construction on
passing tracks and in yards. The length of
the trolley poles is sufl!icient for two cross-arms
at the top on which are carried the direct-
ctirrent feeders, the 4400-volt signal wires.
and the power limiting and indicating system
wires. A supplementary 4 0 negative feeder,
which is tapped to the middle point of every
second reactance bond, is carried directly on
top of the poles without the use of an insulator.
The positive feeder is tapped to the trolley
wire at everv seventh pole, or approximately
ever>- 1000 feet.
Power for the operation of this division is
supplied by the Inter-Mountain Power Co.
which in turn purchases its energy from the
Washington Water Power Co., and the Puget
Sound Traction Light and Power Co., both of
which ha-vc large watcrpower developments.
Thus the change to electric operation saves
the coal and oil previously used for steam
operation.
The power supplied b\- the Washington
Water Power Co. is furnished from its Long
Lake Plant northwest of Spokane by a 118-
mile transmission line to the Taunton Sub-
station. The power supplied by the Puget
Sound Traction Light & Power Co. comes
over a ten-mile transmission line from its
Snoqualmie Plant to the Cedar Falls and
Renton Substations. Power to the other
substations is sujiplied by the Railway
Company's own high-tension transmission
lines which connect between all of the eight
substations except the Cedar Falls and
Ronton Substations. The construction of this
ELECTRIFICATION OF THE COAST AND CASCADE DIVISIONS 271
line is similar to the line which has been in (3)
service on the Rocky Mountain and Missoula
Divisions, practically the only difference being
that there are a number of transpositions in
order to reduce as far as possible any induc-
tive interference on the neighboring telephone (4)
and telegraph lines. The transmission con-
struction, except where special construction
is necessary, such as on cur\'es, etc., consists
of 45 and 50-foot Idaho cedar poles with two
cross-arms, on which are carried the 100,000
volt lines on suspension type insulators and
an uninsulated ?/g-inch steel ground wire. The
high-tension conductors are 2/0 stranded
copper, with a with a hemp cor^.
The work undertaken by the Chicago (5)
Milwaukee & St. Paul Railway has been
Development of the High-Speed Cir-
cuit Breaker by which com-
mutating apparatus including loco-
moti\-es can be protected from
injury due to short circuits.
Development and successful appli-
cation of the Twin Trolley Wire
by which the operation of the
heaviest freight trains can be accom-
plished at 3000 volts without any
sparking at the trolley wire at any
speeds within safe operating require-
ments. The wear on the trolley
wires which have been in service for
the last four years is inappreciable.
The development of the Slider Panto-
graph in connection with the adop-
Fig. 14. Overhead Construction Looking East Toward Cedar Falls Substation
notable for the development in electric
traction brought forth, and as these develop-
ments have been in successful operation for
some time, they are here summarized.
(1) Commercial application of Electric
Regeneration by which power is
generated by trains on descending
long grades and is returned to the
line for use on other parts of the
system .
(2) Development of a Power-Limiting
Indicating System by which the
power supplied to a system at a
number of points can be totalized
at one point and the maximum
power available for operation at
all times can be controlled.
tion of the twin trolley wire has
been as notable as the adoption of
the twin trolley wire itself.
(6) The introduction of the compounding
of the synchronous motor exciter
of a motor-generator set used for
supplying the high-voltage direct
current, by means of the main direct
current so that the synchronous
motors operate economically under
all conditions of load and at the
same time compensate in a measure
for the drop in line voltage due to
load.
(7) Development of a High Speed Passen-
ger Locomotive especially adapted to
trans-continental service over heavy
mountain grades and severe curves.
272 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 4
Passenger Locomotives for C, M. dS, St. P. Rwy.
By A. F. Batchelder
Engineer Locomotive Department, General Electric Company
and
S. T. DODD
Railway and Traction Engineering Department, General Electric Company
The Chicago, Milwaukee & St. Paul Railway since 1915 has been operating electrically over the moun-
tain ranges of ^Iontana. The results of this operation have convinced the directors of the railroad of the
marked advantages of electrification; and, as a consequence, they have extended the electrification over the
Cascade Range. When placing orders for locomotives for this extension, the railway company proposed to
purchase locomotives designed strictly for passenger service, their characteristics and equipment to be those
most suitable for this purpose. In the following article the authors discuss the details of the electrical and
mechanical construction ofjthe new passenger locomotives, five of which were completed last year at the Erie
plant. — Editor.
D
lECE.MBER 9,
1915, may be
considered the date of
the initial electrical
operation over the
electrified lines of the
Chicago, Milwaukee
& St. Paul Railway.
During the following
winter the electrifica-
tion was extended
over 440 miles of route
from Harlowton,
Montana, to Aver\-,
Idaho, a section which
crossed the Belt, the Rocky, and the Bitter
Roots Mountains. The locomotives for this
initial electrification were of the geared type,
designed and built especially with a view to
the most economical operation of the freight
serA'ice. The locomotives for passenger ser\'-
ice differed from the freight locomotives only
in the details where it was absolutelv necessarv
A. F. Batchelder
to meet the operating
requirements, such as
changing the gear
ratio to increase the
speed and providing
each with heating and
lighting equipment.
Three years later,
in 1918, the successful
operation of the orig-
inal equipment had
convinced the railroad
company's officials of
the economical ad-
vantages of electric
operation, and they decided to equip an addi-
tional section extending over the Cascade
Mountains between Othello, Washington, and
Tacoma, Washington, a distance of 212 miles.
In choosing the equipment for the new exten-
sion, it was decided to give special emphasis to
the requirements of passenger scr\Mce and to
purchase locomotives which were primarily
S. T. Dodd
Fig. 1. Three-quarter View of New Direct-current Electric Locomotive
PASSENGER LOCOMOTIVES FOR C, M. & ST. P. RWY.
273
'•-.IK-
dcsigned with this in view, taking advantage
of any details whieh would assist in the
proper and economical operation of passenger
trains. For the freight service it was decided
to retain the geared locomotives that were in
use on the Harlowton-Avery Division, chang-
ing the gear ratio where necessary to rneet
freight conditions, and using only locomotives
of the new design for passenger service.
To meet the specifications for the passenger
locomotives, the General Electric Company
has designed, completed, and tested a locomo-
tive which appears to embody the necessary
qualifications and to successfully fulfill the
requirements, both from electrical and me-
chanical standpoints. In designing the loco-
motive, particular attention has been given
to the features affecting safety, reliability,
efficiency, convenience of operation, effect
on track, and cost of maintenance. The
locomotive has especially good riding quali-
ties; it has no apparent effect on the align-
ment of the track, and to a marked degree
it is free from transverse movements or
oscillation which would tend to create lateral
pressure against the rails.
It is the intention of this article to give a
description of this locomotive, which differs
in many ways from the locomotives that are
now in operation on the Harlowton-Avery
Division, to indicate the reason for choosing
this design, and to call attention to some of
the principal features which differ from
usual practice. Briefly stated, the service
requires the locomotive to haul a 95()-ton
passenger train over the mountain divisions
of the Chicago, Milwaukee & St. Paul Rail-
way at 25 m.p.h. up 2 per cent grades, with a
maximum operating speed of (iO m.p.h. on the
level, and to provide regenerative braking on
the down grades at speeds consistent with
safe operation. Fig. 1 shows a view of the com-
pleted locomotive and train. Fig. 2 is an outline
drawing of the side elevation, giving the general
dimensions. Fig. 3 is a section through the ap-
paratus cab, showing the location and arrange-
ment of the principal pieces of apparatus.
It will be seen that the running gear is
composed of four individual trvicks, two
end trucks having three axles each, and
two center trucks having four axles each.
These trucks are connected together by special
articulation joints. The motor armatures
are mounted on the axles and the motor
fields are carried on the truck frames.
The superstructure is made in two sections
of similar design with a third section between
them. The third or central section contains
the train heating equipment, which consists
274 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
of an oil fired steam generator together with
water and oil tanks. This unit is complete
in itself, and is carried over supports attached
to the two middle trucks. It can be readily
removed for repairs without interfering with
any other part of the locomotive. It is
placed between the two operating cabs in
Fig. 3. Cross-Section of Apparatus Cab
order to be easy of access to the engineers'
helper, or fireman, from either end location.
The two end sections are similar to each
other in appearance. The operator's cab in
either section is on the inner end next to the
heater cab just described, in order that the
operator will be convenient to the heater and
in order to allow a maximum space for
apparatus in the apparatus cab or outer end
section. Another advantage of this arrange-
ment of cabs is that the operator can have
access to any section of the locomotive requir-
ing his presence without passing through a
section containing high-tension apparatus.
The engineer's or operating cab contains a
main or master controller, the air brake
valves and handles, and an instrument panel
containing air gauges, ammeters, and speed
indicator. The engineer uses either of the two
operating cabs according to the direction in
which he is running.
A door gives access from the operating cab
to the apparatus section, which extends with
a cylindrical top to the extreme end of the
locomotive. The cylindrical construction
naturally adapts itself to the protection of the
apparatus included; and, in addition, it has
the advantage of allowing a clear vision for
the operator from his normal operating posi-
tion. Contained in this apparatus section are
the resistors and contactors to control the
power circuits of the locomotive. The start-
ing resistors are placed in two rows on each
side of the central passage just above the
floor of the superstructure, and they are
covered at the sides by removable covers
which when opened allow the separate resistor
boxes to be slid out upon the longitudinal
running board outside of the apparatus cab.
The air compressor for the air brakes, the
motor-generator set for train lighting, and
the storage battery for m.arker lights and
emergency control stand upon the same level
as the resistors and can be removed or re-
placed in a similar m.anner. Above the
resistors are located the contactors with their
arc chutes facing a central aisle two feet wide.
This arrangem.ent allows ample arcing space
and room for inspection of the contactors.
Above the contactors is the cylindrical roof
of the locorrotive with trap doors for inspec-
tion of the back connections and insulation,
and for removing the contactors in case
replacem.ent is necessary. The whole design
and arrangement of this apparatus cab lends
itself to a maximum economy of cost and
material, as well as to convenience of inspec-
tion and repair of apparatus.
Motors
The motors are of the well known bi-polar
gearless design which was adopted by the
New York Central Railroad fourteen years
ago. The continuous operation of these
Fig. 4. Bi'polar Gearless Armature and Wheels
motors since that time, in hauling heavy
passenger trains between the Grand Central
station and Harmon, proves them to be of a
design well suited for the service. This motor
has demonstrated its remarkable reliability
and low cost of maintenance.
PASSENGER LOCOMOTIVES FOR C, M. & ST. P. RWY.
275
To insure light weight per axle, flexibility
in control, good truck arrangement for curving
as well as for high-speed running, 12 motors
were chosen, each of relatively small capacity.
They are especially designed to withstand
high temperature, being insulated with mica
and asbestos.
Fig. 4 shows the motor
armature complete, built
directly on the axle with the
wheels pressed and keyed in
place. The continuous rating
of each motor at 1000 volts
and with 120 degrees rise by
resistance is 266 h.p., corre-
sponding to 3500 lb. tractive
effort at the rim of the drivers
at a speed of 28.4 m.p.h.
Forced ventilation is em-
ployed for cooling. The
armature core is provided
with holes for the passage of
the ventilating air. Blowers
are located above each motor
armature and deliver air at the
commutator end of the motor where it
divides, part passing through the armature
and part back through and around the field
coils where it escapes upward and is afterwards
used for ventilating the starting resistors.
This type of motor gives very high effi-
ciency in average operation, it having no
journal bearings or gearing. It lends itself
nicely to simple and compact locomotive
design as the frame is made use of to furnish
the entire path for the magnetic flux. The
pole pieces and field coils are fastened to the
cross transoms of the trucks and the magnetic
flux passes horizontally in series through all
twelve motors, finding a return path through
the locomotive frame. The articulated joints
between the trucks are made in such a manner
that large surfaces are in contact to provide
a low reluctance path for the flux. The pole
faces are made flat in order to prevent them
from coming in contact with the armature
during the vertical movement of the truck
frame on its springs, or when removing or
replacing the armatures. A minimum clear-
ance of }/s inch on each side is allowed be-
tween the armature and the pole piece tips.
The brush-holders are bolted to the transom,
an arrangement which permits the brushes to
move up and down with the fields as the frame
rides on the truck springs.
Control
In choosing the control apparatus special
care has been taken to use those individual
pieces of apparatus best suited to the parti-
cular requirements. Where single independ-
ently operated switches are necessary, as on
the resistance notches, electro-magnetic con-
trol is used. Where several switches are
required to operate at one time, as in changing
from series to parallel motor connections.
J4<!.00a
'
~—
—
—
—
—
'~~
\
\
\
\
\
\
\
■0
\
\
J...
lencu :^.'^otprs
/T ser,es
\
,
,'-
\
\
■^ eoDoo
1
\
\,
\
\
\
\
■^,..
\f
N,
^
\
-"<
-io
1
2im>
»
\
■<i
^;-.
■^
<^
^-<t^,
J*
^
»3l
^^11
r^
=^
^
— ■M'?'-*
'^'-'p^k
—
—
...
■f
Sa
L_
--'■
r-
—
, —
-
p^
—
' 1
7
/-
A
Z5 X ^
'to
45
£0
£S
60
^
Fig. 5. Locomotive Characteristics
banks of switches with electro-pneumatic
cam control are used, thus insuring positive
operation, eliminating interlocks, and simpli-
fying the wiring.
The control for motoring is arranged for
four motor combinations.
The first combination has 9 rheostatic steps,
one full-field step, and one tapped-field step,
with twelve motors in series across 3000 volts.
The second combination has 6 rheostatic
steps, one full-field step, and one tapped-
field step, with six motors in series and two
sets in multiple.
The third combination has 8 rheostatic
steps, one full-field step, and one tapped-
field step, with four motors in series and three
sets in multiple.
The fourth combination has 8 rheostatic
steps, one full-field step, and one tapped-
field step, with three motors in series and four
sets in multiple.
These combinations result in a total of 39
control steps with a choice of eight operating
speeds, exclusive of the resistance steps. The
locomotive characteristics on the various
steps are clearly shown in Fig. 5.
The regeneration of power for braking is
accomplished in a simple manner by using
some of the motors for exciting the fields of
the others, which in turn arc used as generators
to return power to the line.
As a provision against short circuits, or
extreme overloads, there is provided in the
276 April, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, Xo. 4
apparatus cab a quick acting circuit breaker
which will protect the circuit in less than
1 100 of a second.
Mechanical Construction
For flexibility in cur\nng, the running
gear is made up of four trucks, each of a
Fig- 6. Centering Device of Leading Axle
relatively short wheel base. The two middle
trucks have four driving axles each; and the
two end trucks, two driving axles and one
guiding axle each, making a total of 14 axles.
The trucks are connected together with ar-
ticulated joints which allow of no relative
lateral mo\'ement between them, so that
each truck positively leads the following
of limiting the lateral oscillations of the
locomotive structure, which tend to distort
the track, and of minimizing the effect on
the track of such oscillations as ocoir. If a
locomotive were built with a rigid wheel base
as long as the total wheel base of the present
locomotive (67 feet), the lateral oscillations
could not reach any large angular value.
However, on account of the long wheel base,
such a locomotive would be incapable of
taking curves. By articulating the wheel base
the locomotive is capable of accomodating
itself to track cur^'ature; and, at the same
time on account of this articulation and the
consequent guiding effect of one truck on
another, the lateral oscillations on tangent
track are minimized in the same manner as
would be done by the use of a long rigid
wheel base.
To soften any lateral blow that may be
given against the rail, the leading and trail-
ing axles are allowed a movement of one-
half inch, relative to the truck frame, either
way from their central position. This move-
ment takes place against a resistance intro-
duced by wedges above the journal boxes
which tend to hold the box in its central
position and to give a dead beat action
opposing the motion. This wedge construc-
tion is illustrated in Fig. 6. To further
l)rotect the track from lateral displacement
on the ties, the outer end of the superstructure
is carried on rollers, bearing on inclined planes
upon the truck frames; while the inner end
of the superstructure is rigidly bolted to one
of the middle trucks. This construction
Fig. 7. Side View of Locomotive
truck. This is for the purpose of reducing
flange wear on cur\-es and lateral oscillation
on tangent track.
The most important problem that has to
be faced in the design of a locomotive for
high-speed passenger ser\-ice is the problem
tends to hold the leading and trailing trucks
in their central position. When a blow is
delivered by the leading or trailing truck
against the rail head, the superstructure is
displaced laterally across the outer trucks.
In such a sideways displacement, the weight
PASSENGER LOCOMOTIVES FOR C, M. & ST. P. RWY. 277
of the superstructure rolls up on the inclined As a matter of record, it should be said that
plane on that side, and thus transfers weight the first of these new locomotives was deliver-
to the rail that is aflfected, thereby increasing ed to the railway company at Deer Lodge,
adhesion of the rail to the tie. This action Montana, on December 14th, 1919, and
really has two results: It not only increases was put in operation handling passenger
the holding power between rail and tie at trains between Deer Lodge and Avery,
that point, but it introduces a time lag and For convenience of reference, Table I gives
increases the time and distance during which a summary of the principal dimensions and
the pressure is delivered to the rail head. characteristics of this locomotive.
TABLE I
LOCOMOTIVE DIMENSIONS
Total weight 521,200 lb.
Total weight on drivers 457,680 lb.
Weight per driving axle 38, 140 lb.
Dead weight per driving axle 9,590 lb.
Weight per idle axle 31,750 lb.
Dead weight per idle axle 3,560 lb.
Length overall 76 ft. 0 in.
Width overall 10 ft. 0 in.
Height over cabs 14 ft. 11 5^ in.
Height over pantograph, looked down 1 6 ft. 8 in.
Total wheel base 67 ft. 0 in.
Maximum rigid wheel base 13 ft. 9 in.
Diameter of driving wheels 44 ft.
Diameter of idle wheels 36 ft.
Size of journals 6 ft by 13 ft.
Dimensions of operator's cab 5 ft. by 10 ft.
Dimension of heater cab 14 ft. 11 in. by 10 ft.
Heater capacity 4000 lb. steam per hour
Water capacity 30,000 lb.
Oil capacity 6,000 lb.
Compressor capacity 150 cu. ft. per min.
Number of motors 12
Type of motor (Bipolar-) GE-100
Diameter of armature 29 ft.
Clearance between bottom plate and top of rail 554 in.
Working range of pantograph 9 ft. 0 in.
Locomotive Rating v Tapped Field Full Field
Total horsepower, one hour motor rating 3,480 3,380
Total tractive effort one hour motor rating 36,000 46,000
Speed, m.p.h 36.2 27.5
Total horsepower continuous 3,200 3,200
Total tractive effort continuous 32,000 42,000
Speed, m.p.h 37.8 28.4
27S April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
TABLE ]
Motors in
Series
Number
Groups
of Accelerating
Steps
SPEED AT CONTlNUOt;S
BATING
Full Field
Tapped Field
12
1
10
5.0
8.0
6
2
i
12.3
18.7
4
3
9
20.0
29.0
3
4
9
27.0
40.0
Control Equipment of the New Locomotives
for the C, M. C^ St. P. Rwy.
By F. E. C.\SE
Railway Equipment Department, General Electric Company
Power not under control would be useless — in fact would be destructive in most instances. Control is,
therefore, a vital factor in the production and application of power. Since the difficulty in designing control
equipment for a machine increases with the variety of the conditions under which the machine is to operate,
the control of an electric locomotive employing regenerative braking probably presents the most complex
problem. That a satisfactory solution has been arrived at is evidenced by the successful performance record
of the earlier C. M. & St. P. locomotives. The new locomotives are equipped with essentially the same
system of control, the principal modifications being along the line of simplification. — Editor.
nPHE five new pas-
-'■ senger locomo-
tives which were re-
cently delivered by
the General Electric
Company to the
Chicago, Milwaukee
and St. Paul Railway
differ materially in
appearance from the
original type* and the
arrangement of elec-
trical apparatus has
_ also been somewhat
changed. The adop-
tion of twelve bi-polar gearless m.otors per
locomotive instead of eight geared ones, as on
the earlier locomotives, resulted in a different
electrical grouping of the motors, and further
development of regenerative electric braking
permitted a simplifying of the method of
control.
The motors are air blown and each has a
continuous rating of approximately 250 h.p.
and a one hour rating of 270 h.p. The use
of a larger number of motors made it possible
to provide a greater variety of groupings for
motor operation with a corresponding increase
in the number of running speeds.
Motoring
There are four motor combinations, the
motors being connected 12, 6, 4 or 3 in series,
and the fields may be weakened with each
grouping to secure four additional running
speeds. This field weakening is obtained by
tapping or cutting out a portion of the wind-
ing.
The number of accelerating steps for each
of the four motor groupings, and the speeds at
continuous capacity with full and tapped
field, are shown in Table I.
* A full description of these earlier locomotives appeared n
the General Electric Review, November, 1916.
Regenerative Braking
Broadly speaking, the system of regener-
ative bralving is similar to that on the previous
locomotives, in that the series motors when
regenerating have their fields separately
excited to a density higher than would be
obtained for motoring at corresponding speeds.
In consequence, the combined armatures
generate a higher voltage than that of the line
and return power to it. The main difference
is the source of energy for exciting the motor
fields and the consequent arrangement of the
motor circuits. In the original equipments
a separate motor-generator was used for
exciting these fields, but in the new ones part
of the motors are employed as exciters for
the fields of the other motors and no separate
machine is required.
The motors are connected in two virtually
independent groups for regeneration and these
groups may be connected either in series or
parallel, depending upon the train speed
desired. Each group consists of four motors
which generate the power returned to the
line, and two which generate current for
exciting the fields of all six motors. When the
regenerated current returned to the line is
of the same value as the exciting current in
the fields and is also equal to the continuous
rating of the motors, the locomotive speed
is about 1 1 miles per hour for the series con-
nection of the regenerating groups and 23 for
the parallel connection.
CONTROL EQUIPMENT OF NEW LOCOMOTIVES FOR THE C, M. & ST. P. RWY. 279
The amount of regenerated current required
to maintain any constant speed is dependent
upon the train weight and grade, and it is
necessary to vary the field excitation to secure
the proper loading of the motors.
Above about 22 miles per hour it is desirable
to have the two regenerating groups connected
in parallel in order to niake the armature and
field currents as nearly equal as possible.
With these connections power can be returned
to the line at speeds up to more than GO miles
per hour.
The control is so arranged that it is possible
to start regeneration without first passing
through the motor running positions.
With the pantograph lowered, or power off
the line, it is possible in an emergency to make
all the regenerating connections by means of
a storage battery provided on these loco-
motives. The braking effort obtained keeps
the train bunched and the current generated
may be used for operating the main air com-
pressor so that the air brakes can be employed
on the train.
Main Circuit Apparatus
The principal pieces of apparatus in the
main circuit are:
2 sliding contact pantograph trolleys.
3 knife blade disconnecting switches.
1 magnetically operated high-speed circuit
breaker.
2 magnetically operated line contactors.
4 groujjs of magnetically operated resistor
contactors.
5 groups of electro-pneumatically control-
led, cam operated contactors for series-parallel
and regenerating motor connections.
2 electro-pneumatically controlled, cam op-
erated motor reversers.
2 electro-pneumatically controlled, cam op-
erated field tappers.
4.3 cast grid resistors.
Most of the apparatus is essentially the
same as that used on the earlier locomotives.
Pantograph Trolleys
The pantograph trolleys are the same
as those of the previous equipment, long
continued operation in very exacting service
having shown that they needed no important
change. Although there are two trolleys on
each locomotive, one is of ample capacity for
collecting the current, the other being held in
reserve.
The two sets of lubricated copper strips,
which are mounted on the top of the panto-
graph for making a sliding contact with the
two trolley wires, have given remarkably
long ser\dce, many of them lasting for more
than 10,000 miles.
In both installations the trolleys are raised
by admitting compressed air to two pistons,
which extend the elevating springs, and are
lowered by exhausting the air. On the new
Fig. 1. Master Controller. Tlie mnster controller provides for
8 motoring speeds and 2 combinations of motors for regen-
eration. During motoring the locomotive may be operated
with 12, 6, 4, or 3 motors in series and the motor fields may
be tapped in each combination. During regenerative
braking the motors may be operated either 8 or 4 in series
locomotives the raising and lowering is
controlled by means of a small switch, in
either end cab, which operates an electrically
actuated air valve located close to the trolley.
This arrangement permits a quicker operation
in an emergency than the manually operated
valve, owing to the shorter distance between
valve and trolley.
With the pantograph down, the storage
battery provides a source of energy for con-
trolling the raising valve and for operating
280 April. U)2l)
.l| o
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 4
(— ^0 — o o-
Qi/JCtc j^C 7V/VC
1 c/'Ku/T omc^/rc/r
-r, — t
/t'^rj/srZy/f'^
AA--e-WhArnr^--^^^^
/ e \ -z /
t-O-rOH
3 3_
H>=XH
* jT 6 6. J- 4;
-o^ — *V*~°2*~
J — J/ — T — ^1 — 9
-TT — tT^
9H.oK7y
/2 // // /Z
9 3 7 \ 7 e 9
Hi-oH. ^
C C
Fig. 2. First Running Position for Motoring with 12 Motors in Series. Contactors Marked "C" arc Cam Operated
and the Remaining Ones Magnetically
•l| O
I
O |l-
y>*yt//y 3tv/TCM
rftOi-L^ r OIS COAIMCC TIMt
siv/rcH
:^
<iu/c/r AC r//vs
c/zpcu/r B^C^K^^
l-^i.//^c confT^crofrs
^y
h/ I,
T, — 57 — V,
-^SU6}
/rcsiaTO/ts
,oj£.-(D-(|>
EE
-\hX-ar-'!l\fAI\fij^^^fJ^
^..J^^/fl^isi/r^/r^J^^^
Fig 3. Second Riinnintj Position for MuloriuK with 2 Multiple Groups of (> Motors in Scries in Ench Group
CONTROL EOUIPMEXT OF NEW LOCOMOTIVES FOR THE C. ^L c^- ST. P. RWY. 2S;i
M •
f-A€4//^ ^t^/rc^
( ^/^r coffTACToma
-\WWi
IIXL
-.-JWiAAJi
:^y<sKe)
[c^ 1
<>~<.^>^,trqy>~af^^X~*^fli^t<>^\L^
' c I
Ot'Sjr^.O^D
/
Ft^. 4. Third Running Position for Motoring with 3 Multiple Groups of 4 Motors in Series in Each Group
.H o
/V1y^/M 'SkY/rCH
r^POL Lcr D/sc o/y/y^c r/A/Q
Cff^CU/T B^£AH£'^
/JV£ CO/fTAC TORi
y i I, 1 — vrs
-i. i i i
fftsisToin
r^\
syo-stizy^-Mj)
k>~^ (-U^B-y/^lV^|3j*H I •
i
J ^><3KIW--^^^^^^
-ST-T
^ K^U-OjO-^^V^VHh
/
^--Mm
9 a 7 7 a 9
<9K<9K^>
(O-Ji-
Fig. 5. Fourth Running Position for Motoring with 4 Multiple Groups of 3 Motors in Series in Each Group.
Overload Relay is Placed in Each Group.
282 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
a small auxiliary air compressor which pro-
duces a supply of air for raising the trolley if
the pressure in the reserv'oir is inadequate.
On the previous locomotives an auxiliar>'
base with pole is used for making contact
with the trolley wire when the locomotive
Fig. 6. Disconnecting Switch and Line Contactors. The knife
blade switch makes it possible to test the control apparatus
without applying power to the motors. The magnetically
operated line contactors break the motor circuit both when
the master controller is turned oET and when an overload
occurs
has been idle so long that the air pressure
in the reservoir falls below the amount
required to raise the pantograph trolley.
After the air compressor has produced a
pressure of about 50 pounds in the rescr\-oir
the pantograph trolley may be put up and the
pole trolley lowered.
Knife Blade Disconnecting Switches
A knife blade switch, mounted in a weather-
proof shect-stccl box, is located in a con-
venient place near each trolley for dis-
connecting it from the main lead to the
interior of the locomotive in case of damage
*"New Type o( HiRh-spced Circuit Breaker." p. 286.
or during inspection. The switch is of the
double-throw type and is so connected tha
when a trolley is cut out it will be groimded
for safet}' during inspection.
The switch shown in Fig. 6 is provided for
disconnecting the main circuit from the
trolley. It is used when it is desired to test
out the functioning of the different pieces of
apparatus without applying power to the
motors. In the down position of the switch,
connection is made to a coupler contact at
the side of the locomotive. When it is
desired to move the locomotive into a round
house or repair shop where there is no over-
head trolley wire, a cable leading from a
low-voltage supply may be attached to this
contact.
High-speed Circuit Breaker
This protective device for the main circuit
is fully described in another article* appearing
in this nimiber of the Review. Its operation
is much more rapid than anything pre-
viously used for the purpose on locomotives
and in consequence the damage resulting
from a motor flashovcr, or ground, will be
greatly decreased.
Fig. 7. Magnetically Operated Contactors. These coatacton
short circuit the accelerating resistors. They are assembled
in groups as shown in another cut. Similar contacton,
except that they are closed and opened by cams operated
pneumatically, are used in the groups for transposing motor
circuits
When a short circuit or other overload
occurs and the circuit breaker opens, it
introduces a resistance in the circuit which
limits the current to a nonnal amount. A
small switch, which is directh' operated by
the breaker, simultaneously opens the control
CONTROL EQUIPMENT OF NEW LOCOMOTIVES FOR THE C, M. & ST. P. RWY. 283
Fig. 8. Magnetically Operated Contactor Group. These contactors are mounted in conveniently handled groups on steel supports
An overload relay and two other control relays are shown installed below the contactors
circuit of those contactors which cut out the
accelerating resistors and the latter are
introduced in the circuit to reduce the current
still further. The main circuit is then broken
by the line contactors. This sequence of
operation permits of opening the overloaded
circuit with a minimum disturbance.
The circuit breaker is adjusted to carry
the total current for all four motor circuits in
multiple. In order still further to limit the
current in the individual motor circuits an
overload relay, which opens the holding coi^.
circuit of the quick acting circuit breaker, is
placed in series with each combination of
motors. On the geared locomotives the
main circuit is protected by a magnetic blow-
out copper ribbon fuse and the individual
motor circuits are provided with overload
relays which open the contactor circuit in
case of overload.
The circuit breaker is automatically re-
closed by a solenoid when the master con-
troller is turned to the first point. A lever is
also provided for closing the circuit breaker
manually. Opening the main control switch
permits the circuit breaker, as well as other
apparatus operated magnetically, to open.
Contactors
The line contactors are mounted on the
same frame as the main circuit disconnecting
switch and, aside from the magnetic blowout
Fig. 9. Cam Operated Contactor Group. The various motor combinations for motoring and regenerating are effected by these
pneumatically operatcQ contactor groups. The above illustrates one which has three positions; one for
6 motors in series, one for 4 motors in series, and the third for regeneration
2S4 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 4
and arc chutes, are similar to the contactors
for cutting out the accelerating resistors.
Both forms of contactor are directly operated
by means of electro-magnets from the low-
voltage circuit.
The series-parallel and regenerating con-
tactors are composed of main circuit parts
♦♦vvv»
Fig. 10. Field Tapper. A section of each motor field winding
is cut out of the circuit by two of these pneumatically
operated field tappers and additional speeds for motoring
are thereby obtained.
which are the same as in the resistor con-
tactors but they are operated by a set of cams
moved by air cylinders. The air cylinders are
controlled by electrically operated valves.
This arrangement of contactors was used
with great success on the previous loco-
motives. It permits of the simultaneous
operation of a considerable number of con-
tactors with a minimum of magnets, valves,
or air cylinders. Also in making transitions
from one grouping of motors to another, a
perfect sequence of contactor opening and
closing is assured.
Auxiliary Apparatus
The following arc the princijial pieces of
auxiliary apparatus:
() blowers.
1 main air compressor.
1 auxiliary air compressor.
1 motor-generator set for charging the stor-
age battery.
1 storage battery.
1 lightning arrester.
Blowers
The locomotive is provided with six double
blowers for cooling the main motors, each
half being capable of delivering approximately
3500 cu. ft. of air at 2-in. pressure. Each
blower is driven by a 12-h.p. series motor. The
circuits are so connected by four magnetic
contactors that the motors may be operated
six in series or in two groups of three each
in series on 3000 volts. The motors are
connected directly across the line voltage
when starting without any resistance in the
circuit.
Air Compressors
The m.ain air compressor has a capacity of
150 cu. ft. of free air per minute and it is
driven by a 3000-volt motor requiring a
current of 10.5 amperes when the compressor
is operating at a pressure of 135 lb. per sq. in.
It is started by connecting it across the line
by a magnetically operated contactor with a
starting panel in series. The panel consists
of a resistor and a series contactor which
automatically closes and short circuits the
resistance when the current through the com-
pressor motor has dropped to a predeter-
mined value. The coil of the contactor is
energized by a standard air compressor
governor when the air jjrcssure drops to
123 lb. or below, and is de-energized when
the pressure reaches 135 lb.
The auxiliary compressor which provides
a supply of compressed air, in an emergency,
for the pantograph trolley and other pneuma-
tically operated control apparatus, has a
capacity of approximately 10 cu. ft. of free
air per minute. It is operated from the
storage batter\- and requires about 25 amperes
when compressing at 70 lb. per sq. in.
Motor-generator
The motor-generator consists of a com-
pensated shunt motor of approximately 40
h.p. at 3000 volts and a shunt generator rated
at 25 kw. SO volts. This set provides a low-
voltage source from which the various pieces
of control apparatus may be operated, and
for charging the lighting storage batteries
located on the various cars of the train and
the auxiliary storage battery* on the loco-
motive. The generator voltage is held
constant by means of a single vibrating
regulator relay with its coil across the genera-
tor tenninals and its contacts acting to short
circuit the field of a very small generator, the
armature of which is in scries with the shunt
field of the main generator. This small
generator annature is mechanically con-
nected to the shaft of the generator which it
regulates.
CONTROL EOUIPAIENT OF NEW LOCOMOTIVES FOR THE C, M. & ST. P. RWY. 2S5
The set is controlled by a magnetically
operated contactor. When starting, the
motor of the set is thrown directly across the
line with two starting panels consisting of a
resistor and series contactor connected be-
tween the low side of the armature and ground.
As the shunt field is tapped from between
the two commutators to ground it has
practically double excitation when the set
starts, and it is connected across the low
comniutator when the series resistance is
short circuited. The two panels are so con-
nected that when the current through the
set is reduced to the value for which the
series contactors are adjusted, one of them
closes, short circuits one section of the
resistance, and connects the series coil of the
second contactor into the circuit. The
second contactor closes when the current
has again dropped to the proper value.
With the control arranged in this wav the
set automatically starts up when power has
been returned after an interruption.
Storage Battery
The storage battery is composed of 36 cells
rated at 95 ampere-hours, based on a 4}^ hr.
discharge rate. It is automatically connected
to and disconnected from the generator
terminals by a reverse current relaj'. It is
used for operating the auxiliary' air com-
pressor and other pieces of apparatus requir-
ing low-voltage current when the pantograph
trolley is not raised.
Lightning Arrester
The arrester comprises twelve standard
direct current cells of the liquid type con-
nected in series. Each cell is composed of a
glass jar containing aluminum plates which
are submerged in a liquid electrolyte. Balanc-
ing resistances of a high ohmic value are used
to equalize the potentials across the cells.
The Long Lake Station of the Washington Water Power Company located on the Spokane River.
Washington. This system is interconnected with that of the Puget Sound Traction. Light &
Power Company and furnishes energy to the Cascade Division of the C, M. & St. P. Rwy.
286 April, 1920
GENER.\L ELECTRIC REVIEW
Vol. XXIII, No. 4
Ne^v Type of High-speed Circuit Breaker
By J. F. Tritle
R.iiLWAY Equipment Engineering Department, General Electric Company
The higher the %'oltage for which a direct-current generator is designed the greater is the likelihood that the
machine will flash over and become damaged on short circuit. The problem of protection did not reach an
acute stage, however, until the advent of our latest high-voltage direct-current railway electrifications. The
generators for such systems require a degree of protection far greater than it is possible to secure by good com-
mutating characteristics alone and consequently an auxihary device was developed for the purpose. This first
took the form of a special circuit breaker employing the fundamental principle of a standard breaker but capable
of operating at 15 times higher speed by reason of powerful springs held in leash by a train of latches. The
article below describes a new type of high-speed circuit breaker that operates on an entirely different principle
nad is far superior to the older type. — Editor.
T!
'HE problem of
protecting direct-
current generators,
particularly high-
voltage generators,
from flashover, has
received a great deal
of attention in recent
years. Various im-
provements have
been made in the
comm.utating charac-
acteristics of the
machines, but as yet
J. F. Tritle . , •'
no commercial m,a-
chine has been built which is immune from
flashover under the most severe short-circuit
conditions unless it is protected by some
external device, such as a high-speed circuit
breaker.
Standard circuit breakers operate much
too slowly to prevent flashover on heavy
short circuits. Repeated tests have indicated
that to prevent flashover a circuit breaker
shotild operate, stop the current rise, and
reduce it below the flashing value in some-
thing less than the time required for a com-
mutator bar to pass from one brush-holder
to the next. On a GO-cyclc machine, this
means a speed of approximately eight one-
thousandths of a second. As the standard
circuit breaker operates in about eight to
fifteen one hundredths of a second, it has
less than one tenth the required speed.
A high-speed breaker using a refinement
of the principles of a standard breaker was
developed to protect the 3000-volt gen-
erators that supply power to the Rocky
Mountain division of the Chicago, Mil-
waukee and St. Paul Railway.* One was
installed as part of the equipment in each of
the 14 substations. These breakers have the
required speed of operation and have demon-
* "High-speed Circuit Breakers for Chicago. Milwaukee & St.
Paul Electrification," by C. H. Hill, G. E. RhVIliW, Sept.. 1018.
strated quite conclusively that direct-current
machines can be made practically immune
from flashovers and damage from excessive
overloads and short circuits. These breakers,
however, are rather large and expensive; and
the tripping mechanism, which holds the
Fig. 1.
ISOOampcre 3000-volt Direct-Current High-speed Cir-
cuit Breaker with Covers Removed
breaker closed against ver>' powerful operat-
ing springs, and which consists of a train of
latches and levers actuated by a solenoid,
requires great accuracy in manufacture and
adjustment.
NEW TYPE OF HIGH-SPEED CIRCUIT BREAKER
287
A new type of high-speed circuit breaker
has recently been developed which operates
on entirely different principles from the
original, particularly in regard to the method
of tripping and the arrangement of the
magnetic blowout. All mechanical latches
and triggers have been entirely eliminated.
The breaker is tripped electro-magnetically
instead of electro-mechanically. The spring
power necessar\- to operate the device has
been greatly reduced. The m.agnetic blow-
out has been improved and has a combina-
One of these new type breakers is installed
in series with each of the eight 3000-volt
2000-kw. motor-generator sets in the sub-
stations supplying power for the electrifica-
tion of the Coast and Cascade Division of the
Chicago, Milwaukee and St. Paul Railway.
7=\
j-...S.Z\^
Fig. 2. Cross Sectional View of 1500-ampere 3000-volt High-
speed Circuit Breaker
tion of two powerful magnetic fields and a
narrow arc chute, which increases the speed
of blowout and reduces the arcing space
required. The breaker is, therefore, ver\'
simple and rugged in construction and is much
reduced in size, weight and cost.
PO^i7'y£ BO'S
— t
•/■JP M/G» sP£^o c/f^cu/ r Sf'^'^^^jP
Fig. 3. Schematic Diagram of the High-speed Circuit Breaker
and the Connections for Installing It in the Negative
Side of a Generator
A similar breaker is also installed on each of
the five new bi-polar gearless type passenger
locomotives.
Fig. 1 shows a side view of this new type
of breaker with the covers removed, and Fig.
2 shows a cross-sectional view. Fig. 3 shows
its principle features, together with the con-
nections for installing it in the negative side
of a generator.
In this latter illustration, Fl and F2
represent a laminated field structure some-
thing like that of an ordinary' alternating-
current m.agnet. The poles of Fl and F2
are bridged by a ver\- light armature A
pivoted at P and held in contact with the
poles by a shunt coil SI energized from any
convenient constant voltage source, such as
the exciter circuit or the main bus. A series
bucking bar S2, which electro-m_agnetically
trips the breaker, is located between the poles
of the field m.agnet and in close proximity to
the armature. Thus the current flowing
in the bar produces the maximum change in
the annature flux, with the minimum change
in the flux interlinking the shunt winding SI .
The bucking bar simply shifts the flux from
the arm.ature to the air path at the right of
the bucking bar, thus causing the armature
288 April. 1920
GENERAL ELECTRIC REVIEW
Vol. XXIH. Xo. 4
to release as soon as its flux is reduced a
predetermined amount.
The tension spring attached to the armature
gives the high-speed opening of the contacts
and also provides a means of adjusting the
breaker. On account of the relatively light
armature and the fact that it is not necessary.'
to trip any latches to release the breaker, a
pull of less than 800 lb. is required of this
'iztoo
■^ 2400
^2200
I
^2000
I /SOO
^/tOO
a /■^C'O
1200
1000
800
too
400
ZOO
Fig. 4.
~0 .1 2 3 .4 .5 .« .7 .* S 10 I.I 1.7
Holding Coll Current (^mpcrei)
Calibration Curve for 1500-ampere 3000-volt High-
speed Circuit Breaker
spring for the 2o00-ampere capacity breaker,
which is far less than that recjuired in the
original type.
The main contacts Cl and CZ are of the
solid copper type used so successfully on
railway contactors. Contact Cl is materially
heavier than C2 and, when the armature is
released, Cl follows C2 for a predetermined
distance, but at a much lower rate of speed,
so that Cl and C2 begin to part practically
simultaneously with the beginning of move-
ment of the armature.
The blowout coil 5 is in scries with the main
circuit and is designed to give a vcn,' intense
field but of comparatively small area around
the main contacts Cl and C2. An additional
blowout coil, S2, is provided in the auxiliary-
arc chute and is automatically cut into the
circuit during the time the arc is being
ruptured. When the tips C\ and C2 begin
to part, the series coil 5 blows the arc upward
oflf the tips onto the arcing horns HI and H2.
As the arc moves further upward, it comes in
contact with the ends of the arcing horns
HS and ¥U^, between which is connected the
blowout coil S2. This inserts the coil in the
circuit and divides the arc into two parts,
one of which is blown upward through the
left-hand side of the chute between the arcing
horns HI and H3 and the other through the
right-hand side of the chute between the
arcing horns H2 and H!^. The coil S2 sur-
rounds the iron core F3 to which is con-
nected the field pieces F4 and Fo which cover
practically the entire area of the auxiliary
chute. The auxiHar>' arc chute is hinged at
both ends so that it can be easily swung out
of place for ready inspection of the main
contact tips. The sides of the arc chute are
Fig. 5. Short Circuit on 2000-kw. 3000-volt Motor-Generator
Set Protected by High-speed Circuit Breaker and Flash
Barriers. Line Resistance, 0.0 ohms; Limiting Resistance
1.2 Ohms. Tripping Point. 2250 Amperes. A. line voltage;
B, line currrnT C. 59,5 cycle Timing Wave
Fig. 6. Short Circuit on 2000-kw. 3000-volt Motor-generator
Set Protected by High-speed Circuit Breaker and Flash
Barriers. Line resistance, 0.0 ohms; limiting resistance, 1.2
ohms. Tripping point, 2550 amperes. A^ voltage across
contacts and generator series fields; fi. Line current; C.
Current in generator shunt field
also arranged in a novel way to provide an
arc chute materially narrower than the con-
tact tips, thus increasing the resistance of the
arc stream for a given length and giving the
maximum cooling effect to the vapors.
NEW TYPE OF HIGH-SPEED CIRCUIT BREAKER
289
Means are provided for closing the con-
tacts either manually by means of the handle
L or remotely from the station switchboard
through the solenoid 5>:J, the plunger of which
engages with the lever Ll .
Fig. 4 shows a typical calibration cur\?e for
the high-speed breaker shown in Fig. 1.
From this curve it will be noted that it
requires a current of 0.17 amperes to hold the
armature closed with zero current in the
bucking bar. If it is desired to have the
breaker trip at 2000 amperes load, it is
merely necessary to adjust the holding cur-
rent of SI to 0.7 amperes by means of the
calibrating rheostat R2. The main current
ma\' rise to the trip value slowly as in the
case of overload, or rapidly as in the case of
short circuit. In either case the armature
of the breaker starts to move the instant its
flux is reduced to normal drop-out value by
the bucking bar current. On short circuit,
Fig. 8. Short Circuit on 2000-kw. 3000-volt Motor-generator
Set Protected by a High-Speed Circuit Breaker, a Resistor,
a Reactor and Flash Barriers. Line resistance, 0.175 ohms;
line reactance, 21.6 milU-henrys. Limiting resistance. 1.2
ohms. Tripping point 2470 amperes. A, line voltage; B,
line current; C, voltage across breaker
the current will rise to several
times the normal tripping
point and the flux in the
armature will be reduced to
a very small value or even
reversed. This condition,
however, gives the maxi-
mum speed of operation,
as the armature starts to
move the instant the flux is
reduced to normal drop-out
value, and by the time the
flux reaches zero, the arma-
ture is moving at a fairly
high rate of speed. On ac-
count of the high speed and
steep pull characteristics of
Fig. 7. High-speed Circuit Breaker and 2000-kw. 3000-Volt
Motor-generator Set Under Short Circuit. Record in Fig. 6
the armature, it is not possible for the buck-
ing bar current to rise rapidly enough to
build up sufficient reversed flux in the arma-
ture to hold it closed.
The connection shown in Fig. 3, with the
circuit breaker installed on the negative side
of the generator and operating to introduce
a limiting resistance Rl , gives the maximum
])rotection against flashovers. With this
connection any possible flashover current
from the positive stud to frame has to pass
through the limiting resistance to return to
the armature. Tests have demonstrated that
the breaker will successfully open the cir-
cuit completely instead of only inserting the
Fig. 9. Short Circuit on 2000-kw. 3000-volt Motor-generator Set, Protected by a High-
speed Circuit Breaker, a Resistor, a Reactor and Flash Barriers. Line resistance. 0.455
ohms; line inductance. 11.1 milli-henrys. Limiting resistance, 1.2 ohms. Tripping
point, 2580 amps. A, voltage across contacts and generator series fields; B, line cur-
rent; C, 59.5 cycle timing wave. Current rupture completed by air circuit breaker
290 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 4
limiting resistance; but it has been found in
practice that better protection is afforded the
machine if the current is ruptured in two
steps.
Figs. 5 and 6 show oscillographic records
of dead short circuits on one of the 3000-volt
2000-kw. direct-current motor-generator sets
that it reached its maximum value and
started down in 0.008 seconds and was down
to normal in 0.015 seconds.
Figs. S and 9 show oscillograph records of
3000-volt short circuits through various
amounts of resistance and reactance. While
the speed of the breaker on these tests was
Fig. 10. Short Circuit on C. M. & St. P. Locomotive No. 10250.
When Protected by high-speed Circuit Breaker on Loco-
motive Only. A, Voltage across high-speed circuit breaker
and line contactors: B, Line current; C, 60-cycle timing wave
built for the Chicago, Milwaulcee and St.
Paul electrification; and Fig. 7 shows the per-
formance of the machine and the breaker
during one of the tests. For these tests the
positive terminal of the generator was con-
nected to ground through a circuit closing
Mtra
i
•T
seo.
men
1
— ; —
5150.
wtpcaa
1
AU«
c ■ ■
t
. M
Fig. 12. Short Circuit' on 600volt, 300kw. 60 cycle Sjiichro-
nous Converter Protected by High-speed Circuit Breaker and
Flash Barriers. Line resistance, 0.0033 ohms. Limiting
resistance. 0.666 ohms. Tripping point. 1500 amps. j4.
Line voltage; B. Line current; C. 60 cycle timing wave
necessarily lower than in the tests without
reactance in circuit, on account of the lower
rates of current rise, the flashing at the com-
mutators was practically negligible, as the
current peaks were much lower.
-0957 Sec. -1
Fig. 11. Short Circuit 9600 Feet from Substation on 1500-kw. 3000-volt Motor-generator
Set Prote- d by High-speed Circuit Breaker and Flash Barriers. Single track road,
100 pound . Ills, two 4 0 trolleys
contactor and a high-speed circuit breaker by
means of a 1,000,000 cir. mil cable so that the
total resistance in the circuit including the
resistance of the generator was approxi-
mately 0.095 ohms. Fig. 5 shows that the
current rose to approximately 7100' amperes.
Ver}' extensive tests under every con-
ceivable operating condition were made on
the breaker in connection with the Chicago.
.Milwaukee and St. Paul motor-generator
sets. The acceptance tests alone required
approximately 05 successive short circuits
NEW TYPE OF HIGH-SPEED CIRCUIT BREAKER
291
of various degrees of magnitude. Five dead
short circuits were thrown on the set inside
of 10 consecutive minutes at the conclusion
of the acceptance tests without any flashovers.
No attention was given the brushes or com-
mutator during any of these tests.
Fig. 10 shows a dead short circuit on loco-
motive No. 10,250 standing at the Erie sub-
station. In this case the high-speed breaker
in the station was not in use and the short
circuit was easily cleared by the breaker
and the line contactors on the locomotive.
volt magnetic blowout circuit breaker, and
also shows the theoretical curve for a similar
short circuit, with only the regular circuit
breaker in service. The areas enclosed by the
two curv^es indicate 184 ampere-seconds with
the high-speed circuit breaker in service and
615 ampere-seconds without, which values in
a measure show the relative punishment of
the generator under the two conditions.
The simjilicity, ruggedness and reliability
of this new type of breaker opens for it a wide
field of application in the protection of direct-
current apparatus. The magnetic blowout
and arc chute are particularly effective and
insure the successful rupture of practically
Fig. 13.
600-voIt, 300-kw. 50-cycle Synchronous Converter
Under Short Circuit. Record in Fig. 12
.03 .04 .05
Ti/ne- Seconds
Fig. 14. Short Circuits on 2000-kw. 3000-volt Motor-generator
Set. A, Line current with high-speed circuit breaker in
operation. Plotted from Fig. 6. B, Theoretical line current
with regular 3000-volt magnetic blowout circuit breaker in
operation
Fig. 11 shows a-digld short circuit 9600 feet
from the substation.
Fig. 12 shows a typical oscillograph record
of a short circuit on a 600-volt SOO-kw.
60-cycle rotary converter; and Fig. 13 shows
the performance of the machine when pro-
tected with barriers and the high-speed cir-
cuit breaker. Heavy short circuits on 60-
cycle machines usually cause a slight amount
of arcing at the brushes, but not enough to
cause the machine to flashover. It should
be noted that the short circuit causes only
a momentary drop in the machine voltage,
and that the voltage is very stable after the
high-speed breaker removes the short circuit
and reduces the current to normal.
Fig. 14 shows the plot of an oscillograph
record of a dead short circuit on a 3000-volt
2000-kw. machine protected by the high-
speed circuit breaker and the regular 3000-
any direct current and voltage which it is
possible to obtain from modern commercial
generators. In combination with the flash
barriers, this breaker insures practical im-
munity from flashovers under the most
severe short circuit conditions. There are no
latches, triggers, etc., to get out of adjust-
ment; and, as the tripping mechanism is
simply a straight copper conductor carrying
the main current, the breaker may be ex-
pected to duplicate its performance many
times in succession. It may also be used for
the protection o feeders and other circuits,
as well as main generators.
One particular advantage of the protection
afforded by this type of breaker is that it can
be applied to old or new direct-current gen-
erators, or synchronous converters of any
type or voltage, with no change whatever in
the machine itself.
292 AprU, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 4
Power-limiting and Indicating System of the
C, M. CS, St. P. Rwy.
By J. J. LlNEBAUGH
Railway and Traction Exgixeerixg Department, General Electric Company
The power limiting and indicating system employed on the C. M. & St. P. Rw\-. is one of the unique features
of this electrification. As its name implies, it is designed for two purposes: the first to limit the peak load
demands made upon the Power Supply Company, the second to facilitate the determination of the amount of
power used. The first function results in a high operating load-factor, with mutual benefit to the Power Supply
Companv and the Railway Company; and the second produces on a single meter a reading which is the total
resultant of the power supplied to the railway system at five widely spaced feeding points minus the power
returned by regeneration of the railway locomotives. In addition to the metering system canceling out the
interchange of power within the railway distributing lines, it furnishes the train dispatcher with a continuous
indication of the amount of power his operations are drawing from the Power Supply Company. In the initial
stages of the development of this system very careful investigations were made of the possibilities of various
other schemes. However, eloquent testimonal as to the excellence of the scheme selected is furnished by the
record of its three years of successful operation. — Editor.
J. J. Linebaugh
^ Bar he L ines 24 M.
THE Power Limit-
ing and Indicating
System constitutes
one of the many new
and novel features
developed and in-
stalled as pail; of the
original equipment
furnished by the Gen-
eral Electric Com-
])any to the Chicago.
Milwaukee and St.
Paul Railway for the
electrification of its
Rockv Mountain and
Missoula Divisions. This system has over-
come so many difficult problems and performs
so satisfactorily it is believed a detailed de-
scription will be of interest.
Several different schemes were proposed
and investigated, such as increasing the
frequency of the circuit proportional to the
power input, adding electrical impulses of
different kinds, etc., but the system which
was finally adopted and which will be
described was found to be the simplest,
to require the least nvimber of pilot
wires, and to necessitate ver>- little appa-
ratus in either the substations or the dis-
])atcher's office.
Averij
lowth
Two Dot
DeerLodgt
Dispatchtrs%.
0/fice '
Fig. 1 .
Jonnet/
General Connections of the lOO.OOO-volt System of the Montana Power Company and Transmission Line of the Chicaco,
Milwaukee & St. Paul Railway, Including Location of the 3,000-volt d-c. Railway Substations
POWER-LIMITING AND INDICATING SYSTEM OF THE C., M. & ST. P. RWY. 293
The general requirements specified by the
railway were based on its desire to obtain
an equipment which, with heavy trains
comparatively few in number, would give
the highest load-factor consistent with good
railroading; and on the part of the Mon-
tana Power Company to prevent exces-
sive peaks which might cause serious voltage
variations and require the installation of
excess generating apparatus to take care of
the railway load. The power company was
also very desirous of obtaining means by
which the total power supplied to the railroad
Missoula Division
feeding points and the heavy grades with
regenerative braking. The apparatus
described was designed, built, installed, and
tried out in service on this section before
going ahead with similar equipment for the
220-mile Missoula Division which has only
two feeding points.
The equipment for the Rocky Mountain
Division as first installed was based on
metering the power at the five feeding points
(Two Dot, Josephine, Piedmont, Janney, and
Morel Substations) , but was later changed to
meter the power at the low-tension side of the
Rochy Mountain D/vision
vs
/fOWl7/7a
2-2000 '•mUnils
OoldCreeh
l-2000hw.Units
primrose
2-2000 IfK. Units
2 HmliOOVolt DC.
Motor OeneraiorSets
Train Dispatchers
Office
Deer Lodge
f^orel
2-20G0hW- Units
[fAMv Up
Janney
3-l500hw Units
^M^ "ji
Piedmont
1-I>00I\M Units
Tar/iio
2-2000/M. Units
U^ *li
Eustis
7-2000 Itw.Urits
Orexel
2 2000/1W. Units
Josephine
2-2000hw.Units
fr^
fast Porta/
3-2000 hw.Units
J-
Avery
il500h'^. Units
>^^Mv_ Pj/ot Wire ffheosiat
\ ' (Zero Load Setting Indicated)
Contact Makinq Ammeter which
- Controls DC. Voltaqe of 3000 \/oJt
Motor Generator Units
Un^M/
Loweth
2-2000 Hw.Units
t
IjvNWv
7(V0 Dot
2-2000 km. Units
Fig. 2. Diagram Showing General Connections of the Pilot Wire Circuit and Location of the Contact-making Wattmeters
for the Rocky Mountain & Missoula Divisions
transmission line at a number of different
points, over a distance of 220 miles, could be
accurately recorded at one place and on one
meter; to replace the former practice of
having laboriously to add up records of as
many as five curve-drawing meters, which
are somewhat difficult to synchronize as to
time, in order to obtain proper peak load data
upon which to base the price of power. It is,
therefore, very evident that the accomplish-
ment of these results is of great mutual
advantage and benefit to the railway company
and the power company.
The 220-mile Rocky Mountain Division
was selected for the first installation as being
the most difficult section due to the five
motor-generator set step-down transformers.
This change was decided upon by the two
companies concerned as it was found impos-
sible to prevent the transfer of very large
blocks of power from one of the power
company's lines to the other lines through the
railway company's transmission line at times
of switching or line troubles with resultant
losses not correctly chargeable to the railway
company, added duty to the metering equip-
ment due to the necessity of adding and then
subtracting this power, excess meter capacity,
etc. The Missoula Division with only two
feeding points did not have these objections
and power for this division is metered on the
high-tension side.
294 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
The main features of the power company's
transmission line, railway company's trans-
mission line, feeding points, location of
substations and train dispatcher's office is
shown in Fig. 1. The dispatcher's office is
located at Deer Lodge, Mont., the center of
the 440-mile electrification, and all the
indicating and recording apparatus for both
divisions is installed at this point.
•r
Fig. 3. Curve Drawing Kilowatt Totalizing Wattmeter.
Dispatcher's Office
The complete system comprises the two
separate and distinct functions of limiting
the maximum power demand at the will of the
train dispatcher and of indicating and
recording the total net power at all times.
The combination of these two functions ac-
complishes the following remarkable results :
(1) Independent of the number of feeding
points, it indicates to the train dispatcher at
all times the total net amount of energy being
delivered to his division and it makes a
permanent record for future study and as a
basis for power bills.
(2) It automatically deducts regenerated
power if returned to the power company's
lines or transfer of ]5ower from one line to
another over the railway company's trans-
mission line.
(3) It automatically limits the amount of
power supplied to the division by lowering
the trollev voltage and slowing down the
trains so that the peak load on the system
cannot exceed a certain predetermined
maximum.
(4) Its maximum limit can be changed
instantly, easily, accurately, and directly by
the dispatcher at any time without any
necessity of notifying substation operators.
(5) It is capable of reducing the peak power
demand by 30 per cent.
(6) If desired, the equipment can be
adjusted so that the lightly loaded substations
will not be affected, thereby providing the
highest possible voltage for the operation of
passenger trains.
(7) If desired, the equipment can be
adjusted to reduce the voltage on the hea^^est
loaded substations at the time of peak
demand (above the maximum limit) slightly
in advance of the other stations, thereby
tending to equalize the load on all the stations.
(8) If an excessive demand for power
occurs near any one substation, the voltage
of the nearest substation is automatically
lowered without affecting the voltage of the
other substations, dividing the load between
this substation and the stations on either side.
Fig. 4. Indicating Kilowatt Totalizer. Dispatcher*! Office
(9) The total power fed in at any point or
transferred from one power line to another or
the amount returned due to regeneration can
be easily taken care of by a change in the
ratio of the current iran.sformcrs, or by an
adjustment of the wattmeter rheostats.
Preliminary negotiations between the rail-
way comiiany, the power comjiany, and the
manufacturer were completed in No\omber,
1915; the equipment was completed and
installed for the first division in 1917, and
has been in successful o]icration since that
POWER-LIMITING AND INDICATING SYSTEM OF THE C., M. & ST. P. RWY. 295
time. The equipment for the second or
Alissoula DiAasion ihas been installed and
is now in operation.
The system is essentially an ohm-meter on a
large scale; consisting of a pilot wire circuit
extending the length of the division, connect-
ing in series all of the substations, and the
train dispatcher's office with contact-making
wattmeters and suitable rheostats at the
incoming power points, and contact-making
ammeters with voltage lowering generator
rheostats in each substation.
As each of the divisions was about 220
miles in length, No. 8 B & S. copper wire was
selected as being the smallest wire that for
mechanical reasons should be used. To give
the utmost reliability, a two-wire pilot circuit
was installed in each instance. This pilot wire
is placed on the trolley poles beneath the
.3000-volt direct-current feeders, Fig. 18. It is
very probable that one wire with a good
ground return would be satisfactory, but as
this was the first installation it was thought
best not to take any risk of such an arrange-
ment being unsatisfactory and therefore a
complete non-grounded metallic circuit was
installed. The insulators were selected after a
very extensive investigation of the com-
parative merits of porcelain and glass, leakage
constants, etc., a special attempt being made
to obtain an insulator which would give a
minimum surface leakage under all atmos-
pheric conditions. A 6600-volt porcelain
insulator was chosen. The leakage under the
most severe conditions has been so slight that
the accuracy and operation has not been
affected. No. 4 B & S. wire is used at all
crossings.
A constant source of direct-current poten-
tial is applied across the two ends of the pilot
wire loop at the dispatcher's office, power
being obtained from a 2-kw. 1200-volt direct-
current motor-generator set, the voltage of
which is held constant by a standard voltage
regulator. The voltage applied to the pilot
wire is determined by the length of the
division, the resistance of the pilot wire, the
number of substations, and the power feeding
points. The equipment as finally worked out
only requires a maximum of 1200 volts
direct current for the 220-mile division, or
440 miles of pilot wire.
The indicating and limiting feature is
obtained by inserting or removing a certain
niunber of ohms or resistance for a definite
change in the kilowatt demand which causes
a definite decrease or increase in the current
flowing in the circuit when a constant voltage
is held across the pilot wire.
The contact-making wattmeter resistances
and the pilot wire contact-making ammeters
are connected in series with the pilot wire as
shown in Fig. 2, which shows connections for
both divisions. As the Rocky Mountain
Division has the greatest number of feeding
points, and maximum regeneration, the
equipment of this division will be described
in detail.
Due to the necessity of accounting for the
regenerated power from the locomotives, or
the transfer of power through the 100,000-volt
60-cycle three-phase transmission line of the
railway company, it was necessary to provide
a so-called Zero Center Meter, which made
necessary having either the resistance for the
total power or the regenerated and trans-
ferred power always in circuit at the no-load
position. The total amount of power in both
directions, for which it was necessary to
provide resistance, is shown in Table I.
It will be seen that it would have been
necessary to provide resistance for a total
power input of 70,000 kw. and 21,000 kw. for
regeneration. In order to obtain greater
accuracy it was therefore decided to insert
resistance for increase of power and to have
TABLE I
MAXIMUM KILOWATT CAPACITY FOR INCOMING POWER AND REGENERATION
AT EACH SUBSTATION
Substations
Maximum
Incoming Power
in Kw.
Maximum
Regeneration
in Kw.
Morel
10,000
10,000
10,000
10,000
10,000
10,000
10,000
1,000
6,000
6,000
Eustis
1,000
3,000
3,000
Two Dot
1,000
296 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
Fig. 5. 2-kw.. 1200-volt Motor-generator Set. Dispatcher's Office
the contact-makini,' ammeters arranged to
make contact at minimum instead of at
maximum current.
This arrangement makes it necessary to
have only the resistance for regeneration in
the line permanently, while the resistance for
power input has to be available only at each
wattmeter. This arrangement gives a much
smaller total meter scale with greater accuracy
at the loads usuallv obtained.
Harlowton is approximately 434 miles, with
a total resistance at 75 deg. F. of approxi-
mately 14.50 ohms.
After careful consideration of all the
different factors, the equipment was designed
on the basis of 15 kw. for each ohm resistance
and 125 kw. for each step on the wattmeter
rheostats, giving a resistance per step of
SJ^ ohms. The kilowatt settings, pilot wire
voltage, current in the pilot wire, and resist-
ij^""^**' .nil
^^^^^^HR**^ ^v^Hi^
^^Hll ^Kr' k ^S*' \m
Fig. 6. Switchboard. Dispatcher's Office
A, induction motor switch; B, motor starting resistance; C. induction motor control panel for 2-kw, pilot wire set; D, plus switch for
pilot wire connections; E, hand wheel of kilowatt limit adjusting rheostat; F, kilowatt hmit scale marked from 10.000 to 2o.000-kw. as
noted in Fig. 4; G, hand wheel of rheostat for fine adjustment of regulator voltage; H, pilot wire indicating voltmeter; K. curve drawing
voltmeter; L. pilot wire voltage regulator; M, rheostat for calibrating pilot wire for change in resistance (due to temperature); N. pilot
wire ammeter; P, resistance used when testing out pilot wire for grounds; R, rheostat for hand adjustment of pilot wire voUagc when
regulator is not in use; S, milliammeter used to test for leakage; T, indicating kilowatt totalizer; U, cur\-e drawing kilowatt totalizer; Z,
plug to short circuit ammeter when testing for possible grounds; Y, plug to connect the milliammeter between middle point of the
two generators and ground.
The total length of the No. N B & S. pilot
wire loop from the dispatcher's office at
Deer Lodge to the farthest substation near
ance are given in Table II. from which it is
noted that there are 0.237 amp. flowing in
the pilot wire at the peak kilowatt setting.
POWER-LIMITIXG AND INDICATING SYSTEM OF THE C, M. & ST. P. RWY. 297
The contact making ammeters are designed
to make contact at this current. The appa-
ratus is designed to hold certain definite peak
Umits in 2()()0 kw. steps from If), ()()() to 2o, ()()()
kw. as indicated.
The power-indicating apparatus, with ex-
ception of the contact-making wattmeters
in each substation, is all installed in the
dispatcher's office. The equipment in the
dispatcher's office consists of a 2-kw. motor-
generator set, a milli-ammeter calibrated in
kilowatts, a curve-drawing ammeter also
calibrated in kilowatts, a curve-drawing
voltmeter to give a permanent record of the
pilot wire voltage, and suitable indicating
instruments and switchboard to control the
motor-generator set.
On account of the very small amount of
power available for the operation of the
curve-drawing totalizing wattmeter, a special
meter had to be developed and both the
wattmeter and voltmeters are based on the
well-known Tapalog principle. These meters
Fig. 7. Photograph of Complete Installation in the Dispatcher':
Office, Deer Lodge, Mont.
were built by the Wilson IMaeulen Company,
New York, and one is shown in Fig. 3. The
curve-drawing voltmeter and wattmeter are
exactly alike with exception of the meter
element.
The meter element is a standard Weston
direct-current ammeter for the wattmeter.
and a standard Weston voltmeter for the
voltmeter. A tapping bar actuated by
clockwork and dry batteries, in connection
with an ink ribbon and paper roll, taps the
meter needle at intervals of o seconds making
a small dot on the paper at the point where
the needle happens to be at that time. The
totalizing wattmeter and the indicating
wattmeter work between limits of 0.19U and
0.353 amp., calibrated for the correct
kilowatts. The doors of the meter cases are
equipped with switches so that voltage is
removed when the door is open. These meters
produce a very satisfactory record and have
given very successful operation.
Due to the reasons explained, the indicat-
ing wattmeter reads a maximum at the
lowest amperes and is therefore off scale
above 25.000 kw. or when no current is
flowing.
The motor-generator set consists of two
1-kw. 600-^■olt generators ccnnected in series
for a maximum of 1200 volts direct connected
to a 3-h.p. ISOO-r.p.m. 110-volt 60-
cycle induction motor with 3^-kw.
125-volt exciter. Fig. 5. Power is
supplied by a 3-kw. 2300 /110-volt
transformer. Separate excitation at
125 volts is used with a standard
regulator to obtain very close regu-
lation. The automatic control of
this regulator is somewhat special
as the A'oltage variation covers a
range from 96-4 to 1200 volts and
must be changed at the same time
as the connections of the wattmeters.
The front and back views of the
switchboard with notations giving
the function of most of the devices
are shown in Fig. 6. The gen-
eral connections are shown in Fig. 8.
A view of the comjalete switchboard
as installed in the dispatcher's office
is shown in Fig. 7.
Due to the simplicity of the indi-
cating wattmeters, two of these
meters have been installed for each
division, one on the switchboard and
the other in front of the trick train
dispatcher as shown in Fig. 9.
With this arrangement the dis-
patcher can tell at a glance the exact
amount of power being taken by his
division at any instant and also can watch
the power demand resulting from his orders
to the train crews in charge of trains
ascending or descending the mountain
grades.
298 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
Variation in resistance of
the pilot wire due to
change in temperature is
taken care of in the dis-
patcher's office by a rheo-
stat which can be easily
inserted and the total resis-
tance adjusted to 2000
ohms, the approximate re-
sistance of the pilot wire
loop and the coils of the
contact-making ammeters,
by holding 1200-volts and
adjusting the rheostat for
0.6 amp.
The adjustable watt-
meter resistances in the
substations are automatic-
ally short circuited during
this operation by reversing
the current through the
pilot wire, which action
short circuits the resist-
ance by means of a polar-
ized relay near the top of
the contact-making watt-
meter unit, Fig. 11.
The equipment for each
substation at which power
is supplied is exactly the
same on the Missoula Divi-
sion, only the wattmeters
being omitted from the
other substations ; while
the equipment for the
seven substations on the
Rocky Mountain Division
is identical.
The contact-making
wattmeter equipment con-
sists of a contact-making
wattmeter built along
standard meter lines, Fig.
13, with an indicating
pointer equipped witli
contacts moving between
the two stationary con-
tacts. The spiral spring
of the pointer is connected
to the shaft of the pilot
wire rheostat located im-
mediately above the watt-
meter. This shaft is dri\-en
by the motor-driven clutch
mechanisms at the top of
the supporting framework,
as shown in Fig. 1 1 . When
contact is made on one side,
POWER-LI.MITIXG AND INDICATING SYSTEM OF THE C, M. & ST. P. RWY. 299
Fig. 9. Photograph Showing Indicating Kilowatt Totalizer
in the Trick Dispatcher's Office
Fig. 10. Contact Making Wattmeter Forming Part of Complete Fig. 11. Complete Motor Operated Clutch Driven Contact
Unit Shown in Fig. 1 1 Making Wattmeter with Covers Removed
300 April. 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
due to an increase in incoming power, the
circuit is completed through the clutch coils
causing the clutch to engage the rheostat
gearing and insert a certain amount of
resistance in the pilot wire. At the same time
the wattmeter spring is wound up due to the
movement of the shaft. This action continues
until the torque of the wattmeter is offset by
the torque of the spring when a balance is
removed, while Fig. 12 shows one of the
meters installed in the Janney Substation
The generator field rheostats which are used
to lower the substation voltage are shown in
this photograph located above the wattmeter
unit.
The power-limiting scheme in connection
with the indicating equipment consists of a
contact-making ammeter. Fig. 14, for each
Fig. 12. Complete Motor Operated Contact Making Wattmeter
Unit and Motor Operated Generator Voltage Lowering
Rheostats Located Above the Wattmeter in the Janney
Substation
obtained and llu clutch circuit interrupted
thereby causing the rheostat to come to a
standstill. This operation is continued for any
increase or decrease in the incoming power.
The rheostats forming part of this unit have
the same number of buttons with S}^ ohms
between each button, a sufficient number of
buttons being used to take care of the power
requirements as s])ccitied in Tai)le I.
The complete motor oi)erated clutch con-
tact-making wattmeter resistance unit is
showii in Fig. 11 with the protecting covers
Fig. 13.
Main Negative Shunt U»ed with Relays
Shown in Fig. 14
substation with its coil connected in scries
with the pilot wire circuit so that when
the current in the pilot wire decreases to a
certain predetermined point. 0.287 amp.,
contact is made and resistance inserted in the
exciter circuit supi)h-ing excitation to the
seijarately excited direct -current generators
by means of a motor operated rhetistat. Fig.
12. These rheostats have sufhcient n-sistance
to lower the substation \-oltage to a minimum
of 2100 volts. When contact is n^ade by the
contact-making ammeter, the voltage of the
POWER-LIMITING AND INDICATING SYSTEM OF THE C , M. \- ST. P. RWV
301
substation is decreased and the resulting
slowing down of the trains reduces the total
input of the substation to a value below the
predetermined peak setting. When the total
load becomes less than the peak setting, the
contact-making ammeter will make contact
on the other side and bring the voltage of the
substation back to normal. A secondary
current coil forms part of the contact-making
ammeter and is energized with current from
a direct-current shunt, Fig. 13, in the ground
or negative side of the 300U-volt substation,
so that the heavily loaded substations have
their voltage decreased slightly before those
with lighter loads. If the total alternating-
current input is beyond that covered by the
power contract, or the limit determined by
the train dispatcher, the voltage of all of the
substations will be decreased until the total
input reaches the amount decided upon.
r**/rT- y,£w
Fig. 14. Connections of Contact-making Ammeter Panel
An overload and an underload relay are
also connected across the current shunt.
The underload relay is calibiated to make
contact at about one-half load on a sub-
station so that the limiting equipment is
inoperative until the load is greater than this
amount. The overload relay is set to take
control of the motor-operated rheostats at
three times load and prevents the load going
above this amount by lowering the voltage
independently of the power-limiting equip-
ment which transfers some of the load to the
substations on either side.
If the power demand should be greater
than the peak limit while a locomotive is
regenerating through a substation, the
reverse-current relay, at the bottom of the
panel, Fig. 13, in each substation (primarily
used to give correct field connections of the
synchronous motor exciter) is also arranged
to open one of the control circuits so that the
voltage lowering rheostats are inoperative.
^^"^ r,r*jii
IPeSiSTAnCC O.OilOfif^S
Fig. 15. Connections of Complete Kilowatt
Limit Adjusting Rheostats
With this arrangement the potential is held
constant at 30U0 volts. If the voltage should be
below normal, due to operation of the power-
limiting equipment, and regeneration should
occur, the voltage is automatically brought
back to 3000 volts and held at this value.
The shunt for operating the tmderload and
the overload relays, and the selective coil of
the contact-making ammeter is also shown
in Fig. 13.
The maximum kilowatt peak limit or kilo-
watt setting can be changed at any time by
the train dispatcher to take care of unusual
congestion or other requirements by simply
varying the voltage across the pilot wire in
the definite steps shown in Table I by means
of the handwheel F, Fig. G. The simplicity of
this arrangement is due to the fact that the
higher the voltage the greater the niunber of
302 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 4
ohms which must be inserted to reduce the
pilot wire current to 0.237 amp. This is
clearly shown in Table II.
The different kilowatt settings are therefore
obtained with certain definite voltages which
must be held accurately by the voltage
regulator. This is accomplished by the rheo-
stat handwheel F, Fig. 6, and the connec-
tions, Fig. 15, which change the setting of the
voltage regulator. If the voltage held should
XcLVX slightly, closer adjustment can be made
with rheostat G. These main voltage points are
marked in red on the scale of the indicating
voltmeter to assist in obtaining correct setting.
Due to the necessity of reducing the pilot
wire current to the same value, the kilowatt
totalizing meters, which are ammeters cali-
brated in kilowatts, must record correctly the
total kilowatts although finally carrying the
same amperes, i.e., 0.237. This is accomplished
by gearing the several rheostats together with
a common rheostat handwheel F, Figs. 6 and
the ground after being installed by means of
the standard cur\-e-drawing meters in the
different circuits but this was found imneces-
sary as the check readings, taken with all the
equipment installed exactly as laid out.
indicated that these meters are as accurate if
not more so than the standard meters. Table
III gives a record of a large number of
readings taken before the contact-making
wattmeters were changed to the low side of
the step-down transformers. These readings
show remarkable accuracy as no special pains
were taken to synchronize the clocks of the
different cur\-e-drawing meters which prob-
ably accounts for some of the greater varia-
tions. It should be kept in mind that this
record was made when equipment was first
placed in operation and that the four feeding
points were located along the railway line
over a distance of more than 200 miles, being
the first time that power supplied by more
than one line was added and recorded on one
TABLE II
CURRENT IN PILOT WIRE CORRESPONDING TO VARIOUS POWER LIMITS
Kilowatt
Peak
Limit
25000
24000
22000
20000
18000
16000
14000
12000
10000
Volts
Across
Pilot
Wire
1200
1184
1152.5
1121
1089.5
1058.5
1027
995.5
964
KILOWATT SCALE
0 I 2000 I 4000 I 6000 | 8000 | 10000 | 12000 | 14000 | 16000 | ISOOO | 20000 | 22000 | 24000 25000
Milli-ampcres
353.0
348.5
339.0
330.0
320.5
311.0
302.0
293.0
283.5
339.5
335.0
326.5
317.5
308.5
299.5
290.5
281.5
272.5
327.2
323.0
314.5
306.0
297.0
288.5
280.0
271.0
263.01
315.8, 305.1 i
311.6 301.0
303.5 293.0
295.0 285.0
287.0
278.5
270.0
262.0
253.5|
277.0
269.0
261.0
253.0
245.0
295.0
291.0
283.5
275.6
268.0
260.0
252.0
244.7
237.0
285.5
282.0
274.5
267.0
259.5
252.0!
244.5
237.01
230.01
277.0;
273.3
266.0
2.59.0
251.5
244.0
237.0
230.0
222.5
268.5
265.0
2.58.0
251.0
244.(1
237.0
230.0
223.0
216.0
261.0
253.5
257.5
2.10.0
250.5
243..-)
244.0
237.0
237.0
2.30.5
2:i(i.<)
224.0
223.0
217.0
216.0
210.5
209.5
204.0
243.0
237.0
230.5
224.0
218.0
212.0
205.5
198.0
240.0
237.0
230..T
224.0
21S.0
212.0
205.5
199.0
193.0
237.0
233.7
227.5
221.0
215.0
209.0
203.0
196.5
190.5
Resistance (ohms).. 3400 3533 3667 3800 3933 4067 4200 i 4333 4467 4600 4733 4867 5000 5067
15. This handwheel which changes the voltage
through the regulator by definite steps also
changes at the same time, by definite incre-
ments, the resistance across the coils of the
two kilowatt meters, thus altering the current
required to give any definite scale indication
in the ratio of the change made at the same
time in the pilot wire voltage. By this
means 0.237 amp., which is the point at
which the contact-making ammeter makes
contact, can be made to represent 10.000 kw..
12,000, up to 2.'), 000 kw. by simply turning
the rheostat handwheel to definite points
plainly marked on the escutcheon, correctly
connecting the three different circuits.
It was thought that it might be necessary
actuallv to calibrate the kilowatt meters on
meter over such a great distance. It is there-
fore e\-ident that the power supplied by an\-
number of transmission lines over practically
any reasonable distance can be accurately
indicated and rcconled in this manner.
The cur\-e-drawing kilowatt totalizing
meter reaches correct readings more quickly
than the standard curve-drawing switchboard-
type wattmeters in the substation and
consequently gives a better detailed record
of the load.
The lowering of the trolley voltage in the
substation is accomplished slow enough, In-
proper speed of the motor-operated field
rheostat, as not to affect the operation of the
trains objectionably, the only result being a
gradual slowing down of the train.
POWER-LIMITING AND INDICATING SYSTEM OF THE C., M. & ST. P. RWY. 303
Additional power limiting is also obtained
by instructing the freight engineers to drop
back to the series connection of the locomotive
motors if very low trolley voltage is indicated
by the voltmeters in each locomotive cab.
Several different peak settings have been
tried out from time to time during the last
two years to ascertain the correct peak limit
for different service conditions. Some of the
lower settings slowed down the trains to such
an extent as to be objectionable on account
of overtime of train crews or delay in passen-
ger trains. It was found that peak settings
could be obtained which would prevent
excessive peaks and still maintain good
operating voltage practically all the time,
giving load-factors which have never before
been obtained for similar service in electric
TABLE III
COMPARISON OF CURVE-DRAWING KILOWATT TOTALIZING METER INDICATION WITH
SUMMATION OF READINGS OF THE SUBSTATION CURVE-DRAWING
WATTMETERS, APRIL 16, 1918
- - -
- -
- -
- - — -
Summation
Curve Drawing
Time
Morel
Piedmont
Josephine
Two Dot
Curve Drawing
Meters
Kilowatt
Totalizer
8:00 a.m.
0
1000-
2.500 —
5800
2300
2100
8:30
200
200 —
3400 —
5800
2500
2800
10:00
0
1.500
2800 —
10000
7700
77.50
10:30
1000
800
2800 —
9000
8000
8000
11:30
2800
1000
2400 —
10.300
11900
11800
11:30
3000
1600
2600 —
10000
12000
11800
12:00 noon
4000
3000
2800-
9600
13800
13300
12:30 p.m.
1000
2000
2600 —
9000
12000
11700
1:00
1500
1600
2800 —
10800
11100
11700
1:30
2600
1800
2500 —
9800
11700
11.300
2:00
5400
3000
3000 —
11700
17100
1 7.500
2:30
2800
0
3200 —
10800
10400
10900
3:00
0
.500 —
2600 —
11000
7900
7800
3:30
400 —
500 —
3200 —
7900
3800
3500
4:00
1500
500 —
3000 —
5900
3900
3900
4:.30
1500
500-
2400-
6900
5500
6200
5:00
1700
0
2200 —
8100
7600
7500
6:30
4000
1500
2600-
10100
1,3000
13500
7:00
1000
1600
2800 —
7100
6900
7600
7:30
3100
500
2200 —
7000
8400
6600
8:00
4000
0
2600 —
7000
8400
8.500
9:00
4000
400 —
3000 —
7900
8500
9000
10:00
4000
0
3000 —
7500
8500
9000
11:00
2000
0
2200 —
8900
8700
7000
1 1 :.3n
3200
0
2000 —
8700
8900
9400
12:00
3600
0
3000 —
7500
8100
8000
Average
8792
8775
— Indicates power fed back to the power company.
TABLE IV
SUMMARY OF PERFORMANCE OF POWER-LIMITING AND INDICATING SYSTEM
FOR SIX MONTHS^ROCKY MOUNTAIN DIVISION
Date
1919
April
May
June
July
August
September. . .
Average .
Time Peak
Limit Hours
43.6
32.6
6.1
4.6
26.7
65.8
Per Cent
Peak Time
of Actual
Running Time
6.4
4.6
1.6
0.77
4.1
9.5
Load-factor
.59.3
56.1
56.5
55.6
54.7
58.8
56.8
304 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 4
-^-1--
__
^
i_
1
1 ,_■ ;
r^
\ .
V
q
r
,: ' ■
i
■ A
1
L
y' '
1 1 ■ [ i..
T 1 II-
1
T '• -
1 1
' >< .
T P . n . n. «
•^ -
o
1
!■
;'.
*
Al
* .
, «
']■/■
• %. 1
U c
E -
*l
^ a
* c
S =
■oE
Si «
Al
u
E £
<% »
o
0 £
i •
. JZ
POWER-LIMITING AND INDICATING SYSTEM OF THE C. M. & ST. P. RWY. 305
■
9f
1
i
>
m 3
>
i
' ■ 1
n
If). o.
■
1
-
■ ■ -. ,
"0 S
■>-'■ e
^j 11
[_
L_
1
-]
>
4
n
1
'
1'
1
';
.'.
'\
in
3
'
S
{
}
\f
>
t
1
:
1
.■-.
;'
C
>
\
S .2
o *-
u 'oil
o o
CO
. a
13 n
i «:
2b
306 April, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, Xo. 4
railway operation. The peak limit was set at
14,000 kw. for the Rocky Mountain Division
on April 1, 1919, and operation has been very
satisfactory- on this basis. The average load-
factor per month for six months is given in
Table IV, or a total average load-factor of
56. S per cent, which is a unique showing for
railway operation and which confirms the
wisdom of the railway company in specifpng
and installing the equipment described in
this article.
The railway company pays 0.536 cents per
kilowatt-hour for 60 per cent of the peak
irrespective of whether this amount is
actually used. The load-factor maintained is
so near 60 per cent that the increase in cost
of power or the cost of power not used is
verjr slight. With increase in the ntunber of
trains, the load-factor will be raised and no
difficulty should be e.xperienced in holding a
load-factor of 60 per cent or better.
Short lengths of the curve-drawing totaliz-
ing wattmeter record are shown in Figs. 16
and 17. These records were taken with a
peak setting of 14,000 kw. and show how
close the peak power consumption is kept to
this point. One section shows an input of
14,000 kv,-. for several hours, var\-ing from this
amount to zero reading or reversal of energ^-,
all of the stand-by losses being supplied by
power regenerated by the locomotive.
If the power-limiting feature is removed,
peaks as great as 7000 to SOOO kw. above the
14,000 kw. limit result.
One of the great indirect benefits obtained
is the valuable assistance the indicating
equipment gives the train dispatcher in
dispatching trains in such a manner as not to
give excessive peaks and thereby lowering the
voltage due to the power-limiting equipment.
By careful train dispatching so that one train
is ascending the mountain grade while
another train is descending, it is possible to
assist the automatic equipment in maintaining
a good load-factor ver\- materially and to
greatly increase the efficiency of the general
operation of the railroad.
Great credit is due Messrs. E. S. Johnson,
J. R. Craighead, J. B. Taylor, and E. J. Thiele
for valuable suggestions, improvements, and
assistance in working out the details of the
great number of new and untried features.
Fig. 18. View taken on the Missoula Division of the C. M. ti St. P. Rwy.
The lower cross arms on the poles at the left carry the pcwer.
limiting and indicating pilot wires
307
Electrification ot the Hershey Cuban Railway
By F. W. Peters
Railway and Traction Engineering Department, General Electric Company
The unqualified success that has attended the operation of high-voltage direct-current railways in the
United States is attracting more and more attention abroad. As reported in another article in this issue, the
high-voltage direct-current system very favorably impressed the French Mission recently sent to this country
to study railway electrification. The article below describes such an electrification now being put into operation
in Cuba. — Editor.
A T all large Cuban
-^^ sugar mills, rail-
roads for transporting
cane extend in various
directions to tap the
areas where cane load-
ing stations are lo-
cated. Two wheeled
o.\ drawn carts are
used to gather cane
in the fields and haul
it to the loading sta-
tion where it is placed
aboard especially con-
structed cane cars
which are later made up into trains and hauled
to the inill. The necessity of grinding cane
shortly after it is cut, in order to obtain a
maximum sugar yield, renders desirable the
maintenance of a reliable railway system to
F. W. Peters
Fig. 1. Map Showing Route of Hershey Cuban Railway
supply the mill with a continuous flow of
cane, thereby eliminating cane "shortage"
shut downs which prove so costly to the
sugar operator.
The industry has assumed such proportions
that the mills command attention not only
for their size, intensiye operation, and
efficiency, but also for the supplementary
industrial actiyities necessary to the support
of the mills during that five-month period of
24-hours per day cane grinding when nothing
but a break down or an important holiday is
deemed sufficient cause to stop operations.
Hershey Central, a beautifully situated
town overlooking the Gulf of Mexico, is
located on the north coast of Cuba practically
midway between the cities of Havana and
Matanzas, some 56 miles apart. The major
activity, at this as well as numerous other
Centrals on the island, is the manufacture of
sugar. This mill is now ser\^ed by the Hershey
Cuban Railway, a steam operated road having
approximately 35 miles of single track. The
present motive power consists of seven steam
locomotives ranging from 20 to 40 tons on
drivers. Both coal and oil fired types are in
use, which, on account of
the very high cost of fuel
in Cuba and the inefficient
operation of such engines,
constitute an expensive
item in overall operation
and preclude an efficient
expansion of traffic such as
outlined herein.
In keeping with the broad
plans of the management,
the road is being electrified,
and extensions which will
comprise the main line are
being completed to Havana
on the west, and to Ma-
tanzas on the east. Branch
lines between Havana and
Coj imar, 4}^ miles , between
the main line and Bainoa,
7 J/2 miles, and between the
main line and Santa Cruz,
A]/2 miles, are completed.
These with numerous short spurs and sidings
will total SO miles of electrified single track.
The road is built over a private right of
way through a rolling country in which the
ruling grade is 2}/^ per cent. The track
is standard gauge with 85 lb. per yard run-
308 April, 1920
GENER,\L ELECTRIC REVIEW
Vol. XXIII. Xo. 4
ning rails rock ballasted over the greater
portion.
The service to be maintained tipon inaugur-
ation of electric operation will consist of cane
and sugar transportation besides through and
local commodity freight, express ser\dce, and
Fig. 2. Transferring Cane From Bullock Carts to Railway Cane Cars
Fig. 3. Sugar Mill at Central Hershey
struction was chosen, to be suspended largely
from bracket arms on creosoted pine poles
which carry in addition the steel cored alu-
minum transmission circuits and the 795,000
cir. mil. aluminum 1200-volt direct-current
feeders.
Locomotives
The motive power furnished
for operating the foregoing cane
and general freight sen-ice con-
sists of seven 60-ton four-m.otor
1200-volt direct-current electric
locomotives arranged for multi-
ple unit operation when neces-
sar\-. They are equipped with
swivel trucks, steeple cab type
super structure, and are designed
to meet American standards
throughout. The control pro-
vides for connecting the motors
in series or series-parallel, and
consists of two master control-
lers (one located at each driving
position in the main cab) with
resistors, dynamotor blower set,
solenoid contactors, and other
auxiliaries mounted principally
under the end cabs. Power for
operating the control equipment
is obtained at 600 volts from
the dynamotor. A pantograph
type trolley is mounted on top
of the main cab with provision
for the convenient use of pole
trolleys, to provide for opera-
tion over adjoining electric rail-
ways necesitating such types of
trolley. Combined straight and
automatic air brake equipment
is used with two 35-cubic foot
displacement per minute air
compressors placed in the main
cab and operated directly from
the 1200-volt trollev wire.
multiple unit passenger train service operating
on one-hour headway between Havana and
Matanzas.
The 1200-volt direct-current electric rail-
way system was selected by the railroad
management after a thorough investigation
of various types of electrified roads, as
being that which would fulfill to the best
advantage the present conditions of electrical
operation, as well as provide for efficient
expansion incident to anticipated growth.
Ten-point catenary type trollex' wire con-
Motor Car Equipments
This ckiss of rolling equipment consists of
ten straight passenger cars, three combination
l>assenger and baggage cars, and two com-
bination express and mail cars. The passenger
cars scat 50 persons, have a free running speed
of approximately 40 miles per hour, and will
weigh coniplctely equipped about 29 tons.
Four motors per car are provided with
automatic electro-pneiunatic double-end mul-
tii)le unit control equipment arranged to
connect the motors in series and series
ELECTRIFICATION OF THE HERSHEY CUBAN RAILWAY
309
parallel. Power for the control circuits and
car lighting is obtained from a 32-volt con-
stant potential generator driven bv a ]2()()-
volt direct-current motor operating from the
trolley circuit. Pantograph type
trolley and bases for pole trolleys
are mounted on the car roof.
Line Work Car
The line work car mounts at
one end a short cab, on top of
which is an adjustable insulated
platform for use when working
on the live trolley. The other
end of the car floor carries a hand
crane. Four motors similar to
those on the passenger cars, but
geared for lower speeds, are used
with a Type K four-motor con-
trol for connecting the motors in
series or series-parallel.
35-kw. turbine exciter 1
.50-kw. motor-generator exciters 1
OOO-h.p. oil-fired steam boilers 4
.'iOOO-kv-a. step-up transformer banks 2
.'JOO-kv-a. station auxiliary transformer banks . . 1
Fig. 4. 60-ton, 1200-volt Direct-current Electric Locomotive
Power Generating and Substation Equipment
The power station and substation eciuip-
ment selected to operate the railroad and to
furnish commercial power to Matanzas and
smaller towns along the right of way consist
of the following :
Gener.\ting Station Number
2500-kv-a. turbine alternators 3
Switchboard 1
Spray pond 1
M.\IN' R.\IL\V.JiV SuBST.\TION
1000-kw. 1200-volt d-c. synchronous converter
groups 2
1050-kv-a. step-down converter transformer
banks 2
Railway switchboard 1
Fig. 5. Passenger Car Seating 50 Persons, Four D-C. Motors
310 April, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, Xo. 4
Each of Two Outlying Auto-
matic Substations
1000-kw. 1200-volt d-c. syn-
chronous converter groups. . 1
1050-kv-a. step-down con-
verter transformer banks . . 1
500-kw. 600-volt d-c. spare
converter 1
350-kv-a. single-phase step-
down spare transformer ... 1
Automatic control equipment . 1
The architectural design of
the power station as indicated
in Fig. 6 is such as to per-
mit of readily building ex-
tensions and making additions
to apparatus to provide for
future enlargement . C are has
been taken to obtain the maxi-
mum ventilation and light
which, with the symmetrical
arrangement of eqtiipment,
affords a pleasing and efficient
working combination.
The steam pressure adopted
was 250 lb. at 150 deg. F. super-
heat, with boiler capacities and
arrangement to permit of the
most efficient operation. The
main generating voltage is
2300 three-phase 60-cyclefrom
which step-up transformers
between the main 2300-volt
bus and the high-tension bus
distribute power at 33,000
volts three-phase to the out-
lying substations and points
of commercial distribution.
Power for station auxiliaries
and shops is obtained from
transformers stepping down
from 2S00 to 4S0 volts. This
latter voltage, largely used in
sugar mill work, was selected
to permit a direct tie-in when
necessarj^ with the main bus
of the sugar mill power house
which is close to, but distinct
from the new railway station.
The railway synchronous
converters located in the main
station consist of two groups
in parallel, each group com-
prising two 500-kw. 600-volt
machines connected in series
for 1200 volts. These receive
their power from the main
2300-volt station bus through
step-down transformers.
ELECTRIFICATION OF THE HERSHEY CUBAN RAILWAY
311
mm
Surface condensers are used with the
turbines and hav'e motor-driven circulating
pumps receiving cooling water from a near-
by spray pond. Air washers, feed water
heaters, emergency feed water supply, piping,
oiling system and pumps- have been
chosen throughout with a view to
reliability and economy.
Because of the mild climate, the
boilers have been located out of doors
adjacent to the sugar mill boilers with
only a roof over them for protection
against the tropical rains. This ar-
rangement affords the most agreeable
working conditions for the men, and
has the advantage of lowered initial
building cost and reduced operating
expense since one boiler house organi-
zation can serve both the sugar mill
and railway boilers.
For oil firing a steam atomizing
system is used with exhaust steam
surface heaters arranged to heat the
oil to the right viscosity for proper
atomization. Two 7500-gallon ca-
pacity auxiliary fuel oil tanks are
located near the boiler room, each of
which holds approximately one day's
supply based on the estimated load
for the near future, while some dis-
tance away are the main oil storage
tanks having a 500,000-gallon ca-
pacity. No attempt was made to
utilize bagasse, the refuse from ground
cane, as fuel for the railway power
station since the quantity produced
by the grinding rolls is practically all
consumed by the sugar mill boilers.
The stack is constructed of radial
brick similar in design to that used
for the sugar mill. It is eleven feet
inside diameter and reaches 200 feet
above the level of the boiler room
floor. A lined steel breeching con-
ducts the burned gases from the boil-
ers to the stack.
Provision has been made for con-
veniently installing coal burning machinery
without disturbing the boiler settings or
auxiliaries should a readjustment in the
relative price of coal and oil necessitate the
use of coal for economy.
A spray pond constructed of concrete is
located 600 feet distant from the power house
and is connected by two 36-inch concrete
pipes, one of which connects to the intake
and the other to the discharge wells, in the
generating room, used for the condenser
circulating water. Three motor-driven 4600
g.p.m. high efficiency pumps which force the
discharged circulating water through the
spray nozzles are located in the pump house
at the spray pond.
^m
Coowerter NoZ (Sp«r«l
Converter No 3
Auto<n>tic S»itchbo»rd
—^///A
Plan and Elevation of Substation
Substations
The two outlying automatic substations,
one of which is located near Havana and the
other near Matanzas, are duplicates and each
contains one 1000-kw. group of synchronous
converters consisting of two 500-kw. 600-volt
machines connected in series. A third 500-kw.
600-volt spare converter is provided with
change over switches so that it may be
conveniently substituted for either the high
or low machine of the group. Three 350-kv-a.
312 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, \o. 4
single-phase 33,000-volt high-tension self-
cooled transformers having double secondary-
windings are regularly emijloyed for operating
the converters with a fourth transformer
supplied as a spare. The switching equipment
is completely automatic in operation and is
Fig. 8. 500-kw., 600-volt Synchronous Converter
Located in Automatic Substation
similar to those which during the past few
years have proven ven,- successful in many
parts of the United States. No regular
attendants are required for the operation
of the equipment since it starts automatically
on a power demand and stops when the
demand ceases. During operation the equip-
ment is protected from injury due to excessive
overload by the use of flash barriers and load
limiting resistors on the direct-current side
of the machine. All irregularities emanating
from disturbance on the high-tension lines
or improper functioning of the equipment are
fully protected against, so as to promote
reliable operation. Two feeder circuits leave
each substation to allow the trolley and line
feeder cable to be sectionalized in front of
each station.
Transmission
Provision is made for carrying two three-
phase transmission circuits on a single line
of poles between the power station and
Matanzas. These will ser\-e the railway
substation on the Matanzas Division as well
as certain railway and commercial power
applications in the city of Matanzas. On the
Havana Division, immediate provision is
made for carr\-ing one three-phase trans-
mission circuit to ser\-e the Havana Division
railway substation as well as any commercial
power adjoining the right of way. Should
occasion demand, however, arrangements are
such that an additional three-phase circuit
can be conveniently added to the existing
pole line.
Forty-five foot creosoted poles on l.jO-foot
spacing have been used o\-er practically the
entire distance. The pole line carries at the
top a Vj inch galvanized steel strand ground
wire and either one or two num.ber 1 0 B.&S.
steel reinforced aluminum 33,0(tO-volt three-
phase transmission circuits mounted on pin
type insulators and creosoted wood cross
arms.
The 1200-volt direct-current feeder con-
sists of a 79.5.U0U cir. mil. standard aliuninum
cable carried over practically the entire right
of way. It is supported on pin type insul-
ators mounted on creosoted wood cross arms
located below the transmission circuits. For
approximately four miles each side of the
sugar mill, where the steepest grade and
heaviest sen'ice is encountered, a second
795,000 cir. mil. feeder is used.
A catenary- type construction is employed
for suspending the 4 0 B.&S. gauge grooved
trolley wire. The messenger wire is yg inch
galvanized steel strand carried on pin type
insulators and galvanized tie iron bracket
arms. On sidings and at special work, how-
ever, cross span suspension construction is
used to support the messenger wire. Steel
terminal 4''0 B.&S. copper strand acetylene
gas weld rail bonds are used throughout.
To protect against the rapid deterioration
of exposed ferrous metals, so prevalent in
tropical countries, all iron parts employed in
the transmission system are protected by hot
dipped galvanizing or sherardizing. It has
also been found that a greatly increased life
may be expected from wood poles treated
with creosote, which in this case has led to
their adoption entirely.
HIGH-VOLTAGE DIRECT-CURRENT RAILWAY INSTALLATIONS IN THE UNITED STATES AND CANADA
Supplement to General Electric Review, April, 1920
PKRMANEKT WAY |
OVERHEAD CONSTRUCTION
,ow..
ROLLING STOCK
Previous
System of
Operation
Compiled by
1
o
1
1
||
4
So
1
III
TroUey or Third Rail
Line
Generating Sutions
Substations
Pass
Trai
ers
1
6
3
Locomotives
Control (Pass.)
Healers
Air Brakes
CBNERAL ELECTRIC COMPANY
Riilwir and Traction EnginMring
Departtnenl
Sthtttdady. N. Y. ^
Much 1.1920 W.D.B.
Type ^
■c
>
>, Number, Capacity
Voltage ^ and Type of
5 2 Generator Equipment
£ 0.
z
Type, Capacity and Voltage
of Apparatus
P.
'a
^1
Principal Motor
Equipment
.1
c
II
|(-
1-
1
1
1
Motor
Eciuipment
0
?
dg.S
t-'-iIZM
Multiple Unit
Operation
Single or Double
End Operation
Speed on Half
Voltage
Type of
bX
•3S
H(J
>
Date
Suncd
,High
VolUge
Ope rat 10
Chief Points
jL-wtook Valley Ry- Co.. Presque Isle.
Uuoe
32
38
0
S-70
Direct Suspension 12Q
10 11,000 60
3 Power purchased from Maine
& N. Brunswick El. Pr. Co.
2
(4) TC-6-2O0-12OO-6O0/1200 H.R. Trans-
formers G.E.
2
2
1 Ex.
28
32
33
(4) G.E. 217-600 Volts
(4) G.E. 217-600 Volts
(4) G.E. 205-600 Volts
30
50
Fgt.
1
40
(4) G.E 206-600 Volts
M-NA
(2) K
Yes
D Half
1200
St. & St.
Auto. G.E.
C.P.-22
C.P.-29
24
600
1200
July.
1910
Steam ft
New Extcn.
Sweden
Buttt -loaconda & Pacific Ry.. Bulte,
Moot.
30 1
14
0
S-85
11 Point Catenary 24C
Trolley
10 100,000 60
3 Power purchastd from Mon-
tana Power Co. (Hydro-
Elec.)
2
(7) 1000-kw. Syo. M.G. Sets. 1200/2400 Volts
G.E.
0
28
3
80
40
(4) G.E. 229-1200 Volls
(2) G.E. 229-1200 Volts
2400
Elec.
Hot Air
W. St. &
Auto.
C.P.
26
100
ouo
"1013
Steam
Butte, Anaconda.
Mines ft Smelters
Cinidiaa Nurihtm Ry.. Montreal, Can.
10
30
0
S-90
Catenary Double 24(
Trolley
K) 11.000 60
3 Power purchased from Mon-
treal Lt., Hi. & Pr. Co
1
(2) loOO-kw. Syn. M.G. Sets. 1200/2400
Volts-11,000 Volts G.E.
8
80
(4) G.E. 239-1200 Volts
50
70
6
83
(4) G.E. 229-1200 Volts
M
NA
Yes
D Half
2400 Volt
Elec.
Hot Aif
Comb, St.
ft Auto. G.E.
C.P.
33
100
2400
1918
New Steam
Terminal
Montreal, through Tunnel to Ml.
Royal
Cmtiil CiMorni!. Traction Co.. Stock-
ton. Cil-
S-75
Third Rail and 12
Catenary
HI 60,000 60
3 Power purchased from Western
States Gas & Elec. Co.
3
Ill 300-kw.. and (2) 500-kw. Syn. M.G. Sets.
1200 Volts. G.E.
8
38
(4) G.E. 205-1200 Volts
45
46
.50
2
3
41
46
(4) G.E. 205-1200 Volts
M
NA
Yes
D Half
1200
West.
G.E. C^mp.
C.P.
22
24
000
'r^
New
Stockton ft Sacramento
ChMles Citv Wtstcm Ry., Charles City,
low*
25
30
0
S-70
Direct Suspension 12
» 2,300 60
3 Power purchased from Cedar
Valley Pr. Co.
1
(2) 300-kw., 1200-volt M.G. Sets, 2300-volt
G,E.
4
11
21
28
(3) (2) G.E. 217
(1) t4) G.E. 217
30
34
52
*
35
14) G.E, 205-500 Volts
K
R-200
No.
D Half
1200
St. & Auto.
Straight G.E.
C.P.-29
C.P..3U
27
1200
'"&,
Steam
Charles City ft Marble Rock
Ctoago, Milwaukee & St. Paul R. R.
(Moolana Elfctrification)
140
j91
0
S-90
Catenary Double 30<
Trolley
0 100.000 60
3 Power purchased from Mon-
tana Pr. Co. (Hydro-Elec)
14
123J 2000-kw. M.G. Sets
9-1500-kw. M.G. Sets, I,W0/3000 Volts
30
•12
4
10
282
300
70
275
(8) G.E. 253-1500 Volts
(4) G.E. 255-1500 Volls
(6) W. 348-750/1500 V.
Twin Arm. Quill
Oil Fired
Steam
W. St. ft
Auto.
a.
34
J4
3000
Dec.
1915
Steam
Harlowton. Montana to Avery. Idaho
217
280
0
S-90
Catenary Double 30(
Trolley
JO 100,000 flO
3 Power purchased from Wash-
ingtom Water Power Co.
and Puget Sound Tr., Lt. &
Pr, Co.
8
18) 2000-kw. M.G. Sets, 1500/3000 Volts,
(6) 2060-kw. M.G. SeU, 1500/3000 Volts.
West.
0
5
265
(12) G.E. 100-1000/3000
Volls, Gcarless
Oil Fired
Steam
W. St. &
Auto.
C.P.
34
150
3000
■^""920
Steam
lC»stidt Eletinfication)
W«<h.
Chicago. MiUvaukee & St. Paul R. R.
4
7
0
S-90
Catenary Trolley 15
30 C.600 60
3 Power purchased from Mon-
tana Pr. Co, (Hydro-Elec)
1
1-300-kw., 1500-volt M.G. Set. 6600-volt
G.E.
1
50
(4) G.E. 207-750 Volts
W. St. &
Auto.
(21
C.P.
29
J7
150O
Feb..
1910
Steam
(Grtal Filli Ternunal)
Station
Diveopiir. i Muscatine Ry., Davenport.
lows
30
30
4
S-70
Catenary Trolley 12
OO 33,000 60
Y
3 Power puichised from Moline-
Rock Island Mfg. Co.
2
(4) 300-kw., 1300-voU M.G. Sets. G.E.
6
lEs.
32
(4) G-E. 217-600 Volts
39
52
0
0
K
No. G
Half
West.
D-2-K
W.
18
1200
Aug..
1012
New
Davenport. Muscatine
20
20
4
S-70
Catenary 12
130 None
Power purchased from Beau-
mont & Ft. Arthur Lt. &
Pr. Co.. Sub. in P. H.
2
2-Syn. M.G. Sefi, 60-cyele, 2300-voit, 1200-
volt, d-c. G.E.
6
1 Ex.
30
(4) G.E. 233-600 Volts
45
46
4
0
K
No. a
Half
W.
2S
600
1014
Now
Tms
Port Dc-JEt. D« Moines Be So. Ry.,
Boone. r^«a
120
145
5
S-70
Direct Suspension 12
00 22,000 25
3 6500-kw. Curtis Turbine
4000-kw Curtis Turbine
6
1 Port
(8) 400-kw.. 600-voIt Syn. Conv. (2 in Series)
(1) 300-kw., 1200-volt Syn. Conv. G.E.
10
42
(4) G.E. 205-600 Volts.
45
50
8
7
4
40
60
(4) G.E 206-600 Volls
<4) G.E. 251-600 Volts
K
Yes S
Full
Hot
Water
West, St.
ft Auto,
C.P.
28
23
600
"Tm
000 Volu
Port Dod8«. Boone. Des Moines ft
Rockwell City
HoeldngSunday Creek Traction Co..
Kelwnvillc. Ohio
14.8
15.2
0
S-70
Direct Suspeniion 12
00 11,000 60
3 Power purchased from Hock-
ing Pr. Co.
1
(4) 200-kw,. 600-volt Syn. Con. West
3
1
23.5
35
(4) W-324-600 Volts
42
48
0
0
HL
Yes D Half
Hot Air
W.-A.M.M.
W.
D.-2K.
-5.3
1200
May,
500 Volu
Ioili»oaT)..Us A Louisville Traction Ry.
Co., Scoiiiburu. Ind.
41
41
Traffic
S-75
Direct Suspension 12
00 None
(4) 300-kw., 600/1200-volt
D-C. Engine Driven Gen-
erators
0
Direct Feed
10
34
(4) G.E. 205-600 Volts
53
50
0
0
M-NA
Yea S
Pull
Hot
Water
Emergency
Straight G.E.
C.P.
22
600
Oct.,
1907
New
Seymour ft Sellenburg ft Indian-
apolis ft Louisville, via 600 VolU
Ions Ry. 4 Lt. Co., Cedar Rapids, Iowa
28
28
S-70
Direct Suspension 12
00 16,500 60
2 Eria. Drive A-C. 7000-kw.,
Turbine Units 10,250-kw.,
2300-volt, 2-phasc
2
(1) 500-kw., 2-unil, 600/1200-volt Syn. M.G.
Set
(1) 500-kw.. 600/1200-volt Syn. Coo. G.E.
5
45
(4) G.E. 205-800 Volts
41
^
1
50
(4) G.E. 207-600 Volts
M-K
Yes D Full
G.E. Straight
Ait
C.P.
28
26
000
Operated
at
000 Volts
600 Volto
Cedar Rapids. Iowa City ft Toledo
Kitiiis Ci!> . Ciay Co. & St. Joseph Ry.
72
74
g
S-70
Catenary 12
00 33.000 25
3 Power purchased from Kan-
sas City Rys. Co,
3
(6) 500-kw.. 7S0-r.p.m,. 1200-voU Syn. Con.
22
5
41
(4) G.E. 225-750 Volts
(4) W-327.750 Volts
00
66
0
0
M-A
HL-A
Yes D Half
Hot
Water
Comb. St. ft
Auto, G.E,
C.P.
29
25
1200
Jan..
1913
New
Kansas City, St. Joseph ft Excelsior
Spring!
L^ Erie S Northern Ry., Ontario.
64
5S
1
S-85
Catenary Trolley U
«0 26.400 25
3 Power purchased from Hydro-
Elec. Pr. Commission of Ont.
3
(3) 500-kw.. 1500-volt Syn. Con. (1 Portable)
Cand., G.E.
6
40
(4) W-85-h.p., 750 Volts
45
70
3
2
flo
(4) W-562-D-5, 750 Volls
Wc«.
A-0
Yes D Half
1500 Volu
West.
Automatic
Dyna.
Com p.
25
1600
Feb.,
1910
New
Gait, Dranttord, Simcoe ft Pt. Dover
Londan* PortStanley Ry., London, Ont.
24
30
0
S-70
Catenary 1
00 13,200 25
3 Power purchased from Ontario
Hydro-Elec. Pr. Com.
2
(4) .WO-kw.. 1500-volt Syn. Con.. West.
8
51
(4) G.E. 225-750 Volts
50
56
10
3
63
(4) G.E. 251-750 Volts
M-NA
Yes D Half
Hot
Water
Comb. St.
ft Auto. G.E.
C.P.
29
2fi
1600
Aug,.
1916
Steam
London, St. Thomas ft Port Sunley
*feryland Electric Ry. Co.. Annapolis.
25.3
32.3
0
S-80
Catenary 1
00 13.200 25
3 Power purchased from Consol.
Gas, Elec. Lt. & Pr. Co.
2
(5) 300-kw. Syn. Con.. 1200-volt, d-c.. West.
12
1 Ex.
40
(4) W-317.A^, 600 Volts
45
52
0
1
52
(4) W-362-600 Volts
H-L-P
NA
Yes D Half
Hot Air
W.-A.M.M.
D.-2K.
J5.3
1200
"1914
06OO Volls, 1«
& Steam
Rapidi, Alltgan-Battle Creek
92
OS
4
S-80
Third Rail Over- 1
running
Catenary
00 70,000 30
3 Power purchased from Con-
sumers Pr- Co., Jackson,
Mich.
3
(8) 500-kw., 1200-volt Syn. Con. G.E.
i
4
70
65
65
(41 G.E. 239
(2) G.E. 239-1200 Volts
(4) G.E. 254-600 Volts
70
50
52
Fgi,
4
Esp.
55
(4) G.E 239-1200 VolM
M-NA
Yes S
Full
Hot Air
Comb. St.
ft Auto. G.E.
C.P.
28
25
00
May.
1616
New
Kalamacoo. Grand Rapids. Allegan
ft Battle Creek
49
40
3
S-80
Third Rail Over- 1
running
Direct Suspension
200 20,000 6(1
3 Power purchased fron Con-
sumers Pr Co., Jackson,
Mich.
2
(4) 500-kw., 1200-volt d-c.; 5000-volt a-C.
Syn. Moior-Generator Sets G.E.
12
40
(4) W-333-600 Volts
60
55
0
0
HL
Yes S
Full
1200 Volts
West.
A.M.M.
Dyna.
Comp.
25
1200
May,
1914
Now ft
000 Volts
Flint, Saginaw ft
45
77
5
s-eo
s-70
Direct Suspension 1
200 20.000 3t
3 Power purcli.ised from Con-
Bumcra Pf. Co.
1
(2) 500-kw.. 1200-volt Syn. Con. G.E.
10
35
(4) W-333-600 Volts
60
55
0
0
n"a
Yes S
Full
Hot
Water
West.
Dyna.
Comp.
25
1200
May.
1910
600 Volu
Grand Rapids. Holland ft Lake
ResorU Macatawa and Saugaluck
lUpids
Milwaukee Electric Ry. & Lt. Co., Mil
waukee. Wit.
135
135
47
S-80
Catenary Trolley 1
200 13,200 2
38,000
3 Part of an Eitlcnsive System
6
(12) 300-kw. \ Syn. Con.. 600-volt. 2 in Series
(5) 500-kw. J G.E.
11
15
43.5
39.S
(4) G.E- 205-000 Volts
(4) G.E. 207-600 Volts
45
64
59
0
•■
M
NA
Trailers D
Only
Full
Hot
Water
WeU.
D.-3 N
600/I2O0J
Mar..
1010
1300 Volti
Singlc-p hasc
Milwaukee. E. Troy,
Burlington ft Wntertown
Nashi-iile-Gallatm Inter. Ry., Nashville
Tenn,
27
27
3.
S-70
Direct Suspension I
200 33.000 6(
3 Power purcSased from Nash-
ville Ry. 4 Lt. Co.
1
(3) 200-kw.. 600-volt Syn. Con.. 2 in Series,
G.E,, 1 spore
4
lEx.
38
(4J G.E. 205-aoO Volts
40
50
Q
0
K
No S
Half
Hot Air
G.E. Straight
C.P.
29
J5
1200
Apr..
1915
New
Nashville ft Gallatin
0-klM.d. Antioch & Eastern Ry.. Sar
Prancisco. Cal.
lis
118
4
5-70
Catenary 1
200 11.000 6
3 Power purchased from Great
Western Pr. Co.
5
1 Port
(6) 730-kw.. I3lK).voU Syn. M.G. Sets;
11.000-volt. a-c., West.
(1) 300-kw.. 1200-volt Syn. M.G.. G.E.
3 G.E.
15 W.
40
14) G-E- 205-1200 Volts
(4) W-322-600 Volts
55
SO
IS
2
2
2
49
62
35
(4) W-30a-B. 600 Volu
(4) W-321-000 Volu
(4) G.E. 205-1200 Volu
M-NA
W-HL
Yei D
Full
200 Volts
We«t.
Dyna. 2125 1200
Comp.
"ffifa
New
aakland. Antioch. Sacramento ft
San Francisco
Oregon Electric Ry.. Portland, Ore.
154
180
25
S-70
S-75
Catenary
200 60.000 3
3 Power purchased from Port-
land Ry. & Lt. Co.
6
I Port
6
(0) 500-kw.. 600/1200 Syn. Con.
(7) 500.kw., 1200-volt Syn. Con. G.E.
53
8Ei.
1 W.
45
14) G.E. 2O.0
(4> G.E. 222-600 Volts
(4) W-32I-600 Volts
45
62
28
6
4
60 (4) C-E. 212-600 volt* M-A
50 (4) G.E. 207-600 Volu
Yes D
Full
Hot
Water
1200 VolM
Weal. D-3 35 1200 1 Aug.. 800 Volls PortUnd, Salem. Albany ft Hugene
1S12 ft New
_ . ■
Pacific Elearic Ry.. Lo» Angele*. Cal. 157 91.& I I
7 S-70
1 S-75
Catenary
200 15.000 5
Q 3 Power puiY:hjUcd from So. Cal.
Edison Co.
4
1
(4) 1000-kw., 1200-volt Syn. M.G. Sets;
15.000-volt. a-C. G.E.
(1) 1000-kw., 1200-voU M.G. Set. A. C. Co.
24
5
44
3
54
44
(41 G.E. 2M-000 Volu
(4) G.E. 222-600 Volts
(4) W-33S-«)0 Volu
(4) W-557-600 Volts
60
eo
0
10
65
(4) W-308-600 Volu
M-A
HL-A
Ye. n
I Pull 1300 1 WmI 1 Dyno. ,' JIfi 1 1200 May. 600 Volts Los Angeles. San Bernardino «
I Automatic I Corap. 1 ibH Rivcnide
-l-J
Piedmont & Northern Line*. Charlotte. 125 hiS i lU 1 S-80 1 Catenary llfiOO 2.400 Ifl
1 IhI
em Pr. Co. iHjdro-Elec.)
■ pI,,
(12) SOO-kw.. ISOO-voU. :i-UQit M.G. SeU
I (6) 250.kw. Syn. Con.. 2 m Scries, for 1500-
volt. Woil.
31
2
42
(4) W-32i-750 Volu
12) G.E. 217-7fiO Volu
45
SO
0 6W.
S G.E.
55
M
(4) W-308-750 Volu
U) G.E. 212-750 Volu
Wot.
IIL
R-200
Yes S Half 1 WmI. VariabU Dyna. -iS Heoo/TBOi May. l New ft / Charlotte. Greenville. Spartanbuiv
" "'" Release Comp. 1 1 ifiia 1 Sinitle-pha.e | ft And^n =^ *~ -Hi
Pituburgh. Mar» & Butler Ry. Co.. 33 33 G Uu.auin. Catenary
Pittiburgh. Pa. 1
1200 22,000 25 3 (1) 1500-lcw.. 13» 750-kw..
Turbine Genetator S«U
6000 volu. Wm-
2
(4) 30O-kw., 1200.VO11. Syn. Coo.. G.E.
13
3S
(4> G.E. ZU-eoO Volu
47
48
0
0
U-NA Ym D
&S., Xi,"""'" "if- " "~ ife
P.ttabur^. Hwrnony. Butler * N
„.,
'
4 U...2J,
Double Trolley
i2oo| i:t.200
(» 3| (2) lftOO.kw.,Turbin« Gcner
1 alors. Vi 300 G-B also pur
4
^61 -lOO-kw.. 1200-roU. 3-umt Syn. M.C. SelA
(1) SOO-kw.. laoO-volt. 3-uail Syn. M.G- Sets;
^&=-
35
32
(4) G.E, 205-aOO Volu
H) W-322-eOO Volu
4S
flO
51
F«t.
41
pr
Yes S
h 1
Half
Pull
Wal*r Strol«ht 22 WJ
000/1200
1908
* Pittsburgh. Butler. Harmony ft
Ellwtwd City. Beaver Falls ft
sas City Rvi. Co.
West ■ "*^' ■'
5
(4) W-327-750 VolU
HL-A
Waler
Comb. St. ft
Auto. G.E.
C.P.
25
1200
"la,-,
New
Kwaas Cty, St, Joaeph ft Ex«U,or
^ Brit ft Sorthwo Rr- Ont«io.
CM.
3 Power purchased irom Hydro-
Elec. Pr. Coniioissioo ofOnt.
{3) 500-kw., 1500-voll Syn. Con. (1 Portable)
Cand., G.E.
6
40
(4) W-85-h.p., 750 Volts
45
70
3
2
60
(4) W-562-D-5, 750 Volts
West.
A-B
Yes
D
Hall
1500 Volts
West.
Automatic
Dyna.
Comp.
IT
1500
Feb..
New
Gall, Brantford. Sioaeoe ft Pt. ttow
Lt^daoi Pott St»oIey Ky-. l^aon. «Jw-
Catenary
3 Power purchiitJ from. Ontario
Hydro-Elei Hr. Com,
(4) 500-kw.. loOO-volt Syn. Con.. West.
8
51
(4) G.E. 225-750 Volts
50
56
10
3
63
(4) G.E. 251-750 Volu
M-NA
Yes
D
Half
Hoi
Water
Comb. St.
ft Auto. G.E.
C.P.
29
25
1500
Steam
(|»rTl«.d Eleclnc Ry. Co.. Ann.polis.
Catenary
3 Power purcho^i'J from Consol.
Gas. Elec. Lt, &. Pr. Co.
2
(5) 30O.kw. Syn. Con.. 1200-volt, d-c. West.
12
lEx.
40
(4) W.317-A-4. 600 Volts
45
52
0
I
52
(4) W-562-600 Volw
H-L-P
NA
Yes
D
Half
Hot Air
W.-A.M.M.
D.-2K.
25J
1200
Jan.,
1914
MOD V^ts. I*
Baltimore ft Annapolis
ffiS. ASK«.-B.itle Ctwk
running
Catenary
3 Power purdia^ed from Con-
sumers Pf. Co.. Jackson,
Mich.
(8) 500-kw., 1200.volt Syn. Con. G.E.
8
2
4
70
65
65
(4) G.E. 239
(2) G.E. 239-1200 Volts
(4) G.E. 254-600 Volts
70
50
52
Fgt.
4
Exp.
55
(4) G.E 239-1200 VolU
M-NA
Yes
S
Pull
Ho( Air
Comb. St.
ft Auto. G.E.
C.P.
28
■zT
60
1^15
New
Kalamaioo. Grand Rands Alleoaa
4 Battle Crock ^^
Michigwi Ry.. Flitn-B»y Uty
Third Rail Qvet-
ninning
Direct Suspension
3 Power purchased fron Con-
sumers Pr. Cc, Jackson.
Mich.
(4) 500-kw., 1200-volt d^:.; 5000-vott a-c.
Styn. Motor-Generator Seu G.E.
12
40
(4) W-333-aOO Volts
60
55
0
0
HL
Yes
S
Full
1200 VolU
West.
A.M.M.
Dyna. [25
Comp. 1
1200
May.
1914
New ft
flOOVoIU
Flint. Saginaw ft
Bay City
Michjgia R>-., Norlhirest D.v.. Grand
S-70
sumersPr.Co.
(2) 500-kw.. 1200-volt Syn. Con, G.E.
10
35
(4) W-333-600 Volts
60
55
0
0
M
NA
Yes
S
Pull
Hot
Water
West.
Dyna.
25
1200
May.
600 Volts
Grand Rapids. Holland ft Lake
nokec. WU.
Catenary Trolley
38.000
3 Part of an Extensive System
U2) 300.kw. ISyn.CoD.. 600-volt.2inSeries
(5) 500-kw. / G.E.
11
15
43.1
39,5
(41 G.E. 205-600 Volls
(4) G,E. 207-600 Volts
45
61
5U
0
M
NA
Trailers
Only
D
Pull
Hot
Water
West.
D.-3 N
~
MW/1200
Mar..
1910
33U0 VoUs
Single-phase
Milwaukee. E. Troy.
N»*b«lJe-Gallaiu) Inter. Ry.. Nashville.
27
Direct Suspension
33.000 60
Y
3 Power purcriased from Nash-
ville Ry- & Lt. Co.
1
13) 200-kw.. 600-volt Syn. Con., 2 in Series,
G.E., 1 spare
4
lEi.
38
(4) G.E. 205-600 Volts
40
50
0
0
K
No
S
Halt
Hot Air
G.E. Straight
C.P.
29
25
1200
Apr..
1015
New
Nubvillc&GaUatin
Fractuv^i. Cal.
IS
18
4
S-70
Catenary
200
11.000 60
3 Power purchased from Great
Western Pr. Co.
5
1 Port
(6) 750-kw.. 1300-voIt Syn. M.G, Sets;
11,000-volt, a-c. West.
(1) 300-kw., 1200-volt Syn. M.G.. G.E.
3 G.E.
15 W.
40
(4) G.E. 205-1200 Volta
(4) W-322-aOO VolU
55
50
18
2
2
2
49
62
35
(4) W-308-B, 600 Volts
(4) W-321.600 Volts
(4) G.E. 205-1200 Volts
M-NA
W-HL
Yes
fi
Pun
1200 Volts
West.
Dyna.
Comp.
2i2S
12O0
Sept..
1913
New
Oakland. AnUoch. Sacramento ft
San Francisco
Oregan Etacmc Ry.. Portlwid. Ore.
M
80
25
S.70
S-75
Catenary
200
60,000 33
3 Power pur hased from Port-
land Ry & Lt, Co.
0
I Port
6
(6) 500-kw., 600/1200 Syn. Con.
(T) 500-kw., 1200-volt Syn. Con. G.E.
53
SEx.
1 W.
45
(4) G.E. 205
(4) G.E. 222-600 Volts
(4) W-321-600 Volts
45
62
28
6
4
60
50
(4) G.E. 212-600 volti
(4) G.E. 207-600 Volls
M-A
Yes
D
Full
Hot
Water
1200 Volts
West. D-3
3J
1200
Aug..
1912
600 VolU
ft New
Portland, Salem, Albany ft Eugene
Pwafic EtKiric Ry.. Los Angeles. Cal.
67
gi.5
1.7
S-70
S-75
Catenary
200
15.000 50
3 Power purthiSKd from So. Cal,
Edison Co.
4
1
14) 1000-kw.. 1200.volt Syn. M.G, Sets;
15.000-volt, a-c. G.E.
(1) 1000-kw., 1200-volt M.G. Set, A. C, Co.
24
5
44
2
54
44
(41 G.E. 254-600 Volts
(4) G.E. 222-600 Volts
(4) W-333-600 Volts
(4) W-557-600 Volts
60
60
0
10
65
(4) W-3US-600 VolU
M-A
HL-A
Yes
D
Full
1200
West
Automatic
Dyna.
Comp.
Sj
1200
May.
1914
600 VolU
Los Angeles, San Bernardino ft
Riverside
Piedmonl i Northtto Lines. Charlotle.
25
2.1
10
S-80
Catenary
1500
2.400 60
13,200
2.000
44,000
3 Power purchased from South-
em Pr. Co. (Hydro-Elec)
1 Port
(12) 500-kw.. 1500-volt, 3-unit M.G. Sets
(6) 250-kw. Syn. Con., 2 in Series, tor 1500-
volt. West,
31
2
42
(4) W-321-750 Volts
(2) G.E. 217-750 Volts
45
60
0
6 W.
fi G.E.
55
64
(4) W-308-750 Volts
(4) G.E. 212-750 VoUs
West.
HL
R-200
Yes
S
Half
West. Variable
Release
Dyna.
Comp.
25
1500/750
May.
1&12
Newft
Single-phase
Charlotte. Greenville, Spartanburs
ft Anderson
Piiwburgb. Man 4 Ballet Ry. Co..
Pitta burati, Pa-
33
33
5
5 tt. 2 }i in.
80
Catenary
1200
22,000 25
3 (1) 150(>kw.. (2) 750-kw..
Turbim. Geneiator Sets,
6600 vo.ts. West.
2
(4) 300-kw., 1200-volt, Syn. Con., G.E.
13
38
(4) G.E. 225-600 Volts
47
48
0
0
M-NA
Yes
D
Full
Hot
Water
G.E Straight
Air
C.P.
28
•'
600
"iliis
6600 VolU
Sinale-phasc
PitUburgh ft Butler
PittsbiTBh. Hiimony. BuDfT & New-
castle Ry.
71
82
4
M,.2j«i„.
Direct Suspension
Double Trolley
1200
13.200 60
3 (2) 1.5004.W., Turbine Gener-
al irs, 1 ;,200 G-E also pur-
chase Diiquesne Lt. Co.
4
(61 400-kw., 1200-volt, 3-unit Syn. M.G. Sets
(1) 500-kw., 1200-volt, 3-unit Syn. M.G. Sets;
G.E.
(2) 250-kw. Syn. Con.. 2 in Series
(2) 500-kw.. 1200-volt. 3-unit, and
(2) 200-kw.. 2-unit, 600-volt Syn. M.G. Sets,
2 in Senes. West,
(2) 500-kw.. 1200-voU 2-unit M.G. Sets, G.E,
22 G.E.
7 W.
2 G.E.
35
32
(41 G.E. 205-600 Volts
(4) W-322-600 Volts
(4) G.E. 225-600 Volts
45
60
51
Fgt.
0
M-A
P.C,
Yes
S
ft
D
Halt
Full
Hot
Water
G.E. Emer.
Straight
C.P.
22
30
28
Hi
000
flOO/1200
'"lfc8
Nc«
Pittsburgh. Butler, Harmony ft
Ellwood City, Beaver Palli ft
New Castle
Salt -ake. Garfield ft Western Ry.
20
20
1.25
S-60
Direct Suspension
1500
44.000 80
3 Power purchased from Utah
Pr. it Lt Co.
2
(2) 600-kw., 1500-volt, Ind. Motor Gen. Sets,
G.E. Automatic Control
6
45
(4) G.E. 240-750/1500 V
45
56
18
0
P.C.
Yes
D
Full
Hoi
Water
Comb. St.
ft Auto, G,E.
C.P.
28
3--,
1500
Aug..
1910
Steam
Salt Lake City ft Saltair Beach
Salt Lake ft Utah Ry.. Salt Lake City.
Utah
77
77
1
S-75
Catenary Trolley
1500
44,000 eo
3 Power purchased from Utah
Pr. & U. Co.
4
(12) 2o0-kw. Syn. Con., 750-volt. 2 in Series.
1 spare, West.
2Ez.
14
2
35
44
lU
(4) W-562-A-5-750 Volts
(4) W-334-E-750 Volts
(2) W-530-750 Volts
50
62
0
2
50
(4) W-562-A-5-750 Volts
HL-NA
R-200
Yes
D
Full
1500 VolU
West. St, ft
Auto.
Dyna.
Comp,
37
1500
July.
1914
New
Salt Lake City, Provo ft Payson
Shore Line Electric Ry. Co.. Saybrook.
Conn.
56.4
5S
10.3
S-70
Catenary and Direct
Suspension
1200
11.000 25
3 (2) 1500-kw., Turbine Gener-
ator, 11.000-volt. G.E.
Saybrouk
2
(8J 200-kw., 600-volt Syn. Con., 2 in Series,
14
8 W.
30
(4) G.E. 217-600 Volts
(41 G.E. 205-600 Volts
(4) W-327-600 Volts
43
44
0
0
M-NA
HL-NA
Yes
D
(14) Half
(8) Full
1200 VolU
Emer. St, ft
Auto. G.E.
C.P.
29-28
25
000/1200
Sept..
1010
Now
New Haven & New London
So. Cambria Ry. Co., Johnstown. Pa.
27
28
0
S-70
Direct SuBpenaion
Double Trolley
1200
None
(2) 300-kw., 600/1200-volt,
(2) 500-kw.. 600/1200-volt,
Engine Dnven Gen.. G.E.
0
Direct Feed
10 G.E.
38
(4) G.E. 205-600 Volts
(4) G.E. 217-600 VolU
43
50
0
1
35
(4) G.E. 205-600 Volts
M-NA
IV.
D
Pull
K„
Emer. St.
Air G.E.
C.P.
22
24
000
1010
New
Johnstown, Ebcnaburg. South Fork,
and Nani-Y-Glo
So. Illinois Ry. & Pr. Co.. Harrisburg, 111.
15
17
0
S-80
Catenary Trolley
1200
33,000 60
3 (2) 1000-kvv., 2300-volt Curtis
Turbine at Muddy
1
(2) 300-kw, Syn. M.G. Sets. 1200-voU.
located in P.H., G.E.
5
40
C4) G.E. 205-600 VolU
41
46
2
0
M-NA
Yes
D
Half
Hot
Water
Comb. St. ft
Auto. G.E.
C.P.
29
I'o
1200
'?9i3
New
Eldorado, Harrisburg, Camrrs Mills
Southern Pacific Co., Electric Division.
Portland. Ore
146
102
i
S-76
s-eo
Catenary
1500
13,200 60
60,000
3 Power ourchased from Port-
land Ry-, Lt. & Pr. Co.
5
1 Port
(7) .500-kw., 1500-volt Syn. M.G. SeW. G.E.
(4) 500-kw., 7S0-volt Syn. Con. West., 2 in
Scries
30
SEx.
3 Mail
51
(4) G.E. 205-750 Volts
.50
60
U
3
60
W-308-750 Volts
M-A
Yes
D
&
S
Pull
1500 VolU
West.
Dyna.
Comp.
3.'j
(k)0/1500
"1914
Steam ft New
Portland, Oswego, McMinville.
Forest Grove, Independence ft
Corvallis
So. Padfic Railroad (Oakland, Alameda
ft Berkeley Div.)
118
las
0
S-80
Catenary
1200
13.200 25
3 (2)5000.kv.Turb.Gen.West.
(1) 2c.00An.. 60/25-cycle Freq.
Ch. Scl G.E. Pr. also pur.
Gi, Western Pr. Co.
3
(20J 750-kw., 600-volt Syn. Con., 2 in Series,
81
10
3=S'
(4) G.E, 207-600 Volts
(4) W.337-600 Volts
40
27
88
116
52
40
0
M-A
HL
Yes
D
Half
West.
A.M.M.
West.
arj
1200
Apr.,
1011
Steam
Oakland, Alameda ft Berkeley
Southwestern Traction & Pt. Co., New
Iberia. La.
14
14
0
5-60
Catenary
1200
None
(1) 200-kw.. 1200-volt d-c.
Gen. Engine Driven, G.E.
0
Direct Feed
3
26
(4) G.E. 217-600 Volts
36.5
46
0
0
K
No
D
Half
Stra-Bht Alt
G.E.
C.P.
29
1200
May.
I9l2
New
New Iberia ft Jeanerette
Teias Electric Ry.. Dallas. Texas
158
158
0
S
fS
Catenary
1200
66,000 60
3 Purchased (torn Texas Pr. &
Lt. Co,
6
1 Port
(4) 400-kw.. 1200-volt Syn. M.G. Sets
(6) 400-kw., B00/1200.volt Syn. M.G. Sets,
G.E.
22
6 Ex.
41
(4) G.E. 225-600 VolU
65
56
12
2
25
(4) G.E. 225-600 Volts
M-NA
No
5-D
23 -S
Full
1200 VolU
West. A.M.N.
C,E Comp.
C.P.
29
600
Oct,,
1913
New
Dallas, Waco ft Corsicatia
Tidewater So. Railroad Co.. Stockton,
Ca).
33
33
0
S-65
Catenary
12O0
16.500 60
3 Power [lurchued from Sierra
& San Frar.cisco Pr. Co.
2
(4) 200-kw., 600-volt Syn. Con.. 2 in series.
3
30
(4) G.E. 201-600 Volts
43
40
0
1
40
(4) G.E. 207-600 VolU
M-A
Yes
D
Half
1200Volti
West.
Dyna,
Comp.
35
600/1200
Nov.,
1913
New
Stockton ft Modesta
Toronto Suburban Ry.. Toronto, Can.
49
GQ
3
S-70
Catenary
1500
25,000 25
3 Power [mrchiised from To-
ronto I'r. Co
3
(41 500-kw., 1500-volt Syn. Con., G.E.
a) 500-kw.. 1500-voU Syn. Coo., Cand., G.E.
6
45
(4) G.E. 240-750 Volts
52
56
0
0
M
NA
Yes
S
Half
Hot
Water
Comb. St, ft
Auto. G.E,
C.P.
28
■^
750
Nov..
1916
New
Toronto, Ccorgelown ft Guclph
""■"■'< il^£.. Portland. Ore.
20
-"
0
iSS
Catenary
1200
60.000 33
3 Panicc puechaiEiI from Port-
land Lt. ft P[. Co.
1
(2) 500-kw., eOO-volt Sya. M.G. Sets in
Series, West.
7
40
(4) G.E. 73 and 205
(4) G.E. 205-600 Volts
45
62.
3.
1
40
M-A
Ye*
D
FuU
1200 VolU
West.
Auto.
Weat.
600/1200
J''?913
Utah-Idaho Central RatUoad Co.
97
97
0.2
S-70
Catenary
150C
44,000 60
3 Power purchastd fron Utah
Power & Ll. Co.
3
(5) 500-kw.. 1500-volt Syn. M.G. Sets,
2.^00■voll a-c. West.
(1) 500-kw.. 1500-volt, 3-unit M.G. Set,
2300-volt a-c.. West.
19
45
(4) W-334-750 Volts
47
56
6
5
1
50
30
(4) W-662-A. 750 Volls
(4) W-334-760 VoUs
West.
HL-NA
Yes
D
Full
1 500 VoUs
West.
A.M.M.
Dynn.
Comp.
36
1600
Apr.,
1915
New
Ogden. Logan ft Preston
WaEhingtun. Baltimore & Annapolis
Elec.^. R.. Baltimore. Md.
61
103
14
S-80
Catenary
120C
33.000 2
3 Power puithased from Polo-
mac Elec. Pr. Co.
4
(16J 300-kw.. 800-volt Syn. Con., 2 in Series,
53
40
(4) G.E. 205-600 Volw
(4) G.E. 233-600 Volts
~45~
54
a
3
45
(4) G.E. 207-600 Volts
M-A
Yes
D
Half
Hot Air
West.
G.E. Comp.
C,P.
29
25
1200
Feb..
1010
0600 VolU. 1^
ft Steam
Washington. Baltimore ft Annapolis
Waterloo. Cedar Palls ft Northern
Waterloo. Iowa
60
65
S-72
Catenary
120(
40,000 2
3 (2) 1500-kw., (1, 3000-kw.
Turbine Genentor
4
(4) 500-kw.. 1300-volt. Syn. Con.. Allis
Chalmers Co.
8
5
47
32
(4) W-333-600 Volts
(4) W-317-600 Volts
60
52
5
60
(4) W-308-600 Volts
HL
Yes
S
Full
Hot
Water
West.
A.M.M.
Dyna.
Comp.
38
1200
Sept.,
1014
Newft
600 Volts
Waterloo ft Cedar Rapids
Willamette Valley Southern Ry.
1 ^
30
32
0
S-flO
Catenary
120(
57,100 6
3 Power purchased from Port-
land Ry. & Liidt Co,
2
(2) 500-kw.. Syn. M.G. Sets. 1200-volt, West.
3
34
(4) G.E. 240-600 VolU
40
58
:i
1
50
(4) G.E. 207-600 Volu
M-NA
Yes
D
Pull
200 Volt*
West.
Auto. Air
Dyna.
Comp.
25
1200
"1914
New
Oregon City, Beaver Creek, Mt,
Angel
FOREIGN HIGH-VOLTAGE DIRECT-CURRENT RAILWAYS
GENERAL ELECTRIC
Bethlehem Chile Iron Mines Co., Tofo,
Chile
15
24
0
S-100
Catenary
2400
2,300
60
3
(2) 3500-k« . 2300-voU Tur-
bine GtQcraitr. General
Electric
1
(2) 1000-kw., 3-unit, 2400-voh d-c. Syn.
M.G, Sets. 23O&-V0U a-c., G.E.
0
3
120
(4) G.E. 253-1200 VolU
West.
St. ft Auto,
C.P.
34
150
2400
19)8
New
Tofo Iron Mines ft Cnis Grande
Hershey Cuban Ry., Cuba
75
81
0
S-85
Catenary
1200
33,000
60
3
(3) 2000-kw. G.E. Turbinea
3
(10) 600-kw.. 600/1200 Syn. Conv. G.E.
16
29
(4) G.E. 263-600 VolU
lie
50
0
7
00
(4) G.E. 251-«00 VolU
P.C.
Yes
D
Half
None
Comb. St, ft
Auto.
C.P.
29
26
1200
=%o
Steam
Havana ft MaUnias
Imperial Railways, Japan
20
40
4
42 in.
Catenary
1200
11,000
25
3
(3) 150q-kw,. n.OOO-voIt Gas
Eng.Allcmaion Dick. Kerr.
I
(2) 1000-kw., 3-unit. 1200-volt Syn. M.G.
Sets, Siemens
40
40
(4) G.E, 244-flOO VolU
45
45
14
M-A
Yea
Full
G.E. Comb.
St. ft Auto.
IDIS
Steam
Tokio. Yokohama
South Uanchurian Ry., China
25
43
0
S-50
S-80
Catenary
1200
2,300
60
3
(3) 400-kw., 1200-voU Syn. M.G, SeU, G.E.
4
35
(2) G.E. 205-1200 VolU
25
3
10
40
56
(4) G.E, 206-600 VoUs
(4) G.E. 207-600 Volu
M
No
D
Half
C.P.
2fi
1200
1914
Steam
1 Victorian Rys.. Melbourne, Australia
150
325
0
85
5 ft., 3 in.
Catenary
1500
20,000
25
3
(fl) _lO.000-kw. PaiMns Tur-
15
Syn. Con., 750 to 3000 kw. each. Total.
400
50
(4) G.E. 237-750 VolW
52
70
4O0
0^
M-A
Yes
D
Half
G.E. Comp.
C.P.
29
27
1500
May.
1919
Steam
\ • To be transferred to Caicad
eDiviti
on for
reigbt
«rvic«.
313
Summary of High-voltage Direct-current Railways
By W. D. Bearce
Raii.wav and Traction Engineering Department, General Electric Company
T'
W. D Bearce
'HE movement
toward higher
direct-current volt-
ages began with in-
terurban railways in
1007, when the Indian-
apolis and Louisville
Traction Rwy ■ started
operation at 1200
\'olts. This installa-
tion was followed
shortly afterward by
the Pittsburgh, Har-
mony, Butler & New
Castle Rwy. and many
others. In 1913, the Btitte, Anaconda &
Pacific Rwy. adopted 2400 volts direct current
for a 30-mile steam-road electrification; and,
after this successful demonstration, 3000 volts
direct current was selected for the main line
electrification of the Chicago, Milwaukee &
St. Patil R. R.
The universal success of the higher di-
rect-current trolley voltages is due in a
large part to its logical development from
existing well-tried 600-volt equipment. The
first 1200- volt car eqtiipment used 600-volt
motors, two in series, followed later by straight
1200-volt motors on the Central California
Traction lines. From this point it was only a
short step to 1200/2400-volt and 1500/3000-
volt motors for steam-road electrifications.
Even less difficulty was encountered in
building substation equipment for the higher
\'oltages. Synchronous converters are oper-
ating at 1500 volts on 25 cycles with the same
success as 600-volt machines, while 2400-volt
and 3000-volt motor-generator sets are giving
unquestioned reliability under severe service
conditions.
During the past few years there has been
little progress in the construction of inter-
urban railways, due to adverse financial con-
ditions, so that comparatively few new high-
voltage direct-current installations have been
made. The accompanying table is a revision
of a similar tabulation published in the
General Electric Review, November,
1916, and contains information on additional
equipment and new roads. Notable addi-
tions to this table include the Othello-Seattle
Tacoma Division of the Chicago, Milwaukee
& St. Paul R. R., with 217 miles of road, which
has been electrified with 3000-volt direct-
current; the Hershey Cuban Railway in
Cuba, at 12U0 volts; and the Salt Lake,
Garfield & Western Railway, at 1500 volts.
Below is a summary of high-voltage lines in
the United States and Canada grouped
according to trolley voltage.
System
Number of
Installations
MILES
ROLLING STOCK
Route
Single Track
Cars
Locomotives
1200 volts
1500 volts
2400 volts
3000 volts
32
9
2
2
1847
596
40
657
2082
630
144
871
604
134
8
0
62
29
37
61
Total
45
3140
3727
746
189
314 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 4
Control for 1200 and 1500-volt Car Equipments
By R. S. Beers and C. J. Axtell
Railway Equipment Department, General Electric Company
WITH the advent
of the high-
speed interurban lines
it became necessary
to increase the trolley
voltage in order to
reduce the sub-station
and distribution sys-
tem investment. The
higher voltages which
have been used arc
750, S50 1000, 1200,
and 1500. The last
two voltages have be-
come standard in
America, and foreign
adopting them.
With these higher voltages the control
differs only in the addition of "breaks" or
contacts, increased insulation, and auxiliary
circuit apparatus. The direct control of the
motors is accomplished with either a drum
R. S. Beers
countries are rapidly
opening the contact-
ors, thereby entirely
eliminating the elec-
trical interlocks re-
quiredfor this purpose
on individually oper-
ated contactors. This
sequence is accom-
plished by mounting
the cams on a shaft
which is actuated by
a single air cylinder
with two pistons, the
air being governed by
two magnet valves.
In addition to the cam operated contactors,
the complete controller contains line breakers,
reverser overload and accelerating relays,
thus embodying in one piece of apparatus
all of the motor control parts except the
resistors required to accelerate, series-parallel,
and reverse the motors, as well as to rupture
C. J. Axtell
controller or a multiple-unit controller, the
same as is used in standard 600-volt practice.
On small cars and where train operation is not
essential, the drum controller has the same
advantages that it possesses for (JOO-volt
work. In a large majority of cases, the capac-
ity of the cquif)ments and the necessity of
train operation require multiple-unit control.
The latest development in multiple-unit
control systems is one in which the individual
contactors are closed by cams.* The use
of cams is advantageous in that it secures an
absolutely definite sequence of closing and
♦ This method of control is commonly known as the
"Type PC."
the motor circuit under normal and over-
load conditions.
The 1200 and 1500-volt car equipment
presents many interesting features on account
of the possible variations, due to operation on
existing 600-volt lines, in the source of energy
for the auxiliary circuits, such as headlights,
lights, comj^ressor, and control.
The simplest control is that for operation at
one voltage. With the exception that two
motors of each pair are permanently con-
nected in series, a 1200-volt equipment is
similar to one for tiOO volts. Each motor is
essentially a (iOO-volt motor having its wind-
ings insulated for 1200 volts.
CONTROL FOR 12UU AND 1500- VOLT CAR OUIPMENTS
315
While this is the simplest method of
operation, in the majority of cases it is also
necessary to operate at some lower voltage,
usually (300 volts, to enter cities over existing
systems. When operation on the lower volt-
age is at half speed the main motor circuit
is unchanged, though it is necessary to change
over the auxiliary circuit connections.
A third motor-circuit combination arises
when it is required to operate the equipment
at full speed on both the high and low-voltage
sections of the system. Under these con-
ditions the simplicity of the other two equip-
nients is lost to a considerable degree, for it
becomes necessary to commutate the motors
to obtain full speed. When the motors are
commutated, the current capacity of the other
parts of the circuit must be increased to com-
pensate. This, in detail, means additional
length and double cross-section of cables,
and the commutation of motor resistors to
obtain a smooth acceleration and of the line
breakers for capacity, as well as commutating
the overload device to obtain the same degree
of protection with both voltages. At the same
time it is necessary to commutate some of
the auxiliaries.
In many cases an analysis of the require-
ments shows that the simpler equipment
works out most advantageously in both city
and interurban operation. This is due to the
fact that the speed for which the interurban
cars are normally geared cannot be economic-
ally or safely used in city operation. Thus,
the 600/1200-volt equipment operating at
half speed on 600 volts automatically accom-
plishes, without complication, that which
(Fig. 5). also that a further simplification will
be effected by omitting the commutating con-
nections of the compressor and motor genera-
tor if operation on but one voltage is required.
On 1200 and 1500-volt equipments it is the
practice to use a low-potential source for
Fig. 2.
Commutating Relay for Auxiliary Circuits When
Operation is at Half Speed on 600 Volts
some of the auxiliary circuits. Economy,
safety, and reduction in size of apparatus have
been the deciding factors in this respect. _ The
auxiliaries to be provided for are: lights,
headlights, control, compressor, and some-
«k\«|'.
Fig.
3. Commutating Switch for Main and Auxiliary Circuits
When Operation is at Full Speed on 600 Volts
some operators have added considerable
apparatus to a 600-volt equipment to obtain.
Figs. 4 and 5 show in a simple manner the
connections of the two types of equipments.
It is self-evident that the equipment for half
speed on 600 volts (Fig. 4) is much simpler
than that for full speed on both voltages
times heaters. Power for these auxiliaries
has been derived in the following ways:
Direct from Trolley
Lights, headlight, compressor, and heaters.
"Potentiometer" or Resistance Method
Control.
316 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
Storage Battery
Lights, headlight, and control.
Dynamotor
Lights, headlight, control, and compressor.
Motor Generator
Lights, headhght, and controL
Supplying the car lights and headlight
with power directly from the trolley requires
ten lamps connected in series, thereby causing
a loss of one third to one half of the illumina-
-/^Overl
The compressor and heaters can be operated
most advantageously direct from the trolley;
any other method means a large increase in
rotar>' transforming apparatus, as the com-
pressor and heaters are a very large percent-
age of the total auxiliary load. Where it is
essential that the heaters or compressors
deliver full output on 600 volts, they may
have their circuits changed over like the
motor circuits.
With the "potentiometer" or resistance
method of obtaining a source of power for the
Rcvcrscr
600/1200 Volt Chanqe-over Connections
- of Compressor and Motor Generator
Fig. 4. Circuit Connections for Half Speed
Operation on 600 Volts
Fig. 5.
Circuit Connections for Full Speed
Operation on 600 Volts
tion if a single lamp burns out. With a
headlight of the ordinary luminous-arc type
there is an energy loss of over 4 kw. in the
headlight resistor. The fixtures and switches
for both the lighting and headlight must
necessarily be larger and more expensive than
the same devices for lower voltages. Up to
the present time the use of a high-power
incandescent headlight operated directly from
1200 volts has not been feasible, due to the
inability to arrange the filament for both
illumination and safety.
control, certain disad\-antages are inherent.
Among these disadvantages are: a waste of
energy in the resistor, a var>-ing voltage of
the control circuit due to the var>-ing load, an
increased size of the master controller, due
to rupturing the high-voltage circuit through
the resistor whenever the controller is turned
off, and a train line through the coupler at
trolley potential. Furthermore, this method
does not provide any means of obtaining
low-voltage current for the other auxil-
iaries.
CONTROL FOR 1200 AND l.-jOO-VOLT CAR EOUIPMK TS
317
The second method of obtaining a low-
voltage source of current which has been used
to a limited extent is a storage batterv. A
small storage battery which will furnish
enough power for controlling the pneumatic
cam motor controller can be charged in series
with the air compressor motor. A 32-volt
storage battery has usually been selected in
order to keep the number of cells at a mini-
mum. Such a battery will not furnish suffi-
cient energy to light the car, as the energy avail-
able from the compressor motor is limited.
A storage battery supplying energy for
control, car lights, and headlights must be
of considerable size, as during the winter
season lighting is required for 10 to 14 hours
per day. Aside from the high first cost and
worked out satisfactorily, due to the fact
that cars do not arrive at terminals at
regular inten-als and can not be held over
long enough to obtain the proper charge.
With a storage batten,' on the car, some
means of automatic regulation for maintain-
ing a constant voltage on the lights and head-
lights must be provided, as the regulation of
the batteries from full charge to discharge
is too great for the satisfacton' operation
of the lamps. Furthermore, a storage batten,'
must be properly maintained by skilled labor;
and even then the maintenance for such
sen-ice will be high com_pared with other
methods of obtaining power.
A dynamotor used as a source of auxiliary
power, while operating satisfactorily and
Fig. 6a. 1200-voIt Receptacle
Fig. 6b. 32-volt Receptacle, Lamp, and Reflector
heavy weight of these batteries, the serious
problem involved is to find some method of
charging. This can be accomplished in one of
three ways; by connecting the battery in the
grounded side of one motor, by putting a separ-
ate motor-generator set on the car for charg-
ing, or by charging the batteries at the ter-
minal stations. If the first method of charg-
ing is used, a ven.- high ampere capacity
battery- is required in order that the high
accelerating current of the motor does not
cause rapid deterioration of the plates. The
second method involves a considerable com-
plication of apparatus, aside from the motor-
generator set, such as relays, etc., to charge
the battery properly. The method of charg-
ing at the terminal has been employed to a
certain extent on Pullman cars, but has not
providing current for lights, headlights, con-
trol, and possibly compressor, has the dis-
advantage that to be built of an economical
size it must divide the trolley voltage in the
ratio of about two-to-one. This necessitates
operating the lamps in the car connected five
in series which does not afTord as flexible or as
economical an arrangement of lighting as
when the lamps can be connected in multiple.
Also, to obtain a headlight satisfactory for
high-speed operation, a considerable amount
of energ>' is wasted in a headlight resistor.
The voltage variation of the dynamotor is
practically the same as the voltage variation
of the trolley line, since the machine does not
provide any means of maintaining uniform
illumination of the car with van,-ing trolley
voltage.
318 April, 1920
GENEIL\L ELECTRIC REVIEW
Vol. XXIII, Xo. 4
With pneumatic cam control equipments,
the most satisfactory source of power is a
small motor-generator, mounted on the car,
furnishing power for lights, headlights, and
control. The set furnished on such equip-
ments is 1.5 kw. capacity, the motor being
interior lighting that has a stronger filament
than those for a higher voltage. The 32-volt
incandescent headlight has been adopted
as the most satisfactory- one obtainable,
as it provides not only a powerful direct
beam but the necessarv diffused illumination
Fig. 7. 1 Jj-kw. Motor-generator. 600 1200-volt Motor, 32-volt Generator
designed with two windings and two com-
mutators with provision for connecting these
windings in series when operating on 1200
volts and in parallel when operating on (500
volts. A novel design of generator provides
inherent regulation to hold a practically
constant potential of 32 volts on the generator
with any normal variation in trolley potential.
A potential of 32 volts for the auxiliary cir-
cuits was selected after a careful study of the
various voltages used in steam railway prac-
tice. This voltage is sufficiently high so that
troubles from loose and dirty contacts, inher-
ent with much lower voltages, are not exper-
ienced. It permits the use of a lamp for the
* This lighting equipment was completely described in the
article "An Improved System for Lighting Interurban Trolley
Cars," by W.J. Walker, General Electric Review, Feb.. 1918.
at the sides of the track. The 32-volt 250-
watt incandescent headlight is equal in
illumination to a 4-amp. luminous arc head-
light at its best and never has the unstead-
iness of beam inherent with any arc lamp.
With this lower voltage lighting in the cars
all lamps are connected in multiple, so that
the burning out of a single lamp does not affect
the illumination of any of the other lamps and
permits the use of any number and size of lamps
desirable for car, vestibule or signs.
The inherent regulation of the motor-gen-
erator means a unifoma brilliancy of illumina-
tion from the lamps and the headlight which
is most agreeable to both the passengers and
crew. It is obtained without any moving parts
other than the motor-generator.*
ISOOvoIt Controller. Cover in Place
319
The Public Trusteeship of the Boston
Elevated Railway
By Edward Dana
General Manager, Boston Elevated Railway Company
Several years ago our city transportation systems called to the public's attention the fact that increasing
cost of operation on the one hand and fixed remuneration for service on the other were assuming the charactei
of "the devil and the deep blue sea" and also that the intervening gap within which the companies could
remain financially sound was becoming constantly narrower. The engineer responded to the call of distress and
designed more efficient equipment to retard the approach of the devil, but the deep blue sea resisted all en-
treaties to recede until the public was made to realize that the funeral of the traction companies would be its
own as well. Various methods have subsequently been used in readjusting the rates of fare. In the following
article Mr. Dana first describes the events which led up to the Public Trustee plan employed in operating the
Boston Elevated and then appends a brief of the provisions of the Trusteeship. — Editor.
Events Leading Up to the Trusteeship
N'
Edward Dana
'UMEROUS arti-
cles have been
WTitten in the past
two years relative to
the appointment of
the Ptiblic Trustees of
the Boston Elevated
Railway Company,
and consequently it
may be interesting to
go back for a period of
ten years and briefly
state a few of the
salient facts which
were brought forth
when the road was privately operated, and
which ultimately led to the necessity for
placing the road imder public operation.
In response to an order adopted in June,
1908, by the Massachusetts Senate, the Rail-
road Commissioners subsequently reported
that: "A careful study and comparison of
systems of street railways in Massachusetts
and elsewhere show no grounds for serious
criticism of the service rendered in this
Commonwealth. The street railway com-
panies here are fulfilling their functions quite
as well as any similar utility elsewhere."
At the same time, early in 1909, one of the
officials of the Company showed that the
capital invested had increased 58 per cent in
five years, whereas the increase of earnings
was but 3.4 per cent and that "a study of the
foregoing figures shows clearly that under
the present system of fare and transfers the
company is rendering a greater ser\'ice than
it can afford for the compensation received."
Therefore, as early as 1908, although the
street railways were satisfactorily performing
their duties to the public, it was beginning
to be recognized that the capital investment
was increasing at a rate all out of proportion
to the increase in return to the companies.
Again in November, 1910, the President
of the Compan}% at a hearing before the
State authorities, stated that the Company
had made large contributions toward trans-
portation facilities; that whereas, when the
West End property was taken over bv the
Elevated, in 1898, there were $26,000,000
invested, in 1910 there were upward of $81,-
000,000; that the company had undci-taken
to e.xpend in the following four or five years
$31,000,000, making a total investment for
1914 of $112,000,000; that the demands for
improved facilities had increased far beyond
the increase in revenue; and that the Com-
pany was in no position to asstune new
burdens in the way of subways and tunnels
with the uncertainty existing as to the future
rettirn. This in effect was a protest against
the constant agitation on the part of the
public for extensions of the subway and
tunnel facilities, on all of which the Company
had to pay the entire interest on the cost,
as well as to contribute a certain percentage
on the cost each year toward an amortization
fund.
In the years 1910 to 1914 constant agitation
was made for additional facihties until, in
February, 1915, the President of the Company
before a committee of the Massachusetts
Legislature protested that: "The people of
Metropolitan Boston in street cars are getting
more than they pay for, or as I prefer to put
it, they ought to pay more for what they are
getting — and the investors are not getting a
fair return."
No relief was forthcoming and on Alay 22,
1916, the Company asked the Governor of
^Massachusetts to request the Legislature
to make provision for the appointment of a
commission to report to the next General
Court as to whether in its opinion "it is
advisable for the State to take any action,
either by way of legislation or otherwise, with
a view to enabling the Company to obtain a
320 April, 1920
GENER.\L ELECTRIC REVIEW
Vol. XXIII. No. 4
net revenue adequate for its corporate and
public purposes; and, if so, what action."
In September, 1916, the Company through
counsel submitted an elaborate report as to
its financial condition to the Special Com-
mission, and in it said: "The present finan-
cial condition of the Ele\-ated Company is
due to the following three causes.
(1) The increase in cost of materials and
labor has made it impossible to
secure the reductions in operating
costs which should otherwise have
resulted from improved methods
and increased efficiency.
(2) The enormous increase in the per-
manent investment has been at a
much greater rate than the growth
of the business has warranted.
(3) Owing to the extension of the length
of rides in connection with the free
transfer system, the revenue has
not been increased in proportion
to the service rendered."
This statement furtlier said: "Whatever
action is recommended by this Commission,
it should include some arrangement for a
period of years, say until the ex])iration of
the present subway leases, by which the
Company may be assured of six per cent
dividends so long as it is properly managed
and properly performs its functions as a
public agent."
The result of the investigations of this
Special Commission of 19 Hi was the passing
of an Act by the Legislature providing for an
investigation of the Company- b\- the Public
Service Commission; and at the same time
another commission was appointed by the
Legislature to investigate the situation on
all the street railways of the State, to look
into the situation all over the country, and
to report as to what might be accomplished
toward relief of the street railwa\- jirop-
erties.
In appearing before the Street Railway
Investigation Commission in 1917, the Presi-
dent of the Company said: "I suggest to
your honorable Commission the expediency
of recommending to the next Legislature the
enactment of legislation to accomplish the
result of iiermitting street railway companies
to fix what they believe to be fair, i)rt)pcr
and necessar\- taritTs, subject to corrective
supervision."
* This section of the article was prepared by Mr. H. C. Clark
for the purpose of outlining the details of the Act, and is here
being reprinted from Aera^ October, .1919.
Early in 1918 the Special Street Railway
Investigation Commission reported to the
Legislature recommending among other things
a service-at-cost act for the street railways,
and shortly thereafter the Public Service
Commission recommended as a result of its
investigation of the Elevated Railway that
the Company should be placed under public
control under a Board of Trustees of five
members.
The Company at this time was in dire
straits financially. The country was at war,
the cost of material and labor were mounting
with rapid bounds, the cost of new subway
and tunnel extensions together with the
greatly increased capitalization of the prop-
erty itself were bringing the results pre-
dicted; and finally on March 1, 19 IS, Samuel
W. McCall, Governor of Massachusetts, sent
a special message to the Legislature stating:
"I am convinced that immediate action on
this subject is necessary in the interest of the
public, as well as those who own and operate
the Elevated Railway."
The following month a bill was reported
from the Committee on Metropolitan Affairs
to the Legislature containing the Trustee
plan, and on May 22, 1918, the plan as
finallv appro\-ed was passed under Chapter
l.'>9 o'f the Special Act of 1918. A Board of
five Trustees was appointed b>' the Governor
and took office on July 1st following.
It is interesting to note that over nine
years elapsed from the time that the officials
of the Company first began pointing out the
impossibility of increasing capital invested
for impro\-ed facilities way out of proportion
to the increase in revenue obtained, and that,
entirely apart from the foregoing brief
sketch of salient points leading up to the
establishment of Public Trustees, the Com-
pany and its affairs were constantly in the
public eye. There has been nothing hasty
in the action of placing the Boston Elevated
Railway under public control. It is merely
the careful working out of a situation that
had been developing for a niunber of years
previously.
Provisions of the Trusteeship*
1. Life
The period of public operation
specified in chapter l.'^9 of the special
acts of the Massachusetts Legislature
of 1918 is ten years from the date
when the act took effect, but luiless
tenninated by the State continues
indefinitely. {Stxs. 1 and /i.)
THE PUBLIC TRUSTEESHIP OF THE BOSTON ELEVATED RAILWAY Sm
2. Rcneii'als
Public operation and management
shall continue after the expiration of
the ten-year period until such time as
the Commonwealth shall elect to
discontinue it. {Sec. 12.)
'.i. Forfeiture
By appropriate legislation, passed
not less than two j-ears before the date
fixed for termination, the Common-
wealth may tenninate public manage-
ment, either at the expiration of the
ten-year period or at any time there-
after. {Sec. 12.)
Municipal Purchase
1. By the City (in this instance bj^ the
Commonwealth or any political sub-
division thereof) :
(a) When Purchas ^ Can be Made :
Under provisions of the act, at any
time during the period of public
management and operation; under
the State's power of eminent domain,
at any time. {Sec. 16.)
(b) Terms of Purchase :
Upon the assumption of the Com-
pany's outstanding indebtedness and
liabilities and the paA-ment of an
amount in cash, equal to the amount
paid in cash by its stockholders for its
stock then outstanding. {Sec. 16.)
Readjustment of this provision
in order to meet conditions which
would arise through the purchase,
the terms of which are already
provided by law previously enacted,
of the West End Street Railway
Company, now leased by the Boston
Elevated Railway Company is pro-
vided for, but the principle involved
is the same. {Sec. 16.)
2. By License of City (in this case by
the State or any political subdivision
thereof) :
No provision for purchase by license.
Control
1. Corporate Autonomy
The Company practically surrenders
its corporate autonomy. A Board of
Directors is retained, elected by the
stockholders, but the President, Treas-
urer, Clerk and all other officers of the
Company are appointed and may be
removed by the Public Trustees. The
Directors shall "have no control over
the management and operation of the
street railway system, but its duties
shall be confined to maintaining the
corporate organization, protecting the
interests of the Corporation as far as
necessary, and taking such action
from time to time as may be deemed
expedient in cases, if an3% where the
Trustees cannot act in their place."
{Sec. J,.)
The Trustees shall allow the Board
of Directors from each year such siun
as may be deemed reasonable to pro-
vide for the corporate organization
and enable the Board of Directors to
perform its duties. {Sec. 4-)
2. Of Service
(a) Within Municipality (This Act
provides for State control regard-
less of municipal divisions) :
The control of service lies entirely
with the Board of Trustees. The act
provides that they "shall determine
the character and extent of the ser-
vice and facilities to be furnished,
and in these respects their authority-
shall be exclusive and shall not be
subject to the approval, control or
direction of any other State Board
or Commission." {Sec. 2.)
(b) Outside of Municipality :
Powers of Trustees extend over
entire system.
3. Extensions, Betterments and Permanent
Improvements
(a) Definitions :
The act contains no definition of
Extensions, Betterments or Per-
manent Improvements.
(b) Within Municipality (This Act
proA'ides for State control regardless
of municipal divisions) :
The State's control is complete,
except that contracts for the opera-
tion or lease of subways, elevated
or sm-face lines, or extensions thereof
beyond their present limits, may not
be made if they involve the pa^Tnent
of rentals or other compensation
by the Company, after the period
of management and control by the
State, unless consented to by the
Company's Board of Directors. How-
e\-er, surface lines may be construct-
ed, or purchased, beyond the limits
of existing lines, even should the
Board of Directors refuse if, after
a public hearing, the Board of
Trustees decide that public necessity
and con\-enience requires their con-
struction or operation. This power
322 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 4
lapses, when the Commonwealth has
passed legislation providing for the
termination of pubHc control and
operation. (Sec. 3.)
(c) Outside of Municipality:
Powers of Trustees extend over
entire system.
4. Capitalization, Finances and Accounts
(a) Ordinary- Expenses:
Control complete. {Sec. 2.)
(b) Securities:
The Trustees have authority to
make contracts in the name of the
Company and to issue stocks, bonds
and other evidences of indebtedness
in its behalf. {Sec. 3.)
In spite of the fact that the Com-
pany, by the acceptance of the act,
has consented to this power being
lodged in the Board of Trustees, the
Board of Directors are required by
the provisions of the act to take such
action as they may be requested to
by the Board of Trustees to validate
its acts in relation to the issuance of
securities. {Sec. 4-)
(c) Bookeeping:
Control in hands of Trustees.
(d) Methods and Practices:
Control in hands of Trustees.
5. Use of Tracks, etc., by Other Companies,
No pro\asions covering these matters
in the act, but as Sec. 2 takes juris-
diction away from the Public Ser\-ice
Commission, that body is now without
power to order joint use of track.
6. Machinery of Control
(a) Power, Where Lodged:
All control is lodged in a Board
of five Trustees, appointed by the
Governor, with the advice of his
Council. Their term is ten years,
the fixed period of public manage-
ment and control. If this period is
extended, their successors may be
appointed for a like term, but not
for longer than public management
and control shall continue. They
shall own no stock, or other securities
of the Company, or companies leased
or operated by it. They receive
§5,000 a j-ear, each, paid by the
Company. They may be removed
for cause by the Governor, with the
advice and consent of the Council.
Vacancies are filled by the Go\^emor
with the consent of the Council.
The Trustees are relieved by the
act from the legal inhibition against
the emplo>"ment by the Company of
any person at the instigation of
public officers, in so far as it might
apply to them as public officers.
In other respects they are subject
to the laws of the State governing
public officers, as are the Directors
of the Boston Elevated Railway
Company. {Sec. 1.)
In the manangement and opera-
tion of the Company, the Trustees
shall be deemed to be acting as
agents of the Boston Elevated Rail-
way Company and not of the Com-
monwealth, and the Company is
liable for their acts as if they were in
Company employ; but the Trustees
shall not be held personallv liable.
{Sec. 2.)
A majority of the Board con-
stitutes a quorum for the transac-
tion of business. {Sec. 2.)
(b) Administration:
The affairs of the Company are
administered by the Board of Trus-
tees, a majority of whom shall
constitute a quonun. {Sec. 2.)
(c) Powers and Duties of Administra-
tive Body or Officer. The Board
of Trustees shall:
Manage and operate the Company
and the properties owned, leased or
operated by it. {Sec. 2.)
Exercise all the rights and powers
of the Companv and its directors.
{Sec. 2.)
Appoint, and remove at its direc-
tion, the President, Treasurer, Clerk,
and all other officers of the Companv.
{Sec. 2.)
Fix and regulate fares, including
the issue, granting and withdrawal
of transfers and the imposition of
charges therefor. {Sees. 2, 6, 7, 10.)
Determine the character and ex-
tent of the ser\-ice and the facilities
to be furnished. {Sec. 2.)
Receive and disburse the income
and funds of the Company. {Sec. 2.)
Make contracts in the name of and
in behalf of the Company. For limi-
tations, see Control, 4 (h). {Sec. 3.)
Issue stocks, bonds and other
e\-idences of indebtedness for the
Companv. For limitations, see C
4 (.b). (Sec. 3.)
THE PUBLIC TRUSTEESHIP OF THE BOSTON ELEVATED RAILWAY 323
Collect from the Commonwealth,
at stated intervals, sums sufficient
to make up deficiencies in the Re-
serve Fund, caused by the failure of
revenue to pay the cost of service.
{Sec. 11.)
Repay to the Commonwealth,
when the condition of the Reserve
Fund pennits it, moneys received to
make up deficiencies. {Sec. 11.)
Borrow needed sums in antic-
ipation of pa}Tnents by the Com-
monwealth to make up deficiencies
in Reserve Fund. {Sec. 11.)
Maintain the property of the Com-
pany in good operating condition and
provide for depreciation, obsolescence
and rehabilitation. {Sec. IS.)
6. Arbitration
(a) Machinery for:
No provision is made for arbitra-
tion. In the event that the Trustees
desire to make extensions to, con-
struct, or purchase surface lines
beyond the limits of existing lines
and the Board of Directors of the
Company refuses consent, on the
ground that it entails rentals, or
other obligations, upon the Com-
pany after the period of public
management and control, the Trust-
ees are required to hold a public
hearing. After such hearing they
may, however, decide that public
necessity and convenience requires
the construction of the proposed
line, under which circumstances
they may proceed with its exten-
sion, construction or purchase, de-
spite the failure of the Board of
Directors to consent. {Sec. 3.)
(b) Powers of Arbitration Boards:
No arbitration boards provided
for.
(c) Penalties:
No arbitration provided for.
(d) Expenses of Arbitration:
No arbitration provided for.
Return
1. Initial ]'alne
No initial value is fixed. The act pro-
vides for the payment of rentals interest
on all indebtedness, fixed dividends on
preferred stock, and dividends on
common stock at stipulated rates. The
capitalization of the Company at the
time of the taking effect of the act
was thus recognized. {Sec. 6.)
2. Added Value
No provision is made for added value.
The Trustees have the power to issue
stocks, bonds and other evidences of
indebtedness and may fix the rate of
return thereon, excepting that the re-
turn on common stock is limited by
the provisions of Section 6. (See
Return, 4, Return on Common Stock.)
{Sec. 3.)
3. Deductions from Value
None.
4. Rate of Return
On Rented Property — rents stipu-
lated in lease. {Sec. 6.)
On Indebtedness — interest fixed by
securities or other evidences of in-
debtedness. {Sec. 6.)
On preferred stock — fixed dividends.
{Sec. 6.)
On special issue of preferred stock
authorized by act to provide $2,000,000
for betterments and improvements
and .SI, 000, 000 to provide a Reserve
Fund — fixed dividends not to exceed
seven per cent. {Sec. 6.)
On Common Stock — five per cent
for the first two years of the ten-year
period of public management and
control; five and one-half per cent
for the next two years, and six per
cent thereafter. {Sec. 6.)
5. Additional Alloivances
None.
6. Assurance of Return
If, on the last day of June, or the
last day of December, in any year,
the amount in the Reserve Fund shall
be insufficient to make good any
deficiency in the cost of service, the
Trustees shall notify the Treasurer and
the Receiver General of the State of
the amount of such deficiency, less
any amount remaining in the Reserve
Fund, and the State shall thereupon
pay over to the Trustees the amount
so ascertained, which shall be used for
the purpose of paying such deficiency,
{Sec. 11.)
Pending the payment of this simi by the
State, it shall be the duty of the
Trustees to borrow such sums as will
enable them to meet all deficiencies,
including dividend payments. {Sec. 11.)
If, on the last day of June, or the
last day of December of any year,
the Reserve Fund shall exceed
the original $1,000,000, the Trustees
324 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 4
shall apply the excess, so far as
necessarv', to the reimbursement of
the State for the money advanced to
to the Trustees to meet deficiencies.
(Sec. 11.)
The Treasurer and Receiver General
of the vState may borrow, if necessa.-y,
the money with which to pay the
deficiencies ascertained by the Receiver
General. {Sec. 11.)
The amounts so paid to the Trustees
shall be assessed upon the cities and
towns in which the Company operates
by an addition to the State tax ne.xt
levied, in proportion to the number of
persons in said towns and cities using
the service of the Company at the time
of the pa\-ment, this proportion to be
ascertained b}' the Trustees and cer-
tified to the Treasurer and Recei\^er
General. {Sec. 14-)
Cost of Service
1. Definition
The cost of the ser\-ice includes:
Operating expenses,
Taxes,
Rentals,
Interest on indebtedness,
Depreciation,
Obsolescence,
Losses in respect to property sold,
destroyed or abandoned.
All other exi:)enditures and charges
which, under the laws of the
Commonwealth now or here-
after in effect, may be prop-
erly chargeable against income
or sur]ilus.
Fixed di\-idends on preferred stock:
Di\idends on par value of com-
mon stock, at five per cent,
for first two }-ears of period of
public control and manage-
ment; five and one-half per
cent for next two years, and six
per cent thereafter. {Sec. 6.)
2. Allowances
(a) Operating:
No allowances are fixed by the
Act. Expenditures are made in the
judgment of the Trustees.
(b) Maintenance, Repair and Renewal :
Nc allowances fixed by the Act.
Expenditures are made in the judge-
ment of the Trustees.
(c) Depreciation:
The allowance for depreciation is
specificalh- left to the judgement of
the Trustees. They are, however,
required to pro\'ide for obsolescence
and "losses in respect to property
sold, destroved or abandoned." {Sec.
6.)
The Trustees are further required
to maintain the property in "good
operating condition and to make
such pro\-ision for depreciation, ob-
solescence and rehabilitation, that,
upon the expiration of the period of
public management and operation,
the propert}- shall be in good operat-
ing condition." (Sec. 13.)
3. Special Tax and Impost Features
No special taxes or imposts are
pro\aded for, but the Act contains the
following declaration : " Nothing here-
in contained shall be held to affect the
right of the Commonwealth or any
subdivision thereof to tax the Com-
pany or its stockholders in the same
manner and to the same extent as if
the Company had continued to man-
age and operate its own propertv."
{Sec. 2.)
Fares
1. Schedule of
The Trustees were required, within
sixty days of the taking effect of the
Act, to put into operation rates of
fare, which in their opinion ivere
sufficient to pay the cost of the ser-
vice, and within sixty days thereafter
to adopt a schedule of eight different
grades of fare, four abo\-e and four below
the rate first established, and to at all
times keep the schedule, so that there
shall be four grades above and four
below the rate in effect at the time.
{Sec. 7.)
The Trustees may at any time
change the schedule so as to alter the
rates, or the method and basis of
charges for fares and transfers. {Sec.
7.)
2. How fixed
The Comi^any was required, before
the Act took effect, to ])ro\-ide the
sum of Sl.OOO.dOO. through the sale
of its preferred stock, to be used as a
Reserve Fund. {Sec. o.) This fund
can be used only for making good any
deficiencies in cost of ser\-ice, or for
reimbursing the Commonwealth for
moneys advancetl to meet .such defi-
ciencies. (5t'f. S.)
THE PUBLIC TRUSTEESHIP OF THE BOSTON ELEVATED RAILWAY 325
Into the fund is paid any surplus
remaining after the cost of service is
paid, and from it is taken any amount
needed to meet deficiencies in the cost
of service. (Sec. 9.)
If, at the termination of the period of
pubHc manaj^'cment and operation, the
Reserve Fund shall contain less than
the amount contributed thereto by
the Company, the vState shall pay to
the Company the deficiency. If, on
the other hand, there is a surplus in
the fund, it shall become the property
of the State, which shall distribute it
among the cities and towns served by
the Company, in proportion to the
number of people in such cities and
towns using the Company's service
at the date of the tennination of the
period of jniblic management and
operation. {Sec. 13.)
If, on the last day of any Septem-
ber, December, March, or June, the
amount in the Reserve Fund shall
exceed by thirty per cent the amount
originally established and, for the pro-
ceeding three months, income shall
have exceeded the cost of service, the
Trustees are required to put into effect
the next lowest rate of fare in the fare
schedule; and, if on, the same dates the
amount in the fund shall be less than
seventy per cent of original sum and,
for the preceding three months, income
shall have been less than the cost of
service, they are required to put in
effect the next higher fare. In like
manner the fare shall continue to be
increased or decreased, as the case may
be, on succeeding quarterly dates.
In determining the state of the Reserve
Fund, money received from the State
and paid therein, and for which the
State has not been reimbursed, shall
first be deducted. {Sec. 10.)
Transportation of Freight, Express, Etc.
No provisions are made for the trans-
portation of freight, express, etc., such
matters being under the jurisdiction of
the Trustees. {Sec. 2.)
Special Provisions
1. When Grant Expires
When the period of public manage-
ment and operation expires, the Com-
pany is given the right to fix its own
fares, so as to provide for the cost of
service, including a six per cent return
upon the par value of its common
stock outstanding, and may establish
an automatic scale of fares; but the
Commonwealth is released from its
obligation to make up deficiencies in
the cost of service. The Company
shall, under such conditions, be subject
to such regulation as the Legislature
may decide upon, but such regulation
shall not be exercised so as to reduce
its income below the reasonable cost
of the service as defined in the Act.
{Sec. 15.) But this provision is de-
clared not to be a contract binding upon
the Commonwealth. {Sec. 18.)
326 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 4
Operating Costs of Various Types of City Cars
By J. C. Thirlwall
Railway and Traction Engineering Department, General Electric Company
The shrinking margin of profit in city traction gave birth to the light-weight, one-man safety car as a
means of reducing operating costs. The many illustrations of this type of car have already demonstrated
its merits. The question now arises as to what extent it can displace the larger and heavier cars. The
author believes that the use of the safetj- car can be extended to all city surface lines and that, at least for
the all-day runs, it will be the more efficient. In substantiation of his claim, he furnishes tabulations of
comparative costs. — Editor.
T'
J . C. Thirlwall
'HE tremendous
success of the
light-weight, one-man
car under a great
\'ariety of operating
conditions has led to
considerable discus-
sion among operators
as to just how far the
use of this type of
vehicle can be extend-
ed upon our urban
transportation sys-
tems. A general con-
sensus of views, at
present, seems to be that in cities of less than
100,000 population there are few, if any,
routes for which the safety car is not pre-
eminently suited; that for larger cities only
a limited part of the ser\-ice can be handled
by this size of car; and that all of the heavier
lines in the largest chics will have to continue
to operate large capacity double-truck two-
man cars, singly or in trains. But, while this
is perhaps the general opinion of the industry,
there are many experienced operators who
disagree with the majority and who feel that
there are no surface routes in any city that
cannot use the safety car for at least the all-
day runs, and that the use of this car will
afford greater efficiency than any other type.
The writer, who has watched and studied
the performance of the safety car since its
inception, has come to agree with this
minority, and some figures are presented
herewith bearing on the question as to which
type of car is most suitable for extremely
heavy traffic in the largest cities.
A number of typical double-truck cars of
modern design have been selected; and in
order to determine the relative advantages
of each type, a study of their operating costs,
based on handling similar numbers of pas-
sengers at various hours of the day, has been
made and the results are presented in tabu-
lated form. It is not the writer's idea that
this comparison holds absolutely true for
every property, but the purpose of this
article is to suggest a line of thought which
can be developed by the transportation
engineers of any railway whenever new cars
are to be purchased.
It is of course obvious that several items of
operating cost are dependent upon the weight
of the rolling stock used, that the power con-
sumption varies in about direct proportion
to the ton-miles operated, and that the main-
tenance, both of track and of equipment,
is largely influenced by the weight factor. We
will assume certain costs per ton-mile for these
items, which will varv between different
Fig 1 S:ifcty Car, BrookI>-n Rapi.l Transit
cities but which will enable us to make a
comparison of the efficiency of these various
types of cars. The figures used are based on
the average costs of o])eration of eastern
electric railways for August, I','!!', as reported
in the Acra magazine.
The power cost for these roads is 5.45
cents per car-mile. As the average weight of
cars operated in eastern cities is about 20
tons, the cost per ton-mile is 0.27 cents.
OPERATING COSTS OF VARIOUS TYPES OF CITY CARS
327
(This is checked by the known fact that, in
frequent stop ser\-ice, energy consumption at
the central station is about 200 watt-hours
per ton-mile, and power costs in steam plants
run from 1.25 to 1.5 cents per kilowatt-hour.)
The average cost of track maintenance is
5.7 cents per car-mile. Of this, probably
one half, or 2.8 cents, is directly affected by
the weight of the cars used; and we will
assume, therefore, 0.14 cents per ton-mile for
this item.
The maintenance of equipment costs 3.7
cents per car-mile, and of this we believe
75 per cent or 2.S cents is governed by car
size and weight, or 0.14 cents per ton-mile.
For the items of power and maintenance,
therefore, we have a total of 0.55 cents per
ton-mile.
For crew costs, we have assumed 55 cents
per man-hour for the double-truck cars, and
Gl cents per hour for the one-man safety cars
(these figures being rather under than over
present wages).
In frequent stop service in the larger
cities, schedule speeds including lay overs
do not average over 8.5 m.p.h. A 19-hour
TABLE I
CAPACITY, WEIGHT, ETC., OF TYPICAL CITY CARS
Designation
Car
Length
Feet
Seating
Capacity
Maximum
Load
Number
Motors
Car
Weight
Tons
Where Used
A
41
44
100
2
18
New England; Philadelphia
B
41
44
100
4
22
New England; Brooklyn
C
46
54
125
4
25
Chicago
D
46
58
125
2
19
New England; Brooklyn; Chicago
E
49
58
125
4
22
Boston .
F
51
58
150
4
22
Cleveland (Longitudinal seats)
F-1
104
118
300
4-0
35
Cleveland train (Longitudinal seats)
G
51
58
150
4
17
Buffalo, Rochester (Longitudinal seats)
H
28
32
65
2
8
Brooklyn; New England
H-1
28
32 ■
75
2
8
(Longitudinal seat Safety Car)
TABLE II
OPERATING COSTS FOR POWER, MAINTENANCE, AND CREW WAGES
Costs
Power and maintenance per ton-mile . . .
Crew wages per car-hour (two-man cars)
Crew wages per car-hour (one-man car) .
Crew wages per train-hour (two cars) . . .
Car-hours per annum
Car-miles per annum
Rush Hour
Only
ANNUAL
COST FOR ABOVE ITEMS
, EACH TYPE
OF CAR
Type
of Car
ALL-DAY OPERATION
RUSH HOUR ONLY
Power and
Maintenance
Crew
Total
Power and
Maintenance
Crew
Total
A
$5825
S7645
$13470
81315
$1720
$3035
B
7125
7645
14770
1610
1720
3330
C
8100
7645
15745
1830
1720
3550
D
6160
7645
13805
1390
1720
3110
E
7125
7645
14770
1610
1720
3330
F
7125
7645
14770
1610
1720
3330
F-1
2565
2580
5145
G
5510
7645
13155
1245
1720
2965
H
2590
4240
6830
585
955
1540
H-1
2590
4240
6830
585
955
1540
32S April, 1U20
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 4
run, operated 3(io days per year at this
schedule speed, requires 6950 car-hours or
59,000 car-miles annually. Rush hour extras
or trippers, in general, will not operate over
5 hours daily, and for only 313 days per year,
so each car so used makes annually 15()5 car-
hours or 13,300 car-miles.
In cities of 200.000 to 500,000 population,
the rush-hour service ordinarily doubles the
all-day service; in the largest cities it fre-
quently triples it; in other words, from a half
to two thirds of the cars owned are operated
only in rush-hour traffic. This fact, of course,
is responsible for the introduction and develop-
ment of cars of large seating and standing
capacity, and, in several cities, for the adop-
tion of train service, either of motor car and
trailer or of two motor cars with multiple unit
control.
Various tyjies of these cars are indicated
in Table I. Cars A and B are the most
typical of the older double-truck designs;
units seating 44 and capable of carrying
about 100 as a maximum load, and weighing
from 18 to 22 tons depending upon whether
two or four motors are used. Cars C, D, and
E are of the more modem, larger capacity,
cross-seat types, representing the latest de-
signs used by Chicago, New York, Brooklyn,
and Boston. Type F is the Witt car used in
Cleveland, and F-1 is the Cleveland motor
car and trailer; Car G is the lighter weight
Witt type used in Buffalo and other cities,
but not arranged for train operation. Due to
the use of longitudinal seats the maximum
capacity of the Witt car is somewhat larger
than that of the preceding types. Car H is
the standard Birney safety car, and H-1 the
same car with longitudinal scats.
The operating costs of each of these
types when used in all-day ser\-ice and when
operated as "trippers" onlv is shown in Table
II.
There are few lines in even the largest cities
that operate on less than three-minute
headwavs outside of rush hours. To make
TABLE III
LENGTH OF ROUND TRIP 8.5 MILES: RUNNING TIME 60 MINUTES
NUMBER REQUIRED
Scats
per Hour
Normal
CAPACITY
PER HOUU
COMPARATIVE COSTS PER YEAR
Type
Car
Normal
90-second
Rush
60-second
Rush
60
92
28+41
90-second
Rush
60-second
Rush
All-day
Cars
Extras
90-second
Rush
60-$econd
Rash
A
H
A-l-H
20
28
28
40
61
28 + 21
880
Kim
896
4000
40(1(1
4000
6000
(iOdO
6000
$269,400
191,(m()
191.000
$60,700
.50.900
63.700
$121,400
98..5(H)
124.01H)
B
H
B-l-H
20
28
28
40
61
28+21
60
92
28+41
880
896
896
4000
4000
4000
6000
6000
6000
295.400
191.(X)0
191.000
66.600
50,900
70,000
133.21X)
98.500
136,500
C
H
C-)-H
20
34
34
40
77
34+22
60
115
34+42
1080
1090
1090
.5000
.5000
5000
7500
7500
7500
314.900
232.000
232.000
71.000
66.200
78.000
142,000
124.500
149,000
D
H
D + H
20
36
36
40
77
36+21
60
115
36+41
1160
11,52
11.52
5000
.5000
.5lR)0
7,500
7.500
7.500
276.100
246.(X)0
246.000
62,2(M1
63,1(K)
65.500
124.400
121.500
127,500
E
H
E + H
20
36
36
40
77
36+21
60
115
36+41
lltiO
1 1 52
11. -12
5000
.5000
.501)0
7.5(H)
7.500
7.500
295,400
246,(HH)
246,.5(H)
6ti,('>00
tl.i,l(H)
7(),(HH)
1.33.200
121.500
136..500
K + F-1
H-1
F-l+H-1
20
36
36
*40
80
36 + 11
*60
120
36+21
lUiO
1170
1170
()0()0
6000
600(1
9000
90(H)
9000
295.4(H)
246.0(H)
246.0tH)
36,300
63.100
56.600
87,750
124,500
108,000
G
H-1
G + H-1
20
36
36
40
80
36+22
CO
120
36+42
1160
1170
1170
6000
6000
6000
9000
9(HH)
9000
263.100
246.1M)0
246.(HH)
.59,300
63.100
65, 1(H)
118,600
124..5<X)
124.2(K)
H-1 all day
F-1 rush extras
30
30
9
30
18
960
2250
2700
22.50
.5400
204,900
'48.'8(>o'
97.600
* 20 and 30 two-car trains respectively.
OPERATING COSTS OF VARIOUS TYPES OF CITY CARS
329
ii ^.^^im^i^m
1'!': 1111 iisiai
iisjijiiii^^ifSi^ ijiniiii' igs^^^ippffiriii ;i;|
Wi 5 II ■•'•
... .v|[ ff H ... 1
mm
Fig. 2. Two-car Train. Low-wheei, Light-weight Cars. Buffalo and Lake Erie Traction Company
the comparison on very heavy traffic routes,
therefore, we will assume two city lines
giving three-minute normal sen.'ice; on one
we will reduce this to yU-second headways for
the rush hours, and on the other to (il)-
second with single cars or to two-minute head-
ways with two-car trains. Few, if any, routes
can secure anywhere near a seated load for
large capacity cars on a three-minute head-
way, outside of one or two trips in the morn-
ing and evening; and there are very few
which would not provide seats for all pas-
sengers outside of rush hours if the Birney
safety cars were used on j^resent headways.
But, to make the comparison as severe as
possible, let us assume that the Birney cars
must at all times furnish equal seating
capacity per hour and equivalent maximum
capacity per hour or for any fraction of the
hour at the peaks. Table III shows how
many cars would be required, as compared
with each of the larger types, if used exclu-
sively; and also if used for the all-day runs,
the larger cars being employed for trippers.
The respective cost of operation, based on the
figures in Table II, is also shown.
To properly analyze the figures in Table
III, we will summarize the operating cost
TABLE IV
INVESTMENT, FIXED CHARGES, AND OPERATING COSTS
Type Car
A
H
A-l-H
B
H
B-l-H
C
H
C-t-H
D
H
D-(-H
E
H
E-fH
F-1
H-1
H-H-F-l
G
H-1
H-l-l-G
H-1 all day
F-1 rush
Total No.
40
61
49
40
61
49
40
77
56
40
77
57
40
77
57
20T
80
47
40
80
58
30
■ 9
60
92
69
60
92
69
60
115
76
60
115
77
60
115
77
SOT
120
57
60
120
30
18
Purchase Cost
$500,000
366,000
431,000
560,000
366,000
462,000
600,000
462,000
.534,000
520,000
462,000
489,000
560,000
462,000
510,000
460,000
480,000
469,000
.520,000
480,000
.502,000
180,000
207,000
.$750,000
552,000
682,000
840,000
552,000
742,000
900,000
690,000
834,000
780,000
690,000
748,000
840,000
690,000
790,000
690,000
720,000
699,000
780,000
720,000
761,000
180,000
414,000
Fixed Charges
.175,000
55,000
65,000
84,000
55,000
69,000
90,000
69,000
80,000
78,000
69,000
73,000
84,000
69,000
76,500
69,000
72,000
70,500
78,000
72,000
75,000
27,000
31,000
$112,500
82,500
102,100
126,000
82,500
111,000
135,000
104,000
125,000
117,000
104,000
112,000
126,000
104,000
118,500
103,. 500
108,000
105,000
117,000
108,000
114,000
27,000
62,000
Operating Cost
$330,000
242,000
255,000
362,000
242,000
261,000
386,000
298,000
310,000
338,000
309,000
312,000
362,000
309,000
316,000
3.32,000
329,000
303,000
322,000
329,000
311,000
} 254,000
$391,000
290,000
315,000
429,000
290,000
328,000
4.57,000
357,000
381,000
400,000
368,000
374,000
429,000
368,000
382,500
383,000
391,000
354,000
382,000
.391,000
370,000
302,000
Net Annual Saving
by Safety Cars
1 108,000
85,000
149,000
116,000
109,000
86,000
38,000
31,000
68,000
53,500
27,500
1,000
1,400
$131,000
86,000
183,000
116,000
131,000
86,000
45,000
31,000
83,000
54,000
*12,500
27,500
* 18,000
1,500
* Increased cost with Birney safety cars.
330 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
figures and add a comparisioii of the amount
of investment and the annual fixed charges (or
cost of capital) covering interest, depreciation,
taxes, and insurance (which will be about 15
per cent). These data are given in Table IV.
These figures indicate (neglecting for the
moment the question of track capacity or
saturation) that the Birney safety car can be
used under even the most extreme conditions
of surface traffic, and that in first cost and in
operating cost per passenger handled it will
be more efficient than most types of double-
truck cars now used and equal to the best of
the latter. The only question that can be
raised is whether schedule speeds would be
seriously interfered with by attempting to
operate the small cars on 30-second head-
ways. The writer believes there would be
no difficulty on this score. All of his observa-
tions and experience have indicated that the
reduction in stops made and in the number
of passengers handled per stop would enable
the Birney car on a 30-second headway to
maintain schedules and spacing better than
would the larger trains on a two-minute
headway, or the large single car on a one-
minute inten.'al.
But, if such short headways should produce
serious interference in sections where several
routes operate over the same track, the
combination of large cars or trains for the
rush-hour extras with the Birney cars on
the all-day runs will be found to afford a
smaller initial investment and a lower operat-
ing cost than can be secured by the exclusive
use of any type of double-truck car; and will
give a wider rush-hour spacing than the
exclusive use of Birney cars, and a shorter
all-day headwaj^ than is afforded by the larger
types. While this might not be productive
of any such increased riding as results when
shortening five-minute or longer headways,
still it is probable that some increase in
traffic could be looked for. This is an addi-
tional argument for the selection of the
Birney safety type car.
But the most decisive factor is the fact
that in everj- city there will be available, as
Birney cars come to be used more and more
for the all-day service on various routes, an
increasing amount of displaced large capacity
rolling stock, which while inefficient for all-
day service can be profitably employed in
rush hours ; and these cars can and will be used
for tripper ser\-ice to supplement the Birney
types and to produce the big economies of
all-day operation that can be made with a
minimum investment in new equipment.
1
m
1
%
1
*m
"vi^^H
»..a.
• .... . > • -f ;•( III! K'
i iBI •"" IHi/m I* ■■■■ '
Fig. 3. Peter Witt Car. Schenectady Railway
331
Motor Busses or Trackless Trolleys
By H. L. Andrews
Railway and Traction Engineering Department, General Electric Company
The trackless trolley, which is a vehicle practically unknown in this country, is making an exceptionally
good record for earnings and service abroad. A comparison of its operating cost with that of the motor bus,
and a balancing of each against that of the most modern street railway practice, shows that on an equal service
basis at an equal fare the motor bus cannot compete with the street car but that the trackless trolley can.
Therefore it will not be long before the American street railway operator must seriously consider this trackless
conveyance as an adjunct to his own equipment or as a competitor of it. Because the propulsion equipment of
this new type of vehicle and the power distribution to it are essentially the same as that of the street car, and
the vehicle's qualifications make it preeminently suitable as an auxiliary to a street car system, the author of
this article earnestly recommends that the railway operator adopt it and thus in one move secure its benefits
and eliminate its competition. — Editor.
WITH the grow-
ing use of gaso-
lene motor busses as
feeders to street rail-
way' system.s and also
as competitors to es-
tablished street rail-
way system.s, it seems
desirable that some
ana]}-sis be made of
the relatiA'e merits of
gasolene and electric
power for this type of
>tiblie conveyance
and that the operat-
ing costs be compared with those of an electric
car.
Gasolene motor busses have been given
thorough trials by a number of railway
companies which have operated them as
feeders to their established street railway
systems or in connection with their systems
to ser\'c a portion of the city not serA'ed by
existing street car lines.
Experience in the operation of the gasolene
propelled bus, so far as it has been developed,
has proved :
(1) That they cannot compete in operat-
ing costs with an electric street car
and cannot maintain an equal service
at an equal fare.
That they are an excellent type of
vehicle to operate as feeders, or to
connect up street railway routes.
That they are unsuitable for dealing
satisfactorily with heavy town traffic.
That they are not adequate for deal-
ing with peak loads.
That the>' have no advantage over the
electric car as regards schedule speed.
There is no question but that the gasolene
motor bus has come to stay and that its use
will increase. Rather than meet it as a
competitor, the street railway people should
(2)
(3)
(4)
(5)
handle and develop this bus as their experience
will enable them to make the bus a successful
auxiliary to their established business.
While it has been demonstrated by actual
application that the gasolene propelled bus
cannot compete in operating costs with the
street car, it has also been demonstrated that
the gasolene bus can be a worthy competitor
of the street railway s>-stem by providing a
higher class of service and charging a cor-
respondingly higher fare.
The operation of the gasolene propelled
busses in Baltimore is an excellent example of
the faihire of this vehicle to compete success-
fully with the street car on an equal fare basis,
but the Fifth Avenue busses of New York
Fig. 1. Typical Trackless Trolley Bus LIsed in England
City and the Chicago busses are illustrations
of successful operation by providing a higher
class of sen'ice and charging a higher fare.
The success of this latter type of installation
is best shown by the fact that the operation
of gasolene busses in Chicago is to be extended
and that gasolene busses are to be installed
in Detroit. The City of New York is fur-
nishing servace by means of gasolene operated
busses to the former patrons of those street
332 April, 1920
GENER.\L ELECTRIC REVIEW
Vol. XXIII. No. 4
car lines which have been discontinued due
to their inability to earn operating costs.
The gasolene propelled bus, due to its cost
of operation, will undoubtedly never displace
the street car; but the bus with very little
change from standard automobile construc-
tion can be converted into a trackless trolley
driven by railway motors supplied with power
from two trolley wires. In this converted
form the motor bus, or trackless trolley, ma}-
prove to be a very worthy competitor of any
street railway; and it can compete in operating
costs with the Safety car, which is the most
efficient means of transporting passengers on
steel rails.
Fig. 2. Trackless Trolley Bus Operation on Tecs
Side System, England
An analysis of the operating costs of the
gasolene bus immediately suggests to the
railway operator the reduction of this cost
by the application of a railway motor as the
power unit. Xcarly 30 per cent of the operat-
ing cost of the gasolene bus is for power,
maintenance, and depreciation. The costs
of o])eration of two gasolene bus lines are
approximately as given in Table I.
.•» < TABLE I
Maintenance of equipment 10.57
Gasolene and oil 4.84
Conducting transportation i:}.88
General and niiscellaneoiis 4.3
Traffic expense 0.04
Taxes 1.6o
Total :!.5.28 32.73
Depreciation 3.22 Ij.oil
Total .38.50 : 39.32
These operating costs are actual figures
from typical installations of gasolene busses,
and reference to columns ^4 or B indicates
that the cost of maintenance, gasolene, and
depreciation is 45 to 50 per cent of the total
operating expense. In comparing these three
items with the most recent and most efficient
street car, the Safety car. we have the results
given in Table II.
TABLE II
GASOLENE Bt S
ELECTRIC
CAR
Cents per Bus Mile ^'^^y/Jf;
Maintenance of
equip-
ment
10.57
6.51
•1
Gasolene
4.84
5.09
—
Power
l.o
Depreciation. . .
3.22
6.59
2
Total...
is.ii:',
IS. lit
.~>..5
These operating costs, which are for the
same general type of vehicle as regards seating
capacity and ser\'ice rendered, indicate a
reduction of approximately 13 cents per mile
in favor of the electric street car.
The operating costs as given in Tables
I and II for the gasolene operated bus are
based on two-rran oi)eration. While there
arc gasolene propelled buses in ser\-icc with
only one operator, no attempt has been m.ade
to operate these busses in congested districts;
and the gasolene bus as it is developed today
cannot be operated by one man with the same
degree of safety and efficiency as a trackless
trolley or a Safety car. The successful opera-
tion of more than 2000 Safety cars in over
200 cities in the United States has proved that
one ojjerator can successfully handle con-
gested traffic, provided the car is equijiped
with safety features designed to minimize
labor and protect i^assengcrs. These safety
features could be adapted to the trackless
trolley, and thus equiiJi^ed the vehicle could
be successfully handled in heavy traffic by
one o|)erator with the same degree of safety
as a Safety car.
A true comijarison of the relative operating
costs of the gasolene motor bus, the trackless
trolley, and the Safety car. assuming that the
gasolene bus can be operated with one man, is
best represented by Table III.
By giving the motor bus and the trackless
trolley the benefit of their comparatively
lower capital expenditure, which will vary
with the frequency and headway of ser\-ice,
we have the comparison of operating costs
MOTOR BUSSES OR TRACKLESS TROLLEYS
.333
TABLE III
CENTS
CENTS PER
BUS MILE
PER CAR
MILE
Gasolene
Trackless
Safety
Bus
Trolley
Car
Maintenance of over-
head
0.5
0.5
Maintenance of way . .
1.5
Road taxes
0.75
0.75
Maintenance of equip-
ment
8.54
3.0
2.0
Platform expenses
8.0
8.0
8.0
Traffic expenses
0.04
(1.04
0.04
Power
4.54
1.8
1.8
General
3.54
3.54
3.54
Depreciation
(3.59
2.0
2 0
Total
32.00
19.63
19.38
pari.son of operating costs this tax has been
included.
In comparing the maintenance of the gaso-
lene propelled bus with the trackless trolley,
consideration must be given to the cost of
maintaining the gasolene engine, clutch,
gear box, differential, radiator, magneto, and
lighting set as against a railway motor,
worm drive with differential, controller, and
two trolley poles.
The maintenance costs of railway motors,
controllers, and trolley poles are well known
figures. The maintenance of the gasolene
pro])elled bus has not been as definitely
determined, but all information available
indicates that the maintenance figure used
in the foregoing tabulations for the gasolene
propelled bus is a conservative one.
Fig- 3 Trackless Trolley Bus on Shanghai Tramways, China
as given in Table IV. This comparison
indicates that the operating cost of the track-
less trolley is approximately (iti per cent of that
of the gasolene propelled bus, and is approxi-
mately the same as that of the Safet}' car.
It will be noted that the comparison in
Tables III and IV includes a road tax for
the gasolene bus and for the trackless trolley.
It may not have been the custom to. charge
the gasolene propelled btis for the use of city
streets, but if this bus should come into
general use and provide a regular service on
a specified schedule, a tax would probably be
imposed; and in order to make a true com-
The body maintenance of the gasolene
propelled btis and of the trackless trolley will
TABLE IV
CENTS PER
BUS MILE
CENTS
PER CAR
MILE
Gasolene
Bus
Trackless
Trolley
Safety
Car
Operating costs
Capital expenditure- . .
32.0
1.85
19.63
2.85
19.38
3.37
Total
33.85
22.48
22.75
334 April, 1920
GEXER.\L ELECTRIC REVIEW
Vol. XXIir, Xo. 4
be no more than that of a street car, and the
truck maintenance incurred in street railway
practice will be almost eliminated. Against
this lower body and truck maintenance must
be balanced the relative cost of rubber tires
and steel wheels. From all data available,
the cost per bus-mile or per car-mile for tires
and wheel wear is in favor of the steel wheel.
The relative figures are given in Table V.
TABLE V
Tire Life
in Miles
Cents per
Bus Mile
Wheel Life
in Miles
Cents per
Car Mile
19,000
28,000
13,500
1.00
1.51
1..35
45,000
0.24
22,166 ave. , 1.28 ave.
This tabulation indicates that the cost per
bus-mile for rubber tires will be approxi-
mately five times the cost for steel wheels; or
the cost of rubber tires on a gasolene propelled
bus or a trackless trolley will be one cent per
bus-mile greater than the cost of wheel wear
on a Safety car. Taking into consideration
the higher cost of rubber tires per bus-mile,
and knowing the cost of maintaining a Safct\"
car, it seems consen,'ative to estimate the
maintenance of the trackless trolley as 50
per cent greater than that of a Safety car,
particularly if the trackless trolley is equipped
with a single motor without gears, axle lin-
ings, or gear case.
The trackless trolley could be built with
approximately the same sealing capacity as
the Safety car for a weight not to exceed
12,000 lb. or approximately 75 per cent of the
weight of the present Safety car. A single-
motor drive with necessary' control can be
supplied which will permit of the adoption
of all the safety features now standard for
the Safety car. For a trackless trolley the
power consumption will be ajiproximately
the same as for a Safety car, as the weight
will be 75 per cent of that of the Safety car
and the frictional resistance of a rubber tired
vehicle on a good asphalt, wood block, or
smooth brick pavement is ver\- little higher
than on steel rails, particularly where there
is used a grooved rail laid in paving.
On first thought, it would appear that the
gasolene propelled bus has an advantage over
the trackless trolley as regards unlimited flex-
ibility. It is true that a gasolene propelled
bus can operate on any route and can readih'
have its route changed without incurring any
expenditure for changing overhead con-
struction. There is little question, however,
but that with the introduction of the gasolene
propelled bus the city authorities would insist
on a definite route and a definite time table
and while the gasolene propelled bus has
unlimited flexibility as regards routing it
would be as definitely bound to a specified
routing by ordinances or legislation as a
trackless trolley would be by reason of the
overhead construction. It is questionable,
therefore, whether either type of vehicle has
any actual advantages as regards flexibility
for both vehicles can pass other traffic.
With proper overhead construction and proper
collectors, the trolley bus can have a range of
operation of 15 ft. either side of the trolley
wires, which is ample to permit passing other
traffic.
These estimates would illustrate that a bus
similar to the gasolene propelled bus could
enter the urban transportation field and
become a worthy competitor of the street
railway system, giving equal ser\nce for equal
fare. This is particularly true where the
officials of railway companies have not
profited by the experience gained in the
application of the Safety car and have not
applied its principles to their transportation
problems. The trackless trolley having lower
operating cost than the gasolene propelled
\-ehicle will be successful where a gasolene bus
could not operate.
Sooner or later the street railway industr>-
is going to meet the gasolene bus or trackless
trolley in competition, and it seems desirable
that railway operators study their trans-
portation problems with a view to utilizing
the trackless trolley as an auxilian,- to their
present transportation system rather than to
meet it in competition.
335
Improvements in the Design and Construction
of Railway Motors
By E. D. Priest
Engineer, Railway Motor Department, General Electric Company
The modern light-weight railway motor for the same weight as its predecessors is capable of handling,
under ordinary conditions, a car about twice as heavy per pound of motor. These improved motors are the
result primarily of developmental and research work and secondarily of the fact that there are available today
better materials and methods of manufacture than heretofore existed. In the following article Mr. Priest
details a number of the more prominent features in the design and construction of the latest General Electric
railwav motors. — Editor.
r
N the November,
1913, number of
the General Elec-
tric Review, the
writer ijubhshed a
short article entitled:
"The Develoinnent of
the Modern Direct-
current Railway Mo-
tor." This article was
a brief review of the
subject. Sinceitspub-
lication there have
been many substan-
tial improvements in
the design and construction of railway motors ;
and it is the purpose of this article to supple-
ment the earlier one in a measure and to de-
scribe briefly some of these improvements.
A marked advance has been effected in the
design of railway motors. This has been ac-
complished by the use of higher grade mate-
rials, refinements in design, increased ventila-
tion, higher armature speeds, increased gear
ratio, and reduction in weight made possible
by these changes. If it were not for these
E. D. Priest
improvements, the present manufacturing
cost of railway motors, to jjerform a given
service, would be much higher.
Heat-treated alloy steel is now used for the
armature shafts. The steel in the smaller
motor shafts is substantially the same and,
for like sizes, is equal to that used in the
crank shafts of the " Liberty motors " designed
for use in airplanes.
The quality of steel in gears and pinions
has been improved and improved methods
of heat treatment have been developed. The
highest grade materials are now used for
railway motor gears and pinions. New ways
have been found of tempering cast-steel
gears which produce qualities substantially
equal to forged gears.
Bearing metals are now of the highest
quality obtainable. All babbitt is genuine
tin-base babbitt. This is the most expensive
babbitt manufactured and long experience
has shown it to be the best. The highest
grade bronze is used in the linings.
In some instances key stock is heat treated
to secure hardness and is ground to size to
insure close fits and freedom from wear.
Fig. I. A Modern Light-weight Railway Motor, showing Axle Side
336 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
Heat treated carbon steel bolts are quite gen-
erally used in the construction of motors and in
some motors heat-treated alloy steel is used.
For brush-holders, expensive high-grade
bronze castings are used exclusiv^ely and
carbon brushes are of the highest grade
obtainable.
All castings other than bronze are either
malleable iron or steel, no cast-iron being em-
ployed in the construction of railway motors.
In general, the quality of materials now used
is the best, and no inferior substitutes are
emplo}'ed. Operating conditions are so severe
that maximum all round economy can be
obtained only by the use of the best materials.
Much study and research has been devoted
to producing higher grade varnishes em-
ployed for insulating purposes, and in the
be driven into place. In boring the heads
for armature linings and in turning the
linings, ven,' close tolerances are required in
order to secure the proper pressing fit of the
linings in the heads. A tolerance of plus
0.001 to minus 0.000 is used in the bore of
solid gears.
The thread fit for frame-head bolts and for
screws is made so close that special taps and
dies are required to insure tight fitting
threads. Throughout the whole construction
of the motor, limits in workmanship are very
close as it is found that imperfectly fitting
parts rapidly loosen and wear in the abnor-
mally hard service to which railway motors
are subjected.
Armature shafts in bearings are ground to
size and rolled, a process which produces a
Fig. 2. Suspension Side of the Light-weight Motor Shown in Fig. 1
past few years there hax'C been dc^•eloped
greatly superior varnishes which have higher
insulating values and slower ageing qualities.
As with materials, so with workmanship;
the best workmanship has been found to be
the cheapest since reliability in service is of
far more importance than first cost. While
the rough exterior of a railway motor suggests
quite ordinary workmanship, as a matter
of fact it is doubtful if any other line of
machinery manufactured has closer fits and
more accurate workmanship.
Some of the tolerances in armature shaft
fits are plus 0. 00025 to minus .00000. For
frame-head fits in box-frame motors, plus
0.002 to minus .000 are allowed. The fit
must be so close as to require that the heads
hard smooth polish having an ideal bearing
surface. Equal care is taken to secure a
hard smooth surface on the babbitt in the
bearing linings.
In order to prevent vibration due to
armatures being out of balance, the detail
parts of the armatures arc lialanccd .sepa-
artely before being assembled on the shafts,
and after assembly the completed core is
balanced.
Aside from material and workmanship,
substantial improvements have been made
in the design of motors. Box-frame motors
have come into almost universal use, this
construction being greatly superior to the
split-frame type in slurdiness ami reliability
of operation.
IMPROVEMENTS IN DESIGN AND CONSTRUCTION OF RAILWAY MOTORS 337
The ventilation of motors has been much
improved so that multiple ventilated motors
have largely increased service capacity. The
continuous capacity in some instances is 70
per cent or more of the Imurly rating. Venti-
lating fans have been strengthened so that
trouble from breakage has been largely
reduced.
^^^.^•^^'^VX-^Wx^N ■■^^ V V^\\^>^^N^^^^j^^?y!^:^^^g^^^^§>y^^
Figs. 3 and 4. Sectional Drawings of the Light-weight Railway Motor shown in Figs. 1 and 2
338 April, 1920
GENER_\L ELECTRIC REVIEW
Vol. XXIII. Xo. 4
A superior construction of armature bars
applicable to large sizes of motors has been
developed. This construction permits the
use of thin folded crossed bars which make
it possible to obtain greater capacity with a
gi\-en size of armature core without increasing
eddy current losses due to hea^^^ copper
sections.
A method of connecting bars at the back
end of one-turn armatures has been devised
which eliminates the use of soft solder that is
liable to melt if motors are subjected to
excessively heavy overloads which sometimes
occur in locomotive service. The improve-
ment consists in using electrically brazed
joints in place of soldered joints.
In armature windings of more than one
turn per coil, wire of rectangular section has
come into more general use. The s]:)ace
factor with rectangtilar wire is materially
higher than with round wire. This results
in an increase in capacity of armatures for
given core sections.
Taking greater advantage of the possibili-
ties of employing commutating poles, the use
of two turns per coil in armature construction
has been extended to much larger motors than
formerly thought possible, thereby decreasing
the weight and cost of the motors.
Sheet steel gear cases have been developed
to a higher point of perfection so that they arc
proving more reliable in service than sheet
steel cases of earlier designs.
A much desired improvement has been
brought about in the method used to pre-
vent rotation of axle linings in large sizes of
motors. The construction consists in the us5
of a long key set in the bore of the magnet
frame for the lining, along the lower edge of
the split in the lining. The lining is not
materially weakened since it is at the point
of separation of the two halves. This con-
struction has been found to hold linings
very securely.
Spring gears have been developed, the use
of which in heavy work prexx'nts excessive
shocks on gear and pinion teeth, resulting
from imperfections in the teeth or rough
service conditions. When twin gears are
used spring gears tend to equalize the work
on the two sets of gearing.
Motors have been designed for largely
increased potentials and 3(J()()-volt direct-
current railway motors ha^•e been in most
successful operation for a number of years,
handling the severest of service.
Higher armature speeds have been made
possible by the use of stronger material in the
shafts and in the pinions and gears, and by
improved shape of gear and pinion teeth which
permits the use of a finer pitch gearing, a
smaller pitch diameter of pinion, and a smaller
number of teeth in the pinion, without a
reduction in the strength of the teeth as
compared with coarser pitch gearing with
inferior shaped teeth.
For many years the standard gear used in
street railway service had three pitch 14j^
deg. angle teeth. By changing the angle to
20 deg. approximately 25 per cent stronger
teeth have been secured, and by lengthening
the pinion teeth addendum and shortening
the dedendum with a corresponding shorten-
ing of the gear teeth addendimi and lengthen-
ing of the dedendum it is possible to change
a three pitch to approximately a 3J2 or 4
pitch without sacrificing strength, and with an
incidental possibility of increasing the gear
ratio. The shortening of the dedendum of
pinion teeth and the use of a finer pitch
permits a reduction in the number of teeth
without a reduction in the thickness of the
metal between the base of the teeth and
the bore.
An increase in strength of the jiinion and
shaft has been effected by reducing the depth
of the keyway in the pinion so that metal is
not cut away at the large end of the bore and
by shortening the keyway in the shaft so
that it does not extend to the inner end of the
pinion hut is stopped inside the pinion fit at
a point where the shaft is supported by the
shrink fit of the pinion.
The maximum armature speed for a given
car speed is of course fixed by the gear ratio.
Consequently an increase in gear ratio makes
it possible to design a lighter and cheaper
motor for a given service. Increased arma-
ture speed not only reduces the size and cost
of motors due to increase in speed, but also
makes jxissible a further reduction because
of increased ventilation resulting from in-
creased si)eed.
The miniminn number of teeth in pinions
for a given \ntch and tooth shape is limited
by the diameter of pinion bore. Sufficient
metal for amjilc strength being allowed be-
tween the base of the teeth and the bore, it is
obvious that the higher the grade of arma-
ture shaft stock used the smaller the pinion
fit and ]jinion bore can be made. Therefore
the size of a motor is fundamentally affected
liy the grade of material used in the annature
shaft and the grade of material used in the
l)inion and gear as well as by the pitch and
shape of the ]>inion and gear teeth.
IMPROVEMENTS IX DESIGN AND CONSTRUCTION OF RAILWAY MOTORS ;339
An improvement has been made in the
design of pinions of small diameter by-
making them slightly bell-mouthed at the
large end of the bore for a distance of ^ to
]/2 inch from the end of the pinion. By
relieving the pinion in this way, so that for
this distance it has no bearing on the shaft,
the metal in the body of the pinion at the
large end of the bore is stressed less when the
pinion is driven and shrunk on the shaft.
Consequently, there is less danger of failure
from breakage both in the body of the
pinion and in the teeth. Incidentally, this
permits of a better design of shaft, since
it is possible to use a fillet with a larger
radius on the shaft between the pinion fit
and the journal bearing.
In the modern light weight motor, used
on safety cars, very careful consideration has
been given to the design of the motor with
particular reference to armature shaft and
gearing. The construction of the motor in
other particulars is also worked out to secure
maximum strength, reliability, effective venti-
lation, and lightness. This has resulted in the
development of motors with continuous ratings
equal to that of earlier types of non-ventilated
motors of three to four times the weight.
In practical operation these light weight
motors do not have increased service ca-
pacity in full proportion to their increased
continuous rating. This is due to the fact
that there is a larger short-time thermal
capacity in heavy motors than in light
motors, the heat generated being absorbed
in the mass of material and slowly dissipated
during periods of light load. However, the
modern light weight motor is capable under
ordinary operating conditions of handling
a car about twice as heavy per pound of
motor and of doing this with a much lower
temperature rise. In fact, the service tem-
perature of a modern safety car motor does
not usually exceed 40 deg. rise as compared
with older and heavier motors which are
ordinarily run at a temperature of 60 to 65
deg. rise.
Street railway motors are now so effi-
ciently ventilated that there is generally
no suiastantial advantage in using heat proof
insulation, since the losses are so effectively
dissipated that it is questionable whether
there is economy in operating at higher
temperatures with increased losses and de-
creased power efficiency. Good ventilation
has made it possible to use with economy
non-heat-proof insulation which is cheaper
in material cost and in application and is
also more impervious to moisture.
Some of the railway motor improvements
which have been briefly outlined are the
most marked and far reaching that have been
made during the past half dozen or more years.
It would be possible to enumerate other
improvements. Only the "high spots" can
be touched in a short article and doubtless
the writer has not mentioned all of these.
Railway motor problems are being given
constant study and further improvements
will surely be made. However, the pre-
diction of the writer, in the article referred
to at the beginning, that "A pound of material
will be made to do more and better work " has
already been fulfilled in large measure.
340 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, X^i. 4
:t'. 1 ^n
i
>
341
Importance of Simplicity in Locomotive Design
By A. F. Batchelder
Engineer Locomotive Department, General Electric Company
T!
A. F. Batchelder
'HAT it is econom-
ical and practical
to operate many of
our tnmk line rail-
ways electrically is
no longer open to
question. It has been
proven beyond a
doubt by actual ex-
l)criencc that, under
limiting physical con-
ditions where the
traffic is dense or is
on severe grades, the
electric locomotive, in
point of reliability and cost of operation, is
superior to its steam competitor. The ques-
tions which are now assuming increased
importance are the method of financing and
the design of equipment, rather than the
question of the advisability of electrifica-
tion.
The matter of financing is one for financial
interests to consider, but the matter of design
of the equipment must be determined by the
engineers. On them will rest the responsibility
of maintaining high standards of reliability
and economy of operation; and for this
reason, in connection with the design of elec-
tric locomotives, we should emphasize the
importance of adopting locom-otive designs
which are fitted to perform the particular
service requirements with the least amount
of complication in the construction and the
operation of the mechanical as well as the
electrical equipment.
The simple design and construction of the
locomotives that are in use on the direct-
current systems of this country is the funda-
mental explanation of m.uch of the economic
advantage that has been demonstrated by
our experience in heavy railway electrifica-
tion. If a careful analysis is made of the
time out of sen.-ice and the cost of main-
tenance of the locomotives which are in
operation at the Baltim-ore Tunnel, the
Detroit River Tunnel, the New York Grand
Central Terminal, the Butte. Anaconda &
Pacific Railway, and the electrified zone of
the Chicago, Milwaukee & St. Paul Railway,
all of which are handling heavy railway equip-
ment, and if these figures are compared with
similar figures from other lines which are
handling similar traffic, but using locomotives
of more complicated design, the advantage
of adopting the more simple designs will be
shown very definitely.
It is also important to reduce the weight
of the locomotive to such a minimum as is
consistent with the requirements for traction
purposes in order to reduce the tonnage
movement and the power requirements to a
minimum. This makes it desirable in con-
sidering locomotives for freight service to
design them with all of their weight on
drivers. It is possible with the direct-current
motor to build a locomotive which has all
the weight of the locomotive on driving
wheels and which has a continuous electrical
capacity sufficient for any railway service.
Experience has shown that, for freight service,
locom.otives m.ade in this manner give satis-
factory operation, the maintenance of both
the locom.otive and the track being low.
Our observations indicate that there is no
real necessity of providing idle axles for any
of our low-speed locomotives. At the Detroit
Ri\'er Tunnel there are such locomotives
weighing from 100 to 120 tons, some of which
have been in operation for 10 years. At the
Baltimore Tunnel there are locomotives of
similar design and weight which have been
in operation for the same length of time.
The Butte, Anaconda and Pacific has 28
locomotives, weighing SO tons each, of the
same general design which have been in
operation for seven years, and some of these
have been hauling passenger trains at speeds
of 50 miles per hour.
All the locomotives mentioned are of the
two-truck articulated type, illustrations of
which are shown in Figs. 1, 2, 3 and 4. The
freight locomotives of the Butte, Anaconda
& Pacific are capable of developing contin-
uously a tractive effort of 25,000 pounds at
the rim of the drivers. The total weight of
the locomotive is 160,000 pounds, and as a
consequence the total continuous tractive
effort capacity is 15 per cent of the weight
on drivers. It is possible to build freight
locomotives of almost any desired capacity
to operate trains at speeds as high as 30 miles
per hour with a tractive effort of 15 per cent
of the weight on drivers, and in some cases
as high as 20 per cent, and still maintain a
simple and rugged construction throughout.
342 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 4
Modern Devices and Control for Automatic
Rail^vay Substations
By Cassius M. Davis
Railway and Traction Engineering Department, General Electric Company
In view of the number of descriptions of automatic substations that have appeared in the technical press,
the author in preparing the following article presupposes that the reader has a knowledge of the fundamental
scheme of operation employed. Two of the earlier articles appeared in this magazine: "Automatic Railway
Substations," October, 1915, page 976, and "Give the Operator a Job," November, 1916, page 1030. The
present article deals with the improvements which have been made in the controlling devices and describes in
detail the automatic operation of the substation; including the starting up, shutting down, protection, and
adjustment of the equipment. — Editor.
Cassius M. Davis
SINCE the intro-
duction of the
automatically con-
trolled substation by
the General Electric
Company several
years ago, many sta-
tions have been con-
verted from manual
to automatic control
and many entirely
new automatic equip-
ments have been
placed in operation.
The development of
any new device or new scheme of operation
necessarily changes rapidly in detail but
usually slowly in fundamental principles.
The automatic railway substation has been
no exception.
It will be the purpose of this article to
record the principal improvements and changes
which have been found desirable; also to
describe, more or less in detail, the scheme
of operation and the functions of the various
individual devices in an up-to-date installa-
tion. It is assumed the reader has a general
knowledge of the principle of the automatic
substation.
Experience with over fifty operating equip-
ments has shown the fundamental scheme of
operation to be rational in its conception and
successful in its application. This same
experience has shown also the limitations of
variotis types and designs of individual
apparatus on the one hand, and the entire
suitability of other individual devices on the
other.
Without attempting to present in any
chronological order the changes and improve-
ments which have been made, we will con-
sider more from the point of view of im-
portance the new eqtiijjment which has been
designed.
It might be said in passing that it is one
thing to design a mechanism which operates
only at infrequent intervals, say two or three
times a day. and quite another to design a
mechanism to do the same work twenty or
thirty times a day. This statement applies
to practically all the apparatus used in an
automatic substation and can be made as a
general statement, rather than one applying
specifically to the oil switch mechanism, but
is particualrh' true with reference to mechan-
isms which are designed to operate high-ten-
sion oil circuit-breakers, and which must not
contain too many small and delicate parts.
Any device which has a number of small and
relatively delicate parts which must be kept
in close adjustment is bound to have a shorter
life and give more interruptions than one corn-
Fig. 1
AC Motor Mechanism for Operating Oil
Circuit Breakers
posed of a few and relativch" rugged parts.
So it has been with the apparatus under dis-
cussion.
The mechanism which now forms part of
the standard equipment is illustrated in
Fig. 1. It was designed after a most careful
MODERN DEVICES AND CONTROL FOR AUTOMATIC RWY. SUBSTATIONS 343
and ]3ainstaking study of the conditions to be
met and was accepted for the service only
after having proved its abihty to withstand
successfully over 1 (JO, 000 operations. The de-
vice which has proved satisfactory under these
conditions consists of an operating motor which
is of the alternating-current, single-phase type,
the motor operating through a mechanism
affording a gear reduction and also through a
spring to move the circuit-breaker to the cir-
cuit-closing position. When the circuit-breaker
is moved to its closing position, the circuit of
the motor is then interrupted and the circuit-
breaker held latched in its closed position.
After the circuit-breaker is closed, there is
considerable tension exerted by the spring
which, thereupon, functions as a source of
power to return the motor and its auxiliary
mechanism to their initial position, leaving
the switch mechanism in such a position that
the mechanism is adapted when tripped open
to move freely to its open position.
Fig. 2.
Motor Mechanism Assembled with 15,000-volt
Breaker in Cell
All parts of the mechanism are made with
a very large factor of safety. There are
practically no small moving parts and such
levers and links as are used are made from
heavy stock capable of withstanding the
forces applied to them. A sheet metal cover
encloses the moving parts and the complete
device forms a weather-proof unit which is
suitable for outdoor work where necessary.
Fig. 2 shows one of these mechanisms as
connected to a 15,G00-volt oil circuit breaker
mounted in a cell.
The early installations were of 300- and
500-kw. capacity, hence, at 600 volts, the
Fig. 3. Shunt Type Relay used for Reverse Current
and Underload
Fig. 4. Shunt Type Relay
Used for D-C. Overload
Protection
Fig. 5. Time-limit Relay
Used to Obtain Delayed
Action When Starting
and Stopping
50 per cent overload current did not exceed
750 amp. and 1250 amp. respectively. When
handling circuits of this cajjacity it was com-
paratively easy to build direct-current relays
having series coils. However, as soon as
larger capacities were encountered it became
344 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 4
evident that more suitable relays should be
used. It was possible in some cases to operate
relays in multiple but it was recognized that
this was onh^ begging the question and it
would soon be necessary to devise relays which
could be conveniently employed on circuits
of any commercial capacity.
In order to meet this condition a line of
shunt-type direct-current relays was de-
veloped. The coils have a normal current
capacity of approximateh- 100 amp. but
receive their current from shunts connected
in series with the direct-current load circuit.
Fig. 6. AC. Conlrol Relay
It is usual ])ractice to wind coils with
copper conductors. If the coils were con-
nected across the ordinary type of shunts
such as used for ammeters, watthour meters,
etc., there might be considerable variation in
the calibration due to temperature changes
in the coil; the coil having a high jjositive
temperature coefficient and the meter shunt
a very small temperature coefficient. The
obvious way of taking care of the situation
was to use the same conductor material for
both the shunt and the coil. This was done
and both are now made of copper. The shunt
is simply a co]Dper sheet of such dimensions
as to carry the current and give the neces-
sar\- voltage droj) for operating the relay.
The use of shunt-type relays made it possible
to use the same relay for any cajiacity circuit
by varying the size of the shunt. In the early
equipments, the overload relays which con-
trolled the steps of the load limiting resist-
ance were operated from current trans-
formers connected in the leads to the slip
rings of the synchronous conA-erter. While
such an arrangement permitted the use of
the same relay for any capacity circuit it
was subject to the disadvantage that any
change in field setting necessitated a cor-
responding change in the relay setting.
Consequently, with the advent of the shunt-
type relay, it was possible to place the over-
load control on the direct-current side of the
machine, which, of course, is more satis-
factory from ever\- point of view.
The standard equipment now includes
shunt-type direct-current devices for the
reverse-current relay 29, the underload relay
37. and the overload relays 23, 24, and 25,
in Figs. 3, 4 and 11.
The first time-limit relays which were
emj^loyed were equipped with small oil dash-
pots. These frequently gave trouble, due
primarily to the fact that the dashpots were
so small it was difficult to get the time setting
desired, especially on relay 3 which is
used to delay the shutting down of the
equipment. It was also found that the con-
tacts were not suitable for the service. These
relays were therefore entirely redesigned,
bellows were substituted on relays 23, 24,
and 25, and an entirely new design of dash-
pot was applied to relay 3. The contact
mechanism of all of them was radically
changed and improved.
The latest type of 3 relay is shown in Fig. 5.
In this, the oil dash])ot has been made much
larger and the stroke of the piston longer.
This permits the use of a greater volume of
oil and reduces the necessity of small clear-
ances. The dashpot is so attached to the
body of the relay as to form practically an
integral part. The piston is self -aligning and
has a long bearing surface. The needle valve
for adjusting the time has undergone a
radical improvement, making it possible to
adjust the setting of the relay with con-
siderable accuracy. Since the cylinder prac-
tically forms a part of the bcniy of the relay
itself the oil is kept at nearly uniform tem-
perature, due to the heat generated by the
coil. While the use of oil is open to the
criticism that it changes its viscosity with
change in temperature, yet with the type of
relay under discussion, this change is reduced
to a minimum. A variation in time setting
is to be expected between summer and winter
but the ordinary daily changes of temperature
will have vcr>- little effect. Furthermore,
since it is not necessar\' that this relay be an
accurate timing device the slight changes in
viscosity which take i>lace from day to day
:M0DERN devices and control for automatic RWY. substations 345
are unimportant. That is to say, if a certain
relay is set to operate in five minutes it will
have no noticeable effect ujion the operation
of the substation if on a hot day it opens in
four minutes and on a cold day in six minutes.
The contacts of the relays 3, 23, 24, 25,
and 30 have received a great deal of attention.
After many exhaustive tests both on the
alternating-current and direct-current
circuits involved it was found that
silver to silver contacts behaved the
best, both from point of view of wear
and burning due to arcs. The con-
tacts are still operated by a toggle
mechanism but it has been so changed
as to provide a quick make and a
quick break feature. With this type
of mechanism it is impossible for the
contacts to open or close part way,
thereby tending to hold an arc. The
toggles are so arranged that when the
contacts start to move they are
forced over center by springs to com-
plete the motion to the end of the
travel. All contacts which handle
direct -current circuits are provided
with small blowout coils to decrease
the time of rupturing the arc. The
travel of the contacts, however, is
sufficient to break the arc even with-
out the blowout coils.
In the early stages of the applica-
tion of automatic control a need was
felt for a simple and ruggedly con-
structed alternating-current relay
which was self-contained on its own
base, could be easily mounted in a narrow
space on the panel, and would handle cur-
rents of somewhat larger capacity than any
of the standard relays then available. To
meet these requirements an entirely new
relay was designed which is shown in Fig. (>.
It will be noted this is constructed along the
lines of a contactor but is provided with an
adjustable stop, thus making it possible to
adjust the pick u\> point over a moderate
range. The contacts will easily carry ten
amperes alternating-current and momen-
tarily much heavier currents. It is of very
rugged construction and easy to keej:) in
adjustment.
The first few installations were designed for
converters without commutating poles. To
adapt the equipment for the commutating
pole type it was necessary only to provide
some means for mechanically raising and
lowering the brushes. A special motor-
operated mechanism was developed for this
purpose, using a repulsion-induction type
single-phase motor driving the brush mech-
anism through a gear reduction. Suitable
limit and auxiliary switches were incorjjorated
in the device.
Only minor changes have been found
necessary in the motor-driven controller.
The original controller was built with a
Fig. 7. Motor-driven Controller
Fig. 8. Motor-driven Controller
vertical cylinder. Later this was changed
to a horizontal cylinder and minor mechanical
changes made to adapt the parts for this
type of construction. For some time a
dynamic brake was used to bring the con-
troller to rest. The braking was accomplished
by throwing a heavy load on the small exciter
connected to the driving motor. This was
later changed and a solenoid brake employed.
This has the advantage of stopping the con-
troller a little more accurately and also pre-
\-ents any possibility of overloading the
driving m.otor. The views shown in Figs.
7 and S give a clear idea of this mechanism.
However, they do not show the cover which
shields the entire device from dirt and dust.
An improved type of bearing thermostat
has been in use for some time. It consists of
a metallic bulb connected to an expansible
chainber through the medium of a capillary
tube. The btilb, tube and chamber are
filled with a liquid which volatilizes at a
346 April, 1920
GEXER.\L ELECTRIC REVIEW
Vol. XXIII, Xo. 4
definite temperature. As the chamber ex-
pands, it operates a relay toggle to which the
contacts are assembled, Fig. 9.
The bulb is inserted in a hole in the lower
half of the bearing, while the relay part is
mounted on a bracket conveniently placed
on the bearing ])edestal. The hole in the
bearing is drilled parallel to the shaft and as
close to the babbitt surface as mechanically
possible. The depth of the hole depends upon
Fig. 9. Bearing-temperature Relay
the length of the bearing, the bulb being .
placed as near the center as possible. The
bulb is thus located advantageously for
registering the highest temperature through-
out the bearing length. For ^•e^^• long
bearings, two temi)crature relay's are used,
one inserted from each end of the bearing.
The relay contacts are normally closed, and
open due to excessive temperature. Having
once opened they must be reset by the in-
spector after ascertaining the reason for over-
heating.
A number of minor moditications have been
made here and there to imjmn-e the indi\-idual
devices or the general scheme of operation
and protection. For example, the j)olari7,e(l
relay 36 has been somewhat changed in
design to make it conform in appearance and
construction to the reverse-current and under-
load relays. Also, the speed-limit device IJ
has been provided with an additional set
of contacts. This device now has a normalls-
closed and a normally open set of contacts.
On overspeed, one set closes the circuit to
the shunt trip on the line circuit breaker
(which is othenvise non-automatic) and the
other set opens the control circuit. When
the speed-limit device trips, both contactor
18 and the line breaker open, thus doubly
insuring that the machine is cleared on the
direct-current side.
SCHEME OF OPERATION
In what follows is gi\-en a detailed de-
scription of the operation of a typical auto-
matic railway synchronous converter substa-
tion. The general scheme is almost identical
with that used in the first installations.
The motor-driven controller may be con-
sidered the "brains" of the outfit, with the
contact-making voltmeter and the underload
relay acting as the "eyes."
The voltmeter registers the load demand
and starts the controller. The controller
then fixes the sequence of events and closes
or opens the various control devices in the
proper order and at the i^roper time during
the starting ojierations. It ser\-es at once
electricalh' and chronologically to interlock
the various breakers, contactors, relays, etc.
Fig. 10. Contact-nuking Voltmeter
After the machine has been delivering load
and the economical demand ceases, the under-
load relay acts to shut down the station and
advance the controller to its initial j)osition
ready for the next start.
These remarks, while ap])lying to the
control of a s\nchronous con\-orter, are also
MODERN DEVICES AND CONTROL FOR AUTOMATIC RWY. SUBSTATIONS 347
348 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 4
applicable to that of a motor-generator set
starting from transformer taps or compen-
sator. The description refers specifically to
the wiring diagram shown in Fig. 11.
With the high-tension line energized to
its full voltage and the lever switch 8 closed,
relay 27 closes its right-hand contact, ener-
gizing the coil of £7-X and closing it. When
£7-X closes, it seals itself in through the
circuit completed by its contacts.
Fig. 12- A-C. Inverse Time-limit Overload Relay
Starting Up
A load demand on the equipment is
indicated by a reduction in the trolley voltage
at the substation. The contact-making
voltmeter 1, Fig. 10, connected between the
direct-current bus and ground, utilizes this
voltage reduction to start the converter.
As long as the trolley voltage is u]) to normal
the contacts remain closed and keep relay
2 energized thereby holding the contacts of
the latter open. As soon as the trolley voltage
falls to the setting of the voltmeter, the fol-
lowing sequence of operation takes place:
(1) The contacts of the voltmeter / open and
de-energize relay 2. The contacts of 2 do not
immediately close owing to an adjustable time
setting. If the reduced trolley voltage persists
continuously, during the lime setting of this relay,
its contacts close and close relay S. A circuit is
then established from the upper alternating-current
control bus through the contacts of 27-X, the con-
tacts of S, coil of Ji, auxiliary switch on the circuit
breaker, contacts of 26, speed-limit switch, bearing
thermostats, and hand reset switch on the oil
circuit breaker mechanism back to the lower alter-
nating-current control bus.
(2) Contactor ^ closes, establishing a circuit
from the upper control bus through one of its
contacts to segment IS on the controller thence to
segment Ui, upper contact of auxiliary switch on
brush-raising device, and to the operating coil of
contactor 6', back to the lower control bus.
(3) Contactor 6 closes and starts the motor
driving the controller.
(4) Segment 15 on the controller makes con-
tact, closing the circuit to the operating coil of
contactor 5.
(5) Contactor 5 closes, establishing a circuit
through one of its contacts to segment / on the
controller and, simultaneously establishes a circuit
from the same contact to the closing circuit of the
oil switch motor mechanism.
(6) The oil switch closes, energizing the power
transformer and therewith the coils of both relays 3;?.
(7) Segment 14 on the controller makes con-
tact, completing a circuit through the auxilian,-
switch on the oil circuit breaker, one of the contacts,
of S and the operating coil of 5. This operation
thus establishes a holding circuit for contactor -5
as soon as the controller advances bevond segment
15.
(8) Segment 2 on the controller makes contact,
establishing a circuit through the contacts of both
relays 32 and the operating coil of contactor 10.
(9) Starting contactor 10 closes, placing reduced
voltage upon the collector rings of the converter
from the transformer taps. The converter starts.
(10) If the converter has come up to synchronous
speed by the time the first gap in segment 16 is
reached, a circuit is established from segment H
through the contacts of 13 to segment 20 and thence
to segment IS and the operating coil of contactor
6". This holds contactor 6 closed until the gap in
segment 16 is past. If the converter has not come
up to speed by the time the gap in segment 16 is
reached, the circuit to the operating coil of 6 con-
tactor is broken and the controller comes to rest
until synchronous speed on the converter is reached;
i.e., until IS closes.
(11) Segment S makes contact, closing the cir-
cuit to the operating coil of field contactor SI.
Fig. 13. A-C. Low-voltaKC Relay
(12) Contactor 31 closes and connects the
fields of the converter to the 2.")0-volt exciter on
the controller, giving proper polarity to the con-
verter. As the converter is brought to the proper
polarity, relay S6 closes its contacts.
(13) Segment S breaks contact, opening con-
tactor SI.
(14) Segment 4 makes contact, energizing the
operating coil of full-field contactor / (.
■MODERN DEVrCF.S AND CONTROL FOR AUTOMATIC RWV. SUBSTATIONS 349
(15) Contactor 14 closes and places the field of
the converter across its own armature for self-
excitation.
(16) Segment 2 breaks contact, opening start-
ing contactor 10.
(17) Segment S makes contact energizing the
operating coil of running contactor 16.
(18) Contactor 16 closes and puts
full alternating-current voltage on the
collector rings of the converter. At the
same time, relay 30 closes due to the
establishment of full voltage across the
armature of the converter.
(19) Segment 26 makes contact,
establishing a circuit through the upper
contacts of the limit switch on the
brush raising device.
(20) The motor of this device starts
lowering the brushes.
(21) If the brushes reach their lowest
position and the lower contact of the
auxiliary switch on the brush raising
device is closed before the controller
runs off the second gap in segment 16,
a circuit is established from segment 17
through the lower auxiliary switch of
the brush raising device to the operating
coil of contactor 6, thus holding 6 closed
and permitting the controller to con-
tinue to revolve.
(22) If the controller runs off seg-
ment 16 before the brushes are in their
lowest position, the operating coil cir-
cuit of 6 is opened and the controller
stops until the lower auxiliary switch on
the brush-raising device closes and com-
pletes the circuit from segment 17 de-
scribed above.
(23) Segment 7 makes contact, giv-
ing direct-current potential to segments
S, 9, 10. and 11.
(24) Segment 26 breaks contact, de-
energizing the circuit to the brush-
raising device.
(2.5) Segment S makes contact,
establishing a circuit through one of the
contacts of contactor 4, the contacts of
polarized relay 36, the contacts of relay
30, the electrical interlock on contactor
16, and the operating coil of contactor
IS.
(26) Contactor iS closes, connecting
the converter to the bus through all
three sections of the load limiting
resistance.
(27) Segment 9 makes contact,
establishing a circuit through the
operating coil of 21 contactor and the contacts of
relay 2.5.
(28) Contactor 21 closes, short-circuiting a sec-
tion of the resistor.
(29) Segment 10 makes contact, establishing
a circuit through the operating coil of 20 and the
contacts of 24- r'
(30) Contactor 20 closes, short-circuiting^ a
section of the resistor. !
(31) Segment 11 makes contact, establishing a
circuit through the operating coil of 19 and the
contacts of 23.
(32) Contactor 19 closes, short-circuiting the last
section of the resistor. The machine is now con-
nected directly to the bus and delivering load.
During the last several operations mentioned above.
after IS contactor closed, the contacts of relay 37
close, short-circuiting the contacts of relay 2.
Simultaneously, the voltage of the bus has been
brought up to normal and the contacts of 1 again
close. This action opens the contacts of 2 but
relay 3 still remains energized, due to the by-pass
circuit through the contacts of 37.
Fig. 14. Panel Containing Control and Converter Field Contactors
Typical A-C. Starting and Running Contactors
(33) Segment 17 breaks contact, opening the
circuit previously established through the lower
contacts of the brush-raising device and the operat-
ing coil of contactor 6.
(34) Contactor 6 opens and the controller comes
to rest at the running position.
Shutting Down
(1) When the load demand decreases and reaches
the reset value of 37, the contacts of the latter open
and de-energize relay 3.
(2) If the load does not increase at any time
during the time setting of relay 3, its contact opens
and interrupts the coil circuit of contactor 4- Should
the load increase before 3 opens, relay ^7 again
closes and re-energizes 3.
350 April, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 4
(3) After 3 has opened, contactor 4 opens, inter-
rupting two circuits simultaneously; one, the alter-
nating-current supply to the controller segment 13,
and the other, the direct-current circuit including
the operating coil of contactor IS.
(4) The holding circuit for contactor 5 through
segment 14 and the auxiliary switch on the oil
circuit breaker is broken.
(5) Line contactor IS and contactor -5 open.
(6) The opening of contactor -5 interrupts the
supply to segment 1 on the controller and establishes
a circuit through its electrical interlock to segment 19.
Fig. 16. Typical D-C. Line and Resistor Contactor
(7) Contactors 16 and 14 open, disconnecting
the converter from the transformer and discharging
its field. The operating coil of contactor 6 is ener-
gized through the electrical interlock on contactor
o and segments 19 and IS. The controller motor
starts, 30 opens and contactors 19, 20, and il open.
(8) Segment 24 makes con-
tact, energizing the trip circuit
of the oil switch mechanism: also
segment 25 makes contact through
the lower limit switch on the
brush-raising device.
(9) The high-tension line is
disconnected from the trans-
former, de-energizing relay S2,
which opens. The brushes are
raised from the commutator.
(10^ Segments IS and 19
break contact and controller
comes to rest at the off position.
In the meantime, the motor of
the brush-raising device continues
to operate imtil reaching the end
of its travel when the lower limit
switch is opened, breaking the
supply to the motor.
(11) As the voltage on the
converter armature dies down,
after contactors 14 and 16 are
open, relay 36 opens.
Where individual load-
limiting feeder protection is
orovided, usually onlv two
steps of machine resistance are used, while
each feeder is protected by one section of
resistance. Thus overload relay 23, contactor
19 and the corresponding resistor are trans-
ferred to one feeder and similar equipment is
applied to each of the other feeders. In addi-
tion it is customarA', in important substations,
to provide isolating feeder contactors which
entirely disconnect the feeder in case a con-
tinued overload is sufficient to heat the feeder
resistor above a safe operating temperature.
PROTECTIVE FEATURES
Direct-current Overload
Relays 23, 24. and 25 are calibrated at
successively higher overloads. They re-
ceive their actuating current from a shunt
in the machine circuit. The contacts of each
relay control a corresponding contactor, 19.
20. or 21. These, when closed, short circuit
sections of the load-limiting resistor. When
they open on overloads they insert respec-
tively the several sections of the resistor and
limit the output of the machine.
Alternating-current Overload
In case of trouble on the alternating-cur-
rent converter, or in the transformer, pro-
tection is afforded by relay 2S, Fig. 12. This
is an inverse-time-limit device with a definite
minimvim setting and is energized by the
high-tension current transformer. The cur-
rent setting is well above the corresponding
Fig. 17. Synchronous Converter Equipped for Automatic Control
MODERN DEVICES AND CONTROL FOR AUTOMATIC RWY. SUBSTATIONS 351
current settings of relays 33, 24, and 25.
When relay 28 operates, it trips the oil circuit
breaker and with it the hand reset switch.
The opening of the hand reset switch inter-
rupts the coil circuit of contactor 4- Simul-
taneously an auxiliary contact on the oil
circuit breaker breaks the holding circuit
of contactor 3. The machine is thus dis-
connected first from the high-tension alter-
nating-current side and then from the direct-
current and low-tension alternating-current
sides.
After the oil circuit breaker has been
tripped in the above manner and the hand
reset switch opened, the station will not
start up again until the hand reset switch is
closed by the inspector. Consequently,
relay 28 is set very high and is expected to
operate only in cases of severe trouble where
the attention of an inspector would be neces-
sary in any event.
Low Voltage
Relay 27, Fig. 13, provides the alternating-
current low-voltage protection. When low
voltage occurs, the left-hand contacts of
27 are closed which short-circuit the coil of
relay 27-X, opening it and interrupting the
supply through the contacts of relay 3 to the
coil of contactor 4. Relay 29, in a certain
sense, performs the function of an alternating-
current low-voltage relay whenever the con-
\-erter is running, since, should the alternat-
ing-current voltage fall too much, the con-
verter would invert and supply power from
the trolley to the alternating-current system.
Relay 29 would then open, interrupt the
holding circuit of contactor 4 and shut down
the machine.
Overspeed
The syjeed-Hmit device, 12, on overspeed
closes the circuit of the shunt trip of the
direct-current circuit breaker. When this
opens, the auxiliary switch on the circuit
breaker interrupts the supply to the coil of
contactor 4 and the equijjment shuts down.
An additional safeguard is also provided by
an additional set of contacts on the speed-
limit device which, on overspeed, opens the
coil circuit of contactor 4 also.
Underspeed
The speed control switch iS is a centrifugal
device, the contacts of which remain open
until approximate synchronism is reached.
Sequence
The sequence of events is primarily fixed
bv the controller. However, in addition to
this, there are electrical interlocks on con-
tactors 10 and 16 as well as the holding
circuit of contactor 5, all of which are addi-
tional safeguards against incorrect sequence.
Furthermore, contactors 10, 16, I4, and 31
are mechanically interlocked.
Polarity
The 250-volt exciting generator direct
connected to the motor of the controller
fixes the polarity of the converter, but, as
an additional precaution, the polarized
relay 36 must be energized in the proper
direction before allowing the line contactor
18 to close.
Temperature
Should either of the machine bearings over-
heat, one or the other of the temperature
relays 38 will open, de-energizing contactor
4 and shutting down the machine. The
relays are hand-reset devices and hence after
functioning require the presence of the
inspector.
When the load-limiting resistor overheats
due to overload peaks, or the machine reaches
its maximum heating due to cumulative
overloads, one or more of the temperature
relays 33 opens. The operation of any one of
these relays de-energizes relay 27 which
latter then closes the left-hand contact.
This action short circuits the coil of relay
27-X and opens the coil circuit of 4- The
machine then shuts down until the tem-
perature of the resistor or machine lowers to
a safe operating value, after which it will
again be ready to start on load demand.
Balanced Polyphase Voltage
This protection is provided on the low-
tension side of the power transformers by
means of the two relays 32 which are con-
nected across different phases. All three
phases of the power transformer must be
excited to approximately normal voltage;
otherwise one or both of the relays 32 will
remain open and prevent the starting con-
tactor 10 from closing.
Position of Converter Brushes
Proper position of these Inrushes is assured
by means of the auxiliary switches on the
brush-raising device.
Reverse Current
Should the machine start to invert, due either
to the lowering or interruption of the alter-
nating-current supply voltage, the reverse
current relay 29 will open and de-energize
contactor 4 shutting down the machine.
The reverse current relav also acts as an
352 April, 1020
GENERAL ELECTRIC REVIEW
VoL XXIir, No. 4
Fig 18. Automatic Substation of the Salt Lake. Garfield and Western Railway. 600 Kw.. 60 Cycle. 1500 Volts
Fig. 19 Automatic Substation of the Pacific Electric Railway, 1000 Kw . SO Cycle
MODERN DEVICES AND CONTROL FOR AUTOMATIC RWY. SUBSTATIONS 353
auxiliar\- to the overspeed device. It is set
to open at the direct-current running-light
current of the converter. Consequently
upon the loss of alternating-current potential
the converter begins to motor, but as soon
as it draws an appreciable current from the
trolley, and before it can reach a dangerous
speed, relay 29 operates disconnecting it
from the line.
Interruption of Alternating-current Supply
An interruption of the alternating-current
supply may occur at any time during the
cycle of operation. Three cases will cover all
contingencies.
First, while segment 15 of the controller
is in contact. The failure of supply at this
point leaves the high-tension oil circuit
breaker closed but all other devices (except
1) are de-energized. When the supply is
re-established, and assuming a load demand
exists, the sequence of operations at once
begins where it left off.
Second, after segment 15 has broken con-
tact and before segment 11 has made con-
tact. The failure of supply leaves the oil
circuit breaker closed, but opens all other
devices which happen to be closed at the
time. When the suppty is re-established,
contactor 4 closes but contactor 5 cannot
close because its holding circuit has been
interrupted. Consequently neither the start-
ing contactor 10 nor the running contactor
16 can close. However, a circuit is closed
through the auxiliary contact of 5 to segment
19, thence to segment IS and the coil of con-
tactor 6. This action starts the controller
and runs it to the "off" position and trips
the oil circuit breaker through segment 24.
It is then ready to start the equipment again
if there is a load demand.
Third, while the machine is running and'
delivering load. The failure of supply causes
the reverse-current relay to operate de-ener-
gizing, through contactor 4. all other devices.
The oil circuit breaker is not tripped. Upon
the return of supply the same action takes
place as outlined under the second case above.
Field Current
Relay 30 is in series with the converter
shunt field; its contacts are in series with the
coil of line contactor 18. Failure of field
current at any time thus prevents the attempt
to carry load.
Converter Flashover
Converters which are to be used in auto-
matic substations are equipped with flash
barriers. These greatly reduce the nimiber
of flashovers and materially decrease the
damage due to flashing. Practically perfect
protection can be obtained by the use of
barriers and a high-speed circuit breaker. In
addition to barriers, all new (iO-cycle convert-
ers have high reluctance commutating poles
and a protected type of brush-holder. These
last features ha\-e pro^•en ^•ery successful.
ADJUSTMENTS
The foregoing description gives only the
bare outline of the various steps by which a
converter is started and stopped. During the
time the substation is delivering load, it
must take care of itself under all conditions.
It is not within the scope of this article to go
into the many fine points of operation,
however, the reader will have already gleaned
a general idea of some of the inore important
characteristics. The description of the pro-
tective features brings out many.
One point warrants further mention;
namely, the matter of adjustments. An
otherwise good equipment may give poor
service unless every de\'ice is in good working
condition and properly adjusted. Unsatis-
factory ser\ace, for example, may result from
the improper adjustment of overload relays
23, 24, and 25. These might be set too high
or too low to give the best operating con-
ditions. Another thing; the ohmic values
of the load-limiting resistor steps require the
careful attention of the engineer in order to
best fit in with the current swings during
normal and rush traffic. Needless to say,
the physical condition of the individual
pieces of apparatus requires intelligent and
regular inspection.
Herewith are given some of the more
important adjustments of which a modern
equipment is susceptible. It should be noted
that most, if not all, of them are matters
of trial after installation. This is necessarily
the case since all the operating pecularities
of a given road cannot be anticipated.
The contact-making voltmeter is adjusted
to make firm contact with normal voltage on
the bus. The contacts should open at a
btis voltage such that, after the station cuts
in, the machine will deliver an economical
load. This value varies according to the
service, from 4.50 volts to .350 volts on a
normal 600-volt system.
The primary adjustment on relay 3 is
the time element. This varies with con-
ditions but the relay is ordinarily set between
three and eight minutes.
354 April, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII. Xo. 4
The overspeed device 12 is adjusted to
function at from 10 to 15 per cent overspeed.
The speed control switch 13 is set to close
its contacts at between 95 and 100 per cent
speed, with the necessary allowance made in
variation of transmission line frequency.
Relay's 23, 21^, and 25 are adjustable both
for time and opening value. The current
setting of these is usually about 130, 150, and
170 per cent of the fuJl-load current of the
machine. Their final setting, however, is
entirely dependent upon the load conditions.
The time setting is so adjusted that, when all
Relay 28 is set for about 225 to 250 per
cent load with a time setting of from three to
four seconds.
Relay 29 should open its contacts when the
converter is running light from the direct-
current side and with the high-tension oil
switch open.
Relays 32 are so adjusted they will not
close when the power transformer is energized
and one or another of the high-tension dis-
connecting switches are open. They, of
course, should close immediately all three high-
tension disconnecting switches are closed.
Fig. 20. Automatic Subsiiiuun uf the Conestoga Traction Company. 500 Kw-. 25 C>'clc
of them have opened on overload, the one
controlling the largest section of load limiting
resistance should come in first in about five
to six seconds, the next one, seven to eight
seconds, and the last one, controlling the
smallest section of resistance, in nine to ten
seconds.
Relay 27 is set just below the minimum
point of the high-tension line voltage
variation. Ordinarily, if this relay is ad-
justed to close its left-hand contacts at
about 15 per cent reduction in transmis-
sion line voltage, it will be satisfactory-.
In this connection, care is taken to see
that the 22()-volt alternating-current control
bus is as close to normal voltage as possible,
changing the taps on the control transformer,
if necessa^\^
The bearing thermostats have a fixed
setting at about 100 deg. C. The ther-
mostats mounted over the load-limiting
resistance are adjustable over a wide range
and the setting depends entirely upon the
method of mounting and distance from the
grids. It is usually set by trial, by opening
relays 23, 24. and 23 and allowing all the
current to feed through the load-limiting
resistor. This condition is maintained until
the resistors reach a temperature well abo\e
the boiling point of water and below a dull
red heat.
The setting of underload relay 37 is de-
pendent entirely upon service conditions anil
should be adjusted by trial so the machine
will be shut down when delivering an uji-
economical load.
N. E. L. A.
TWO DOLLARS PER YEAR
TWENTY CENTS PER COPY
GENERAL ELECTRIC
REVIEW
VOL. XXIII, No. 5
Published by
General Electric Company's Publication Bureau,
Schenectady. N. Y.
MAY, 1920
^
CONVENTION NUMBER
For
Fractional H. P. Motors
WHERE maximum productiveness
coupled with maximum serviceabil-
ity is the dominant idea back of a machine,
freedom from breakdown interruptions
must be an in-built quality. Hundreds
of thousands of high-speed, high-duty
electrical machines are fortified against
shaft-in-bearing friction dangers by the
use of "tiZBIM" Bearings, with their tre-
mendous factor of safety against the de-
structive effects of high-speed operation
over long periods.
See thai your Motors
are "NORmfl" Equipped
THE M^MM/^ C^MIF/^MY
/nlmaM© /3\''®imM(g
femj Ig!laifii(i CK^
Ball, Rol lei". Thrust and Combination Bearinqs
General Electric Review
A MONTHLY MAGAZINE FOR ENGINEERS
Associate Editors, B. M. EOFF and E. C. SANDERS
Manager. M. P. RICE Ed.tor. JOHN R. HEWETT In Charge of Advertising, B. M. EOFF
Subscription Rates: United States and Mexico. $2.00 per year; Canada, $2.25 per year; Foreign. $2.50 per year; payable in
advance. Library and Student Rates: United States and Mexico, $1.50 per year; Canada. $1.75 per year; Foreign. $2.00 per year;
payable in advance.
Remit by post-office or express money orders, bank checks, or drafts, made payable to the General Electric Review, Sche-
nectady, N. Y.
Entered as second-class matter, March 26. 1912. at the post office at Schenectady, N. Y.. under the Act of March, 1879.
Vol. XXIII, N.,. .-. ,yCe,:iTE&^cLpa,.y MaY, 1920
CONTENTS Page
Frontispiece: White Way Lighting on Broadway, Los Angeles 3ol)
Editorial :
From LTnccrtainty to Unprecedented Activity 357
By J. R. LovFjoY
Electricity at the Pasadena Convention 358
By R. H. Ballard
Proposed Changes in Conducting N. E. L. A 359
By Thomas Addison
An Alternative for Outdoor Generators 360
By Henry G. Reist
Intensive Street Lighting 362
By. W. D'Arcy Ryan
Fundamental Principles of Polarity, Phase Rotation, and Voltage Diagrams of Transformers 374
By A. BoYAjiAN
Relative Merits of Connections Employed in High Voltage Generating Stations . . 386
By Ernest Pragst
Sixty-cycle Converting Apparatus 392
By J. L. Buknham
Design of a Super Power Station 399
By. H. Goodwin, Jr., and A. R. S.mith
Some Corona Tests 419
By W. W. Lewis
The Alternating Current Network Protector 427
By H. C. Stewart
Alternating Current Lightning Arresters 429
By V. E. Goodwin
Metallic Resistor Electric Furnaces for Heat Treating Operations 433
By E. F. Collins
A New Type of Arc-welding Generator 442
By S. R. Bergman
A New Ty]je of Gathering Locomotive ^ 446
By John Liston
Self-interest Will Solve the Problems Confronting Electrical De^■elopmen1 . . .451
By A. Emory Wishon
The Mariner: The First Electrically Operated Trawler ........ 455
B>' John Liston
Question and Answer Section 4()4
=4
— 1.
-^.
_^
FROM UNCERTAINTY TO UNPRECEDENTED ACTIVITY
I am delighted to have the opportunity
afforded by this issue of Gexeral Electric
Review of expressing to the members of the
National Electric Light Association in con-
vention at Pasadena, Cal., the deep apprecia-
tion of the General Electric Company for
their considerate and patient attitude toward
us during the trying period since the closing
months of 1918 and throughout the greater
part of 1919,
During this period the manufacturer was
confronted with stoppage of work and heavy
cancellation of orders for apparatus and
material, particularly on account of Govern-
ment contracts, involving accumulations of
raw materials sei-viceable primarily for the
completion of such contracts. The sudden
cessation of work early in 1919 imposed upon
the manufacturer serious problems of redis-
tribution of labor to a more normal con-
dition. Costs of raw materials fluctuated,
credits were strained to the limits of safety,
and business conditions generally were warped
and distorted. It was difficult to deal with
these conditions from day to day and impos-
sible to forecast the future with any degree of
certainty. It was not a period for initiative
and progress; the manufacturer could not
venture to provide additional facilities or
accumulate stocks of materials; in fact, they
"marked time" awaiting a clearer vision of
of the immediate future.
It is most astonishing how rajjidly industry
recovered from the uncertainty. In the early
spring of 1919 there were indications of
improvement and there arose within four or
five months thereafter a demand for electric
apparatus ''and devices that grew to such
proportions in the succeeding months of the
year that all doubts which existed a few
months before were dispelled. This cur\-e
of demand continued uinvards with such
rapidity that the existing facilities were
taxed to their maximum, and in the closing
months of the year 1919, continuing through
January and February of 192U, the electrical
industry, and presumably many others,
were not in a position to accept additional
business assuring reasonably prompt or cer-
tain deliveries. Such is the condition toda>'.
In less than twelve months there was a
complete industrial cycle from uncertainty
to unprecendented activity with possibly
some misgivings as to the future.
Additional facilities are now being pro-
vided as rapidly as possible, regardless of
excessive cost of construction and equipment;
unfortunately, such facilities are not im-
mediatelv available and will not be pro-
ductive for several months. In the mean-
time every effort is being put forth to
stimulate production.
The business of 1919 as it gathered force
disclosed rather unusual features. There
was less than a normal demand for large
generating and distributing apparatus; there
was an increasing demand for small apparatus,
such as induction motors, especially from
industrial plants, stimulated no doubt by the
demand for the product of such industrial
concerns. Tremendous activity developed for
small or fractional horse-power motors and
electric ranges and heating devices for domes-
tic and industrial uses. The aggregate rating
of such motors contracted for in 1919 from all
sources is estimated to exceed a million horse
power and the rated cajjacity of electric heat-
ing units in excess of 1,2()(J,()()0 kilowatts.
Contrary to the general impression there
was an increased and well sustained demand
for car equipments for electric railways. The
electrification of steam roads and terminals
did not progress for obvious reasons. Recently,
however, there has been a ouickening of
thoiight and action on the part of the manage-
ment of these properties as the transportation
systems are being returned to their original
owners and operators.
Throughout the year the central station
management likewise had their troubles in
358 May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. r>
their endeavor to meet the demands for power
in excess of existing facilities, and to pro-
vide, within the accepted Hmits of sound
finance and earning power, funds necessary
for improvements and expansion. The future
possibilities, however, of the growth of the
central station industry are, in my judg-
ment, most hopeful, and the field and scope
of operation are expanding and are most
promising. It would seem to be a new era of
development for the central station industry,
and I firmlv believe that sound financint: will
be forthcoming to support the efforts of
executives and managers of established public
utilities to supply the demand of the public
for light and power. L'nwise legislation and
decisions may retard, but cannot prevent the
ultimate development of the central station
industry-, which is founded on such sound
economic principles and conducted in a
manner best conser\-ing the interest of the
public.
J. R. LOVEJOY,
Vice-President, General Electric Company.
ELECTRICITY AT THE PASADENA CONVENTION
In the economic transition that has taken
place, electricity emerges the dominant
factor in the world's work and as such is
confronted with problems and undertakings
of a scope and urgency beyond all parallel.
The annual convention of the National
Electric Light Association to be held at
Pasadena, California, May IS to 22, comes
at a time when these considerations are of
paramount importance to the future of the
industry itself and to the life and activities
of the nation at large This convention, repre-
senting as it does an invested capital of some
$3,(J00,0()(),()()(), will bring into concerted
action the great construction forces of the
industrs- to deliberate and devise the ])hysical
and practical side of these propositions, and
will attract likewise representation on the
part of the financial interests who will desire
to appraise and compute them as bases of
the imm.ense funding requirements that will
be needed to consummate.
The tremendous demands that are thrust
upon the electrical industry- as a result of
world-wide depletion of other resources are
so immediate and ])ressing as to give them
added pro])ortions. The war and its ravages
Jiave served to show how j^recarious is the
supply of natural fuel deposits, such as coal,
petroleum and timber, unless carefully con-
served. Electricity, provided by Nature as
her own means of conservation, is now called
upon to fulfill this greater destiny and must
be made forthwith to meet and perfomi the
many economic needs that await its universal
application. Money, men and j^roduction
are the requi.sites — and above all action, for
time is the essence.
Money in enormous quantities will be
required to develop water sources during the
next ten vears and turn them into h\-dro-
electric energy which will become instantly
active in the production of foodstuffs which
the world craves, and in turning raw materials
into finished products. The orderly financing
of this new electrical era in a way that will not
be inimical to the security of capital honestly
invested in other processes over which
electricity has taken economic precedence,
is of vital interest to the bankers of America.
The Pasadena Convention is a timely occasion
for acquiring first-hand information.
Power generating companies are being
called upon from the Atlantic to the Pacific
to increase the output of their product
against the advancing price of coal, oil and
other portable fuel. The railway lines which
cobweb the continent and perform the
gigantic work of transportation are probably
the greatest consumers of the oil and coal
in Earth's storehouse. It is apparent that
the electrical industr\- is faced with under-
taking the electrification of some of the steam
railroads and the increasing of their facilities
for generating and carrying this new and
ponderous load. Upon the manufacturers
devolve the necessity of turning out electrical
machiner>- and api)aratus which will facilitate
steam and hydro-electric generation and
conserve the current required for railroads,
factories, lighting and domestic uses.
To make ever\' home in America an
electrical one and to see that apiiliances
designed to do the work of c.ial and oil are
of the highest tyi)e and perfectly installed
will be imjxirtant matters for jobbers, dealers
and contractors — but all of these themes must
be in attune with the one vibrant chord. Serv-
ice. Service with profit; Ser\ice that will
make electricity sui)crlative; Ser\-ice first, last
and always; the Service that ser\-es.
R H. B.\I.L.\RD.
President. N. K. 1.. .\.
EDITORIAL
^59
PROPOSED CHANGES IN CONDUCTING N.E.L.A.
For some years past the Annual Convention
of the National Electric Light Association
has seemed to many to have lost interest,
because of its thousands of delegates, its
numerous meetings, and the extensive and
elaborate programs that were attempted to
be carried out. In other words, the con-
ventions have seemed unwieldly, and to some
extent uninteresting, because of their size and
extent.
Also, during the past, it has seemed to
many of us in the manufacturing end that the
representatives of the manufacturing interests
were not given quite that representation in the
organization, or that interest in its delibera-
tions, to which they were entitled — this,
notwithstanding it has been quite generally
recognized that the manufacturing and the
operating interests were distinctly in partner-
ship in promoting the business as a whole.
The manufacturing interests, with their great
research laboratories, have done much to
advance the state :f the art, perfect ma-
chinery, bring out new devices, etc., in thus
indirectly promoting the more extensive use
of electric current in every branch of our
national life.
It goes without saying that the manu-
facturing interests need the operating interests
as an outlet for their product — but to the
same extent the operating interests need the
manufacturing interests in order to furnish
the machinery, devices, new appliances, etc..
that are necessary to a proper expansion of the
business as a whole. In other words, there is
a distinct partnership between the two
interests, each needing the other, each vitally
and equally interested in the welfare of the
other in extending the business in every
direction.
Fortunately, our president, Mr. R. H.
Ballard, has not only recognized this situation,
but has been able to give the interests of the
Association his undivided attention during
the year of his presidency, and he has sought
to correct matters in the following manner:
First: — In the matter of decentralization
of the national organization, and the building
up of the local associations.
Second: — In adding representatives of the
manufacturing interests to at least one of the
imjiortant committees.
This decentralization of the National and
the building up of the Local Associations will,
it seems to me, greatly extend the influence
of the Association as a whole. It will revive
the waning interest of many important
men connected with the business, and will
bring into the local associations a large
number of new men who would not have
taken any great interest in the national
association. Further, the more frequent
meetings of the local associations will help
to keep alive the interest of a larger number of
important men, and in every way extend the
Association's influence.
Representation of the manufacturing
interests also seems a step in the right
direction and will, in the minds of many,
still further extend the influence of the
association.
Both steps seem to me very necessary and
promising for the future influence of the
National Association. It is hoped and
expected that the meeting of the National
Association at Pasadena in May will ratify
and approve both of these measures, thus
bringing about a much desired and much
needed reform.
Thomas Addison.
Mgr. Pacific Coast District,
General Electric Company
360 -May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. .5
An Alternative for Outdoor Generators
By Henry G. Reist
Engineer, Alterx-\ting-current Engineering Dep.\rtment, Gener.\l Electric Comp.\ny
In an article on Outdoor Generators, by Mr. H. W. Buck, in our March, 1920, issue, credit was given to Mr.
H. G. Reist for having first proposed the construction described therein. Mr. Reist recognizes that, while there
are a great many advantages in such construction from an economical standpoint, there are also some dis-
advantages and these are pointed out in the present article. The principal difficulty is to construct the gener-
ator so that windings will be kept dry, and at the same time be properly ventilated. Condensation on the coils
under certain atmospheric conditions is also a common occurrence and steps must be taken to prevent it. In
cold weather lubricating oils become sluggish and are liable to cause trouble unless the bearings are carefully
watched. To overcome these difficulties Mr. Reist suggests as an alternative that standard generators be used
and housed under inexpensive semi-portable shelters. — Editor.
Many engineers have in the last few years
called attention to the necessity of developing
all available water-powers for generating
electric current to conser\'e our fuel supph\
While such developments are desirable at this
time, it is becoming increasingly difficult to
make them, due to the scarcity of materials
and capital and to the shortage and high cost
of labor.
Dr. Steinmetz has repeatedly proposed
small power plants at intervals along streams,
to minimize the hydraulic development.
Xo doubt in many places, particularly in
rolling land, great saving may be made b\-
utilizing small dams of relatively low head,
rather than by building more massive dams
which give a larger amount of power at one
place. This would probably be the case in a
farming country, where with high dams much
valuable land would have to be flooded.
With the development of automatic stations
the gathering of power from distributed power
houses is much more economical than was the
case when attendants had to be provided
at each installation. Such distributed power
installations can utilize either synchronous
or induction generators. The induction
generator was proposed on account of its
simplicity of operation, but with this type
of generator magnetizing current has to be
supplied from another source, which is not
always convenient.
Methods of operating synchronous plants
without attendants have been worked out and
applied with entire satisfaction; therefore it is
probable that synchronous generators will
generally be used for small plants as well as
for large ones.
One of the methods of saving expense in the
construction of power plant installations is in
the omission of the power-house, utilizing
outdoor generators in the same manner as
transformers and switches are used in outdoor
sub-stations. There is no doubt that we
have spent too much money for houses to
roof over watem"heel-dri\-en generators. We
should, however, not deceive ourselves and
assume that power stations are built wholly
to protect the generators. They also generally
contain transformers, switchboards, busbars,
switches, exciters, repair shops, offices and
other conveniences. Since we have learned
to put many of these things out of doors,
especially transformers and high potential
switches and busbars, which take a good deal
of room, no doubt power-houses can be built
very much smaller along the lines of present
construction. These buildings will still need
to have considerable height to give head-room
for erecting and dismantling the generating
unit, and for the traveling crane. The walls
must also be heavy to carry the load put on
the crane.
Outdoor generators ha\e been ])roposed
frequently, and I believe a few machines have
been installed in this way. An article on the
subject by Mr. H. W. Buck was published in
the Gener.\l Electric Review of March.
1020, in which the advantage of such con-
struction is clearly pointed out. That both
sides may lie presented I wish to call attention
to a few of the difficulties encountered in such
construction. We have always assumed and
believed it to be desirable to keep the wind-
ings of electric machines drj-. It is difficult to
maintain this condition and get proper venti-
lation in a generator exposed to the weather,
without S])ecial construction and additional
expense. If the generator is allowed to
become cold when not in use there is likely
to be condensation of moisture on coils under
certain atmospheric conditions. To i^revent
this, many pieces of electric ai)paratus are
provided with special heaters of some sort
to keep the machine warm when it is idle.
In case machines are exposed in ver\- cold
weather there is danger of difficulty from the
oil becoming too thick. This danger would
apply especially to self-oiling thrust Ix-arings
l)laced at the top of the machines, since these
AX ALTERNATIVE FOR OUTDOOR GENERATORS
361
niik'ht heat at starting before the oil became
sufficiently thin to circulate freely.
Automatic stations, which operate without
an attendant, should be inspected from time
to time by a ]mtrol. In case of very cold
or disagreeable weather it is ijrobablc that
such inspection would be superficial.
On synchronous machines placed out of
doors the installation of exciter units presents
difficulties. In small machines exciters are
frequently belted and on larger machines they
are either direct connected or driven by
motors. Whate\-er arrangement is used con-
siderable inconvenience will be caused by lack
of a protecting roof. In some installations
these parts, together with the governing
mechanism, can be placed in chambers in the
masonry below the generator floor; but in
other cases, due to danger from high water,
this could not be done, and with small low
head machines such masonry construction
would add considerably to the expense.
I would suggest that instead of building
special weather-proof out-of-door generators,
standard generators be used and that simple
inexpensive semi-portable shelters be erected
over them. In the case of several machines in
one installation it will be desirable to have a
low house extending over the line of machines,
with a gantry crane bridging the house — the
house, or at least its roof, to be made in
sections which can be moved on a track
parallel with the line of generators to telescope
with the roof over the adjacent generator, or
a unit roof over each generator may be lifted
bodily from its place by means of a crane.
The crane can, to advantage, be enclosed, as
suggested by Mr. Buck in his article referred
to. There would be sufficient heat generated
to keep such a power-house comfortable, and
by suitably arranging the ventilators, warm
air could be passed over any generator that
was not in use, thus overcoming the difficulties
which might be experienced with the gener-
ators out of doors.
The simple house suggested would provide
a shelter for the exciters, watcrwheel govern-
ors, and the patrol on his visits, and would
seem to offer all the advantages of a more
expensive power-house.
It is probable that a protecting house, as
described, could be constructed for little more
than the extra cost of a generator exposed to
the weather.
I have tried to show how a considerable
amount of the ordinary investment in hydrau-
lic power-house could be saved without losing
any operating advantages. We would save
money which is sometimes lavishly put into
monumental power-houses.
^^{X>;5(»^:-:i> :. .
The Outdoor Generating Station, for which an alternative is proposed
362 May, 1<»20
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 5
Intensive Street Lighting
By W. D'Arcy Ryan
Director, Illuminating Engineering Laboratory, General Electric Company
In 1905 the first ornamental street lighting system was installed on Broadway, Los Angeles. This was
the cluster ball globe standard and has been generally copied throughout the country. In 1911 the first real
"White Way" was lighted in New Haven, Conn. A single-light ornamental luminous arc standard was used.
Similar systems have been installed in many cities. Now comes a new epoch in street lighting, the "Inten-
sive White Way," which has had its inauguration on the Pacific Coast and is rapidly spreading eastward.
The author points out the architectural, engineering, commercial, and protective advantages of such a sys-
tem and gives cost data on the various installations that have been made. — Editor.
(2) The minimized window reflections on
account of the height of the lamps.
(3) The intensity of the illumination and the
Intensive street lighting, which had its
inception on the Pacific Coast during the war,
is here to stay ; and this latest development in
street lighting is rapidly moving from the
Pacific to the Atlantic.
The first installation was made in San
Francisco on Market Street in 1916 and is
known as the "Path of Gold." Notwith-
standing that the cost of such an installation
was far in excess of anything heretofore used,
the results were so successful from every point
of view that one year later the system was
extended to include the entire business tri-
angle; so at the present time San Francisco's
main business district is intensively lighted.
A similar system, designed for ^Iain Street.
Salt Lake, was also put into operation about
the same time. In January of this year the
fourth installation of magnitude, viz., Broad-
way, Los Angeles, known as the Radiant Way.
was illuminated with an intensive system:
and for the first time two designs of standards
were used in order to break the monotony of
continuous repeat. This new note in street
lighting arsthetics is worthy of careful study.
The so-called intensive lighting funda-
mentally differs from the ordinary white-way
lighting in the following respects: Greatly
increased illumination; relatively high lamp
standards; initial installation costs ranging
from S4.00 to $8.00 per front foot in place
of approximately $1.00 to S2.00; and mainte-
nance costs proportionately higher.
In a sense, intensive white-way lighting
might be regarded as general floodlighting ; ad-
vantage being taken of the natural decorative
feature of the unit itself as contrasted with
ordinary floodlighting where the light is con-
cealed in a uni-directional floodlight housing.
Intensive street lighting is in a class by
itself and the expense of installation and
operation is borne, for the greater part, by the
merchants and property owners. The
especially prominent features of this type of
street lighting arc as follows:
(1) The cosmopolitan atmosphere and dignified
;i-sthetic effects of the standards by day as well as by
night.
uniformity of distribution on the street and building
facades, with emphasis on the comers in both the
light and the design of the standards.
(4) The readiness with which features of people
in the street can be distinguished, particularly in
automobiles with the tops up. which demonstrates
the undercutting effect of the light.
(5) The brilliancy and sparkle and good
simultaneous contrast of the luminous-arc units with
the window and sign lighting.
(6) The illumination of the building facades and
sharpness of the cornice lines against the sky.
(7) The increased intensity as compared with the
lighting of intersecting streets, clearly marking the
main thoroughfare.
(8) The golden tone of the glassware which gave
San Francisco's Market Street the name of the
"Path of Gold."
This glassware is used onl\- on the Pacific
Coast and is merely a suggestion of the
Golden West. It is intended for the daylight
effect, as this tone of glassware is less insistent.
However, it is slightly detrimental at night as
it naturally tones down the whiteness of the
light and reduces the simultaneous contrast.
In other words, it is a sacrifice made for the
day appearance and is not generally recom-
mended.
Notwithstanding the relatively high initial
cost, the ().()-ampere luminous arc lamp has
been generally selected for these intensive
systems because of its white light, sparkle,
and operating economy. It imparts to the
street life impossible to obtain in anything like
the same degree with a still light and, as
previously stated, has the advantage of
simultaneous contrast with other lights.
On the other hand, we may look for a
wonderful development in Mazda Intensive
Lighting. The latest example in this line is
embodied in the new system now being
installed in Saratoga Sjjrings. N. Y. The
glassware used is entirely unique and very
graceful in design. One of the iirincipal
advantages is the duo-intensity control with-
out the use of additional series wires from the
jjoint of distribution, making it ])ossible. for
exami)lc, to reduce the lighting from 1000 to
INTENSIVE STREET LIGHTING
363
364 Mav, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, Xo. 5
2.50 candle-power at midnight or other pre-
determined time. This system will be care-
fully watched and if it proves as successful
as anticipated, it is bound to become a ver\'
important factor in street lighting, particu-
larly as it furnishes an intensive system during
the early evening and a moderate amount of
light throughout the balance of the night,
with all lamps burning. This method has
many advantages over the present systems
of turning out more or less of the lights
completely.
There are a number of other points which
might be mentioned in connection with
Intensive Lighting, but they will not be
enumerated, except to say that systems of
this kind are not only a benefit to the mer-
chants and property owners in stabilizing real
estate values and increasing window shop-
ping, but they are of material assistance to
the police and fire departments, a benefit to
the general public, and are of considerable
advertising value to the city.
The following statistics (by A. F. Dickerson)
on some of the recent intensive White Ways
which have been installed in accordance with
plans issued by the Laboratory arc given,
together with photographs of the actual in-
stallations. It should be borne in mind that
these figures represent the costs at the time
of installation. In many cases operating costs
have increased above those enumerated.
Main Street: Salt Lake City
This system consists of 70 standards, each
carrying three (i.O-amp. General Electric
ornamental luminous arc lamps. The stand-
ard envelopes the trolley pole; spacings are
about 100 ft. and the overall height 29 ft.
The svstem was first lighted Si-ptembiT '.]().
1916. '
Total cost .'S28,2::().4()
City's share 2,()8.').91
Property owners' share 2."),.'i;54.l?!l
Taxable property, linear feet (i,;f72.(U)
Total cost per foot 4.4.S
Property owners' cost per foot 4.(11
City's cost per foot (1.42
Operating cost for three years 2;).:{.'{4,().")
Property owners' share for three years. 2."i, 174.04
City's share for three years 4,1(J0.()I
Operating cost per foot for three years. 4. {11
Property owners' share for three years. '.i.iK^
City's share for three years (I.(i.")4
Operating cost per foot front per year. . I. ,14
Property owners' share 1 .:i2
City's share (1.22
Fig- 2. Lightinft Standnril. State Street and
Broadway, Salt Lake City. Utah. 6.6.amp
Luminous Arc Lamps
INTENSIVE STREET LIGHTING
365
The installation on Main Street, Salt Lake
City, was made by the Utah Light & Power
Co., and was paid for as above recorded. The
system was put in under the State Street
Lighting Improvement Act, so that all front-
age on the street was assessed yearly for its
maintenance.
State Street and Broadway: Salt Lake City
The new street lighting on State Street and
Broadway will be virtually an extension of the
Main Street system. There will be .504 Gen-
eral Electric 6.6-amp. ornamental luminous
arc lamps used. The standard will envelope
the present trolley pole and carry two lamps
below the trolle>' wire and one above. The
total cost of this new installation will be ap-
proximately $140,000. The Utah Power and
Light Company will pay for the substation
equipment and the feeders to the street cir-
cuits. With the exception of a small amount
contributed by the city, the remainder of the
installation expense and the yearly mainte-
nance will be borne by the property owners.
Market Street: San Francisco
This system extends from the Ferry to
Seventh Street. There are 137 standards each
carrving three General Electric G.G-amp.
ornamental luminous arc lamps equipped with
eight-panel globes. The standards are spaced
approximately 110 ft. apart and opposite, and
are 32 ft. overall in height. The total instal-
lation cost approximately $100,000 and was
paid for by the Pacific Gas «S: Electric Co
which owns the entire system, with the
exception of the trolley pole part of the
standard which is the property of the United
Railroads. The P. G. & E. Co. entered into
a three-year contract with the Downtown
Business Men's Association and a yearly
contract with the city of San Francisco. The
total operating cost per year is $34,753.48.
Of this amount the city pays $13,2.51.33,
which is the maintenance of the center all-
night lamp. The Downtown Association and
the United Railroads pay $14,92(5.1.5 and
$6, .576 respectively for the two lamps on each
standard which are extinguished at midnight.
The amount paid by the United Railroads is
in accordance with their original franchise
agreement. The money is obtained by the
Downtown Association from voluntary sub-
scriptions. To take care of those who will not
contribute, they are asking $2.00 per front
foot from both property owner and tenant.
On this basis, although they were only collect-
ing for three vears, thev were able to obtain
enough money to operate the lights for five
years. The cost of the all-night lamp is
$96,725 per year, and the midnight lamp
$78,475 per vear. The svstem was first
lighted October 4, 1916.
Triangle Lighting: San Francisco
The Triangle District includes all the
streets bounded by Market, Powell, Sutter,
and Kearny. The installation cost approxi-
mately $85,000 and is the property of the
Pacific Gas & Electric Company. Two
General Electric ornamental luminous arc
lamps similar to those on Market Street are
used on each standard. The arrangement of
standards is staggered, with approximately
one standard to each 55 ft. of street. The
height of the standard is 25 ft. One hundred
and ten lamps, costing $116.80 per year, or a
total of $12,848, are burned all night and are
paid for by the city. One hundred and sixty-
eight lamps, costing $102.20 per year, or a
total of $17,169.60, bum until midnight and
are paid for by the Downtown Association
under a five-year contract. All the trolley
poles in this district have been removed; the
trolley wires having been fastened to the
building facades. The Downtown Association
are collecting $1.25 per front foot from both
the property owner and the tenant and have
sufficient funds to carry them beyond the
five-year contract. The system was first
lighted about Januan.- 1, 1919.
Broadway: Los Angeles
There are 134 two-light ornamental 6.6-
amp. luminous arc standards in this instal-
lation. Sixty-seven lamps burn all night and
two hundred and one are extinguished at
midnight. The standards are spaced on an
average of 106 ft. apart and opposite and are
27 ft. high. This system was installed under
the State's "Street Lighting Improvement Act"
and is being paid for by assessing the prop-
erty owners, some of whom are paying the
installation assessment on a ten-year bond
plan. The total installation was approxi-
mately $85,000, or about $6.50 per front foot.
The annual operating cost is $13,700, or about
$1.00 per front foot. The Bureau of Electric-
ity of the city of Los Angeles is supplying
the power and maintaining the system. The
average rate per lamp is $50 per year. The
system was first lighted January 17, 1920.
New Orleans
In New Orleans the electric company has
entered into a ten-vcar contract with the
366 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. .-.
INTENSIVE STREET LIGHTING
3G7
e
a
o
368 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. o
.^^^^^^
kJ
0
* A'
•
. ^ -^ #ir.: ..V.
#
'" -****
"r.'^r-. . . ... nnH
! ;d
*^1
I
^■^^^
!♦♦ -jt* <^ ^' *^^^^te.J^^-'\, i^»'^
«* ■ ^ •* \ ^k
».»: ^
^Bf _ w^' 7*-^ ^,
i-^ «
Fig. 5. Carnival on Market Street. San Francisco. Cal., Inaugurating "Path of Gold" Lighting
Fig. 6. Grant Avenue. Looking South. San Francisco. Cal.. Intensively Lighted by 6.6amp. Ornamental Luminous Arc Lamps
INTENSIVE STREET LIGHTING
36fl
37(1 Mav. 10211
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. o
Fig. 8. Day View of Broadway, Los Angeles, Cal.. showing 6.6-amp. Ornamental Luminous Arc Lamps.
The lamp units on the trolley suspensions were used for temporary lighting during the
installation of the arc lamps. A night view is shown in the Frontispiece
Fig. 9.
State Street, Looking South. Chicago. 111.. Intensively Lighted by Novalux Stippled Glass Globe Umtt
Containing lOOOwatt Multiple Mazda C Lamps
INTENSIVE STREET LIGHTING
371
Fig. 10a. Duoflux Lighting Standard
Installed on Broadway, Saratoga
Fig. 10c. Sectional Drawing
Duoflux Lighting Unit
Fig. 10b. Novalux Lighting Standard
Installed in Congress Park, Saratoga
372 May, 102(i
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 5
Fig. II. Sketch showing Lighting Effect on Randolph Strcrt. Chicago III . by Ornamental Lv.m>nou, Arc Lamp.
INTENSIVE STREET LIGHTING
373
cit_\' for new street lij^hting; the city to own
the system at the expiration of the contract.
During the past few years over a half million
dollars has been spent under this agreement,
which includes over 1400 standards of boule-
vard incandescent lighting, 450 two-light
incandescent standards on the crosstown
business streets, and 3300 jjendcnt luminous
arc lamps. The maintenance of all these
lights is being paid for by the city. Plans
were completed for a very elaborate instal-
lation of luminous arc lamps, five to the
standard, for Canal Street, but the installation
was held up by the war.
Broadway: Saratoga Springs
Construction is now under way and the
system should be lighted June 1 , 1920. Nearly
a mile of street will be lighted by (iO standards.
Each standard has two General Electric
Duoflux units and each unit contains one
1000-c-p. and one 250-c-p. series Mazda
lamp. The Duoflux is an innovation that
will soon be widely advertised. Besides being
of a new and exceptionally pleasing design,
this fixture possesses a distinctive utilitarian
feature. The large lamp in each globe is
extinguished at midnight and the smaller
one is lighted. This arrangement will per-
mit the use of reduced ilkmiination after
midnight, without a duplication of lighting
circuits. The Saratoga installation will cost
about $32,000 and will be installed and owned
bv the Adirondack Electric Power Corpora-
tion. The city will pay the entire mainte-
nance cost of .1; 10,350 yearly.
Randolph Street: Chicago
Proposed plans have been submitted and
approved for the lighting of Randolph Street,
Chicago. The present trolley poles will be
utilized as cores for ornamental enveloping
casings. Each standard will carry two
General Electric 6. (i-amix luminous arc lamps.
It is proposed to extend this system eventually
throughout the Loop District.
South State Street: Chicago
A system has recently been installed con-
sisting of General Electric Novalux fixtures
on trolley pole brackets with 1000-watt
Mazda lamps. Considering the relatively
small installation expense, this system has
been \"erv successful.
Fig. 12. Main Street, Salt Lake City, Utah. Intensively Lighted by Ornamental Luminous Arc Lamps
374 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. .-,
Fundamental Principles of Polarity, Phase Rotation,
and Voltage Diagrams of Transformers
By A. BoYAjiAX
Transformer Engineering Department, General Electric Company
Perplexing problems frequently arise in determining transformer polarity, phase rotation, and angular
displacement, when two or more units are to be arranged for parallel operation. The following article has
been prepared to clear up these difficulties and uncertainties. The author explains the fundamental princi-
ples first as applied to single-phase circuits and then to three-phase circuits. He discusses their bearing on
parallel operation and solves three problems of a practical nature. — Editor.
Polarity and phase rotation are of impor-
tance primarily on account of their bearing
on parallel operation of transformers. It
is desirable, therefore, to treat the subject
in such a way as to make their application
to parallel operation readily intelligible.
S, P, Pi s^
S, P, P2 Si
11 Ti
S, P, P2 Si
♦Jit W It
<
-
_j
Fig. 1
Fig. 2a
Fig. 2b
SINGLE-PHASE CIRCUITS
Fig. I represents a simple single-phase
transformer. It is of interest to consider
the relative directions of windings, currents,
and voltages.
Direction of Winding
It will be obsen'ed that coil S\Si is wound
in the same direction as P1P2 (with respect
to the core), assuming that the first starts
from Si and the second Pi. On the other
hand, SiS; is wound in the opposite direction
to PoPi if we assume that the first starts from
5i and the second from P>. We conclude,
then, that whether two coils are to be con-
sidered as wound in the same direction or in
opposite directions depends on which ter-
minals arc considered as the "start" and
which the "finish." In some simple forms
and combinations of coils, for instance
cylindrical high and low-voltage coils, high
and low-voltage leads at the same end of the
core leg might naturally be taken to corre-
spond to each other; but in more comi^licated
designs, such as interleaved disc windings,
no such "natural" guide would be reliable.
It is good practice, therefore, in comparing
directions of windings, to assume the winding
as starting with the first named terminal and
ending with the second. Thus, in Fig. 1 :
Coils 5i5-. and PiP; are wound in the same
direction.
Coils S\Si and P2P1 are wound in opposite
directions.
Direction of Currents
In a transformer, the load current in the
secondar\- flows in such a direction as to
neutralize the magnetomotive force of the
load current in the primary; and, it is ordi-
narily said, therefore, that primar>- and second-
ar>- currents are opposed to each other. It
would be more accurate to say that primar>"
and secondary ampere-turns are opposed to
each other. Then, if the directions of the
windings are the same, the currents in the
high and low-voltage terminal leads are
opposed; but, if the directions of the windings
are opposed, then the currents are in the
same directions. Figs. 2a and 2b.
In the parallel operation of transformers,
it is only the voltage vector relations that
have a direct bearing, and the current vector
relations need not be considered. Of course
one can always be derived from the other, but
their simultaneous consideration leads to
confusion. Hence, it is advisable to neglect
currents when discussing polarity and phase
rotation.
Direction of Voltages
In speaking of voltages, it at once becomes
necessary to specify whether impressed or
induced voltages are considered. The con-
fusion of the direction of mpressed and
induced voltages j^robably causes more mio-
luiderstanding than any other factor, and
hence, it is essential for clarity to use only
one throughout a discussion. It would be
undesirable to consider the impressed voltage
in the jirimary and induced voltage in the
VOLTAGE DIAGRAMS OF TRANSFORMERS
;i7.')
secondry, since it is not always certain which
is primary and which secondary and, further-
more, polarity and phase rotation are inde-
pendent of which winding is primary and
which secondary. It is the simplest, clearest,
and most logical procedure, therefore, to
consider only the induced voltage relations.
Since the primary and secondary induced
voltages are induced by the same flux, the>'
must be in the same direction in each turn.
Figs. 3a and 3b. However, whether they will
appear in the same or opposite directions
as viewed from the terminals depends on the
relative directions of the windings. Thus,
in Fig. 3a, voltages HiH^ and A'iA'2 have
the same direction, and in Fig. 3b, voltages
H1H2 and A'sA'i have opposite directions.
If we take the order of lettering to indicate
also the direction of voltage, as was assumed
above for the direction of winding, then,
in Fig. 3a,
Voltages H1H2 and A'iA'2 are in the same
direction.
Voltages HiH'i and AVVi are in opposite
directions.
In Fig. 3b,
Voltages HiHo, and A'zA'i are in opposite
directions.
Voltages H1H2 and A'iA'2 are in the same
direction.
\ \
\ \
l>-f'l
2
(
(
(
1
)
)
i — \
\ — \
_1 1_
Fig. 3a
Fig. 3b
Polarity
Since the relative direction of induced volt-
ages, as appearing at the terminals of the
windings, is dependent on the order in which
these terminals are taken, therefore, in order
that "polarity" may have any meaning, it
must be referred to a perfectly definite order
in which the terminals shall be taken. By
common usage, polarity refers to the volt-
age vector relations of transformer leads as
brought outside the case, and both high-
voltage and low-voltage leads being taken
in the same order (from left to right or right
to left) facing the same side of the trans-
former in both cases. Thus, referring to the
tank sketch in Fig. 3a, polarity is the relative
direction of induced voltage from H\ to H2 as
compared with that from A'l to A'2, both
being in the same order (from left to right)
with respect to the tank.
Additive and Subtractive Polarity
When the induced voltages of the high and
low-voltage sides are in opposite directions, as
in tank sketch Fig. 3b, the polarity is said to
be additive; and when the induced voltages are
in the same direction (Fig. 3a), the polarity is
said to be subtractive.
The reason for this nomenclature will be
evident from the following: Referring to the
tank sketch Fig. 3a, if we connect a high-
voltage lead to the adjacent low-voltage lead,
for instance H^ to A'2 and excite the trans-
former on either side, the voltage across the
other leads H\ to A'l will be the difference
of the voltages of the two sides. Following
the voltage from A'l through A'2 to Ho and
then to Hi it is evident that the voltage H2 to
Hi will oppose the voltage A'l to A'2. Hence
the polarity is subtractive.
Referring again to the tank sketch Fig. 3b,
which shows primary and secondary induced
voltages in opposite directions, if we connect
an H lead to the adjacent A' lead, for instance.
Hi to A'2, and excite the transformer, the
voltage across the other leads, i.e.. Hi to A'l,
will be the sum of the primary and secondary
voltages, for reasons explained in the previous
paragraph. Hence the polarity is additive.
Testing of Polarity
The foregoing definition of polarity leads
to two general methods for testing the polarity
of transformers. First: With primary and
secondary in series, one primary lead being
connected to the adjacent secondary lead, the
transformer is excited from an alternating-
current source on either side ; and the voltages
across the high-voltage winding and also
between the free primary and secondary
terminals are measured. If the latter voltage
is found to be less than that across the high-
voltage winding, the polarity is subtractive;
if more, it is additive.
Second: With or without primary and
secondary in series, the transformer is excited
from a direct current source on either side
and a direct-current voltmeter is connected
to the excited side so that a positive deflection
is obtained. The voltmeter leads are then
transferred directly to the adjacent terminals
376 May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. .3
of the other winding without crossing. The
direct-current excitation is then broken and
the inductive kick in the voltmeter observed.
If the needle swings in the same direction as
before the polarity is additive, otherwise
subtractive.
\ — \
"l.Hl »l, ^2
1
», —^ Hz
\ \
\ f
Xz H,H?X,
m
t
RJ
Fig. 4a
Fig 4b
In testing for polarity a fraction of the
rated voltage is sufficient.
Marking of Leads
It would be desirable that lead designations
be indicative of polarity also. This is provided
for by the A.I.E.E. Standardization Rules
in accordance to which high-voltage leads
brought out of a case are to be marked Hy. H->.
etc., and low-voltage leads A'l, X«, etc., the
order being such that "when //i and A'l are
connected together and voltage applied to
the transformer, the voltage between the
highest numbered H lead and the highest
numbered A lead shall be less than the volt-
age of the full high-voltage winding. When
leads are marked in accordance with the
above rules, the polarity of a transformer is:
Subtractive when Hi and A'l are adjacent.
Additive when Hi is diagonally located
with respect to A'l.
To simplify the work of connecting trans-
formers in parallel, it is recommended that the
Hi lead shall be brought out on the right-
hand side of the case, facing the high-voltage
side of the case.
"Transformers ha\-ing leads marked in
accordance with these niles ma\' be ojjerated
in parallel by connecting similarly marked
leads together, provided their ratio, \'ottages,
resistances, and reactances are such as to
permit parallel operation."
The rule can also be stated in this way :
If HiH^ represents the direction of induced
voltage in the high voltage at a given instant,
then A'lA'i must be the direction of the
induced voltage in the low-voltage winding.
Applying this rule to Figs. 3a. 3b, 4a, and 4b,
we find that they are correctly lettered and
that the polarities are also correct as marked.
Relation of Polarity to Potential Stresses
The polarity of a transformer conve^'s no
information as to the arrangement of the
windings or of the internal leads or of the
internal potential stresses. Consider Figs. 4a
and 4b: HiH^ and A'iA'2 are two cylindrical
coils. The coils of Fig. 4a are identical with
those of Fig. 4b and similarly mounted, except
that the positions of the X leads are inter-
changed and therefore their polarities are
reversed. And yet, priman- and secondar\-
are wound in the same direction in both
cases, and the potential stresses are alike.
Polarity, therefore, cannot be taken as indicative
of a higher or lower arrangement of potential
stresses within a transformer.
Standardization of Polarity
Most low-voltage distribution transformers
in use today have additive polarity, while of
power transformers some have additive and
others have subtractive polarity. This situa-
tion has probabh- been very confusing to the
operating companies when attempting to
connect in multiple, or in bank, single-phase
transformers jjroduced by the various manu-
facturing companies. Appreciating the con-
dition, this matter was considered some time
ago by the (general Conference Committee on
Technical Subjects, which committee con-
sisted of representatives from the A.I.E.E.,
X.E.L.A. and E.P.C., which definitely recom-
mended that a uniform polarity be standard-
ized and that this be subtractive polarity.
The reason given for this recommendation is
that although polarity has no bearing on
internal voltage stresses, yet subtractive
polarity has a small advantage over additive
polarity in the matter of the voltage stresses
between external leads as has been explained.
That is, if two adjacent high and low-voltage
leads should accidentally come in contact, the
voltage acrt>ss the other leads would be the
sum of high and low voltages for additi\e
polarity, and their difference for subtracfive
polarity. Furthennore, imder operating con-
ditions with leads insulated from each other,
the potential stress between adjacent high
and low-voltage leads is one half the sum
of the high and low voltages for additive
polarity, and one half their difference for sub-
tractive polarity. This advanta.ge of subtrac-
tive i)olarity, although entirely negligible ordi-
narily, may become appreciable for transform-
VOLTAGE DIAGRAMS OF TRANSFORMERS
377
ers of which both primaries and secondaries
have very hij^'h voltages.
Three-phase Connections
In single-phase transformers primary- and
secondary voltages are either in phase or in
opposition and this is completely specified
by the polarity or the lettering of the leads.
In poly]3hase units or banks, however, these
vector relations, being more complicated,
are represented b>- voltage diagrams because
the mere lettering of the leads does not
indicate the polarity or these vector relations.
showing that there would be an unbalanced
voltage short circuited through the delta.
The chief three-phase connections that are
commonlv used are: delta-delta, Y-Y, and
delta- Y (or Y-delta).
Y-delta Connection
The method of constructing the voltage
diagram of Y-delta connected coils has
already been described. It is evident on
inspection of Figs, oa and ")b that that con-
nection has subtractive polarity, 3()-deg.
angular displacement, and standard jjhasc
H, Hz H3
Fig 5a
X, <^..N
^3
Fig 5b
Fig. 6a
«2
X, X,
H3
^Z
X3
XZ X, Xy
H, H3 Hz
Hz
H, Hj
'^iJ-Tii— xj
Fig. 6b
Fig. 6c
Fig. 6d
Furthermore, polarity alone is inadequate
to represent vector relations in poh'phase
connections; the subject can be more readily
handled by voltage diagrams.
How to Construct Voltage Diagrams
The method of constructing the voltage
diagram for a given design may best be
explained by an example. Fig. 5 represents a
Y-delta connected three-phase unit. Draw
HiHiH., (Fig. ob) representing the induced
voltages of the Y-connected winding. The
voltage diagram of the delta-connected wind-
ing can now be drawn. Coils A'iA'2 and HiN
(Fig. .)a), being wound in the same direction,
their induced voltages must also be in the
same direction. Therefore, we draw A'lA'o
(Fig. .5b) parallel to and in the same direction
as HiN. On the middle leg (Fig. oa), coil
A'jA's is wound in the same direction as H^N ,
and, therefore, their respective voltages are
drawn parallel and in the same direction
(Fig. .5b). Similarly for the third phase, and
the diagram is complete.
If the delta were improperly formed in
Fig. .5a, the delta in Fig. .5b would not close,
rotation. Figs. Ga, 6b, (ic, and 6d have
additive polarity, 30 deg. angular displace-
ment, and standard phase rotation.
The voltage diagrams of Figs, oa, ob, Oa,
(ib, (ic and (id are identical, but the lettering
of the leads is different due to the different
internal arrangements. The different letter-
ing of the leads is thus equivalent to inter-
changing the leads since similarly lettered
leads are to be connected together for multiple
operation. It thus becomes evident that by
interchanging leads identical voltage diagrams
are obtained on Y-delta transformers that
have different internal arrangements, which
was not possible with delta-delta or Y-Y
transformers. This may be further explained
as follows:
Construct a voltage diagram as previouslv
explained for the connection shown in Fig. 7a
which will be like Fig. 7b. Interchange two
leads on the high side {H^ with //.) and two
leads on the low side (A'l with X-,) as shown
in Fig. 7c. Constructing the voltage diagram
for this new arrangement, the diagram of
Fig. 7d is obtained, which is identical in
form and lettering with those of Figs. .5b.
378 May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. o
6b and 6d. We conclude then that all Y-delta
or delta-Y connections can be reduced to the
same diagram by properly selecting the order
of the leads. This, as before mentioned, is
not possible with delta-delta or Y-Y con-
nections.
H, Hz Hi
Hz
Hi
H-5
X3
Fig. 7a
^2
Fig. 7b
Delta-delta Connection
If we try all possible delta combinations
in high and low-voltage coils, we find that
there are only two diagrams that are operative
at all. These are shown in Figs. S and 9.
It is not intended to convey the idea that the
coil windings and combinations shown in these
illustrations are the only ones that will give
Hi Hj Hi
Hi Hj H,
Fig. 7c
Fig. 7d
the indicated voltage diagrams. It is rneant
that these voltage diagrams arc the only delta-
delta diagrams which arc possible or operative
at all.
Polarity
Considering the polarity of these diagrams,
we see that that of Fig. S is subtractive and
that of Fig. 9 is additive. Polarity, however,
does not necessarily sufficiently specify the
phase relation between the high and low
voltages since they are not single straight
lines but polygons, and the angles may not
necessarih- be only 0 or 180 deg. but also
some intermediate value. The phase relation
is called "angular displacement" and is
^?
'■#&; /
H, Hj
Fig. 8a
Fig. 8b
H, X, H-,
Fig. 8c
defined by the A.I.E.E. Standardization Rules
as the angle between the lines H\K (.V being
the neutral point of the diagram) and A'lA'.
The location of the H\ lead is defined as above
for single-phase units, that is, the right-hand
side of the obsen-er facing the high-voltage
side. The location of the A'l lead is fixed so
as to make the diagram fall under one of the
Fig. 9a
Fig. 9b
Fig. 9c
Standardized Groups to be described later.
We see that the angular displacement of Fig.
8b is zero, and that of Fig. 9b is ISO dog.
Phase Rotation
In order that the relative phase rotation
of high and low voltages may have any
significance at all, it must refer to a perfectly
definite order in which the leads are to be
considered. Thus, in Fig. 8b, phase rotation
VOLTAGE DIAGRAMS OF TRANSFORMERS
379
is clockwise in the order hiH-^Hs, but counter-
clckwise in the order //o//)//.,. The phase
rotations of H1H2H3 and A'lAV^'s are the
same; those of H\H«Hi and A'sAVVj are
opposed. In view of the necessity of specify-
ing the order of leads, the Standardization
Rules referred to above provide that the leads
shall be marked in such a way that phase
rotation of high and low voltages in the lead
order H1H2H3 and A'iA'2A'3 shall be the same.
That is, if a three-phase motor were trans-
ferred from the high-voltage circuit to the
low-voltage circuit, transferring its terminals
from Hi to A'l, from H^ to A'2, and from H3 to
A's, its direction of rotation will be the same.
Considering Fig. Sb, H1H2H3 and XiX-^Xs have
the same rotation and, therefore, correct phase
rotation. Considering Fig. 9b, phase rota-
tions H1H2H3 and A'lAVAs are the same, and
therefore, also correct. It will be interesting
to note that while Figs. Sb and 9b have oppo-
site polarities and different angular displace-
ments, yet they have the same phase rotation.
It is evident that a voltage diagram indi-
cates only the relative phase rotation of primary
and secondary, and gives no information
as to the actual phase rotation on either side,
this being determined by the supply circuit.
Clockwise or counter-clockwise lettering of
the primary voltage diagram, also its location
on the paper (pointing one way or another),
are of course entirely arbitrary. It is evident
M, H, H3
M-I-K
H, Hz H3
■*>,
'-L
^z,
.^
:=L
■ 'u
Fig. 10a
H, Hj
Fig. 10b
also that this relative phase rotation of the two
sides refers to a definite sequence of leads.
It will be seen that changing the lettering
or interchanging the leads (leaving coil
connections unchanged) cannot alter the
voltage diagrams of Figs. 8 and 9. That is,
the transformer in Fig. Sa cannot be made to
give the diagram of Fig. 9b by manipulating
its external leads. The onlv wav to effect
such a change would be lu change the internal
connections of the coils.
With delta-delta connected transformers
the lettering of the leads is the same regard-
less of the angular displacement, as seen
from Figs. 8a and 9a.
Y-Y Connection
The method of constructing the voltage
diagrams of Y-Y connected transformers is
Hz
H, H3
Xj X,
Fig. 11a
X2
Fig. lib
the same as that described in the foregoing.
Two connections are possible as shown by
Figs. 10b and lib. The first has subtractive
polarity, zero phase displacement, and stand-
ard phase rotation. The second has addi-
tive polarity, 180-deg. angular displacement,
and standard phase rotation. It will also be
evident that no manipulation with the
external leads will change the diagrams,
although it may change the order of lettering
of the voltage diagrams.
With Y-Y connected transformers, the
lettering of the leads is the same regardless
of the angular displacement, as seen from
Figs. 10a and 11a.
To Obtain Voltage Diagrams by Test
We have described the method of con-
structing a voltage diagram when the design
is given. If the design is unknown, coils
and connections inaccessible, and no vector
diagrams furnished, these can be obtained b>-
test. Polarity and jjhasc rotation tests are
valuable checks when the diagram is given
(or assumed), but are not necessarily sufficient
to enable one to draw it. Voltage diagrams
can be determined by the following method,
neglecting the polarity and phase rotation
tests if desired: Connect one of the high-
A'oltage leads to one of the low-voltage leads,
excite the transformer at a voltage safe for
the low-voltage circuit, measure the voltages
380 -May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 5
between all the other high and low-voltage
leads, and plot them to scale. For instance,
referring to Fig. 8a, if we should connect Hz
to A'3 and make these measurements, we
wotild obtain a diagram like that of Fig. 8c;
or referring to Fig. 9a, if we connect Hz to Xz
we would obtain a diagram like Fig. 9o.
Three-phase Tra/o/or/rrers without Tops \
6roup I
Angular
Oisplocermnt
A A
Hz Xz
XX
H, H, X, X,
Oroup Z
Angular
Disploce/mnt
J SO*
a' 'I ",'
A V
H, H3 X2
Hz X, X,
AV
Hi Hj Xz
Group 3
Angular
0/sp/octawi
30"
Hz X,
H, H, X,
Hz Xz
A-<3
H, H, X,
Hj Xj
A >'■
Hj H, Xj
H, Xz
X >■
H, H, X,
Three phase Transformers with Taps
Group 3
Angular
D/spkKe/mnt
30"
Fig 12a
If we should apply this test to a Y-V
connected unit of the same polarity (Figs. 10a
and 11a), we would obtain the same diagrams
as in Figs. Sc and 9c respectively. That is,
it would not be possible to determine by such
tests whether the internal connection is delta-
delta or Y-Y. However, so far as parallel
operation is concerned, the distinction is
unnecessar\-.
The test will indicate the angular dis-
placement between high and low-voltage
circuits, but cannot distinguish between con-
nections that belong to the same group, that
is, connections which will successfully parallel
with each other.
It will be evident that obtaining voltage
diagrams by such measurements becomes
difficult when the low voltages are \-cry
small comjjared with the high voltages.
PARALLEL OPERATION
In order that two transformers of similar
voltage rating may safely be connected in
multiple, their polarity, phase rotation, and
angular displacement must be the same.
Delta-delta and Y-Y transformers have
correct angular displacement when their
polarity and phase rotation are correct.
This, however, is not necessarily true for
delta-Y (or Y-delta) transformers. In this
case, however, these can be adjusted by
the proper selection of the sequence of
leads.
If the voltage diagrams of the transformers
which are to operate in parallel are available,
it is then only necessary- that these diagrams
coincide and corresponding terminals be
connected together. // is entirely unnecessary
then to raise questions of polarity and phase
rotation, because when the voltage diag,ranis
coincide, leads which are to be connected together
will have the same potential, this being the basic
requirement for multipling; whereas, polarity,
phase rotation, etc., are merely means to
arrive at this condition. When voltage
diagrams coincide, polarities and phase rota-
tions must necessarily agree, although the
converse of this is not necessarily true.
For the purpose of simplifying the con-
necting of transformers in parallel and avoid-
ing the necessity of testing for polarity, phase
rotation, etc., the A.I.E.E. and X.E.L.A.
have standardized the marking of trans-
Group 4
Angt/lor
Displacement
0'
Oroup 5
Angular
Displacement
30'
Anqufar
Dispiocemmnt
30'
3t* phase Transformers without Taps
HZ ^2 -J
Ms X4 Xs
"'^'*
x,-^'.
Sfjt-fihojg rrvnsformvs with Taps
Xj X,
M, ^r~, Xf X}
-• »* "J •«« xt
Fii Ub
former leads (covered in A.I.E.E. Rules.
Sections ()00-0 17) as has been e.\]>laine<l in this
article. Transformers that are marked in
this manner can Ix^ operated in multiple by
simply connecting similarly lettered leads
together. This, of course, is contingent on
the transformers having proper character-
istics, i.e., ratio, impedance, angular displace-
ment, etc.
VOLTAGE DIAGRAMS OF TRANSFORMERS
381
Three-phase transformers are divided into
three groups based on their angular dis-
])lacements as shown in Fig. 12a.
Four of the usual three-phase to six-phase
diagrams are shown as Groups IV and V in
Fig. 12b Their construction involves noth-
ing more complicated than the method
indicated for three-phase to three-phase con-
nections.
To operate in multiple, transformers must
belong to the same group. No interchange
of external leads can change one group into
another. Thus, two delta-delta transformers,
one of Group I and the other of Group 11,
cannot be operated in multiple. If the high-
voltage diagrams be superposed, the low-
voltage diagrams will not coincide. All
Y-delta or delta-Y transformers, however,
can be reduced to the same diagram, and,
therefore, they are classed in only one
group.
Practical Problems
Polarity, phase rotation, etc., are of
interest primarily on account of their bearing
on the parallel operation of transformers.
The operator wishes to know these facts
about his apparatus before connecting in
multiple, as otherwise a wrong connection
subjects the apparatus to short circuit. Some
of the problems that come up in practice will
be discussed.
(1) Transformers lettered in accordance
with the A.I.E.E. rules, i.e., high-voltage leads
marked Hi, H^, etc., and low-voltage leads
marked Xi, A'2, etc.
Single-phase transformers which are so
marked and have like 'ratios and impedances
may be connected together in the order of
lettering regardless of any question of polarity
since the method of lettering takes care of
polarity.
Three-pha-se units also may be connected
together in the order of lettering, provided,
however, that the units belong to the same group,
i.e., have the same angular displacement.
Otherwise they cannot be operated in multi-
ple at all. Angular displacement cannot be
altered by manipulating the external leads
without changing the internal connections.
The same applies to the parallel operation of
six-phase transformers.
The connections of six-phase transformers
to synchronous converters is simplified by
a correct understanding of the manner in
which the windings of the latter are tapped
and brought to the slip rings, and the system
of numbering used. Fig. 13 shows how the
winding is tapped and brought to slip rings.
The slip rings are numbered 1, 2, 3, etc.,
beginning from the bearing and proceeding
towards the armature. The diagram also
shows the actual direction of the physical
rotation of the armature which is counter-
clockwise looking from the slip-ring end of the
machine. The actual electrical i)hase-rotation
is clockwise, i.e., in the order /, S, 3, etc.
Evidently, the transformer must be so con-
nected to the converter that neither the
rotation of the latter is reversed nor any one
l^hase is short-circuited. As explained pre-
viously, when the jihasc rotation on the high-
e .5 4- 5 2 /
3/po/or D/agrarr,
4.
I Mec/7ar7/ca/
] ffotot/oo
■ Siationary
I ■9e,'isre/7ce Po/nL
P/7ase .fotat/on
Fig. 13
voltage side of the transfonner is in the order,
H\, H-i, H3, the phase rotation on the low
voltage side is in the order A'l, A'2, A'3, etc.
Therefore, if the high-voltage supply phases
are correctly connected to the high voltage
of the transformer, the transformer and
converter will operate properly when Xi of
the transformer is connected to ring 1 of the
converter, A'2 of the transformer to ring £
of the converter, etc. Although this is the
standard connection, there are eleven others
or altogether twelve operative connections
which may be used if for any reason they
are found more convenient. Of these twelve
operative connections, six correspond to one-
phase rotation on the primary, and the other
six to the opposite phase rotation on the
primary. Thus, six of the possible con-
382 May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 5
nections for one phase-rotation are as
follows :
Connect A'l to Ring 1 or ^ or 5 or 4 or 5 or 6.
Connect Ao to Ring 2 or 3 or J^ or 5 or 6 or 1 .
Connect A'3 to Ring 5 or 4 or 5 or 6 or 1 or 2.
Connect A'4 to Ring 4 or 5 or 6 or 1 or .^ or 3.
Connect X-^ to Ring 5 or 6 or 1 or 2 or 3 or J^.
Connect A'e to Ring 6 or 1 or ^ or 3 or i or o.
z 5 I. d
XX. A A
r 34 6 c' *<7 ^^ ^e
Fig. I4a
Fig 14b
A V V D>D>
Fig 15a
Fig 15b
Fig. ISc
Each vertical row constitutes one operative
set. Connections must not be made partially
from one vertical row and partially from
another. If on connecting to the supply,
converter rotates in the wrong direction, it
can be corrected by reversing one phase on
the high-voltage side.
It will be observed that when transformers
are lettered in accordance with the A.I.E.E.
rules, and also their angular displacement
given, a vector voltage diagram is not neces-
sary to be able to connect them properly
although it usually is given by some manu-
facturers as an added safeguard.
(2) Transformers not lettered in accordance
with the A.I.E.E. rules, hut their voltage
diagrams available.
When voltage diagrams arc available no
questions of polarity or phase rotation need
be asked. It is necessary and sufficient to
determine whether the diagrams of the two
units (or banks) will coincide; i.e., whether
they have the same angular displacement
between the primary and secondary. For
instance, a transformer of which the voltage
diagram is shown in Fig. 14a can be paral-
leled with one whose diagram is as shown in
Fig. 14b. Superposing the two diagrams, both
high and low-voltage lead potentials coincide;
thus / with c, 2 with b, 3 with a, 4 with /,
■T with d, and 6 with e. There can be no
difficulty about telling which leads are to be
connected together. The fact that the volt-
age lines of the two diagrams do not coincide,
one being a Y diagram and the other a delta
diagram, is of no consequence, since the
points which are to be connected together
coincide and must therefore have the same
potential.
Confusion is sometimes experienced when
voltage diagrams are shown in different
positions, as for example, in Figs. 15a, 15b,
and 15c, where identically the same voltage
diagram is shown in three different positions.
What a voltage diagram indicates is not the
actual potential of the terminals, but the
voltage vector relation between the two
windings. This relation is identical in the
above three figures. This can be seen still
better if we rotate the high and low-voltage
diagrams of Fig. 15b through either 00 or 180
deg., when it becomes identical with 15a.
The same refers to 15c which would have to
be rotated counter-clockwise 90 deg. or clock-
wise .30 deg. to coincide with 15a.
In rotating a diagram, care must be taken
not to alter the relative position of the
primary and secondary voltage diagrams with
respect to each other. This being done, the
Fig. 16. Quarter-phaM Connections
rotation of a diagram through any angle is
permissible. It will also be seen that diagrams
which are alike can be sui)erpo.sed and made
to coincide in three dilTorent jwsitions. For
instance, Fig. HbcanbesuperposedonFig. 14a,
making / coincide with c. b. or ii. There are
VOLTAGE DIAGRAMS OF TRANSFORMERS
383
"J
j_Ll
\»2
Ht '3 *.
A., V
|W/ l»! I
'/Z
J Additive \
\Polantif]
-"j
"Z
/'H,Hz\
iiubtractlr^
\Po/anti/j
X <
Figs. 17 to 22. Three phase Connections
384 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol XXIII, Xo. 5
thus three ways of connecting three-phase
units for multiple operation.
(5) One or both units not lettered hi
accordance with the A.I.E.E. rules and no
voltage diagram available.
With single-phase units the procedure ma.\-
be either to test their polarity prior to con-
necting them in multiple, or they ma}' be
multipled for a trial through a fuse or volt-
meter and when the connection is found to
be O.K. the fuse or voltmeter mav be short
leads, connect a pair of low voltage leads
together (fused or unfused); the second pair
of low-voltage leads should now be connected
together through a voltmeter. If the volt-
meter indicates no voltage, it may be short
circuited, taken out and used to test the third
phase. It will be appreciated that the volt-
meter must be capable of withstanding twice
the line voltage on the leads to which it is
being connected, since, if the polarity is wrong
the phase voltages will add in the voltmeter
»i
x.
Hz
A
Hi MS
X6 Xi
Threc-phasr to Six-phase Connections
circuited In important cases both pre-
cautions may be taken.
In multipling two three-phase units for
trial at least two of the three jihases should be
fu ed, and preferably three. The fuse should
be connected between the leads which are to
be connected together, preferably on the low-
voltage side, and the excitation applied to the
transformers should be small. Preliminary
tests for voltage diagrams, as has been ex-
l)]ained, would be very desirable ; however, trial
multiple operation through fuses is fre-
quently found verv sim])le. A better scheme
is to substitute a voltmeter for the fuse. For
instance, having multipled the high-voltage
and give a large deflection instead of neutral-
izing each other and giving a zero deflection
as would happen if the connection were
correct.
If the two units which are being tested are
in V-delta or delta-V connection, and having
first connected the high-voltage sides in
parallel, it is found that no combination of the
secondary leads is operative, one phase on the
high voltage of one of the units should Ix-
reversed, and then an operative combination
of the low voltage leads can be found. This is
a characteristic of the V-delta and delta-V
connections and sometimes shows itself in a
puzzling manner. For instance, in connecting
VOLTAGE DIAGRAMS OF TRANSFORMERS
385
two identical V-dolta units in multiple, if one
phase of one unit is reversed on the primary
side, the secondaries cannot be multipled
refjardless of any reversals that may be
attempted on the secondary sides. If two
phases are reversed on the primary-, then
an o^ierative combination of phases can be
(a) a reversed phase on the primary of the
former can be made good by a reversed phase
in the secondary, and (b) with a jjiven connec-
tion of primaries in multiple, if an operative
combination of secondary leads cannot be
found, no other combination of primary leads
will make parallel operation possible for them.
Figs. 25 and 26. Three-phase to Six-phase Connections
found on the secondary sides. Delta-delta
and Y-Y connected transformers differ from
the Y-delta or delta-Y transfomiers in that;
Figs. 1(3 to 26 show the connection of single-
]ihase imits of different polarities in various
single and polyphase banks.
386 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No
Relative Merits of Connections Employed in
High-Voltage Generating Stations
By Ernest Pragst
Power and Mining Engineering Department, General Electric Company
In planning a system of connections for a central station consideration must be given to the factors of
personal safety and protection to transformers and generators; also a degree of flexibility comparable with
the importance of the station should be provided, so that facilities will be available for operating the station
economically at all times whether at full load or fractional load. The author presents and discusses eight
different systems of station connections, ranging from the simplest to the most elaborate. — Editor.
In the evolution of the modem high-voltage
generating station a number of commonly
accepted arrangements of interconnection
between the generators, transformers and
Hnes have come into use, each having its
advantages, disadvantages and particular
field of application. In this article the writer
proposes to set forth a number of these basic
or fundamental systems of connections (limit-
ing himself solely to their application to
generating stations, where power is stepped
up in potential and transmitted over high-
tension transmission lines), analyze them,
and attempt to determine their particular
field of application.
Any system of switching, along with power
transformers, transmission lines, etc., is
selected with one primary object in view;
namely, to transport the electrical energy
available at the generator terminals to one or
a number of sources of load. With a natural
realization of this object, it is at once recog-
nized that the system of connections con-
templated for a generating station is in-
fluenced not only by the selection of the
prime movers with their corresponding gener-
ators, but also to an equal if not greater
extent by the character of load, its geograph-
ical location in relation to the generating
station, and, in the case of existing systems,
the relation of the generating station under
consideration to the existing system.
A detailed study of the interconnection of
apparatus within the station, with an ever
watchful eye on the influence of these con-
nections on those external to the station,
will show that any successful arrangement
should fulfill the following conditions:
(a) It should not afford an undue risk
to the operating force, particularh-
when conducting switching opera-
tions as a result of such abnormal
conditions as electrical failures of
lines and apparatus.
(b) It should permit of the economic
operation of apparatus.
(c) It should be simple in principle.
electrically.
(d) It should lend itself to simple and
rugged mechanical arrangement and
construction.
(e) It should have a reasonable degret- of
flexibility.
(f) It should assure a degree of con-
tinuity of ser\-icc commensurate
with the class of load ser\ed.
It might be well to elaborate on condition
(b). Here it is meant that the connections
should be such as to permit the operation of
apparatus at a load as near as possible to that
corresponding to maximum efficiency or full
load, as desired. To do this, it is usually
necessary to arrange to operate all apparatus
in parallel. In the generating station this
is provided for by paralleling all apparatus
on a low tension or high tension bus or
both. It is also customar>' to operate all
plants in a given location interconnected
or in parallel. This permits of the necessit>-
of carr\-ing but a minimum amount of
spare generating capacity, makes possible
the operation of a system at maximum effi-
ciency, and automatically takes advantage of
any diversity in load that might exist on the
system. Also should the failure of any one
piece of generating ai)paratus occur, if the
system is large in proportion to the capacit>
of the lost generator, system oi)eration in all
probability will not be affected, as the load
suddenly lost by the generator which has
failed is distributed among all units remaining
in service.
Having determined in a general way the
factors influencing any system of station con-
nections and the conditions these connections
should fulfill. let us next consider the means
of making and changing these connections to
accomplish the desired results. We have at
our disposal :
(a) Sim])le disconnecting switches, whose
ap].)lication is limited solely t<> the
CONNECTIONS EMPLOYED IN III(',lI-\-()LTACE (]EN1>: RATirJC, S'|-ATIONS ;iS7
isolation of apparatus. They arc
not applicable for disrupting the
flow of current.
(b) Multi-pole (usually triple-pole) air-
break diseonnectinj:; switches. These
are usually manually operated, arc
used for the isolation of apparatus,
and can be employed for dismpting
small amotmts of current such as
the charging current to short lengths
of high-tension line, the exciting
current of transformers of medium
capacities, and light loads. One of
their chief fields of application in
the type of generating station under
consideration is for use in the high-
tension circuits in place of the
simple single-pole disconnecting
switches as a result of the increased
facility and speed with which switch-
ing operations can be performed
when they are employed.
(c) Air-break circuit breakers (automatic
and non-automatic). The ajjplica-
tion of these is limited almost
exclusively to low-voltage (usually
under 600 volts) high-current cir-
cuits.
(d) Oil circuit-breakers (automatic and
non-automatic). These arc used in
all circuits where currents of large
magnitude must be broken, where
quick switching operations must be
conducted, or in circuits which, if
opened or closed by other forms of
switches, might present an undue
hazard to the operators.
Hand-operated knife switches, fuses, special
combinations of fuses and switches and the
like, have been omitted from the foregoing
tabulation, because their application in a
modern generating station is extremely
limited.
As a rule more than 95 per cent of service
interruptions are the result of insulation
failures external to the generating station;
hence the control of the outgoing lines from
the station should be such as to permit of
carrying on switching operations with rapid-
ity. Although failures of this class are
extremely numerous, annoying and costly to
the power consumer, the value of the apparatus
lost in this manner represents a small item
when compared to that lost as a result of
apparatus failure in the generating station.
Even though failures within the generating
station arc relatively rare, when they do occur
their results arc often far rcacliint; both in
intcrmiition to scr\-icc and expense of replace-
ment. Therefore, whereas the method of
controlling outgoing lines should be designed
with a view to quick switching operation, the
control of the circuits within the station
should be designed with ])articular reference
to the prevention or localization of ]30ssible
failures of apparatus.
When considering the switching arrange-
ment of a generating station as a part of a
svstcm or of the individual pieces of apparatus
in their relation to the station as a whole, one
must appraise the relative importance of the
part under consideration in its relation to the
whole. For example, should the station
represent but a small and unimportant part
of a large system, a station which can be
dispensed with for a short time without
materially affecting service, then we are natu-
rally justified in cmplo\ing an inexpensive
switching scheme , if a generator in a station
can be spared, one oil circuit breaker for its
control is sufficient; and the same holds in
the case of transformer control, busses, etc.
Having now a clear vision of the objects
we wish to accomplish and a knowledge of the
limitations of the apparatus at our command,
let us investigate some of the more commonly
used station connections.
Fig. 1. shows one of the simplest forms of
station connections and represents, perhaps, a
minimum expenditure for switching material.
Here each generator is rigidly connected
to and used as a unit with its corresponding
transformer bank. The station auxiliaries
are fed from an auxiliary bus, which in turn
1,
1,
/
i i-
J
J J-
SUlion
Bus
C
Fig. 1. Simple Station Connections Employing Generators and
Transformers Rigidly Connected as Units
can be served from any one or all generators
through the circuit breakers a. The chief
criticism of this arrangement is its lack of
flexibility. Each generator must be used
as a unit with its corresponding transformer
bank: failure of cither or of the conductors
388 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. .3
between them will result in the shut-down
of both. Also, as in the case of most sinplc
bus arrangements, a failure of the high
tension bus will result in a complete shut-
down of the plant until such time as repairs
can be made. As long as no trouble occurs
within the station, normal switching oper-
ations may be conducted without incon-
venience or hazard; the generators and trans-
formers being placed in and remo\'ed from
sen'ice b}- means of the oil circuit breakers b,
and the lines controlled through the oil
circuit breakers c. This arrangement finds its
particular application in the case of small
stations supplying loads where continuity
of service is not of primar\^ importance and
where the cost of the installation must be
held to a minimum at a sacrifice in flexibility
and assurance of ser\-ice continuity
Fig. 2 represents, no doubt, the most
commonly used system of connections. Here
all generators and transformers are connected
to a common low tension bus, while the high
tension side of the transformers and out-
going lines are connected to a common high
tension bus. The capacities of generators,
transformers and lines need not bear any
definite relation to each other as in the case
of Fig. 1. For switching under normal con-
ditions, and the protection of apparatus in
case of failure, this arrangement will meet
every requirement; that is, for failure within
a generator, circuit breaker a may be opened,
and in case of trouble within a transformer
/ {
T
T
X»
Dus
/
r
Fig. 2. A Commonly Used Arrangement Employing
Single High and Low Tension Busses
bank, the bank can be isolated through the
opening of circuit breakers h and c. The
criticism of this arrangement is its inflexi-
bility. A failure of any generator circuit
breaker results in the forced withdrawal of
the corresponding generator from service;
and similarly with the transformers, should a
failure of oil circuit breakers 6 or c occur:
and with the lines in case of failure of oil
circuit breakers d. Should a failure of either
bus occur, a complete shut-down of the station
will naturally result. This is often partially
^
£ £ i?
1 Bu««a
Fig. 3. An Elaboration of Fig. 2. Using Low and High Tension
Busses with a Single Oil Circuit Breaker and Selector
Disconnecting Switches in Each Circuit
guarded against by the introduction of
sectionalizing disconnecting switches in the
busses. Because of the expense of high
])0tential oil circuit breakers, it is often
customary to substitute for oil circuit breaker
c and its corresponding disconnecting switches
a triple pole air break disconnecting switch,
with the result that operating flexibility and
assurance of sen-ice continuity are lessened.
Though open to these criticisms, the arrange-
ment of Fig. 2 is well adapted to the require-
ments of small and medium sized instal-
lations, as the protection to apparatus is good
and the personal hazard to operators is small,
also, as the probability of failure of properly
selected oil circuit breakers and well con-
structed bus structures is ver>- small, con-
tinuity of service is quite well assured.
Fig. 3 shows a diagram of connections which
is an elaboration of Fig. 2, double high and
low tension busses being employed. Approxi-
mately twice the amount of bus material and
number of disconnecting switches are re-
quired as for the arrangement of Fig. 2.
This arrangement will practically eliminate
the possibility of a prolonged shut-down as
the result of a bus failure. It also permits of
maintaining service when working on either
bus. However, it will not eliminate the neces-
sity of withdrawing apparatus from ser\-ice
in case of trouble with its corresponding
CONNECTIONS EMPLOYED IN HIGH-VOLTAGE GENERATING STATIONvS 3S9
circuit breaker The arrangement has one
marked advantage over those of Figs. 1 and 2;
namely, should a feeder trip out it is possible
first to test it out, thus avoiding the risk of
again tripping it out or causing surges on
other feeders by placing it back in service
f ^
Fig. 4. Another Elaboration of Fig. 2. Using Single Main High
and Low Tension Busses with Transfer Busses for Utilizing
a Reserve Circuit Breaker in Each of the High and
Low Tension Circuits
when it is still short circuited or grounded.
For example, suppose generators Nos. 1, 2
and 3 are in operation with transformer banks
Nos. 1 and 2 and connected to low and high
tension busses .4, with generator No. 4, trans-
former bank No. 3 and busses B in reserve.
If one of the feeders trips out from some
unknown cause it will be an easy matter to
test it out with generator No. 4 and trans-
former bank No. 3 operating through busses
B. Bus tie circuit breakers e and / are often
included to facilitate this class of switching
operation. Should the tested line prove good,
it could at once be placed in ser\ace by closing
either circuit breakers e or /, or both. With
circuit breakers e and/ closed and, of course,
the ,4 and B busses in synchronism and at
the same potential, the transfer of a circuit
carr\'ing power from one to the other bus can
be effected without danger or interruption to
service by means of the disconnecting switches.
Fig. 4 is a further elaboration of the con-
nections of Fig. 2. The arrangement of Fig. 3
provides for the failure of a bus, but makes no
improvement in the operating limitations of
the oil circuit breakers over the arrangement
of Fig. 2, while Fig. 4 provides for the with-
drawal from ser\-ice of anv oil circuit breaker
without interrupting the operation of the
corresponding apparatus, but makes no pro-
vision against bus failure. The amount of
switching equipment required for the scheme.
Fig. 4, is almost identical with that of Fig. 3,
and the two systems of connections are an
improvement over that of Fig. 2, Fig. 3 pro-
viding against a possible bus failure and Fig. 4
against a failure of circuit breakers only.
Under normal operation all apparatus will
operate from the main or ^4 busses. Should
it be desired to withdraw from service a
circuit breaker, it can be done without
interruption by connecting the corresponding
piece of apparatus by means of its dis-
connecting switches to the transfer or B
bus and connecting the two busses A and B
together through the bus tie switch e or /, as
the case may be, after which the circuit
breaker in question can be withdrawn from
service.
Fig. 5 is included primarily to illustrate a
method of high tension connection which has
been very commonly used in past years for
high tension stations. The low tension
arrangement may be as shown or similar
to that of Figs. 3 or 4 without materially
affecting the high tension system. Here
transformer banks are operated as a unit
with the lines. As shown, three high tension
oil circuit breakers c, d and e are employed
for each group of transformer and line.
Because of the expense of these high potential
circuit breakers, triple-pole air-break dis-
7
7
Hiqh
-lenS'on
Bus
_J
y.
-Tension
Fig. 5. A Wiring Arrangement Based Upon Operating a Trans-
former Bank as a Unit with a Line
connecting switches are often substituted.
When this substitution is made, circuit
breakers c or d or both are those usually
replaced. In the case of line failure, it is
customary to trip the line along with its
corresponding transformer bank by means
.J'JU Alav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo 5
of the low tension circuit breaker b. The
chief advantage of this arrangement lies in
the fact that when operating in this manner
the magnitude of the surges resulting from
switching operations on the high tension
system is reduced to a minimum. On the
other hand the arrangement presents
numerous disadvantages. It does not
usually lend itself well to connection
in a network; and is uneconomical
in those cases where the generating
station supplies widely separated
loads, as in this case the transformers
in all probability must be of different
capacities to meet the load require-
ments, which is an undesirable feature
from a construction and maintenance
point of view. Also, when tripping __^_
transformers with their transmission — r~r"
lines, there is a marked increase in ) }
potential drop over the remaining LqJ
circuits as compared with the case 3
where all transformer capacity con-
tinues in operation supplying power
to the lines remaining in service. Fig. 6.
The significance of this potential
drop will readily be appreciated particularly
in the case of low power factor loads,
when it is recalled that the reactance
of a transformer bank is ver>- often the
equivalent of that of the line, while the
combined reactance of both the generating
station and substation transformers might
be twice that of the line. With the present
state of the art of designing and building high
potential api^aratus, when it is possible to be
reasonably assured that the apparatus if
properly applied will withstand all of the
usual abnormal stresses experienced in practi-
cal operation, it is doubtful whether the
advantages gained by reducing the surges
incident to switching by resorting to low
tension switching are sufficient to outweigh
the limitations of this method of connection
and operation in most cases. The arrange-
ment finds its i^articular application in those
cases where jjower is generated at one station
and transmitted over a number of lines to a
single substation. In such cases it affords an
effective and economical arrangement.
Fig. G represents a rather novel scheme
which contemplates the operation of a
generator and transformer bank as a unit
during normal conditions and practicallv
eliminates all of the usual low tension circuit
breakers. It is particularly interesting in that
all normal switching ojierations are to be
performed on the high tension side of the
station. Although almost all oil circuit
breakers have been omitted from the low
tension circuits, provision has been made to
operate any generator with any transformer
bank in case of trouble. A careful study of
this arrangement will reveal the fact that
1
^fi/fi, :
K=«
^ V ^ ^-
vJLm vjiw vvivw "*«.
Hiqh
Te'\sion
Buses
An Arrangement On^itting Low Tension Circuit Breakerst Throughout
when it is necessary to do any low tension
switching, involving the use of disconnects
a, h, c and d, great care must be exercised,
necessitating perhaps the withdrawal of appa-
ratus from service to avoid a possible sc\'cre
accident should a mistake be made in switch-
ing. Besides the chance of accident to the
operating force and bus structure when doing
low tension switching, the time required to
make the change-over, and the fact that
considerable capacity must be withdrawn
from ser\'ice during such time, makes it
appear that this jiarticular arrangement
has but limited application except in
those stations where there arc installed a
large number of units and sufficient spare
capacity to make the necessity of operating
a generator with other than its corre-
sponding bank of transformers an unusual
procedure
Fig. 7 shows an arrangement with a com-
plete dui)lication of switching equipment and
busses. Such an arrangement will fulfill all
the technical requirements of a well designed
switching scheme for the majority of cases.
As a matter of fact, the strongest criticism
that can be advanced against such an arrange-
ment is the expense involved; and for this
reason such an arrangement can be adopted
only in large cai^acity stations where con-
tinuity of scr\Mce is of |)rimar\' importance and
where its assurance will justify the ex])ense.
CONNECTIONS !• MPLOVKI) IN HICxII-VOLTAGI': GENKRATINC^. STATIONS
;!!il
Fif,'. 7 is made up on tlic basis of usiiiK on tlio
low tension side a main and auxiliary bus.
The main bus has been segregated by means
of bus sectionalizing circuit breakers /.
Current limiting reactors k are also shown, as
they are invariably required in the case of
Fig. 7. An Elaborate Arrangement Employing Double Busses and Double
Selector Oil Circuit Breakers in Both the High and Low Tension Circuits
large capacity stations, and are usually
]jlaced as shown in such a bus arrangement as
this. Only a simple single auxiliary bus is
used without sectionalizing switches or re-
actors on the theory that it will never be
necessary to withdraw from service more than
one main bus section at a time. With such
a bus arrangement all normal switching
operations, and those of an abnormal nature
as well, can be performed by the station
operator from the main control board with a
minimum loss of time. The arrangement also
lends itself very nicely to a station feeding
])Ower at several potentials, in which case all
generators and low tension sides of transform-
ers can be paralleled on the common low
tension bus as shown, utilizing the necessary
number of independent high tension bus
structures. To reduce the cost of the switch-
ing equipment, a common and well worth
considering alternative to the high tension
arrangement as shown is sometimes em-
ployed, namely, the use of a single oil
circuit breaker and selective disconnects as
shown in Fig. 8. To facilitate switching
operations in this case, the selective dis-
connects often take the form of trii:ile-pole
manuall}- operated air-break disconnecting
switches, particularly in the case of stations
operating at the higher potentials.
Fig. N is another very common arrange-
ment, used in large stations stepping up
all power to one potential. The scheme
contemplates operating each gen-
erator as a unit with a transformer
bank, and paralleling all generators
on a common low tension transfer
bus. Where it is necessary, as with
large capacity stations, to install
current limiting reactors, they are
usually placed as shown /. This
arrangement is less expensive than
that of Fig. 7, but does not give the
same assurance of continuity of
service in the ca.se of a circuit breaker
failure, although should the failure of
circuit breaker a,b or c occur, it could
be cut out of service by means of a
jumper placed around it and further
switching operations carried out with
the remaining two breakers with but
slight inconvenience.
In addition to the arrangements
shown, there are an almost unlimited
number of others, but on analysis
they prove to be as a rule nothing
more than an elaboration or a slight
modification of those considered here.
It is not possible to give any set rules govern-
ing the selection of a switching scheme; each
Fig 8. A Popular Arrangement Using the Generators as Units
with Corresponding Transformer Banks and Paralleling on
Double High Tension Busses Through Single Oil Circuit
Breakers and Selector Disconnecting Switches
case should be
careful analysis
\'olved
decided separately after a
of the various factors in-
392 May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 5
60-cycle Converting Apparatus
B}- J. L. BURN'HA.M
Engineer, Direct-current Engineering Department, General Electric Company
The popularity of the synchronous converter is attested to by the fact that 2}^2 kilowatts of this type of
apparatus are built for every kilowatt of motor-generator. The 60-cycle converter for voltages up to 300 has
proved very successful, but for voltages of 500 and 600 it has been very susceptible to flashing at the com-
mutator when subjected to short circuits or quick changes in load. Because of this trouble the whole phe-
nomenon of flashing was thoroughly studied, and as a result several devices have been developed which, when
employed together, effectively prevent flashing even on complete short circuit. This freedom from flashing
is secured through the use of high reluctance commutating poles, a special form of brush rigging, screened
flash barriers, and the high speed circuit breaker. Another disadvantage of the synchronous converter is the
inflexible ratio between alternating-current and direct-current voltage. Direct-current voltage regulation is
therefore usually effected from the alternating-current end by means of the synchronous booster. In conclu-
sion the author makes a comparison between the synchronous converter and the motor-generator on the bases
of efi&ciency, reliability, flexibility, costs and floor space. — Editor.
The great increase in use of 60-cycle
generators and extension of 60-cycle transmis-
sion systems in recent years has resulted in a
corresponding demand for 60-cycle converting
apparatus to supply direct current. This
demand has not been difficult to meet
successfully with motor-generator sets or with
converters delivering up to 300 volts direct
current, but has emphasized the difficulties
with synchronous converters for 600 volts and
over.
The inherent sensitiveness to flashing of
60-cycle railway converters has always been
more or less annoying to operating com-
panies. Changes in conditions of operation,
such as lengthening of feeders, ha^■e given
some relief but still there is a need for more
stable characteristics.
By the use of commutating poles it was
possible to increase the output per pole thereby
giving higher angular speeds. At about the
same time designs with greater spacing or
pitch of brushes and poles were made to
increase the flashing distance and reduce
the voltage on the commutator adjacent to
the brushes. Bridges were improved to give
greater stability and many minor improve-
ments added, but only recently has the most
promising development been accomplished.
Several years ago a special study of the
causes of flashing and remedies was under-
taken and has reached a stage where, with
certain equipment now developed, the (iO-
cycle railway converter may be made immune
from flashing at the commutator as a result of
direct current short circuits.
In a paper entitled " Protection from Flash-
ing for Direct Current Apparatus," presented
at the A.I.E.E. convention in June, 191S,
experimental results were given showing that
protection by high speed circuit breaker and
flash barriers would give comjilctc protection
against flashing. Since that time further im-
provements in both the converters and the
new type of high speed breaker have been
made to give still greater margins of safety.
Changes in the 60-c\cle railway converters
are principally in the form of commutating
pole and its windings and in the brush rigging.
Commutating Poles
The commutating pole construction which
is now applied to our line of standard (>(i-
c\xle railwa}- con\-erters makes use of non-
magnetic material in the place of steel for a
large portion of the pole next to the magnet
frame. The resulting increase in reluctance
of the commutating pole magnetic circuit
requires much higher excitation and thus an
increased number of turns in the winding to
gi\e the same flux density neccssar>' for
commutation.
Fig. 1. Oscillogram Showing Values of Altcrnalins an J Dirrtl
Current in 1000-kw., 500-voU. 60-cycle Transformer When
Subjected to Six Times Full Load and Tripped
with Ordinary Type of Breaker
The principal advantages arc •
1. The excitation may be increasinl to a
value in excess of the direct curuMit annaturc
reaction.
fin-CYCLE CONVERTING APPARATUS
393
2. Reduction of the effect of saturation.
3. The commutating field responds more
quickly to changes in load.
1. When a converter is suddenly loaded
the direct current increases more rapidly
proportionately than the alternating current
for a short period and then the reciprocal
relation is established on the reverse swing of
pulsation, etc. Thus the balance between the
alternating and direct current reaction is
different than for steady loads. The oscil-
logram of Fig. 1 is a record of the relations of
direct and alternating current when about
six times full load is thrown on and tripped off
a lOOO-kw., (iO-cycle, (lOO-volt converter with
a breaker of ordinary speed.
Fig. la is calculated from the value of cur-
rent in Fig. 1 and shows excitation of the
commutating pole resulting from current in
the field winding and annature reaction com-
bined. The straight lines show the required
excitation to give best commutation for all-
steel poles and for high reluctance poles. The
sequence of relations after application of the
load is given by the arrows. It will be seen
that the maximum percentage departure from
best excitation is nearly four times greater
with the all-steel poles than with high
^00 300 400
Percent Load
Fig. 1-a. Curves Plotted from Oscillogram in Fig. 1, showing
Effect on Excitation of Commutating Poles
2. As the greater part of the excitation is
used to overcome the reluctance of non-
magnetic material, the effect of saturation
is lessened, and thus more nearly can the
commutating field strength be maintained
proportional to load during the heav}' over-
Fig. 2. Radial Brush-holder Unit with
Outer End Insulation Removed
load when it is most needed. For the .same
reason the effect of hysteresis is reduced.
3. The greater speed in establishing the
required field with the greater magnetizing
force is self-evident.
Brush Rigging
The most vulnerable points of lirush rigging
are at the outer end of the commutator and
the leaving side of the brushes. It is there-
fore desirable to avoid overhanging parts in
the direction of rotation, or outward from
the end of the commutator. This is jmrtic-
Fig. 3. Spring and Pressure Adjusting Slide for
Radial Brush-holder, Shown in Fig. 2
reluctance poles. A severe short circuit
might give three or four times greater load
which would reverse the all-steel poles,
but would only weaken the high reluctance
poles as the field winding excitation of the
latter is stronger than the armature reaction.
ularly so with (illO-volt, (iO-cycle converters,
which inherently have a short sj)ace between
adjacent sets of brashes of opposite polarity.
The new rigging recently ap])lied to railway
machines, which is designated as radial unit
type, accomplishes the desired end 1)\- Iiaving
394 Mav, 1920
GENERAL 1':LECTRIC REVIEW
XXIII, No
no overhanging parts for springs or supports
(Fig. 2). The spring is made of a slightly
bowed strip of steel placed radially over the
brush and having radial adjustment for
pressure (Fig. 3). The attachment for sup-
porting the set of brushes is at a radial point
^—^fSZOAMf'SC/J-^XFVLJ. LO/U>
83S0AHPS
AC CutfKEur
!__,
the arc could not be confined, if we could not
dis]30se of the large volume of gas generated
at considerable pressure. A barrier consisting
of a box fitting closely to the commutator
around each set of brush-holders is worse
than no protection at all, as it confines the
gases, making them highly conducting, and
the resulting concentration of energy is very
destructive. The problem of disposing of the
hot gases has been successfully solved by the
use of metal screens inside of the bo.x structure.
Next to the commutator a scoop-shaped
member of the box is arranged to deflect
the gases from the commutator into the
screen and thus cool and condense them.
Barriers of this type are successful when
kept clean and properly fitted with a small
clearance between them and commutator.
However, the closing in of the brush rigging,
rendering it less accessible for inspection and
adjustment, is undesirable. Further study is
now being made to overcome this feature.
nes OF cornKTicMS
t
ocrrnrEH bhh^iies
eoo
• <»1-fo»(t5.
va-TSDC
o ff.e.Aripcifcs.
> ,0(MC.
1
\
A JC.CU/>ff£NT.
--WsAAAAAA/
V^v^v,«»-„VS^
Figs. 4 and 5. Oscillograms Showing Conciitions During Short
Circuit with Only High Speed Circuit Breaker. Switch.
and Slower Breaker in Circuit. In Fig. 4 high speed
circuit breaker shunts n resistance, and in Fig. 5
the breaker is connected directly in series
near the inner end of the commutator where
an arc from flashing has the minimum tend-
ency to form. This type of rigging has the
further advantage of simplicity when it is
desired to adapt fire-proof insulation which
gives still further protection to the rigging
from burning by reducing the exposed
metallic surfaces' on which the arc might
play and produce more conducting gas.
Flash Barriers
As a further protection for extreme con-
ditions, flash barriers of a special fonii have
been (k'vclope<I. It was early recognized that
High Speed Circuit Breaker
In parallel with the study of the converter,
•a search for improvements in external pro-
tection was carried on. A higher speed breaker
first presented itself as offering the greatest
possibility. It was realized that such a
breaker would have to be radically different
from those available in that it must be many
S/civ Sp«e<^BreoAer
Timo
Fig 6. Curves Showing Relative Effects Tending to Product
Flashing, with High Speed Circuit Breaker and
with Usual Type of Breaker
times faster. It was anticii)ated that the cir-
cuit should be ojiencd in less than 1 120
of a second. The objective was therefore
])laced at about ().0(U> second, which is within
a half cycle, or the lime in which a commu-
tator l)ar passes from one brush U> the next
()()-CYCLE CONVERTING APPARATUS
395
r
>
l.jrush of opposite polarity. In this time it
was anticipated that the arc would not be
carried completely across and the energy
absorbed would not be sufficient to cause
serious pulsation.
Fifjs. 4 and 5 are records of direct current
and voltage and alternating current
during short circuit, with the high
speed breaker, switch, and slower
breaker only in circuit. In Fig. 4
the high speed breaker is connected
across a resistance that reduces the
current to 80 per cent load in 0.007
seconds from the beginning of the
short circuit. It will be noticed
that this load is steady until the
slower breaker opens and that the
d-c. voltage then returns to a
steady normal, indicating no pulsa-
tion or flashing. Fig. 5 is the same
except that there is no resistance
across the high speed breaker which
opens the circuit completely in 0.01
sec. from the beginning of the short
circuit. Both records show that
the flashing load (about 5 times
full load) was on for a small frac-
tion of the total time, being barely
reached in Fig. 4. With a stand-
ard breaker the load would have
reached about 25 times full load
for a much greater time. A com-
parison of the relative effects which
tend to produce flashing with the
high speed and usual speed circuit
breakers is shown in Fig. 0. The
relation of areas indicating relative
causes for flashing is very striking.
Without further details of the
development it may be said that two types of
breakers have been made to give the desired
speed, and the simpler one of these is now
available. A number of tests for all values of
short circuit up to the maximum that could be
obtained with only the necessary connections
for switches and circuit breaker have been
made with the improved design of converter
and high speed breaker with absolute freedom
from flashover. It may therefore be said that
the problem of preventing flashing of GO-cycle
600-volt converters from direct current short
circuit has been solved, and the necessary
equipment is now available for the first time.
Voltage Control and Commutation Regulating
Devices
The voltages at the collector and at the
commutator of the simple converter have an
approximately fixed ratio. To regulate the
direct current voltage it is therefore necessary
to have some means for regulating the a-c.
voltage. A special arrangement of poles in
the split pole type produces a change in ratio
within the machine, but few of these machines
jerf£t>^*rr
^
^A±J
ca^^£Kiraja
ca^t'e^r£s
x.
5
-i__P
CSfTJ^ifr^rftAf ^a/^nmoA ^£iwy
Fig.
7. Connections of Synchronous Condenser Employing Synchronous Booster
for Regulation of Voltage. Commutation is automatically controlled
by two-element contact-making relay
are now being built on account of complication
with the application of commutating poles.
The a-c. voltage is now regulated mostly
by means of the synchronous booster except
for the smaller ranges (under 10 per cent)
or where accurate adjustments are not needed.
About 10 per cent regulation is possible by
field control of the converter, with proper
proportions of field and armature windings.
With 15 per cent to 20 per cent reactance
the power factor at full load need not be
less than 95 per cent nor the corresponding
wattless current materially exceeded at other
loads.
The usual arrangement is to drive the
synchronous booster from the converter, in
which case it has the same number of poles
as the converter. In a few instances with
large machines it has been advantageous to
396 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 5
drive the booster with a separate motor of the
same number of poles, thus making a much
smaller high speed booster and eliminating
control equipment for commutation.
When the booster is driven by the converter
to raise the a-c. voltage to be applied to the
^"SL^Sn^r^
Fig. 8. Another Method of Maintaining Satisfactory Commutation When
Employing Synchronous Booster. This method employs rheostats
in the booster field and in the auxiliary commutating field
converter armature, it acts as a series gener-
ator requiring additional motor current
through the converter armature to drive it.
Converseh-, when the booster sen-es to lower
the a-c. voltage it acts as a motor and drives
the converter armature as a generator. These
additional motor and generator currents in
the armature give reactions on the com-
mutating pole which would seriously affect
the commutation if proper correction were not
applied. Several schemes for cancelling these
reactions are in use. Two of them are:
( 1 ) Since the additional current in the
rotary armature is proportional to the kilo-
watts that the booster is carrying, that is,
the product of volts and amperes, a two-
element balanced contact -making device
responsive to kilowatts in the booster was
devised. The wattmeter element contains
two coils, one carrying current from a series
transformer in the main load circuit and the
other receiving the voltage of the booster
across the corresponding phase. Balanced
against this element is another element carr}--
ing commutating field current in one coil and
constant source of excitation in the other.
When properly adjusted the correct balance
between kilowatts on the booster and auxiliav
commutating field current is obtained by the
operation of a motor-driven rheo-
stat in the commutating field cir-
cuit controlled by the contact-mak-
ing mechanism. Connections are
shown in Fig. 7.
( 2 ) Rheostats in the booster field
and auxiliary commutating field
circuits are mechanically connected
so that they keep the two excita-
tions proportional. This gives the
voltage element of kilowatts. The
current element of kilowatts is
obtained by another rheostat in
series with the auxiliary- commu-
tating field which is driven from
a contact-making relay balanced
against load amperes. Connections
are given in Fig. S.
Comparison of Motor-generators and
Synchronous Converters
The great amount of work that
has been done to meet operating
requirements with converters, in-
stead of motor-generators, must be
backed \y< good reasons other than
simplicity. The following com-
parison on efficiency, reliability and
flexibility, cost and floor space,
therefore seemed desirable to assist in arriving
at a choice of apparatus for a given sen-ice.
Efficiency
As there are many combinations of a-c.
and d-c. voltages between which conversion
is desired there are also different cfticiencics
of conversion for a given size of machine.
Cur\es were plotted showing the amount by
which the efficiency of the converter with its
necessary transformers exceeds that of the
motor-generator without transformers for
ilifforent a-c. voltages up to i:{,2(MI, the
assumed limit for motor voltage. Converter
curves included those with synchronous
boosters to give 20 per cent voltage adjust-
ment for lighting ser\ice, the range in d-c.
voltage being 240 300. The cur\-es for the
various conditions were all combined and lie
within the areas.
At 240 volts, or 10 per cent buck, the
converter efficiency may Ix- about 1 jK^r cent
less than given by the lower edge of the
GO-CYCLE CONVERTING APPARATUS
397
efficiency area (Fig. 10). For a-c. voltages
exceeding 13,200, which requires a trans-
former for the motor-generator, the difference
in efficiency will be about 2 per cent greater
than shown. The comparison of 6()0-volt
railway machines does not include so many
variations, in that the converter onlx' without
a booster is considered (Fig. 9).
Methods as defined by the A.I.E.E. niles
were used in the determination of these curves.
3
-
7
^^i
1
^
§6
\
V//
—
(~
—
1
4
^
y
.0
r?*
■^
^
p
777
7rr
h
iatt L
S
r
~
^
5x3
.1)^
>4
k
>».
^
/-
ullLodd
.V)
■u.
n
-
s
.5
.c
o^
/
n
1 1
^00
800 1200 1600
hw. Output
Fig. 9. Gain in Efficiency of Synchronous Converter with
Transformers ever Motor-generator Set for 600-volt
Direct-current Operation. Gain in efficiency varies
over shaded area for alternating-current line
voltage from 2300 to 1.3.200
compared to the usual higher voltage con-
nections of the motor
The greater the d-c. voltage range the less
efficient will the converter become and at
more than 25 per cent range will lose much of
its advantages.
Reliability and Flexibility
A simple converter has the same windings
as a generator, but reqviires a large collector.
The collector probably involves less risk to
interruption of service than the motor of a
motor-generator set. When a booster is
~"
,
\
7
'A
I
\
'/>
1
^
\
////A.
'//////
'/=
^
i"!
6
y///////
';^
-
-
U,
^///////////?.
W
^
-Halt Loaa
5
\
'i^////////////
^^^tU^
^
-^/////////////////yT^
<.
K//
'V/////////////////\
'///
A
■
■C
K^f^^^<-^rri
'-iM
■u.
4
v/Z^
'
'
-
...
H
yy//////
\a
>-.
I'll
—
-
\
'Zy////////777}
7-^
rr-
9-7"
W^},y.
u
&\
^
t^
^
w^izzz//
'/2/////777k
'"
~
~
F
^
^
(V^
<r
;^
[-
fi
^
aooo a'Kic
•^00
800 IZOO I600
Hw. Output
aOOO 2400
Fig. 10. Gain in Efficiency of Synchronous Booster Converter
with Transformers Over Motor-generator Set. For 240/300-
volt direct-current operation. Gain in efficiency varies
over shaded area for alternating-current line voltage
from 2300 to 13.200 with and without boost-
ers. Figures at 270 volts
Attention is directed to the fact that indeter-
minate losses omitted in the efficiency
calculation of the booster converter would be
greater than those of a motor-generator, so
that the actual gain given for the booster
converter would be somewhat less than
shown.
There will also be a small extra loss in
the heavy conductors from a converter to the
low voltage secondary of its transformer as
added to the converter it involves the same
risks as the motor of a motor-generator set,
and the converter outfit then has the added
risk of a large collector. Rather sharp dis-
tinctions must be drawn to show the advan-
tage of either. The converter will require
more attention to its large collector than is
necessary with the small collector of a
motor, but if this is given reliability is not
affected.
39.8 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No.
The direct connection of the armature
winding of the converter to both the a-c.
and d-c. systems is generally not as desirable
as to keep the systems separate, as with a
motor-generator. The ability of the motor-
generator to control power factor and d-c.
voltage independently is sometimes an impor-
tant feature.
In general, voltage delivered by a motor-
generator is subject to less sources of variation
+4U
—
—
—
—
—
1
tJO
\
\
\
%20
\
5
r
y/y/y
^
.t;
yyAV//
\
^HO
^Z^^A
k
X
^/ZV/
^,
?i
\
0<>yy//.
X\
Z4O/300 Volts OC.
,ii
-
\
v//yy/yy>>.
i ' < '
■^^
\
{////yY/A
^
<
y//yy/Xyyy.
^
A^^-^
-ii
<
yyAyxyxxA^y^yCA-z^^^A
f!
<y
(^xy>y^yyy>yy>9LA
^-10
s^
A6yycAy:>cyy/XA
s^
i2yyyyy>9zyXA
l""
"Y^
ZAVAlA
1
1
\
^Y-fi^
-20
^
U"
^7i^/777777777:tr,
■^
U-/
y/////////////////
^
V////////////////////A
-
-
-
-
-
^/ZA
-JO
^
'f^
1 6O0Volt OC.
^
1
1
400
800
IZOO 1600
Kw Output
2000 2400
11. Per Cent Difference of Cost of Synchronous Converter
with Transformers Compared with Motor-generator
Set (100 per cent). Cost difference varies over
shaded area for alternating-current line
voltage from 2300 to 13.200
generators. As the range in transformer
costs is rather wide, depending on efficiency,
method of cooling, and voltage, no attempt
has been made to give the additional per-
centage.
Floor Space
Fig. 12 is for a-c. voltages up to 13,200.
At higher voltages the floor space for motor-
generators with transformers would be about
15 per cent greater.
Conclusions
From the foregoing data the saving of power
by using the synchronous converter is quite
no
400
800 1200 '6O0 2000
Mw Output
Z400
Fig. 12 Per Cent Difference of Floor Space of Synchronous
Converter with Transformers Compared with
Motor-generator Set ilOO per cent!
than that from a converter. With steady
frequency the a-c. voltage fluctuations and
line drop do not affect the d-c. generator volt-
age.
Cost
The comparison of costs for both railway
and lighting machines (Fig. 11) is based on
the same combinations of apparatus as were
considered under Efficiency. For higher
voltages than 13,200, the cost of trans-
formers would be added to the motor-
evident and especially attractive when the
a-c. line voltage is over 13,200. • The cost
and floor space arc generally in favor of the
converter below 13,200 volts, and decidedly
so at higher voltages. Unless the ser\-icc
is very exacting or special, requiring wide
range of voltage control or high d-c. voltage,
the choice would favor the converter. This
is proved by the proportions of this apparatus
built in the past five years, which is about
2' 2 kilowatts of converters for each kilowatt
of motor-generators.
399
Design of a Superpower Station
Steam Turbine Generating Station of 245,000-kw., 300,000-kv-a. Capacity,
66,000-volt Distribution
By H. Goodwin, Jr.
Power and Mining Enginkkkim; Department, General Electru; C'omi'Anv
MECHANICAL DESCRIPTION AND STATION DESIGN
By A. R. Smith
Construction Engineering Department, General Electric Company
The proven economy of large capacity generating units operating at high steam pressure, and the even
greater economies to be obtained when a number of these are located in one station, are leading to the de-
velopment of power in large blocks. Under these conditions the old systems of cable distribution are inade-
quate and the introduction of high voltage cables of large kilowatt capacity is necessitated, introducing
further problems. In this article the authors deal with the design of a station to meet all these conditions.
This has been done by keeping unit design and flexibility as prominent conditions. Therefore, while the de-
sign was made for a particular location, it should be very widely applicable where economy calls for large
generating stations. — Editor.
The general tendency towards the con-
solidation of existing power plants and trans-
mission systems and the probability of the con-
struction of superpower lines will undoubt-
edly result in the erection of large generating
stations which can be operated at practically
full load throughout the year. Whether such
plants be located near coal mines where fuel
is reasonably cheap or in localities where the
transportation cost of fuel is considerable, the
fact that the load factor is high will justify
the construction of a most economical plant.
Aside from the economy of fuel based
entirely on its present value per ton, con-
sideration must be given to its extravagant
use and the possible value many years hence.
Another fact that is often lost sight of is that
the more coal consumed per kilowatt-hour the
greater must be the capacity of boilers, stokers
and coal and ash handling facilities; and the
greater the steam consumption the larger
the piping, condensers and water tunnels.
In brief, the cost of much of the apparatus
that is necessary only for economic reasons
may be largely offset by the reduction in cost
of the essential apparatus because of the
reduced demands on it.
The design herein described was developed
for a particular condition where some of the
fundamental considerations were: high fuel
cost, moderately good load factor, extreme
river floods, and high voltage underground
distribution. This design is of recent origin
and has therefore not been fully developed;
consequently some of the apparatus shown,
particularly the boilers, economizers and
pre-heaters, are proposed designs.
The more unusual features of the design are :
High steam pressure, 350 lbs.
High superheat, 350 deg. F.
Independent power supply for station aux.
Air pre-heaters for stokers.
All electrically driven auxiliaries.
Minimum overhead coal storage.
Simplicity of boiler room building.
Outdoor switch gear for G6,000-volt dis-
tribution.
Means of cleaning circulating water tun-
nels and possible utilization of circulat-
ing pumps in case of flood.
The ratings of the principal pieces of
apparatus are:
Seven main generators of 35,0()0-kw., 0.8-
p-f., 43,750 kv-a. capacity at 13,200 volts
3 phase, driven by steam turbines.
Seven 45,000-kv-a. transformer banks for
stepping from generator voltage to 06,000
volts, each bank composed of three 15,000-
kv-a. single phase units. One generator and
one transformer bank are designated as
reser\'e capacity.
Ten underground and two overhead feeders,
all at G(i,000 volts. The underground feeders
will each be composed of three single con-
ductor underground lead-covered cables and
will have an individual capacity of 45,000
kv-a. The overhead feeder capacity is
approximately" 10,(100 kv-a. each.
Coal Handling Equipment
Coal handling eqtupment has been designed
with the idea that the bulk of the coal will
be unloaded from barges by means of travel-
ing crane towers at the dock and transported
400 Mav. H)2(.)
GENERAL ELECTRIC REVIEW
VoL XXIIL No. 5
E
o
U
■^ii-:^:^^:r^;'>;^-^i\iii^^^^^-7':^-ii^^
r'?'f;i% Yjf '^.f^wy- -jit^ j-C^ i^K-^^ff^. '?" * '^'•y.. ^^fv-^^*^*-^ -■
;:..<i).'-t>^i?'-f-'»JV^»^*v-3».,-^\-^^»'A-^ V V
v^awMMC.-w^
DESIGN OF A SUI'lCRl'oWER-STATIOX
401
^1
I I I
I I I
I I I
J\
1111
by two belt conveyors directly to the four
receiving hoppers and crushers. Fig. 1 shows
the construction quite clearly. The duplicate
belt conveyors provide a large coal hand-
ling capacity when needed, but at the same
time the system is not dependent upon a
single conveyor. It should be noted that
there are no travelling trippers, as all belts
are dead ended. From the receiving hopper
the coal is delivered directly to the out-
side storage or through a crusher and ski])
hoist to the overhead outside bins. These
overhead bins are connected by an emer-
gency belt conveyor so that in ease of failure
of a skip hoist, a crusher or any part of
a receiving hopper tower, the crushed coal
can be transported from the adjacent over-
head bin.
The intention is that one operator located
in a control cab above each receiving hopper
will operate the revolving gantry crane,
the crusher, the skip hoist, etc. Another
operator will be located on each electricalh"
operated larry to transport and weigh the coal
from the overhead bins to each boiler. The
emergency coal storage handled by the loco-
motive travelling crane will be operated onh"
when the excess coal is being stored or
reclaimed.
It is proposed that there be one spare larry
which can be readily run into any one of the
four firing aisles to replace any defective
larry. It will be obser\-ed that the revolving
gantry crane reclaims the coal from the circular
storage without moving the bridge; thus, this
method is very rapid when reclaiming coal,
although in distributing the coal the bridge
will have to be moved slightly from time to
time, but to minimize this movement outside
thoots are shown on the four sides of each
receiving hopper tower. The revolving gantry
cranes overlap so that coal can be transferred
from one pile to another and, furthermore,
the design shown can be partially built and
extended from time to time without interfer-
ing with operation or without changing exist-
ing structures.
Ash Handling Equipment
It is proposed to dispense with all kinds
of ash conveyors, which are at best trouble-
some. The ash hopper under each boiler will
be of such capacity as to contain 12 or 24
hour storage so that ashes need be removed
only once or twice during the day. The ash
hoppers will empty directly into standard
railroad ears which will be hauled by a
storage battery locomotive.
402 .May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXin, Xo. .J
Boilers
The boilers (shown in Figs. 2 and 3) will
have a rating of some 1600 h.p. or 16,000 sq. ft.
of heating surface. The exact rating, of course,
will depend a great deal on the type of stoker
selected, the type of boiler proposed, etc. The
proposed separation of the two banks of tubes
with the superheater in between is suggested
for two reasons: First, to permit using a two-
pass boiler and keep the economizers on the
main floor, which means a greater number of
tubes in height ; second, to get a high amount
of superheat without an excessive amount
of superheat surface. The baffling of this
boiler is simple. All of the heating surface
should be effective, as there are no idle
pockets, and the draft loss, because of the two
passes instead of three, will probably be less
in spite of the fact that the boilers may be
several rows of tubes higher than the usual
standard.
Four boilers per turbine are shown, but as
there will be one spare turbine, there will
naturally be four spare boilers, and some
steam mav have to be transmitted through
the interconnecting steam header, depending
on which boilers are idle.
Stokers
The "extra long " underfeed stoker has been
shown, as the grate area must be commen-
surate with the increased heating surface
resulting from a ver\- high boiler. In this
case the demand for economy was prompted
more by the high price of fuel than by the
high load factor. Where the load is uniform
the stokers must have a greater relative
combustion area.
Blowers and Fans
With the use of economizers and pre-
heaters induced draft fans will be necessary-
because of the increased draft loss and the
low temperature of gas entering the stacks.
These fans might be of the ordinary- plate
type, or possibly of the multi-vane type,
because the temperature is low and the
pressure comparatively high. In other words,
with the introduction of preheaters in
addition to economizers a more desirable fan
Fig. 2. Section of Boiler House
DESIGN OF A SUPERPOWER-STATION
hi:)
H
•a
404 May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 5
requirement for induced draft purposes is
obtained than would be the case if the pre-
heater or the economizer were omitted.
However, the induction type of stack employ-
ing high pressure blowers may be substituted.
Both the induced draft fans and the stoker
blowers are in duplicate for each boiler,
although each would be of only half the
maximum capacity required per boiler; thus,
in event of failure of any fan or motor the
boiler would be operated at a reasonably
high rate with a possible slight reduction in
the fan and blower efficiencies.
Economizers
Wrought tube economizers of the same
construction as the boilers, that is, with
headers inclined with relation to the tubes,
are proposed. The economizers will be
practically the same width as the boilers;
thus, there will be no change in the sectional
area of the flue connecting between the
boiler and the economizers. The economizers
will be cleaned with steam soot blowers
instead of scrapers, and it is anticipated that
there will be no moist soot deposit because
the water entering the economizers will first
be heated to 150 deg. or 160 deg. with exhaust
steam, thus bringing the temperature well
above the dew point of the gases.
Special attention is called to the natural
thermo-siphon flow of water in both the
vertical tubes and the headers, and the
counter-current flow of the gases and the
water in the economizers; also, to the con-
venience of piping the feed water from the
headers in the basement through the econo-
mizers to the boiler drums.
To avoid internal corrosion of the wrought
steel economizer tubes it is proposed to
eliminate as completely as possible all air
from the feed water either in the condenser
or between that and the economizers.
Preheaters
The air from each turbo-generator is dis-
charged into a duct leading from the generator
room to the end of the boiler house. This is
shown clearly in Fig. 3. The tunnel is shown
in section in Fig. 2. The far end of this duct,
being open to the atmosphere, gives a free
discharge for the generators in case no
blowers are in operation, and any air required
for the boilers o\er and above that supplied
by the generator will Ix- taken in at this end.
From this main duct the air passes througli
the heating tubes in each preheater, the prc-
heater being divided into two units per
boiler to make a more practical design. With
this arrangement the boilers nearest to the
turbine room will burn the heated air dis-
charged from the generators, whereas the
boilers at the far end will bum the air from
outside. •
Preheaters have not been in general use in
stationary plants, although they have been
applied for many years on board ships. It is
believed, however, that they are perfectly
practicable, and if the cost of coal is at all
high, as in the case under consideration, they
can undoubtedly be made to show a good
return on the investment.
Piping
(Jn account of the high pressure and high
superheat involved, it is proposed to simplify
the steam piping as much as possible so as to
make the entire piping system more flexible
and to reduce the serious consequences of a
ruptured pipe or fitting. There are no steam
headers in the general sense of the word, but
there is an auxilian,- header or, better named,
a transfer header for equalizing pressures and
transferring steam between boiler rooms.
The omission of all steam-driven auxiliaries
except the house or auxiliar>' turbines, of
which there is one for ever*- two main units,
greatly simplifies the steam piping and
materially reduces the cost of the plant.
The boiler feed piping would resemble in
design the steam piping, inasmuch as there
is a group of boiler feed f)umps for each boiler
room. This piping, therefore, can be segre-
gated in a most advantageous manner and the
sizes of the pipes kept very small.
Circulating Water Tunnels
Fig. 4 shows a cross section of the turbine
room and below this the circulating water
tunnels. The elevation and design of water
tunnels depend upon the water level of the
river, conditions and kind of soil on which
the building rests, etc., and therefore would
probably be modified for each locality. On
account of the large size of these tunnels it
appeared best to divide them into two parts.
With two intake and two discharge tunnels
they should by all means be arranged so that
cither one can be taken out of service for
cleaning or for repairs. This statement
applies most forcibly to the intake tunnels,
which invariably fill up with sand or silt.
Sluice gates have tlierefore been shown
connecting between the intake tunnels and
a center chamber formed b\ the two tunnels.
There is one center chamber for each main
DESIGN OF A SUPERPOWER-STATION
405
turbo-generator; thus, any
one or all of the circulating
pumps can get suction from
cither tunnel. In the event
of a flood with the water
entering the power station
basement any one or all of
the circulating pumps could
pump this drainage by
simply closing the sluice
gates on both sides of any or
all of these center chambers.
The discharge tunnels are
connected together to reduce
the head loss, but if found
advisable these two tunnels
might be provided with iso-
lating sluice gates.
House Turbines
The very high steam tem-
perature involved with 320-
deg. F. superheat at 325-lb.
gauge at the turbines prac-
tically precludes the use of
any small steam-driven aux-
iliaries, such as boiler feed
pumps, etc. Since the water
in the boilers would be
evaporated to a dangerous
level in a very few minutes,
were the boiler feed shut off,
it is evidently necessary that
a most reliable source of
power be provided for the
supply of boiler-feed water.
The installation of low pres-
sure boilers to operate the
boiler feed pumps only might
be considered, but this would
have the disadvantage of de-
feating the unit-design of the
plant, and unless this auxil-
iary supply were made of
practically double capacity
it would not give the neces-
sary reliability. Further,
the necessity for the use of
some steam for the heating
of the feed water would
make a most complicated
arrangement were the auxil-
iaries to be operated from a
separate set of boilers and
their exhaust used for heat-
ing the main feed water.
Power for driving the feed
pumps is not considered
406 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. o
sufficiently reliable when supplied from the
main busses.
In order to provide for a very reliable
supply of power to drive all essential station
auxiliaries electrically and thus obviate the
necessity of emplo\-ing many small turbines
adapted to high pressure high superheat
steam, it is proposed to use one 2500-kw.
auxiliary or house turbine for each pair of
main turbines. Such a unit can be admirably
adapted to the steam conditions contem-
plated and will have sufficient capacity to
supply all such auxiharics as boiler feed pumps,
stoker blowers, stoker drive, induced draft
fans, and condenser auxiliaries.
Each house turbine will be provided with a
low jet condenser which will normally pro-
duce about fifteen inches of vacuum. The
circulating water for this condenser is the
condensate from the main units. It is pro-
posed to make this condenser design such
that the discharge water would have a
temperature as close as possible to the
temperature corresponding to the vacuum.
Of course, any vacuum desired can be
maintained, but where there are economizers
in the station it is most economical to heat
the feed water up to 1,50 dcg. F. or 170 deg. F.
by reducing to a minimum the steam con-
sumption of the house turbine.
The auxiliary power and therefore the load
on the house turbo-generator will not be
proportional to the load on the main units;
consequently, with a fluctuating quantity of
circulating water the vacuum will tend to
vary through quite a wide range. It is there-
fore intended that the house alternator be
paralleled with the main bus so that a portion
of the auxiliary load can be shifted automati-
cally or manually from the house alternator
to the main alternators to maintain a constant
vacuum under all conditions. This electrical
interconnection would be so made that a drop
in potential or a lowering of the frequence,
due to disturbances on the main system,
would automatically disconnect the two and
keep the auxiliaries connected to the house
alternators.
In order to provide for a constant flow of
water through the condenser of the house
turbine, some of the water may be recir-
culated. In other words, if a condenser is
designed for a ciuantitv of water equivalent
to •*4 load on two main units and only one
main unit is in operation, the circulating
pump for the house turbine condenser would
recirculate half of the water. This con-
denser obviouslv serx'cs as a feed water
heater. The tank shown just in front of
the condenser (Fig. 4), is a storage or surge
tank for the boiler feed supply. This per-
forms an important function because it is
impossible to feed the boilers at the same
rate as the condensate is being returned from
the main condensers. This surge tank there-
fore equalizes the discrepancy and prevents
o\'erflowing of hot distilled water and the use
of excessive cold, raw, or treated water.
There is not space here to dwell upon all of
the merits of a house turbine condenser.
Main Condensers
There is little to be said in connection with
the main condensers because they are of
ordinaPi- standard design, each being supplied
with two circulating pumps, two hot well
pumps, and two air pumps. The hot well
pumps are each of full capacity. The others
are of half the maximum capacity. It is
expected that with a reasonably tight con-
denser system only one air pump will have to
be operated at a time and in case of ver\'
cold water one circulating pump will be suffi-
cient. It should be borne in mind that one
circulating pump may give (30 per cent to
70 per cent of the capacity of two pumps
due to the reduction in condenser and pipe
friction resulting from the reduced flow of
water.
The condensers will be mounted on springs
to take care of expansion and to avoid
introduction of an expensive and undesirable
ex])ansion joint between the turbine and the
condenser.
Auxiliary Power Supply
All of the i)ower supply for the more or
less non-essential auxiliaries, such as cranes,
coal larrics, conveyors, lighting, miscel-
laneous pumps, etc., will be from trans-
formers connected to the main bus and
-suiJi^lied through a switchboard located in
the first gallery. All of the essential auxil-
iaries, -such as c ondenser pumijs. feed pumps,
blowers, etc., wi\I be controlled from switch-
boards located o n the main turbine room
floor at each auxiliary turbine. The oi)erator
at the switchboard will have immediate
control of all of the turbine room auxiliaries
within his vision and the control of the supply
of power to all of the boiler room auxiliaries
supiilying that jiarticular section of the
turbine room; thus, there will be four switch-
board operators on the turbine room floor in
addition to the main switchboard operators
in the galleries.
DESIGN OF A SUPERPOWER-STATION
407
Electrical Design
The first consideration in the electrical
design is the determination of the funda-
mental connections between the generators,
busses and feeders, that is, the "backbone" of
the system under normal conditions. This is
influenced by many conditions, chief of which
in a large station is the concentration of power
which it is considered advisable to allow at
short circuit. If this limit is vcr\- high the
rupturing capacity of switches available
must be considered, and in any case the value
of simplicity of arrangement must be balanced
showed that a switch ha\-ing a rupturing
capacitv of .'),s;{,(i()(l kv-a. would best suit the
scrA'ice.
At this point it should be noted that the
capacity of a generator is practically equal to
the capacity of a transformer bank and that
one feeder working at full capacity' would dis-
tribute the entire load of a generator and
transformer bank.
Use of Reactors
Preliminary calculations pro\-ed the neces-
sitv of sectionalizing the station bus by
ya^t tf Ttt ttt
temmnM^s^
ttnttCT'Ofrs
rxj
ttt ttt ttt _ii
^i^ — k \L —
wyv/v
vwvv,
vyvVw
^^''J^^ \'\y\^ v^-y^A >A'v\^
o o 0 6
U.1
■v^/vv\A/
www waAaa/
-K T T T
r-Rjo '^srof? "^er/rs •*~s:ooo ftfVfi
frr/fcr^/^cr
newr/acflcw cfT^r/^/fro^s /?f^fcroms
r;ffiw<p/fwKw (SS4aw727/fV fisficroms
AtrAc r^j^c r
07^
3i7,C£C <*A
•4-iS'O. 000 fCVA
^^7ecc KV ^
s%
^0*1 OOO/CVA
4-^ (OC Kkf.
/3%
■^^7 aaoKVA
^S<i,OOCK\ ff
■5/3_O0Of<Vfi
53ff_0COKVP
/0%
ss^./faa/cyfi
^TlCCCK^f^
/6l
^^O OOOKVA
IS «
u 1,
U3 s
/O /*
A^s-:c\ta
/S%
S'e4:Ot>OKI/H
sis.otfx^p
/e%
te/. OOO AT/9
^-{•6 Oi>0 KVf
13%
•*?or ooo /n/A
SOS.eUCKTiA
'S.Slt
lel,
■S'i)4; teoKYfi
10%
64^,OOOK\/M
Fig. 5. Comparison of Different Methods of Using Reactors
against the greater cost of switches to handle
heavier short circuits. On the other hand, if
the use of the largest switches available is not
considered, the A-alue of protective apparatus
to reduce short circuit intensities must be
balanced against the reduction in cost of
switches.
In this particular case one of the specified
conditions was distribution by means of
()(),0{)0-volt underground single conductor
cables. It was decided that the concentration
of energy in a .short circuit on these cables
should be limited to .')()(),()()() kv-a. on account
of the i)ossible resulting damage to adjacent
cables. Comparing this with the rupturing
capacity of available (i6,000-volt oil switches
reactors in order to approach the 500,000-
kv-a. limit which had been set. Both low
tension and high tension busses were con-
sidered, but the use of a low tension bus was
not only found unnecessary but it made the
problem much more difficult. The reactance
of the transformers added directly to the
generators before either were connected to the
bus would assist in reducing the short circuit
intensities very considerably. Many arrange-
ments of busses and reactors can be con-
sidered, but if sufficiently simple for ])ractical
operation all are reduced fundamentally to
the two forms shown in Fig. .">. The objection
may be raised that a ring bus is not shown.
Obvioush- a ring bvis would not change the
408 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 5
diagram shown at the left. Consider a short
circuit on a feeder supplied from the central
section of the bus in the diagram at the right.
It is evident that one generator supplies
current directly to this and that the three
generators on each side contribute equally.
Therefore no connection, either directly or
with reactors between the two ends of the
bus, would change the short circuit intensity
in this case. But such a connection would,
however, make each feeder in turn a "center"
feeder, whereas in the diagram shown the
short circuit intensities on the end feeders
would be considerably smaller than on the
center. Thus if a center feeder is considered
in the calculations the results will be good
whether the bus is made in a ring or not,
and the connection in a ring can later be
decided from the point of view of flexibility.
The above is true for an odd number of
machines; were there an even number of
machines the short circuit intensity would be
increased slightly by forming the bus in a ring.
The tables given in the lower half of Fig. 5
show all of the different conditions assumed.
The transformers in any case would have a
reactance of 8 per cent and it would not
appear to be advisable to consider increasing
this value. The generators would have a
minimum reactance of 12.3 per cent, and
under certain peculiar conditions this might
be increased to 28.5 per cent. Therefore, the
calculations were carried through for various
values of generator reactance. Inspection of
the table shows the minimum size reactor
which could be used in connection with any
generator reactance in order to limit the short
circuit intensity to approximately 500,000
kv-a. This reveals the fact that the syn-
chronizing bus shown on the left side of the
figure is necessary in order to limit the reactors
to a reasonable size. The reactors should be
rated to earn,- the total output of a generator.
Detail Arrangement of Busses and Switches
Having decided on the fundamentals of the
electrical system, the other features can
now be considered. These include arrange-
ment of switches to give the necessar\-
flexibility and reliability; the location of
transformers, reactors, switch gear and high
tension busses indoors or outdoors; arrange-
ment or location of main benchboard for
visibility of turbine room and outdoor switch
gear; provision for station auxiliary power.
All of these subjects have to be considered
in turn and their effect on each other con-
sidered in order to arrive at the final con-
clusion. Their relation to the mechanical
section of the station is also involved. It
was desired to keep a unit arrangement right
through the station, and this meant that the
switch gear for each generator and its group
of feeders should not occupy more length
than would be required for the corresponding
mechanical equipment of generator and
boilers. Fortunately these all work together
very well and allow enough space on each
generator stub-bus for the feeders required.
Main Bus Arrangement
The calculations considered above showed
the use of reactors to be necessar>', and also
the fundamental arrangement which should be
used. Fig. 6 shows the final complete solution.
The neutral point of each generator will be
grounded directly and positively. Since
there is no low tension bus, grounding of all
generators cannot cause circulating har-
monic currents between the generators. The
main leads of each generator arc connected
directly to the corresponding transformer
bank and through a main oil switch to a
stub-bus.
The transformer connection is delta on the
low tension side and V grounded on the high
tension side. It is particularly necessary that
the ground resistance be made very low so
that in case of short circuit on one of the
single conductor cables the neutral may not
be distorted, thus placing an increased volt-
age stress on all of the cables of the other
phases.
From each generator stub-bus three con-
nections are made; one through the reactor
to the synchronizing bus and the other
two to double feeder stub-busses. The three
selector switches connecting to the busses are
non-automatic. The main switch is arranged
for automatic opening in case of internal
failure in the generator or transformer. This
is accomplished by the use of relays dif-
ferentially connected around the generator
and around the transformer bank. Oil
switches are shown for breaking the syn-
chronizing bus at two points so that a section
may be readily cleared for cleaning, extension
or repairs. The stub-busses arc also arranged
to be connected together so that at times of
light load a small number of generators may
carry the load of the whole station without
feeding through the reactors. Referring to
Fig. (5: connection between the stub-busses
in the lower hne is provided for by oil switches,
and in the upper line by hom-typc air-break
switches. The oil switches would Ix" operated
DESIGN OF A SUPERPOWER-STATION
409
410 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. .3
from the main bench board and the air break
switches locally by hand throu<,'h permanent
levers.
Each feeder is equipped with two oil
switches to select either one of the two stub-
busses. Each would be equipped with induc-
tion type overload or other suitable type of
relay, depending on the detail connection and
interconnection of the distribution system.
It is thus seen that a ver\- complete and
flexible arrangement is provided. Xormally,
the generators are operated in parallel through
the reactors and synchronizing bus, but it is
possible by proper interconnection of the
feeder stub-busses to transfer loads from one
generator to another in almost any manner
desired.
lator. With this method no main field rheo-
stat is necessary. The exciter will be provided
with a field rheostat. For emergency excita-
tion two motor-generator sets are proposed,
supplying a sectionalized bus. These spare
exciters w-ill also be provided with automatic
regulators. The motors driving the exciters
will be supplied from the main station power
board but will be controlled from the main
auxiliary board opposite the bench board.
Indicating lamps on the power board will
show the position of the switches so that there
may be no danger of the operator of the
station power board pulling disconnecting
switches or otherwise interrupting ser\'ice to
a motor-generator exciter set while in opera-
tion.
Excitation
Each generator will have a direct-connected
exciter controlled bv a TA automatic regu-
Bench Board
For the control of the apparatus shown in
Fig. G, the bench board shown in Fig. 7 is
i+
cQQQa
oQQQd
a D D Q
D
III III III III
III III III III III
O 30 O O O
D
QQQ
D
D
III m
III III III III Ul
OOP opo
gnnng
D D D D
D
III iti
Ul III Ul m lit
OOP ooo
uQQQo
C I: Jl- Jr.- J G
D D D D
mm
D
m in
m in
ooo
Ul m
lu m
ooo
vW vW
D
w tu
III in m m ni
]QQQ[
UuJ:.Lk.J:
D
oQQQd
dQQQo
dQQQd
Wv VW
D
Fig.
7 Main Bench Board for Control of Main Gcneratora, Transformers and Fee *ers. Above: Front View. Center: Plan
of Bench Showing Mimic Bus. Below: Back Vcw
DESIGN UP A SUPERPOWER-STATION
411
proposed. This shows the front \-ie\v, the
plan of the bench with the mimic busses,
and the rear view. Opposite to this would
be placed a vertical au.xiliary board on which
would be mounted the automatic regulators
and field control switches.
Auxiliary Power Supply
As prcvioush- outlined, it was determined
that turbines for auxiliary power would be
necessary and that these should be connected
with the main busses through a central
auxiliary switchboard and transformers. A
detailed study showed that one auxiliary tur-
bine could best be used in connection with
two main turbines and that these auxiliar>'
units should be rated at 2500 kw., 0.7 p-f.,
;]o70 kv-a.
For the control of the auxiliary
and the supply of essential auxiliary
was found necessary to provide a swi
near each auxiliary turbine and
general auxiliary power a central swi
located on the switchboard gallery
cables to the four auxiliary units.
Central Auxiliary Switchboard
Fig. (i shows two banks of station power
transformers. It is proposed that each ban
be composed of three single-phase 12.5()-kv-a.
units, making a bank capacity of .'JToO kv-a.
412 May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 5
connected delta-delta. These transformers
would be located out of doors. The high
tension switches would be controlled from the
main bench board and the low tension switches
from the auxiliary power board.
Fig. 8 shows a single line diagram for the
central auxilian.- power board. The react-
ance of the auxilian,- transformers must be
considered in connection with the reactance
of the main units in order to limit the con-
centration of energy on the station power
bus to a value which can be handled by
reasonably small switches. Reactors are
required in each of the four feeders con-
necting to the turbine auxiliary switch-
boards to limit the short circuit intensity
central auxiliary board wotdd be a vertical
board with mimic busses and electrically
operated remote control oil switches, as
shown in Fig. 9.
Station Lighting and Low-voltage Power Supply
Fig. ID shows a single line wiring diagram
of the lighting system proposed. Three 7o-k-w.
single-phase transformers controlled inde-
pendently and connected for 3- wire, 115/230
volts would form the supply. It is proposed
that all lighting feeders be 3-wire, 115 230
volts, and be arranged in three groups which
will normally operate on different trans-
formers, but which in case of failure of a
transformer can be manuallv thrown to one
Fig 9. Front View of Central Auxiliary Vertical Control Board
and permit the use of small rupturing capac-
ity switches. It is proposed that these tie
feeders to the turbine auxiliary boards be
protected by differential relays connected to
current transformers at each end of each line.
In case of trouble on a cable it will be auto-
matically disconnected and will allow the
auxilian,^ turbine to continue supplying the
essential auxiliaries without interruption.
Reverse energy protection at the turbine
auxiliary boards is also provided so that in
case of a severe drop in voltage or frequency
on the main system, the auxiliary turbines will
be automatically disconnected and allowed to
run independently, carrying the essential
auxiliaries.
Other feeders are shown for the supply
of the pumps for the transformer cooling
water, coal handling apparatus, spare exciters,
station lighting and low voltage power. The
of the other transformers. Fig. 10 also shows
a connection from the 125-volt d-c. board,
which will control the storage batten.- for
circuit breaker control, so that in case of
failure of service on the a-c. lighting system
certain emergency lights will be automatically
connected to the battery.
Fig. 1 1 is a single line diagram of the low
voltage power board, supplied by two 150-
kv-a. transformer banks and operating at 240
volts, 3-phase. This board is arranged for
miscellaneous power supply, such as the
machine shop, turbine room cranes, house
supply pumps, etc.; also for the supply of
motor-generator sets for charging the storage
battery and a spare motor-generator exciter
set for the auxiliary- turbines.
Fig. 11 also shows the storage batten.'
proposed and the control circuits for operating
the oil switches.
DESIGN OF A SUPERPOWER-STATION
413
\J2
I ! I
r^
U U U
Turbine Auxiliary Boards
Fig. 12 gives a single line
diagram of boards proposed
for Nos. 1, 2 and 3 auxiliary
turbines, each of which will
operate in connection with
two main units. Fig. 13 is a
single line diagram for board
for No. 4 auxiliary turbine,
which will operate in connec-
tion with a single, main unit.
Each auxiliary generator is
to be protected differentially
just like the main generators.
The tie tothe central auxiliary
power board has been pre-
viously discussed.
Two circuits are provided
for the blowers and fans for
each set of four boilers. As
each boiler has two blowers
and two fans, one of each would be connected
to each circuit. This should insure the oper-
ation of at least half of them, which would
allow the boilers to operate at a high per-
centage of their rating at all times.
Since stoker motors are of small size, it is
o
o
o
fOOA r££0£fi&
\ \ I
i i i i i i
ff-
t t I t t "
I t f T f r
~]
i Hi
1— r
t t t t t
tooA rc£ro£fPs
- ra cMrmsf^cY .
4
£
70 irs V oc
Fig 10. Single Line Wiring Diagram of Station Lighting System
proposed to install two transformer banks, each
consisting of three 37.5-kv-a. single-phase
transformers to supply them. A double circuit
arrangement quite similar to that proposed
for the blowers and fans is suggested with a
double-throw switch at e^-ch stoker motor.
VOL Tfocrrft
jrv/TC^* J
o
T
1
f?£C£^fr^CLr
1
LOW yOLTAG^
^<fi BUSSES
I I I I I ' I I !! I I II.
i i i i i i i i i i i i i i i
CO/^PCMSA TO/f
ftoTO/t ecf/cgA rofr
GO GO
i
o
oi <o
^W'TCMf
o
yoL Tf^rrfft
o
I
o
1 I I f I ! !
I I 1 ! I I I I I 1 I If I T
1 (((((((((((
CO/^TROL BOARO
-i-« ccuvTPoi- c//eci//7S--~
-^
^cer r/G to.
Fig. 11. Single Line Wirinjj Diagram of Low Voltage Power Board and Direct Current Control Board
414 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 5
ps
I
-El3w*r
U
1/ ^' "
If ^
.JaUgi
?
^
I
s
2
3
<
I
•s
i
S
I
»
£
r^oo
i ^
DESIGN OF A SUPERPOWER-STATION
415
i
41(j Alav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. .5
It is proposed that the circulating pumps
be driven by 2300-volt motors with the
neutral brought out for differential protection
and that this protection be provided from the
oil switch on the switchboard to the neutral in
the motor, thus also including the cable sup-
plying the motor. This system will insure the
operation of the circulating pump motors
under all conditions of voltage and frequency
fluctuation. For the various other auxiliaries
for the turbines, two circuits are proposed
with overload protection, the individual
auxiliaries to be provided with no overload
protection. For the smallest auxiliaries a
lower voltage will probably be necessan,^ and
it is suggested that adjacent to these auxili-
aries for each turbine a bank of transformers
be installed for stepping down to 240 volts.
It is proposed that boiler feed pump motors
be made for 2300 volts and protected dif-
ferentially, including their cables, in a manner
exactly similar to that employed for the
circulating pumps, so that as long as there is
any power, either from the main bus or from
the auxiliarv- turbine, it will be possible to
maintain feed to the boilers.
Boiler Room Auxiliaries
It is to be noted in the proposals above that
complete alternating current drive has been
provided for the boiler room. Some of the
most recent installations have adopted this
method.
Elimination of apparatus for conversion
from a-c. to d-c. should increase the reli-
ability. For the blowers and fans, brush-
iMftnw^a ^Mftf^rth
9^^0- — I /&a.a—
r-'^r*.
KfiAmva
m^
\
Fig. 14. Various Sections of 66.000-volt Outdoor Switch Qcnr
DESIGN OF A SUPERPOWER-STATION
417
shifting motors are suggested. For the
stokers, four-speed multi-speed motors arc
suggested, and if a stoker with mechanical
arrangements for two speeds is used, six
speeds of the stoker may actually be obtained.
If it were decided that direct current for
the stoker motors is necessary, it is suggested
that two synchronous converters be installed
in connection with the main power board,
and that duplicate feeders be run to each of
the auxiliary boards for supplying the stokers
at 230 volts. If direct current is necessary
Fig. 3 shows the main and auxiliary units
in greater detail, and also shows the location
of the auxiliary board adjacent to the auxil-
iary unit. These points are again shown in
Fig. 4, which is a section of the turbine room.
This figure also shows a section of the gal-
leries. At the top there is a space reserved
for a load dcspatcher's room. Below this are
the main bench board, station service board,
switches, transformers, etc.
Figs. 14 and 15 show the outdoor station
in greater detail. Fig. 15 shows the section
f
Fig. 15. Plan of Unit Section of 66,000-volt Outdoor Switch Gear
for all of the boiler room auxiliaries, a s}n-
chronous converter could be installed in con-
nection with each auxiliary turbine an 1 one
spare converter in connection with the main
power board and connected to the others by
emergency feeders. This system would also
supply 230 volts at the motors
Arrangement of Electrical Apparatus
Fig. 1 is a general layout of the whole
station, showing the main generators, and
alongside of each pair the auxiliarv' turbine
generator. It shows the extension to center
of the building to provide operating galleries
for the benchboard, main auxiliary board,
etc. The general arrangement of the busses
and switching apparatus outside is also shown.
corresponding to one generator and the
characteristic arrangement for an overhead
feeder, an underground feeder, station service
transformers, and tie to the next section of
each bus. Fig. 14 shows sections through the
outdoor structures at these various points.
Reactors
The reactors to be used are of particular
interest on account of the high voltage and
use out of doors. For these reasons an oil-
cooled type of reactor is recommended. This
consists practically of a set of transformer
coils of the proper number of turns and current
capacity, arranged and held just as they
would be in a transformer, except that the iron
core is omitted. Such units would be water-
4 IS Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 5
cooled like transformers, but the quantity of
water required would be comparatively small.
Test Bus
The high voltage of the cables proposed
necessitates particular arrangements for test-
ing them on installation and after repairs.
Also, it is considered that it would be better
practice to test any cable before it is put back
in sen'ice, after the switch controlling it has
opened automatically. This means that
arrangements must be made for making quick
tests. To accomplish this a testing set is to
be placed in one of the galleries and from
this a connection made to the test bus out-
doors, shown in Fig. 15. Disconnecting
switches are located at convenient points
along this test bus so that flexible cables can
be carried easily from one of them to the
cable to be tested. This arrangement, con-
sidered in connection with the fact that the
operator can view the whole switch yard,
should make testing of cables and restoration
of ser\'ice most expeditious.
Outdoor Structures
The outdoor structure has been planned
with a view to greatest convenience of opera-
tion, low maintenance cost and low first cost.
Figs. 14 and lo show the structure in sectional
plan. It is proposed to use concrete poles
with structural steel members connecting
them at the top. This design results in a very
rigid structure from which the wiring is
supported at frequent inten-als. The break-
ing of an insulator will not result in dropping
a bus or a connection and such insulator
can be replaced with the greatest facility.
The few structural steel members will not
require frequent painting. To provide for
reaching the disconnecting switches easily
a plank walk raised slightly above the ground
is proposed. This will make it possible to
reach all of the disconnecting switches with a
standard insulated switch hook, and it can
easily be kept free from snow and ice in the
winter.
Railroad tracks are run through the vari-
ous aisles to facilitate the removal of trans-
formers and switches. In this connection it
is to be noted that the switches are also to be
provided with trucks similar to transformer
trucks, so that a whole switch may easily
be slid out from its position and taken to the
repair shop in the station. To facilitate
inspection of switches it is planned to have a
truck equipped with an oil tank, a blower
and a pump so that any switch requiring
inspection may be quickly emptied and the
oil fumes blown out of the tanks.
Conclusion
While it is stated in the introduction that
this station was designed to meet certain
specified conditions, it is felt that the design
described is such as to be ver>- generally
applicable. This is important for local
conditions, as any plant may change slightly
from time to time, and the more flexible and
generally adaptable the arrangement the
better able it will be to meet any slight
change in conditions. This same point also
has the advantage, when combined with unit
arrangement as in this case, of making the
design applicable to other changes in details
such as higher transmission voltage, and in
general, for the conditions of almost any other
location requiring such a large station.
419
Some Corona Loss Tests
By W. \V. Lewis
Power and Mining Engineering Uei-artment, General Elf.ctric Company
This article records the results of corona loss tests on a 150-mile transmission line at potentials up to about
200, ()()() volts. Comparison is also made between measured and calculated losses. Such tests establish con-
fidence in our formulas for calculating these losses and are es])ecially apropos at this time in view of the current
discussion of operating voltages in the neighborhood of 220,000 volts. The e.\tremely high losses found in some
of the tests are evidence of the economic importance of such tests. — Editor.
The theory of corona formation and for-
mulas for its calculation on transmission lines
have been fully and carefully worked out by
Peek.* It is interesting, occasionally, to
measure these losses on an operating system
and to compare the measurements with the
calculated losses.
Many existing systems, especially among
the earlier installations, have considerable
corona loss at the operating voltage. It is
now generally recognized that this loss costs
money, and the later installations are usually
designed to operate above the corona voltage.
.Tunctiort
nam
Granil Rapids
47SMi ,
15 5'- 12- 19 6 Spacinq
N»0 Copper
3-3000 KV-a
Fig. 1. One-line Diagram of Consumers Power Co.
Transmission Line Used in the Tests
Some corona loss tests were recently made
on the 30-cycle 140,000-volt system of the
Consumers Power Co., in western Michigan.
It is believed that these tests were carried to
a higher voltage than has been heretofore
* Trans. A.I.E.E., 1911, 1912. 19i:i.
attained on an actual transmission system,
and in view of the present discussion of 220
kv. or thereabouts as the possible next step
above the present voltage limits, these tests
should be of interest, as they were carried up
somewhere in the neighborhood of that
\-oltage.
T[
t\
fi<in\ —
LiA C
"-L
srn?
Fig. 2.
Connection Diagram of Instruments Employed
in the Tests
The transmission line on which the tests
were made extends from ^ Juiiction Dam to
Grand Rapids and Kalamazoo, Michigan, in
a direction almost due south, a total distance
of approximately l.")(J miles (Fig. 1). The
portion from Junction Dam to Grand Rapids
consists of three conductors, each of seven
strands medium hard drawn copper, total
cross-section 110, (JOO cir. mil. The conductors
are spaced practically in a vertical plane 12
feet apart. The distance is 101.5 miles.
This line has been in service about two yeairs.
The portion from Grand Rapids to Kalamazoo
is older and operated for a number of years at
70,001) volts. It consists! of 47.3 miles of
No. 0 copper arranged mainly in a triangle
with sides, respectively, 13.5, 12, and 19.(5
feet. The line throughout is insulated with
10 disks in suspension and 12 on strain. The
tower spacing is about 530 feet. The height
of the lowest conductor in the vertical arrange-
ment is about 40 feet at the tower and al>nf.t
420 Alav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 5
26 feet at the middle of the span. The
average elevation of the line is about 750 feet
above sea level.
Switches marked .4 and E, Fig. 1, were
open throughout the tests, thus separating
the transmission line from the Grand Rapids
and Kalamazoo busses. Switches B, C, and
^200]
45 1800
40 1600
iS 1400
b JO ^izoo
9- *
^
zs 1000
IS fiOO-
10 400
zoo
60 go 100 '20 , 140 160 /SO,
Lotv Yotto^ Kv. otJunctioniHighyolto^EqwmlcnQ
Fig. 3. Corona Loss, Line from Junction Dam to Grand
Rapids. Transformers on at Grand Rapids
.junction Dam transformer ratio l;i.").000:7.JUO.
Grand Rapids transformer ratio (Tests 9 and 13) 126,000:
7500.
Grand Rapids transformer ratio (Test 17) 140.000:7500.
TABLE I
TEST 9
TEST 13
TEST 17
H.V.
H.V.
Kw.
H.V. H.V.
Kw.
H.V.
H.V.
Kw.
Kv.
Amp.
Kv.
Amp.
Kv.
Amp.
74.6
10.8
84
82.0
11.8
125
81.6
11.8
87
80.7
11.7
101
92.4
13.2
151
84.3
12.2
87
94.6
13.5
142
99.8
14.7
168
101.3
14.6
131
102.0
14.3
162
109.0
15.0
247
109.6
15.0
160
111.0
15.2
188
122.3
15.7
.594
119.2
16.5
192
126.0
15.5
336
136.5
15.4
1225
126.0
16.8
235
137.5
14.3
605
145.3
15.2
1010
137.2
17.1
384
145.4
12.1
964
152.6
12.7
2082
145.3
16.4
730
154.0
10.4
1548
154.8
14.1
13.58
155.8
9.4
1730
160.2
12.7
1800
161.1
8.7
2090
10 Z800\
Z700
65 1600
Z500
60 Z400
Z300
55 ZZOO
ZIOO
50 ZOOO
.900
tS IgOO
non
40 J600
•<I500
55 1400
1.500
50 'ZOO
1100
Z5 1000
onn
I
Gen Junction
Test 10 Test 15 Testis
Date 10/15/19 10/16/19 lO/n/ig'
Bar. Z9.34" Z938" Z9.40'
Temp,. 44 T°F i39°F 54 TF
Hum. 91 40/0 944°/o 91.0%
Weit/Kr Clear f^^'ny Cle^r
600
700
If 600
500
ig 400
5 ZOO
100
0 0
60 SO 100 IZO , 140 160 ISO,
Low Voltage Kv. at Junctionwifh Voltage CgunaletiQ
Fig. 4. Corona Loss. Line from Junction Dam to Grand
Rapids. No transformers at Grand Rapids
Junction Dam transformer ratio 1.'15.0(X):7500.
TABLE II
TEST 111
TK>r !.*>
IKST IS
H.V.
H.V.
Kw.
H.V.
"^' Kw
Amp. '^''
12.6 66
H.V.
"■^' K«
Amp. "^^
Kv.
.\mp.
11.3
Kv.
Kv.
77.8
75.3
51
84.6
11.6 5;{
94.0
14.1
84
94.0
14.1 87
91.1
13.4; 70
103.0
15.2
102
101.0
15.0 114
ia3.9
15.2| 88
110.0
16.1
134
110.4 16.3 250
114.1
16.5
105
118.8
17.3
175
120.2 17.5 .5;i8
125.6
17.7
157
128.5
18.2
336
129.8 18.8 806
130.0
1S.6
270
135.7
18.8
500
143.8 20.7 1,").58
146.9
18.9
805
145.8
19.0
1017
UW.O 21.2 2030
l,5;i.O
18.8
1230
K')4.e
20.1
1655
102.1 21.2 2690
1,58.4
18.5
1600
160.2, 19.8
2112
SOME CORONA LOSS TESTS
421
1 1 1 1 1 1 1 1 1 1 I 1 I 1 1 1 1 L ' ' 1
as J400
1 1 1 M 1 1 1 1 1 J 1 1 1 1 M 1 i7f>j^t/?. \
J
~Pi Lr
U ^ ~~ ~'
H *^ 1 "^ 1
\ <L t
aenJm
-^ >J-r
75 3000
dm
H3t^m3Z00 1
^ ^ I
< <1 1
—iJ nI I
10 2500
^"X
14
Grand Rapids W
65 2600
<H
Date iOJISl/9 lOi
Bar. Z9.3Z'' 2S
Temp. 47$'r 46
Hum 90. Z% 34
WeiUKr Clear Ra
l|
60 2400
If
HZ Test /a u
I6II9 101 161 19 It
5&" 2938" t
55 2100
VF 46 ZT t
SO zooo
H
In
W
n
'^iS^ldOO
f
0
V
n
^
■
^iO 1600
W
thF
g
35 1400
a
n
Ifl
H
30 1200
if
I /estd
1 1 j 1
1 -J*
25 1000
-^--±:^ Test/4
J "T^ /
^ 3b -F -
LineC
urren j '
20 600
■
:^e
^ /S 1^
, t*
It--
I ^ \
" 1 J 5 S.Tes.':.:
15 600
t I Ji -
ff J TrsA/
u 2
10 400
7
' r
Te.s
,sS
5 ZOO
M.*:'
^^'^^
f
■ 1 III
I"
0 c
■" 1 III
t_
I
.0
w
/e
so 100
tage Kv it
IZO 140 ISO 160
Junction (High Yottxe £i^uirjient.
Fig. 5. Corona Loss, Line from Junction Dam to Kalamazoo.
Tests 8 and 12; transformers on at Grand Rapids and
Kalamazoo. Test 14; transformers on at Kalamazoo,
off at Grand Rapids
Junction Dam transformer ratio 135,000:7500.
Grand Rapids transformer ratio 125,000:7500-
KalamazoQ transformer ratio (Test 8) 120,000:7500.
Kalamazoo transformer ratio (Tests 12 and 14) 130,000:7500.
TABLE III
75
70
55
50
r
3200
3000
2S00
ibOO
2400
2200
2000
^laoo
40
35
SO
25
2C
IS
1600
1400
1200
1000
100
bOO
1 1 1 J 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 M 1 1 1 1 1 1 1 1 1 1 1
U ]A\A ' \----.z
M. <J*s.l . t
Gen Junction HoJamazoo Test II
Test 7 Test II
Dote W/15/19 10/16/19
Bon 29.32" Z93S"
Temp f/.ov te.o'r
Hum S3. 9 is 96. ait
\
......._ q.
. . _ - . — . J
::___._:_ . i
1 ^, ,
r Total
... - - , J. .
n
: : :: ; :;. rr_ :
tt
I - -
±
: : ji-
- - -X-
tX--
XX
: : . . ;.. ir:.
. . - I - T,^y
I A'* -r i II
' " TZ~'
± A-t .
if - .
tl
.. : -,*| -
. . .- y J .
: : ::;?.:;/_ _ _
/ C- -
r \
- -4- . i
i - --, t
2 t
7 X
" 4 i.
-- I i
• r
i -i
::: z - --
/.,
, '
t,'
... ,-f
.,ss'-- -
■3 '
60 SO WO 120 14C /60 180
Low i^of Cage Hv. at x/unction (High Voltage Ena/ya/ent)
Fig. 6. Corona Loss, Line from Junction Dam to Kalamazoo,
No transformers at Grand Rapids or Kalamazoo
Junction Dam transformer ratio 135.000:7500,
TABLE IV
TEST 8 1
TEST 12
TEST 14
H.V.
H.V.
Kw.
139
H.V.
H.V.
Kw.
H.V.
H.V.
Kw.
Kv.
Amp.
Kv.
81.8
Amp.
Kv.
Amp.
73.8
16.2
17.6
164
80.0
17.4
115
81.6
17.7
171
92.4
19.4
227
93.3
20.1
177
94.0
19.5
245
99.6
20.6
267
100.8
21.4
233
103.0
20.4
270
112.0
21.8
477
109.4
22.8
329
112.6
21.4
400
120.6
22.0
796
119.5
24.3
705
121.9
21.2
512
127.1
21.9
1058
128.9
25.4
1160
128.9
20.1
658
135.4
21.2
1516
134.6
25.8
1495
135.0
18.2
874
145.2
19.1
2280
144.2
26.5
2200
141.5
15.2
1195
1
152.6
156.0
16.7
15.9
2970
3430
153.0
27.0
3110
TEST 7
TEST 11
H.V.
H.V.
Kw.
H.V.
H.V.
Kw.
Kv.
Amp.
Kv.
Amp.
73.4
16.0
85
80.2
17.5
86
79.2
17.4
89
95.6
21.1
157
93.1
20.7
155
103.6
23.1
248
98.5
21.6
168
114.8
25.3
470
105.8
23.5
189
125.6
27.5
863
116.0
25.4
274
134.6
30.3
1345
124.2
27.4
418
145.0
32.6
2275
131.0
29.1
656
152.6
33.8
3030
141.1
31.7
1510
158.0
34.6
3590
142.7
32.1
1672
146.0
32.7
1880
152.1
34.3
2615
422 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo.
D were open on occasion. The loss readings
were all taken at Junction Dam. Fig. 2
shows the manner in which the instruments
were connected.
Two series of tests were run:
(a) With the step-up and step-down trans-
formers connected delta-delta, which
is their normal method of operation;
and
(6) With the step-up and step-down trans-
formers connected delta low-tension
and Y high-tension, thus allowing the
line voltage to be increased 73 per
cent with the same low-tension voltage
applied.
so 100 120 140 160
Low Vottege Hv. dt Junction (Hr^ Volttjye tgut^aient)
The transformers at Junction Dam and
Grand Rapids are duplicates, each bank con-
sisting of three 5000-kv-a. 30-cvcle units.
The transformer bank at Kalamazoo consists
of three 3000-kv-a. units. The low-tension
voltage of all banks is 7500; the high-tension
140,000 with taps for 135,000 130,000 125.-
000, 120,000. In all tests under (a) the step-
up transformers were connected 135,000
volts delta; and in all tests under (b) the
step-up transformers were connected 1 20 -
000 208,000 Y, thus giving 208,000 volts bv
ratio with 7500 volts applied on the low side.
The results of the test are shown in the
accompanying tables and illustrations. The
tables in general give the high-voltage kilo-
volts (i.e., low-voltage kilovolts times ratio),
the high-voltage amperes (i.e., low-voltage
amperes divided by ratio), and the measured
kilowatts loss. In Table VI the high-voltage
kilovolts at Grand Rapids by ratio is also
given.
(
j
_..
1
1
]
I
■
'
1
1
tap -j
'
■
1
1
1
!
1
/
1
i
1
r
;
, ' j
;
/
f
1 A
y[=
11
1
■ '■
_^
\_
y
w
[_
r
1 ,
£j
)//?/- Tf>tn</hr.'rM'r..LA
/
.
1
1 1 1 t 1 1 1 1 1 1 t 1
^.Jj
■
1
1 I 1 1 ! LI 1 ; 1 I 1 1 1
■ )
— y -
/J ibO '
__u
11
^bciSias^ lOrvVo/Uge i^/es X208OQO/7^O(/^
i
7
.
*
1
/ i
1
1
,
I
i
1
A
4 — 1
,
1
1
;
i
(/
^
/
_,
1
1
-J
^ "
'
-
1
1
'
'
,,.'
^
1
I
1
rJ
■^ />
^
1
.
— .-
/
L
,
*f*l
j^
=<
^
''
a* w -
_
-_
1 *-
^4-
,
-
--
-4=
f*" '
*=4-t-
isr-
r^
^^
— .
~ ^
'
T" ~^
1
H-"
1
' 1
1
J 1 , 1 1
1
1
' )
60
60
10
fer
f:i
M
^
H
Iff
'i
w
210
ZIO
, A
i;
I
Figs. 7-a and 7 b. Excitation Curves of Transformer Bank at Junction Dam
140,000-7500 volts
Trinsformcr ratio, delta connection l.io. 000:7 JOO.
Transformer ratio. Y-connection 1 20.000/208.000 Y:r.TOO.
L OH Yoltage first Junction (Hign \foit3gt Sgumj/tntJ
Three transformer!, 30-cycle, 5000. kv-a
TABLE V
TEST 6
H.V.
L.V.
Kv.
Volts
Delta
Conn.
.-ner,
57.0
3640
65.5
4600
82.8
5310
95.6
5920
106.6
0200
111.6
H.V.
Kv.
Y-conn.
87.8
101.0
127.5
147.3
164.2
172.0
L.V.
Amp.
4.0
5.0
6.0
7.5
10.3
12.6
H.V.
Amp.
Delta
Conn.
0.22
0.28
0.33
0.42
0.57
0.70
H.V.
1
Amp. .
Kw.
Y-conn.
0.14
20 1
0.18
24
' 0.22
39
' 0.27
50
0.37
62
0.45
1
69
H.V.
Kv.
Y-conn.
6470
6950
7520
8000
8280
8480
116.5
179.4
125.0
192.7
135.4
208.4
144.0
221.8
149.0
229.8
152.6
235.2
L.V.
Amp.
H.V.
Amp.
Delta
Conn.
14.8
0.82
20.6
1.15
31.3
1.74
49.4
2.74
63.5
3.5:i
79.7
4.43
H.V.
Amp.
Y-conn.
0.53
0.74
1.13
1.78
2.29
2.87
K«.
77
89
109
138
158
178
SOME CORONA LOSS TESTS
423
1
T^.^f?/}-.^ 1
1 1 1 1 V-^"'
\\\\t42
1 1 I 1 i' 1 1 1
L
{ <i "
t% ^
:> l>'--i:~z
•n Junction
Or^^ncl R^n,n. VW V. ZK'LA
1 M M 1 1 1 1 1 1 1 1 1
Ml 1 in : '
it
1 1 1 1 1 1 1 1 1 M 1 1 1 1
11 1 1 1 ffl r
Test ZO TasL
Oate Oct. 19. ISIS Oct 26
Bar 29.62" 2932
Temp. 51 2° r 41° r
Hum 33 % 94 %
Weather Clear Ciouc
A| 1 J
I9IS T
^0 6000
~---j------
/„ - T
- »
1
i
^-" "
t
TestZO
J«<-
A Jf
7 «
[ ^^.isUZ
1* "^
4. - .e ..^-^
X i
i / -
s
L ^ ■Line
^
P f Current
^
t / '
If
V 4
X
J
+.7
,
/
J L
J
t
i
X
*'-
f -
f
,■'
'
I
i
/
1
T
_
—
-
{L
-
_
0 0~
s =
s
•■
- -
-
-
eo
SO
100
li
0
M
0
160 ISO ZOO Zzl)
Low Volt j^t jt Junction QiiQh Voltage Equiva terit)
Fig. 8. Corona Loss, Line from Junction Dam lo Grand
Rapids. Transformers on at Grand Rapids
Junction Dam transformer ratio 120,000 /208.00Y:7500.
Grand Rapids transformer ratio (Test 20) 14n,000/242,.'JOOY:
7.500.
Grand Rapids transformer ratio (Test 42) 12rj,000/216..500Y:
7500.
Figs. 3 to (), inclusive, allow a comparison
to be made between the losses on a clear and
a rainy day and on two clear days with dif-
ferent temperatures. The marked effect of
the exciting current of the step-down trans-
formers in modifying the line current is
apparent from these curves.
Fig. 7 shows the excitation of the step-up
transformers at Junction Dam, Fig. 7-a
1 1 1 1 : 1 1 1 If
1 1 1 1 1 1 M M 1 1 1
1 1 1 1 'est23 J»
:L^^Mr-
° H~"]f — '"
1 J:
— -~ "\f ^
<.0 6000 - „ 7e5^ 23 [e
fir 1 ~^Ti ^^''<''
- Temp. 4a 8" r. 4
- Hum. 6/% e
- Weather Clear C/
'^'•T''r \t i 1
'?i -1- -if. 4~L -
:)ljdi/ 1 1 j
i 1 1
r
*"
jT
1
/ fest Z3
t ^T '
1
si
ylfe 1J
' : (2
1 ^jT
_ t .^f _
2 T
, Line
I 1
7 ' Current
\t "
2
^7
■jn
' ¥
/ -^ //
4.^ 8
- >
' -'I ■ "
" E^
_,^^ .
- t " ~
-*'
1
t - "
J
T /
t
J
:/
I
) :::-:-:
.„ ^
- a^
^ ^^0 80 100 /ZO
40 160 ISO ZOO Zi
0
LowVoltaqe Hv.at Junction (Hi<jh Voliaqe £quivo/eM)
Fig. 9. Corona Loss, Line from Junction Dam to Grand
Rapids. No transformers at Grand Rapids
Junction Dam transformer ratio 120.000 /2OS,00OY:7500.
TABLE VI
TABLE VII
TEST 20
TEST 42
H.V. H.V.
Kv. Amp.
Kw.
Kv.
at
G. R.
H.V.
Kv.
H.V.
Amp.
Kw.
Kv.
at
G. R.
106.5 1(5.3
72
116.4
109.5
16.0 72
109.1
122.6 19.0
120
133.2
127,4
18.8
168
129.9
140.4 21.3
480
152.6
1.38.0
20.2
528
138.5
154.8 25.9
1945
166.9
152.0
24.5
1775
159.4
163. 6i 29.4
2930
177.2
170.9
32.3
3650
178.4
187.0 38.6
5,'545
201.8
181.0
36.1
4750
188.7
202.5 44.0
7440
219.2
197.0
41.5
6645
206.0
208.0, 45.9
8160
224.3
208.2
45.1
7970
218.2
H.V.
Kv.
TEST 23
TEST 43
H.V.
Amp.
Kw.
H.V.
Kv.
H.V.
Amp.
109.0
17.2
48
93.0
15.1
48
126.2
19.8
72
112.0
18.3
48
140.4
22.2
480
127.6
20.8
72
157.6
27.1
2230
139.2
22.8
648
174.2
33.3
4030
1.57.5
27.8
2330
186.1
37.8
5380
163.6
29.6
2880
198.4
43.2
6960
185.5
37.2
5305
198.5
42.4
6840
424 May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 5
60 260
SO 2^0
ISO
30 160
ft-
20 120
ion
10 80
60
40
.^
^<<27 -
.,\yJriW-
^YvXw
rKra^^i
r Ift^yi-oyV
43 xm^W\
-r- Y •&^^J^
it ts^/<fW^Jl
y w&^/
t/s^^
A ^ %W.^?U-
t Mst-
± yi7\\!i2
y J ./2\l°'
^ W^
/ ,>K^^
^^ ^Z ^
J J^
,^y^"" it
J ^y^i- -
^ ^Zr X
Z^T^r t
Zy^'tf -
Zy?^i°_L
Z^?>1°
't,f.'> -t T
2v° t
t
J -
I
J
T
r--'^
_
80 .00 IZO 140 160 ISO 200 220
Low yo/tage Kv at Junction (H/gh Voltage £i^u/ya/ent)
Fig. 10.
Comparison of Potentials at Junction Dam and
Grand Rapids, Test 20
SO 240
220
40 200
ISO
}0 160
S: ^/40
20 120
to
-- -- 4JJo<UiC -L
- ^^^44ftlt^
xTjf. l„!(iy <fi^ X-
A^ ifSiw
!& -vV^S^ A-
__ A^ ,fCZ(Kt'I
JL .%i4r^€-
_Zr,tojZ%t
/■liw^i'-
.^'JzM}!. _(_
_. -,Lii2?^o*
" itJ^^iV-
^ L yym.
-i'/ ?Tot
- -. W>rr-
l "d- -
4- J(^f
J?'
_±%_ _- -_
- - - > w T
^^tl] _- _- --
.fr-it
W'v*
j( r ^
?T J
~ ~ ~ 5 T
S
~ ~ ^^
-4^ -4-4 / T
f._ _
J
I - _-
M
mTM T ' LU ttM TTT
_ _ —
plotted on the basis of delta high-tension, and
Fig. 7-b on the basis of Y high-tension.
Figs. 8 and 9 show the losses with the
Y-connection. It will be noted that the
readings were taken up to about 208 kv. by
ratio. These curves have been plotted with
the readings as taken, without correction for
instrument transformer ratio or phase angle,
for the reason that station instrument trans-
formers were used and the corrections were
not available. The readings of Figs. 3 to 0,
inclusive, were taken with calibrated instru-
ment transformers and have been corrected
for ratio and phase angle. The corrections
in any event are small and at the higher values
are negligible.
Fig. 10 shows for Test 20 the high-tension
voltage at Junction Dam as calculated by
means of the line current and the known
resistance and reactance of the transformers,
also the high-tension voltage at Grand
Rapids by measurement and transformer
ratio, also the power-factor of the corona
readings. Fig. 11 shows similar data for
Test 42.
A comparison will be made of the tested
and calculated losses for Test 42 (Fig. 8).
In order to do this the true line current is
determined as shown in Table VIII, with the
assistance of Figs. 8 and 11. The true high-
tension voltage is calculated as just explained,
giving the results shown in Column 10, Tabic
VIII. The high-tension voltage at Grand
Rapids is found from Fig. 11.
The corona loss is calculated fur Test
42 in the usual way as follows:
Given: Length 101. .5 miles
Spacing 12 ft. vertical
Conductor 110,000 cir. mil.
Barometer 29.32 in.
Temp. 41 deg. F.
s =182 in.
r =0.190 in.
g„ = .i.3.8 kv. in. efTeclivc
„,„ = 0.8.") (assumed).
17.92/> _ 17.92X29.32
4:)9-|-7~ 409 + 41
"eo 30 00 120 140 160 ISO zoo 220
COM Vo/coocXifotJ'urrction fHifh /oytofe^fuiya/entj
6 =
:=1.0.J2
fo = 2. 303go Wo 5 r /o.Cio-clT. kv. to ncul.
= 2.303 Xo3.S XO.85 X 1 .052 xO. 19 x/og,o,
182
0.19
240
Fig. 11.
Comparison of Potentials at Junction Dam and
Grand Rapids, Test 42
= 21.09X/og958
= 21. 09X2. 98137 = 02. 8.") kv. to neutral.
62. 8,>X 1.732= 108.9 kv. between con-
ductors.
SOME CORONA LOSS TESTS
425
TABLE VIII
DETERMINATION OF HIGH-VOLTAGE POTENTIALS
Conditions of Test 42
1
2
3
*
5
6
7
8
9
10
11
12
L.V.
Volts
L.V.
Cur.
Trans.
Exc.
Cur.
H.V.
H.V.
Volts
Grand
Rapids
L.V.X
Ratio
Avg.
H.V.
Junc-
Junc-
P-F.
e
Sin
■ +Jii
Line
Line
Cur.
Volts
Junc-
H.V.
Equiv.
H.V.
Equiv.
Amp.
!I0 deg.
Lag.
Col .6+Col. 7
Amp.
tion
Calc.
and
Grand
Rapids
100
15.3
0.018
89° 0'
0.99985
0.268 + 15.29J
0.12
0.27 + 15.41J
15.42
102.0
97.8
99.9
no
16.4
0.023
88°41'
0.99974
0.377 + 16.39J
0.15
0.38 + 16.54]
16.54
112.0
109.8
110.9
120
17.6
0.032
88°10'
0.99949
0.5.54 4- 17.59J
0.20
0.55 + 17.79]
17.80
122.1
121.1
121.6
130
19.0
0.046
87°22'
0.99894
0.874 + 18.98]
0.24
0.87 + 19.22]
19.24
132.5
132.5
132.5
140
20.8
0.135
82''14'
0.99083
2.81 +20.6 j
0.25
2.81+20.85]
21.02
142.5
143.9
143.2
150
23.5
0.250
75°31'
0.96822
5.88 +22.8 j
0.26
5.88+23.06]
23.8
152.7
155.1
153.9
160
27.7
0.325
71° 2'
0.94571
9.00 +26.18J
0.26
9.00+26.44]
27.95
163.2
166.2
164.7
170
31.7
0.378
67°47'
0.92576
11.99 +29.32J
0.30
11.99+29.62]
32.00
173.5
177.3
175.4
180
35.4
0.420
65° 10'
0.90753
14.86 +32.12J
0.40
14.86 +32. 52i
35.76
183.8
188.1
186.0
190
38.9
0.453
63° 6'
0.89180
17.60 +34.68J
0.60
17.60+35.28]
39.40
194.1
199.0
196.6
200
42.3
0.476
61°35'
0.87951
20.12 +37.20J
0.80
20.12+38.00]
43.00
204.1
200.5
206.8
210
45.7
0.492
60°32'
0.87064
22.50 +39.80J
1.25
22.504-41.05]
46.80
214.8
220.0
217.4
220
48.9
0.500
60° 0'
0.86603
24.45 +42.38J
1.75
24.45+44.13]
50.50
225.0
229.8
227.4
TABLE IX
CALCULATED CORONA LOSS. TRANS-
FORMERS AT GRAND RAPIDS
Conditions of Test 42
E
Line
Voltage
Kv.
« Voltage
to Neu-
tral Kv.
e -62.85
(e -62.85)=
1.998 X
ie -62.85)!
Kw.
Loss
110
63.5
0.65
0.422
0.884
1
120
69.3
6.45
41.6
83.2
83
130
75.05
12.20
148.8
297.2
297
140
80.8
17.75
314.0
627.7
628
150
86.6
23.75
564.0
1127.0
1127
160
92.4
29.55
872.0
1742.0
1742
170
98.2
35.35
1249.0
2497
2497
180
103.9
41.05
1682.0
3360
3360
190
109.7
46.85
2195.0
4388
4388
200
115.5
52.65
2772.0
5.540
5540
210
121.3
58.45
3418.0
6830
6830
220
127.0
64.15
4110.0
8220
8220
(0.189\
1+ --" Ikv., in. effective
(0.189 \
^+Vl.052X0.19J
(0.1S9\
= 56.61
= 80.5
/) = a(/+2o) (e — eo)-X10"^ kw. per mile sin-
gle conductor.
„ = 3S8 17^ 388 ,M|^3^3^^;^^^^^
\5 1.0o2 \ 182
= 368.8 X. 03233 =11.9:5.
/J= 11.93 (30 + 25) ((?- 62.85)2 X 10-5 kw. per
mile per conductor
= 656 (e — 62.85)2X 10~5 kw. per mile per
conductor
= 1968 (e-62.85)-X10-5 kw. per mile per
three conductors
= 1.998 (c-62.85)- kw. 3 cond. 101.5 mi.
c'» = 2.303 gviihr logia-efi. kv. to neut.
r
;hi. = 0.82 assumed
r„ = 2.303X80.5X0.82X0.19X2.98137
= 86.2 kv. to neutral
86.2X1.732=149.3 kv. between cond.
The results of this calculation are set forth in
Table IX.
In Table X is found the net corona loss;
i.e., the measured loss minus the transformer
losses and I-R line loss, also the calculated
corona loss for the average of the high-
voltage potentials at the two ends of the line.
This is about as near as we can come to the
correct calculated value without resorting to
a great deal of refinement, perhaps more
than is warranted under the circumstances.
In Fig. 12 the net corona loss curve and
the calculated cun'e are compared. It will
be noted that there is a very fair agreement
between these two cur\'es. In general, it will
be found that the measured cur\'es show a
more abrupt bend in the lower part of the
cur\'e and are straighter in the upper part
than are the calculated curves.
An interesting feature of the tests is the
fact that the charging current and the rise
in voltage along the line measure considerably
426 Mav, 1020
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. .5
greater than shown by calculation. For
example, in Test 42 at 21)0 kv. high voltage
(i.e., low voltage multiplied by ratio), the
charging current proper for the lOl.o-mile
line from Junction Dam to Grand Rapids
tests about 39 amp. while calculation gives
1 — 1 1 1 1 1 ' 1 — 1 I 1 1 1 1 1 1 1 1 I 1 1 1 1 ■ . 1 ■
lit
// ^
//v
jlxil
~\ U'^ [
H «
f ^
1 5*
J ^
"fl^
fW
5f'*;
-V/v
s//|s-
A*j/(V
X' '-,-"""- sitf^
- - -- - ws
- -~ -- - "t^tt ^
- - _ _ - ■" ^Ifit-
^/l/s
' /l/^
- - - •fl7^!Ll,ra2^
^nt?Tr^^/y^
s - - - : -3th ^^
* ^ti'^ »*yV/
- - - - :|/Ln«?^'V ' ^
SttTwvo^- + i
-^tt.m^^t.^ ^ "*
----- - - '^Uiw^\^ "^ '" ■
ytjW f f -t ^
tjr\y V 1 1_
i^ .A
F F^ _ _
1 J i i_
^ if
t* h -
~f~ [ " " - ■ -- - -
7 [
/i-
?:: . : ~ "
. .h+Ih infill M-I- i W
inductance. For instance, if the radius of
the conductor is assumed to be increased from
0.19 in. (the true radius) to l.OI in., then the
capacitance will be increased from 1.322X
lfl~* to 1.7oX10~^ farads and the inductance
will be decreased from 219X10"' to 177.8X
TABLE X
Conditions of Test 42
so 100 /20 MO ibo /SO zoo zzo ztc
ton Vo/to^e Kv ot Junction (Hi^h Vo/toge Squiva/ent;
Kv.
Total
Loss
From
Test
Kw.
Trans.
Loss
Junction
Kw.
Trans.
Loss
Grand
Rapids
Kw.
1-R
Line
Loss
Kw.
100
60
25
24
12
110-
75
30
29
13
120
110
34
35
16
130
190
40
41
18
140
675
46
48
oo
1 .50
1.525
52
oo
28
KiO
2500
58
63
38
170
3560
66
72
50
ISO
4680
( o
82
63
I'.MI
5!100
86
97
76
l.>()()
7000
98
113
91
L'lO
SI 30
113
133
108
220
'.1250
133
J 1.
160
125
1
G
-"
S
9
Calc. Corona
Kv.
Total
Trans,
and
Line
Loss
Net Corona
Loss
Col. 2 -Col. 6
Kw.
Calc.
Corona
Loss
Kw.
Loss with
AvK. of
H.V. Potentials
at Boih
Ends of
0
0
Line Kw.
inn
61
0
no
72
3
1
1
120
S5
25
S3
111)
l.JO
'.•;i
91
297
375
141)
116
559
628
780
i.-.o
135
1390
1127
1350
ItiO
159
2341
1742
2060
170
188
3372
2497
29(H»
ISO
220
4460
3360
3925
liU)
259
5641
43,SS
51(X)
200
302
6698
5540
6375
210
3.54
7776
6830
7825
220
4 IS
SS32
S220
9300
Fig. 12. Curves of Comparison of Measured and Calculated Corona Loss
about 29. (J amp. I'^or the same test at ISO
kv. high voltage the rise in voltage along the
line from Junction Dam to Grand Rapids
tests 4300 volts and calculates onlv about
(iOO volts.
These discrepancies ma\- be reconciled b\-
assuming that the corona has the effect of
increasing the size of conductor thereby
increasing the cajjacitancc and decreasing the
10"' Henrys for the Htl.o-mile line. Tlic
charging current will then calculate 39 anij).
and the rise in voltage 3200 volts for the
conditions mentioned in the preceding i)ara-
graph. That such an effect takes jilacc seems
reasonable, when it is considered that corona
is caused by the air surrounding the con-
ductor breaking down and becoming con-
ducting.
427
The Alternating-current Network Protector
By H. C. Stewart
Transformer Engineering Department, General Electric Company
Distribution of alternating current from a network fed by a few large transformers offers decided advantages
in improved continuity of service, reduced losses, better regulation, etc., over the common system of distribu-
tion where a great number of small transformers arc used, each feeding a comparatively short independent
secondary circuit. One factor, however, has prevented the general adoption of alternating current networks,
namely, there is a chance that an internal failure in one transformer may impose an overload, besides the short
circuit current, on the adjacent transformer, thereby blowing the primary fuse. This action will progress from
transformer to transformer until the entire system is disconnected. The device described in this article has
been developed to overcome this difficulty. It has been in use for a number of years on the large network
systems in New York City and its operation has proved entirely satisfactory. — Editor.
The advantages of ring and network dis-
tribution in connection with direct current
systems are well known. The same important
advantages, namely, continuity of service and
decreased cost of the distribution system for
given regulation and loss, can be realized in
alternating current network systems supplied
by transformers, if the transformers arc
equipped with the a-c. network protector.
Contrasting the usual a-c. distribution
system with the a-c. network, the former has
a great number of small transformers, each
feeding a comparatively short independent
secondary circuit, while in the network a
much smaller number of transformers of
larger individual capacity supply an inter-
connected secondary system covering a large
area. Because of the diversity of sennce in
this large area it is usually possible to con-
siderably reduce the total installed trans-
former kv-a. The reduction in cost due to
the smaller kv-a. and the lower cost per kv-a.
of the larger transformers will about cover
the cost of additional equipment required in
the network to make the system workable.
The advantages resulting from network
operation — improved continuity of service,
lower losses, better regulation, etc. — are thus
clear gain. Temporary overloads or short
circuits which would interrupt service on a
local independent circuit will not often affect
the network, as there is sufficient capacity
to carry the overload and generally to burn
off any short circuit which takes limited
energy.
One deficiency only has prevented the wide
use of the a-c. networks. Regardless of how
the transformers may be fused there is always
the possibility of an internal failure in one
transformer throwing an overload, in addition
to the short circuit current, on the adjacent
transformer, whose primary fuses may blow.
This action will contintie progressively until
the entire system is disconnected from the
feeders by blown fuses. The development
of a satisfactory static device which would
eliminate this difficulty renders the a-c. net-
work practical. The General Electric Com-
pany has developed stich a protective device
under patents granted to Messrs. Sprong and
McCoy.
Description and Theory of Operation
The protector is a transformer device with
three sets of windings connected as shown
in Fig. 1. Two of these windings magnetically
oppose each other and the third is arranged so
that its magnetic action is neutralized by the
division of current, so that when operation is
VW DIrec ti'on or Winding
— ^ Direction of Current
Fig.
Load
Connections of Network Protector
normal there is practically no magnetization
of the protector core.
One winding A is in series with one of the
high tension lines supplying step-down trans-
former T . The second winding of two parts,
B and C, is connected in series with the low
428 May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 5
tension lines feeding to the network (the
neutral of the low tension winding of the
transformer T, if used, is brought directly to
the neutral of the network). The ratio of
turns between the two windings A and B-C is
such, and the coils are so wound and con-
nected, that the core is not magnetized.
Fig. 2. .SO-kv-a. Subway Transformer Fitted
with Network Protector
However, the low tension current passing
through windings B and C passes into the
middle point of the ])arts D and E of the third
winding. It will be noticed that the current
divides in passing through these windings so
that the magnetic action of these windings is
also neutralized. The ends of windings D and
E are connected by fuses to the low tension
network. There is no e.m.f. developed in the
windings D and E, since there is no magnetism
in the core under nonnal operating conditions;
consequently there will be no flow of current
through the local circuits rci:)rescnted by the
windings D and E and the two fuses con-
necting the ends of these windings to the
network.
Now, should a fault de\'elop in transformer
T it will draw a very heavy current from the
line through fuses FF' and protector winding
A. At the same time a heavy current will
be fed back into the transformer from the
network through the windings B and C of
the protector. This current is reversed to
normal operation so that there is no longer
magnetic opposition between windings .4
and B-C. The protector core is immediately
magnetized and a heavy current flows in the
local circuits of the third winding D and E
through the fuses. Even though the high
voltage fuses FF' are blown, this heavy cur-
rent through the low voltage fuses continues
because of the transformer action between ■
the windings B-C and D-E. In a ver>' short
time inter\-al the circulating current through
the fuses blows them and disconnects the
transformer T from the low tension network.
The action of blowing the low tension fuses
as described in this paragraph takes place
instantaneously and seemingly simultaneously
with the blowing of the primapi^ fuses FF'.
The windings of the protector are of vcr\-
low resistance and consequently the effect
on the regulation of the circuit is negligible.
Construction and Application
A protector is required for each trans-
former in the network, so that it is desirable
to make the protector a part of the step-
down transformer.
In order to isolate the protector from the
oil in the main transformer and thus elimi-
nate damage which might result from explo-
sion of fuses over oil, it is placed in a separate
box on top of and forming a part of the trans-
former cover. The top of the protector case
is seeurel\- closed b>- a clamped cover which
has a glass window for inspection of the fuses.
The device has been commercially developed
for 50 and 100-kv-a. (iO-cyclc subway trans-
formers for standard voltages, and is made an
integral part of the transformer. Fig. 2 shows
a .jO-kv-a. subway transformer equipped with
the network protector.
Alternating current network systems have
not been extensively used in the past, but the
present trend of engineering opinion indicates
that they arc being viewed more favoralih',
even if expensive oil switches and relays are
re(|uired. The a-c. network protector success-
fully meets all requirements and should
encourage the general use of a-c. networks.
The Brooklyn Edison Company and the
United Electric Light Company of New
York City have had a-c. networks, using this
device, in successful operation for a number
of years, and the performance of the network
protector on these systems has proved that it
functions entirely satisfactorily.
429
Alternating-current Lightning Arresters
By V. E. Goodwin
Lightning Arrester Engineering Department, General Electric Company
For a number of years the standard form of lightning arrester has been the electrolytic cell. This cell has
a high discharge rate and high electrostatic capacity and is ideal for large power stations where attendants are
on hand to look after the charging, etc., that is required. A more recent form of lightning protector is the oxide
film arrester. The electrostatic capacity of this cell is lower than that of the aluminum cell and from laboratory
data the latter would appear to possess better qualities for protection, but in actual service the relative merits
are more nearly equal. The oxide film arrester requires no attention and hence is more suitable for isolated
installations. Neither of these arresters offer protection against high frequency surges, and therefore it has
been necessary to develop a device known as the high frequency absorber, which consists of a static condenser
connected in series with a resistance. High frequency absorbers are installed as auxiliaries to the lightning
arresters and are recommended for installation on busbars on the important stations of a system. — Editor.
The purpose of this paper is to discuss recent
developments in the art of protecting moderate
and high voltage electric stations against
lightning and other high voltage phenomena.
Abnormal voltages which are dangerous
to electrical apparatus are of two general
classes: First, those which exceed the test
voltage of the apparatus. These may be
either high or low frequency disturbances or
single impulses. Second, those of high fre-
quency and low voltage which by virtue of
their rapid changes of potential, may build tip
to dangerous values in indtictive apparatus.
For the first class of these disturbances it is
necessary to have an arrester which operates
instantly upon any abnonnal rise in voltage
and which has sufficient discharge rate, or
conductance, to dissipate the energy of the
disturbance at a rate which is faster than it is
generated and delivered at that point in the
circuit. This question of discharge rate is
one which is often overlooked in selecting
arresters for a particular ser\-ice. Many
people assume, when they see an arrester
spark over frequently, that the arrester is
doing a lot of good work. Possibly it is, but
sensitiveness is only one requisite, which by
itself is of no merit since without discharge
rate the excess voltage would not be relieved.
For the past twelve years the aluminum, or
so-called electrolytic "lightning arrester has
been the standard fonn of protector for large
Fig. 1. Aluminum Cell Lightning Arrester
for Indoor Service, 3000-5000 Volts
Fig. 2. Aluminum Cell Lightning Arrester for
Outdoor Service, 50,000-73,000 Volts
430 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 5
stations. This type of arrester, due to its
film or valve action, combined with its high
electrostatic capacity per cell, has character-
istics which are ideal for this ser\-ice. This
film or valve action of the cells limits the
passage of energy' current at normal voltage
to a small value. If the voltage tends to rise
to abnormal values, the current increases
rapidly; thus the cells act as a barrier to
normal voltage but as a virtual short circuit
to the abnormal part of any excess voltage
disturbance. By this action the aluminum
cell tends to automatically keep the voltage
below a predetermined critical voltage at
which the apparatus can be safely operated.
The high electrostatic capacity of the alumi-
num cell is a highly desirable characteristic
of a lightning arrester as it provides a ready
means for absorbing the energy of any high
frequency or steep wave front disturbance,
and it also tends to modify the wave form of
impulsive voltage disturbances so as to render
them less harmful to the system.
A more recent development in this field
of protection is the oxide film arrester, pre-
viouslv described in the technical press,
(A. I.'E. E., June 191S, and G. E. Review).
This arrester has many of the characteristics
of the aluminum type; namely, the cellular
construction, the film or valve action,
and the high electrostatic capacity.
» While possessing these similar features the
two types differ materially in details as well
as in operation. The cells of the aluminum
arrester consist of aluminum cones or trays,
partially filled with electrolyte which forms a
film of aluminum oxide on the active surface of
the aluminum when current is passed through
Fig. 3. An Element of the Oxide Film Arrester
the cells. These cells are immersed in a tank
of oil for cooling and insulating purposes.
The cells of the oxide film arrester arc self-
contained and consist of two metal electrodes
securely clamped to a porcelain spacer. The
i
Fig.
4. Oxide Film Lightning Arrester for
Indoor Service, S000-7S00 Volts
Fig. 5.
Oxide Film Lightning Arrester for Outdoor Servicr,
50,000-73,01)0 Volts
ALTERNATING-CURRENT LIGHTNING ARRESTERS
431
inside surface in the metal electrodes are
coated with thin insulating films. Between
these two films is the active material con-
sisting of a special grade of lead peroxide.
(See Fig. 3.)
The relative film action of the two types is
theoretically dependent on the thickness of the
films and the quality of the active materials
used in the cells. In practice, the critical
film voltage in the aluminum cell is somewhat
more sharply defined than in the oxide film
cell, but it is more variable due to the dis-
solution of the aluminum film.
As regards the values of electrostatic capac-
ity it is worthy of note that in the present
designs the capacity of the aluminum cell is
greater than that of the oxide film cell. While
this is a question of relative plate areas and
thickness of the film, it is doubtful if the ca-
pacity of the oxide film cell can ever be made
as great as that of the present aluminum cell.
fft
and more important stations where skilled
attendants can look after the charging of the
films and give the daily attention necessary
with this type. Actual operating experience
is the real criterion in determining the value
of any lightning protective device; hence the
actual defines of these fields of application will
have to be determined from more extensive
experience by operating engineers.
A design feature of particular interest in
these two types of arresters is the use of a new
line of interchangeable insulators for outdoor
service. (Fig. C.) There are five sizes of
insulators used for ratings 7500 to 73,000
volts. The caps and pins have similar drill-
ings and are accurately jigged so that all
insulators in a class are identical and all
classes are interchangeable. This arrange-
ment is particularly advantageous for use on
apparatus which is to be operated at various
altitudes or in other places where extra
Fig. 6. Interchangeable Insulators for Use With Lightning Arresters Located Out-of-doors
In comparing the two arresters from a
protective standpoint, we have two sources
of data; namely, laboratory tests and actual
service experience. The data from each of
these sources seem to indicate that both of
these types are superior to any other scheme
of protection yet devised. In laboratory tests
the aluminum arrester shows up better
apparently on account of its higher electro-
static capacity. After four years of experi-
ence with the oxide film arrester in actual
service at from 2300 to 73,000 volts, there
does not seem to be as great a difterencc in
protective qualities as would be indicated in
the laboratory tests.
Looking at the problem from a practical
standpoint it would seem that the two
arresters would fill a somewhat similar field.
The oxide film type is better suited to the
smaller stations where there are few or no
skilled operators in attendance, while the
aluminum type will be installed at the larger
insulation is desired, as it is simply necessary
to select insulators having the desired factor
of safety and substitute them in the standard
design without any change in fittings.
Both the aluminum and oxide film types for
a-c. service are equipped with series sphere
gaps to prevent rapid deterioration. This
arrangement limits their operation to dis-
turbances having voltages sufficient to dis-
charge these sphere gaps. These include all
high and low frequency disturbances of volt-
ages in excess of the spark potential of the
gaps. The sphere gaps introduce the shortest
known spark lag; consequently the arresters
thus equipped are best suited to handle steep
front impulses which are so dangerous to the
insulation on induction windings of apparatus.
For the second class of dangerous dis-
turbances mentioned above, namely, high
frequency low voltage, it is necessary to
use some form of protection which does not
depend entirely on the princijile of over.
432 May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 5
voltage for its operation and which con-
sequently does not have a series gap. Its
function should therefore be to separate and
absorb the energy producing the high fre-
quency disturbance.
The high frequency absorber illustrated in
Figs. 7 and 8 has been developed to meet
this condition. The device consists of a static
condenser with a series resistance. The
condenser acts as an automatic relief valve
for high frequencies and the resistance as a
means of absorbing the energy which is tend-
ing to produce these oscillations. They are
so designed as to pass only small leakage cur-
rent at normal frequency. If the frequency
/
Z =
^ V27r fC/
i? 100 C .01 microfarads/ =60 cycles.
Then Z at normal frequency = 265,000 ohms.
At 13,200 volts the current at 60 cycles would
be .05 amperes and the energy- absorbed b%-
the series resistance would be .05- X 100 =
.25 watts per second.
For comparison, let us assume that the
frequency be suddenly increased to 100,(M)()
cvcles, the other factors remaining constant.
Then Z=1SS ohms.
At 13,200 volts the current at 100,000 cycles
would be 70.3 amperes and the energy-
Figs. 7 and 8. High Frequency Absorber. 15.000-25.000 Volts
should increase from some extraneous source,
such as an arcing ground, the current through
the device would tend to increase nearly in
proportion to the frequency if the energy
supplying this condition were not limited.
This, fortunately, is the case, as the high fre-
quency energ>' is limited in value. Hence as
the current through the condenser and scries
resistance increases, the resistance absorbs the
energy and dampens the oscillations, thus ren-
dering them less dangerous to the system.
The action of the high frequency absorber
as described above can be better understood
from a study of the following calculations of
an actual design.
The impedance of a condenser with a series
resistance is
absorbed would be 70.3- X 100 = 494.000 watts
per second.
The high frequency absorbers are installed
as auxiliaries to aluminum and oxide film
arresters and arc usually recommended for
installation on the busbars of the more impor-
tant stations of the system. The high fre-
quency absorber illustrated in Fig. 7 has been
developed for ser\ice on \oltage from 75(X) to
25,000.
These high frequency low voltage disturb-
ances are most serious on moderate voltage
systems. ]>artic\ilarl\- those where the voltage
is stepped down from high voltage transmission
circuits. On the higher voltage transmission
circuits the long lines seem to act as absorl>ors
and to dampen out these disturbanct>s.
433
Metallic Resistor Electric Furnaces for Heat
Treating Operations
By E. F. Collins
Engineer, Industrial Heating Department, General Electric Company
The advantages of the electric furnace for heat treating are well known. Uniform heat distribution and
automatic temperature control are the most important factors. The furnaces described in this article differ
radically from the ordinary laboratory furnaces both in construction and in operating characteristics. The
heating element consists of bare metallic ribbon uniformly discributed over the interior of the furnace. The
ribbons are exceptionally heavy and, being unmuffled, radiate the heat direct, the resistor therefore remains
at a lower temperature, produces quicker healing, and is much more sensitive to temperature regulation than
the muffled or screened furnace. The sensitive automatic heat control and high rate of heat delivery to the
charge insure high termal efficiency. The illustrations show several different types of these furnaces of both the
vertical and horizontal form. — Editor.
There exists today a growing demand for
heat-treating equipment of large capacity.
A considerable demand comes from manu-
facturers of automobile parts, such as gears,
crankshafts, bearings, axles, etc., which are
produced in large quantities, and all of which
require heat treatment. Many tons of steel
are heat-treated each day; therefore the
demand for furnaces of large productive
capacity and high efficiency.
Electric furnaces have long been recognized
as ideal for this purpose, but until the last
few years they have not been available in such
size and of such rugged design that they could
be considered for carrying unassisted the regu-
lar production load.
In a paper on "Electric Heating of vSteel,"
presented at the American vSteel Treaters'
Society meeting at Chicago last September, I
described some vertical electric furnaces which
were used for a j'ear or more preceding the
armistice for heat-treating gun forgings.
In this paper the method of control, and tlie
general characteristics and advantages of this
type of furnace were discussed.
It is therefore unnecessary to again describe
the furnaces in detail; but since it is the
purpose of this paper to recount some things
which have been accomplished, to give
some results of operation and describe the
application of the furnace in several forms to
the problems of heat treating, it is desirable to
again mention some of the salient features
of the furnace.
This type of metallic resistor furnace
entered the industrial field in 1917 and 19 IS,
and its use has been attended with unusual
success in regular manufacturing production.
It was first tested under the rigorous con-
ditions attending gun-making in war time, and
the results produced were phenomenal.
A basic idea incorporated in the design is
the location of the resistor ribbon, unmufflcd.
in the open heating chamber so that it can
directly radiate the heat generated within it.
These ribbons are very rugged mechanically;
they are sometimes as much as two inches
wide by one eighth inch thick and are
formed into loops. They are supported on
refactory insulating members projecting from
the walls of the heating chamber, the body
portion of the support being imbedded in the
wall.
The resistor is thus free to deliver its heat
by radiation to the charge without the neces-
sity of first forcing the heat through the
walls of a muffie, as has been the practice
in most metal resistance furnaces heretofore
constructed.
In order to force heat through a muffle
at high rate and secure rapid heating of the
charge, a high temperature gradient, and
therefore high resistor temperature are neces-
sary. Hence the unmuffled furnace has
inherently a lower temperature resistor,
produces quicker heating, and is much less
sluggish in point of temperature regulation
than the muffied or screened furnace doing
the same work. The unmuffled furnace uses
direct and reflected radiant heat.
The outstanding features of this furnace
are more rugged resistors (absorbing more
power yet resulting in a low temperature
long-lived resistor), splendid heat distribution
and control, and high thermal efficiency due in
part to sensitive automatic heat control and
high rate of delivery to the charge. A large
resistance furnace is more rugged in design
and is inflnitely more dependable in operation
than the small laboratory furnace that has
heretofore existed. This fact has been well
established from the results of operation in
regular production.
Another feature is that heavily heat-
insulated walls with at least 9 in. of heat
insulation outside a 4-in. refractory lining
434 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 5
Fig. 2. Section of Resistor Ribbon for
Heat Treating Furnace
Fig. 1. Cross Section of Cylindrical Vertical Furnace
for Heat Treating Gun Forgings
I
Fig. 3. Detail of Metallic Resistor and
Support for Vertical Furnace
ELECTRIC FURNACES FOR HEAT TREATING OPERATIONS
435
■
IJlf
■ *^^g^ii.^^ ' ^ "^^i/a^r"
.^i^^^^^Ih^ ^j^^^^P^fl^Hjj^fe
ii-
Fig. 5. War-time Plant for Heat Treating Guns
may be used, and gas or air tightness through-
out may be maintained by the use of an outer
casing. This contribiites to thermal effi-
ciency and an operating chamber from which
air mav be cxchided, or in which a gas mav be
held.
Fig. 4. Showing Interior of Furnace of Fig. 1 and
Method of Supporting Metallic Resistor
Fig. 1 shows a cross section of a cylindrical
vertical furnace of the metallic resistor type
used for the heat treatment of gun forgings.
Fig. 4 shows an interior view, and Figs. 2 and
3 show the details of construction. Note the
thick walls of heat insulating material, the
heavy resistor ribbon and the outlet stud
extending through the wall, to which the line
connections arc made. Note also the welds
at the ribbon splice and at the terminal studs.
Fig. 5 shows a typical war-time plant used
for heat treating gun forgings. It consists of
four furnaces and quench tank. Fig. 6 shows
the elevation section of a plant consisting
of two furnaces and quench tank. This
plant is now being installed in Spain.
All these furnaces have automatic temper-
ature control, a typical control board being
shown in Fig. 7. The instruments which
control as well as record the temperature are
shown on the sub-bases of the panels. A
temperature control chart or record is shown
in Fig. 8.
A very complete set of tests were run on an
installation of these vertical furnaces, the
results of which may be summarized as
follows :
Heating to 1450 deg. F
when charged. Furnaces G
24 ft. high, voltage 44U, GO
400 kw.
Furnaces hot
ft. diameter by
cycles, capacity
I Charge
Total Weight
Including .Hold-
ing Fixtures
Energy in
Kw-hrs.
12 3-in. gun tubes
7 4-in. gun tubes
3 4-in. jackets
21,900 lbs.
22,300 lbs.
21,700 lbs.
1874
1880
2088
Total
65,900 lbs.
5842
436 May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 5
Average lbs. per kw-hr.
Kw-hr. per ton
Energy cost per ton at
65.900
■ 5S42
2000
■11.25
■ 173
■• 11.25
1.0085 per kw-hr. = SI. 52
This shows what may be done with the
electric furnace in moderate scale operations.
The time required to heat these charges was
5H to 6 hours, and the maximum diameter
of forgings was I6V2 inches. The furnaces
were 24 ft. high by (i ft. diameter inside
dimensions, and were exactly like those shown
in Fig. 5; the connected load being 400 kw.
each. The ultimate or minimum radiation
from test was found to be 70 kw. The
predicted radiation from design was 75 kw., a
satisfactorily close agreement.
The actual operating results for the four
furnaces for the month of October, 19 IS, just
before the armistice, when production was at
the maximum rate, including both hardening
at 1450 deg. F. and drawing at 1150 deg. F..
was $2.76 per ton based on the power rate of
10.0085 per kw-hr.
The specifications for drawing did not
allow the charge to enter a 'furnace at full
drawing temperature. It was therefore neces-
sary to cool the furnace somewhat, and then
raise the temperature of furnace as well as
charge. It was also required that the charge
be held at the drawing temperature for several
hours. This of course did not allow of the
highest efficiency in pounds of metal heated
per kw-hr.
If work at 250 deg. F., to be drawn at 1100
deg. F. enters a furnace at 1100 deg. F. a yield
of approximately 24 lbs. per kw-hr. would
be realized. This would give:
2000
24
1100 deg. F.
= 84 kw-hr. per ton for drawing to
Fig. 6. Sectional Elevation of Two Heat Treating Furnaces and Quenching Tank
ELECTRIC FURNACES FOR HEAT TREATING OPERATIONS
437
Fig. 7. Control Board for Automatic Temperature Control
Therefore .the overall operation would
require: To harden, 17S kw-hr. per ton for
heating to 1450 deg. F. To draw, 84 kw-hr.
per ton for heating to 1100 deg. F., making
a total of 2()2 kw-hr. per ton.
The cost at $0.0085 per kw-hr. will be
.1(2.23 per ton. The cost of $2.76 per ton
actually achieved during the month of October,
1918, indicates that the furnaces were handled
exceptionally well. They were of course
operated continuoush- without shutdown of
any kind, and some preheating was
done by putting cold charges in the
cooling pits.
These furnaces were in fact the first
to be built, and they were therefore
\'ery conservatively designed. They
could have been rated 600 kw. insteatd
of 400 kw., which would have irt-
crcased the output and efficiency ami
also shortened the time of heating,
since the radiation loss would be the
same in either case, and radiation is
the only loss which occurs with this
type of furnace. Strictly speaking,
this loss is not radiation, but part
radiation and part convection. We
speak of it as "radiation" for con-
venience.
They are made air tight, and there
is therefore no air passing through
the heating chamber, carrying away
heat as it escapes. This accounts for
the fact that there is practically no
scale on the charge, as well as for the
ease with which the furnaces can be
controlled, because in the absence of
air currents there is no tendency for
the heat to rise toward the top of
the furnace. Heated air will rise, but heat
rays will pass from resistor to charge by
direct radiation, and any heat distribution
may .be maintained indefinitely. This of
course is in accordance wi.h well known
physical laws for radiant heat, but it is in-
teresting to know that they can be applied
with great exactness in practice.
Twenty-two of these furnaces for harden-
ing and drawing, with a total rated capacity
of 7000 kw., were built during the war period
- cffAffec UNfro^'^Ly HeATen-
WtmHAHiiii-
a,
T
Fig- 8. Temperature Chart of Electric Heat Treating Furnace Fitted with Automatic Temperature Control
438 May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 5
Fig. 9.
Box Type Electric Furnace and Control Panel Used for Hardening
Punches. Dies and Cutters
for the manufacture of both field
and naval guns, the largest being
8 ft. diameter by 35 ft. high, with
connected load of 700 kw. Those
shown in Fig. 6 are now under con-
struction and are about the same
size. There is at present also under
construction a furnace 10 ft. G in.
diameter by 105 ft. high with a
connected load of 2850 kw., in
which forgings weighing 320,000
lb. may be readily heat treated.
The horizontal furnace is better
suited to much industrial heat
treating, the use of vertical fur-
naces being restricted to cases where
relatively large or long objects are
to be treated. The same principles
of heating hold whether the furnace
be vertical or horizontal, since the
metallic type of resistor windings
may be applied to any type of
furnace chamber. There are in
operation, or in process of installa-
tion, a large number of furnaces
in various forms in which this type
of heating element is used.
Fig. 9 shows a box type furnace
with its control panel used for
hardening punches, dies, and cut-
ters. Fig. 10 shows a small car
bottom furnace used to anneal steel
castings. This is a very conven-
ient form for annealing. Several
of this type are under con-
struction, one of the largest
which is to be installed in
France for annealing tool steel
bars being 17 ft. long by 8
ft. wide and rated 300 kw.
Fig. 11 shows two rotary
annular ring furnaces for
treating gears and similar
parts. The hearth of this
furnace revolves about the
vertical axis, and is suspended
on ball bearings and driven
by a motor through a worm
gear. Control is by push
button or foot switch, by
means of which the table may
be advanced as desired. The
mean diameter of the hearth
is about 5 ft., making it equal
to a tunnel furnace 15 ft.
long. The motor drive, doors,
counter weights and foot
operated mechanism are
clearh- shown.
Fig. 10. Car Bottom Furnace fur Annenling Steel Caslingt
ELECTRIC FURNACES FOR HEAT TREATING OPERATIONS
439
Fig. 11. Rotary Annular Ring Furnaces for Treating Gears and Similar Parts
This promises to be a very popular furnace
for small parts, such as gears, taps, drills, etc.
It has a connected load of 60 kw., balanced
three-phase automatic temperature control,
and is capable of turning out about 300 lb. of
Fig. 12. Interior View of Box Furnace Used for Carbonizing
steel per hour at 1500 deg. F. quenching
temperature.
Fig. 12 is an interior view of two box fur-
naces under construction showing the resistor
windings supported on the wall. These box
furnaces are used for carburizing.
Another type of vertical furnace
for hardening spindles consists
of a hearth which revolves on its
vertical axis, upon which is mounted
a frame of non-oxidizing metal for
holding the spindles in a vertical
position, there being space for
about 40 spindles. The spindles
are put in and taken out through
two holes in the cover. A small
movable cover operated by a con-
veniently placed handle covers
both holes and may be pushed
aside to put in or remove the
spindles; but only one hole can be
uncovered at a time.
Fig. 13 shows a large furnace
with flat hearth 12 ft. in diameter
which has been in use for several
years for tempering leaf springs for
automobiles. This furnace oper-
ates at 950 deg. F., turning out
440 May, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, No. 5
O
HE
'--DC
TJT U-r
=^=ft<;
TF
ELECTRIC FURNACES FOR HEAT TREATING OPERATIONS
441
about 2000 lb. of springs per hour. The con-
nected load is So kw. three-phase, all the
heaters being mounted in the arch.
Since this furnace operates at a relatively
low temperature, 950 deg. F., it is equipped
with so-called low temperature heaters, which
consist of a number of ribbon wound units
mounted on cast-iron frames, in general
similar to Fig. 16. This furnace is doing
excellent work and in large quantity, turning
out about one ton per hour, as stated above.
There appears to be great promise in the
application of air drawing ovens for heat-
treated parts.
Figs. 14 and ISshowaconveyorfumacewith
automatic quench for a large production of
relatively heavy parts, such as cranks and
axles. Note that there are heaters in the
arch as well as on the side walls. These
conveyors are shown as examples of standard
fuel furnace equipment which may be em-
ployed with electrically heated furnaces.
Such furnaces are usually designed special,
and they can be built for almost any tonnage
desired.
In regard to temperature distribution and
control, it mav be said that in the cvlindrical
walls, and by a proper arrangement of arch
and reflecting walls, as shown in some of the
illustrations.
Fig. 15. End Elevation of Furnace of Fig. 14
furnaces this distribution is perfect, as the
entire inner surface is covered by the resistors,
all of the same size, and all carrj'ing the same
current. This condition is very closely
approached in a hearth furnace by locating
resistors in the arch as well as in the side
Fig. 16.
Low Temperpture Heating Units for Fxirnace
Shown in Fig. 13
Automatic temperature control saves time
as well as power, and gives a constant tem-
perature; in fact, it is considered essential
to successful operation and is always fur-
nished.
Maintenance so far as resistors and refrac-
tories are concerned, is very low, since, due to
automatic control, they never become over
heated and there is no wear or abrasion on
refractories except on the hearth.
The resistors are designed to convert a
certain number of watts per square inch into
heat, which may be calculated with exactness,
and so long as safe working limits of tem-
perature and rate of radiation and absorption
are not exceeded, the resistors are absolutely
dependable and practically permanent.
In conclusion, it should be said that we are
indebted to the manufacturers of ordnance,
who first demonstrated the success of this
type of furnace. Their optimism lead to
the installation of electric furnaces, which
gave us an opportunity to prove that our
claims were correct; and there was inaugur-
ated, to undergo the stress of war demands, a
type of furnace for heat treating steel which
will be found equalh- important in time of
peace to help the manufacturer win many
of his industrial battles.
442 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 5
A New Type of Arc-welding Generator
By S. R. Bergman
Consulting Engineer, General Electric CoMP.\xy
A suitable arc is essential to successful arc welding. The production of this stable characteristic
requires some means of regulation. Where each welder is furnished with an individual generator, the neces-
sary regulation is best effected through special inherent regulation in the generator itself. The machine
which has been developed to embody this characteristic of constant-energy output is completely described in
the following article. — Editor.
A large amount of work has been done to pro-
duce direct-current machines having inherent
regulation; i.e., having certain characteristics
obtained by properties of the windings with-
out the use of any external regulators.
The simplest and most successful machines
having inherent regulation are the compound-
wound direct-current generators and motors.
The compounding may either be acoimu-
lative (boosting) or differential (bucking).
Accumulative compounding is used to main-
tain a constant potential in shunt-excited
generators. It is also applied to motors in
order to give characteristics lying between
those of the shunt and the series motor.
Differential compounding in generators pro-
duces unstable conditions except when the
generator is separately excited. This latter
condition exists, for example, in a form of
generator which is often used for charging
storage batteries at an approximately con-
stant rate. In motors, differential compound-
ing is not used since an increase of the load
causes an increase of the speed, a condition
which is unstable.
The compound winding owes its success
to the fact that it can be applied to direct-
current machines without any structural
changes whatever.
There are, however, a great many problems
in direct-current engineering that require
motors and generators having inherent char-
acteristics which compound windings cannot
give. While a large number of attempts have
been made in the past both here and abroad
to produce machines of such inherent regula-
tion, none of these attempts have met with
any appreciable success. One reason for this
failure lies in the difficulties arising from
commutation. The greatest forward step
in the design of direct-current machines was
the introduction of commutating poles, which
so far represents the most powerful method
of obtaining perfect commutation. It may
therefore be expected that, no matter how we
otherwise construct direct-current machines,
the solution must include proper means for
the application of commutating poles.
Some years ago the writer developed a new
type of direct-current generator having in-
herent regulation which type may be called
"A Direct-current Machine with Dual Mag-
netic Circuits." This design supplies means
for obtaining a great variety of character-
istics and in this article the electro-magnetic
properties of an inherently regulated arc-weld-
ing generator will be discussed. This recently
standardized type of generator possesses the
characteristics shown in Fig. 1. Experience
Amp
Fig. 1. The Dnirable Type of Voltage-current Characteristic
for a Single-operator Arc-welding Generator
Fig. 2. Elementary Diagram of the Conttant -energy Arc-
welding Generator, Showing the Paths of the Magnetic
Circuits and the Poaition of the Bruahe*
A NEW TYPE OF ARC-WELDING GENERATOR
443
has demonstrated that these are the char-
acteristics desirable in the single-operator
type of arc-welding generator.
In Fig. 1 the open-circuit potential is 00
volts and the arc voltage is 20. The arc-
current is designated as A amperes; and the
Series Wind
+ »
Fig. 3. Simplified Diagram of the Generator, showing the Main
and Cross Shunt Windings and the Tapped Series Field
Winding by which the Voltage-current Character-
istic is Varied in Accordance with Fig. 4
generator is laid out in such a manner that
this current can be set or adjusted to different
values depending on the work to be done.
From one generator the following arc-currents
may be obtained at 20 volts: 200, 175, 150,
125, 100 and 75 amperes, the adjustment
being made possible by a system of taps as
will be explained later.
In Fig. 2 is shown a four-pole field structure
and an armature wound for two poles. In
general the armature should be wound for
half the number of poles contained in the
field. In standard designs of direct-current
machines adjacent poles have opposite polar-
ity, but in this machine the poles are paired
in groups of the same polarity. Thus there
60
\
fh
Curve A Made with ZOO Ampere Connection
Curve BMadewith 175 Ampere Connection
Curve C Made with ISO Ampere Connection
Curve DModewith 125 Ampere Connection
Curve CMode with 100 Ampere Connection
Curve F Mode with 75 Ampere Connection
-
-
55
iv
?:>
y
S'^
^
SO
^
S
S
s*
s-
,,
S
N
N
4S
\
s
\
S
s
\
s
N
s
\
40
\
s
s
\
S
\
S
\
\
\
\
\
\
\
s
\
\
S
s
§50
\
S,
\
\
\
s
-
F
\t
\^
\^
\e
V
25
\
\
s
\
\
\
s
s
\
A
\
s
s
s
\
\
S
s
IS
\
s
's
\
s
\
s
s
S
s
to
\
s.
s
\
V
V
s
'si 1
5
\
\
V,
\
s
N
'
*
'
n
_
1
L_
__
0 10 40 60 SO 100 120 140 160 ISO 200 220 240 260 280 300
Amperes
Fig. 4. Voltage-current Characteristics with Various Tap
Connections of the Series Field Shown in Fig. 3
is a group of two north poles followed by a
group of two south poles, etc.
In order to establish a working theory,
assume that the flux distribution is such as
shown in Fig. 2. There exist two fluxes (/)m
and ^c at right angles; i.e., these two fluxes
are displaced ninety electrical degrees in
space. The flux </>,„ will be designated the
main flux and the flux 0^ the cross flux. If
the excitation of the main poles is varied and
Fig. 5.
A Development of the Diagram in Fig. 2, showing the
Addition of the Field Windings
. 6. A Further Development of the Diagrams in Figs. 2 and
5, showing the Addition of Commutating Poles and the
Use of an External Reactance
444 Mav. 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 5
at the same time the excitation of the cross
poles is kept constant, we obtain a change
in the main flux, but the cross flux remains
constant; vice versa, if the excitation of the
cross poles is changed and the excitation of
the main poles held constant, then the cross
Fig. 7. Rear View of Constant-energy Arc-welding Motor-
generator Set Complete with Starting and
Control Equipment
flux is varied and the main flux remains con-
stant. The reason for this independent action
of the two fluxes lies in the fact that the poles
are symmetricall)- located and thus
one pair of poles belonging to one
magnetic circuit lies at points of
equal magnetic potential with ref-
erence to the other magnetic cir-
cuit. Exactly the same reasoning
may be applied to the structure
of a standard commutating pole
machine having the same number
of commutating poles as main
poles. In the commutating pole
machine we can distinguish two
independent magnetic fluxes; viz.,
the main exciting flux and the com-
mutating flux. These two fluxes
are absolutely independent of each
other provided that no saturation
exists in those parts of the mag-
netic structure common to both
fluxes; viz., the field yoke and the
armature core.
The load current of the armature
is taken from the two brushes .4
and C, Fig. 2, placed in neutrals
located between poles of opposite
polarity. As soon as the armature
is loaded an armature reaction is
built up which reaction may be resolved into
two components at right angles; viz., OD act-
ing in the direction of the main poles and OE in
the direction of the cross poles. The compon-
ent OD supports the main flux and the com-
ponent OE opposes the cross flux. The main
magnetic circuit is so designed that magnetic
saturation exists and the component OD.
therefore, cannot force any more flux through
this circuit. Hence, the main flux remains
constant independent of the load. The cross
magnetic circuit, however, is not saturated;
hence, the component OE blows out the cross
flux which thus decreases as the load increases.
If a third brush B is placed in the neutral
between poles of the same polarity it is
obvious that the voltage AB remains con-
stant since the conductors on the arc AB are
cutting a constant flux at a constant speed.
Advantage is taken of this fact to supply the
excitation of the generator from these two
I)oints which possess a constant difference
of potential and thus secure an inherently
stable machine at all loads. The diagram
of connections for this shunt excitation is
shown in Fig. 5.
Referring again to Fig. 2 it may be observed
that the voltage BC decreases as the load
increases, since the conductors over the arc
BC cut at constant speed a flux which
decreases as the load increases. Fig. 2
Fig. 8. Front View of the Wrldmg Outfit Shown in Fig.
Mounted on a Truck for Portable Use
shows that the main flux and the cross flux
are entering the armature between the
brushes .4 and C from the same direction.
Hence, the electromotive force is induced in
the same direction along the arc AB as along
A NEW TYPE OF ARC-WELDING GENERATOR
445
the arc BC. Therefore voltage AC^AB^
BC volts.
Since the voltage AB is constant and BC
decreases with the load, the line voltage AC
must decrease with the load. The generator
is designed to give 60 volts at no load :
hence
AB = BC = :iO volts
AC = AB + BC = m volts.
At a certain load the arraature reaction is
just strong enough to counterbalance the
cross excitation; i.e., the cross flux disappears.
Then neglecting the small ohmic drop AC =
AB = 30 volts. As the load increases over this
\'alue the cross flux reverses and the voltage
BC becomes negative. At a certain current,
200 amp., AC = AB+BC = S0-10 = 'H) which
is the arc voltage.
If desirable to weld with a smaller current,
a series winding is placed on the cross poles
having such a polarity as to support the arma-
ture reaction, which means that this series
field opposes the cross shunt. This series
winding is shown in Fig. 5 and is sufficiently
strong to limit the arc current to 75 amp.
By aid of taps brought out from the series
field, the number of active turns may be
varied in accordance with the diagram shown
in Fig. 3 and anv of the following arc currents
can be obtained"; 200, 175, 150, 125, 100 and
75 amp.
In Fig. 4 are shown the characteristic
performance cun-es of this generator. Each
cun-e corresponds to a series field tap (Fig. 3) ■
viz.,
200 amp. corresponds to tap 1
175 amp. corresponds to tap 2
150 amp. corresponds to tap 3, etc.
In order to obtain perfect commutation,'
commutating poles are added at the points
A and C, Fig. 6, from which the load current
is taken. At the point B there is no need of
any commutating pole since from this brush
only a very small current, the exciting current,
is flowing from the commutator. Observation
has proved that by aid of these commutating
poles perfect commutation exists even at the
highest loads. In Fig. 6 is also shown a
reactance in series with the load. Experience
has shown that it is easier to hold a steady
arc if a reactance is used since this steadies
the current. However, expert arc-welders
can weld without the use of this reactance
but it is a part of the standard outfit since
it has been found desirable for general
applications.
The full-load speed of the generator is
1750 r.p.m., this speed having been selected
in order to make it convenient to couple this
generator to a four-pole 60-cycle induction
motor. The horse power of the motor is suffi-
cient to drive the generator at the maximum
Fig. 9. Oscillogram of Arc Voltage and Welding Current of
a Constant-energy Generator showing That the Inter-
action of the Voltage and Current is Instantaneous
output. In Fig. 7 the generator is shown
direct driven by a standard three-phase
induction motor, complete with all controlling
equipment. Fig. 8 shows the arc-welding
outfit arranged for portable use.
This general arrangement enables the
manufacturer to produce one single type of
arc-welding generator which may be stocked
like any other standard product. With a
standard induction motor or a standard
direct-current motor it will form an arc-
welding set, or it may be belt driven, which
from a manufacturing standpoint gives the
least possible complications.
The generator has the appearance of an
ordinary four-pole machine with two com-
mutating poles. Its manufacture therefore
offers no new problems, standard methods
being employed throughout.
A thorough and extended investigation
shows that a perfect weld can be produced
by aid of this machine. One reason for this
fact lies in the instantaneous action of the
voltage and currents which may be seen from
the oscillograph record in Fig. 9. It should
be borne in mind that the regulation of this
generator is mainly produced by the arm-
ature itself. Since the armature is the seat
of the induced voltage, it is obvious that if
the armature itself is the seat of the regulating
power this action is as intimate as can be
obtained. Experiments have also shown that
446 May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 5
it is easy enough to produce a machine which
for a steady load gives proper regulation
(Fig. 1), but when used for arc-welding
absolutely fails due to slow regulation. Such
machines will not hold the arc properly and
the welds produced are unsatisfactory.
Another advantage of this generator lies
in the small amount of power necessary for
welding. This may be realized from the fact
that a 73^-h.p. motor is sufficiently large to
take care of the highest arc output which is
4 kw. It is also of interest to note that when
operating from an alternating-current system
the power-factor is high, corresponding to
the power-factor of the induction motor used
to drive the generator.
A New Type of Gathering Locomotive
By John Liston
Publication Bureau, General Electric Company
The output per man in the coal mines of the United States is greater than that of any other countrj'. This
result is largely, if not wholly, due to the fact that more machinery is used in the mines of the United States
than in those of other countries. The gathering locomotive is one of the most important factors in securing a
high per capita output, and its development and refinement has been the subject of continuous study by both
mining engineers and electrical manufacturers The particular type covered by this article embodies the latest
developments which practical experience has shown to be most desirable for the attainment of a positive and
.simple system of control which, at the same time, makes it possible to secure maximum operating economy. —
Editor.
The electric gathering locomotive has be-
come such a valuable factor in the economi-
cal quantity production which modern indus-
trial conditions have rendered imperative in
coal mining, that changes in its design and con-
struction tending toward improved control
and operating characteristics, reduced main-
tenance costs, and increased length of service
are of great practical importance to both coal
mine operators and their engineers.
The new type of gathering locomotive
shown in Fig. 1 combines in a single two-
motor unit five new features, all of which have
successfully withstood such severe and long
continued tests in practical coal mine sen-ice
as to demonstrate fully their general utility.
Electric Braking
By means of a new type of controller (Fig. 2)
positive and graduated electric braking is
secured. Heretofore, in gathering work,
there has been a great deal more effort
expended in operating the brakes than with
the heavy haulage locomotives. The haulage
locomotive, as a rule, starts from the same
given point and ends at the tipple or shaft
bottom a considerable distance away. En-
route there are few, if any, stops and the
motorman is therefore seldom required to
operate brakes on the way. In gathering work
the locomotive ordinarily starts and stops
many times on account of the switches to be
thrown at the room necks, and couplings hav-
ing to be made to each individual car; there-
fore, while the gathering locomotive is lighter
than the main haulage locomotive, the sum
total of braking effort expended by the motor-
man is considerably greater.
The new controller was designed with the
view of relieving the motorman of a large part
of this braking and operates so that the
locomotive is stopped by its own momentum.
This is accomplished by providing on the
controller reverse c}-linder a set of connections
that turn the motors into self-excited gener-
ators and the energy- developed by them is
absorbed in the main resistors. The amount
of this energy, and consequently the degree
of braking effort, is governed by the main
cylinder of the controller. The more resist-
ance cut out of circuit, the more quickly
will the stop be made.
The reverse cylinder of the controller is
provided with four points, two for each
direction of motion. For the first of these
points the motors are connected in the
regular motoring position. When it is
desired to stop, the main cylinder is thrown
off in the usual way, and the reverse cylinder
is thrown to the second, or braking point.
The main cylinder is then turned on again
and the motors (or generators, as they are
now) begin to retard the locomotive.
The degree of braking is under the motor-
man's control at all times, for if he finds that
he is stopping too quickly, he merely has to
throw off the main cylinder and permit the
locomotive to coast.
In numerous tests it was demonstrated
that, with the trolley disconnected, the resid-
ual magnetism of the motors, when acting as
generators, was sufficient to insure the maxi-
mum braking effect with no ai)preciable
difference in the time element involved as
compared with the results obtained with the
trolley connected. This is an important
A NEW TYPE OF GATHERING LOCOMOTIVE
447
factor in estimating the all-around ser\-ice-
ability of electric braking for gathering work.
On a level track the motorman can bring
his train to a dead stop without using the
ordinary hand brake at all. He can also
bring it to a stop on a grade, but since there
is no energy developed when the wheels have
stopped turning, the locomotive will start
and continue to roll, stop and start again unless
the hand brakes are set. A runaway is, how-
ever, impossible so long as the train weight
and grade are within the braking capacity of
the locomotive.
With electric braking, the hand brakes,
therefore, need to be used very little, and as
a result there is very great reduction in the
motors sustain a heavy rush of current and
the gearing and other parts of the mechanical
equipment receive very severe shocks, all
of which tends to shorten the life of the
various parts, and runs up the maintenance
costs.
The controller is different in another way
from the ordinary mine locomotive con-
troller. Most controllers at present are
arranged so that the locomotive will start
either with the motors in series or with the
motors in parallel. Here again the indif-
ferent motorman will not use the series posi-
tion when running slow. Instead he will
leave the reverse cylinder in the parallel
position and get slow speed by running on a
Fig. 1. Eight-ton Gathering Locomotive Equipped with Electric Braking Controller
wear of brake shoes and wheel treads as com-
pared with hand braking. Further, with
electric braking, since the braking effect is
zero as soon as the wheels have stopped
rotating, there is practically no skidding of
the wheels, and consequently there will be
very few, if any, flat spots developed from
this cause.
There is another incidental benefit with
this type of controller. With the ordinary
controller, careless or indifferent motormen
do not always use the hand brakes when they
want to stop. In many mines it is a rather
too frequent practice for the motorman to
save effort by reversing the motors when he
wants to stop. Stopping in this way, the
resistance point. This increases maintenance
costs of resistors and, while on slow speed,
consumes twice as much current as if the
motors were in series.
This question of additional current con-
sumption may not in many cases represent a
serious economic loss, but when a number of
gathering locomotives are used the total
amount of energy wasted in this way in a
year of ser\dce is always a matter of serious
consideration to the mine engineer who is
desirous of maintaining a high overall effi-
ciency for the electric system of the mine.
The electric braking controller is a positive
insurance against this particular form of
waste as it is of the series-parallel type
448 May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 5
(Fig. 3) similar to that used on the ordinary
street car and the first point of the controller
is always series-motors. Therefore the motor-
man cannot get to parallel until he has gone
through all the series points.
Fig. 2.
Electric Braking Controller for
Gathering Locomotive
In Fig. 4, the controller is shown with the
arc chutes in normal operating position and
it will be noted that the apertures are greatly
restricted as compared with ordinar\- con-
troller construction. This arrangement was
adopted after exhaustive tests had demon-
strated that by this means the arc could be
extinguished in about one third of the time
required with the more open form of arc chute.
This detail insures longer life for the con-
tacts as for all practical purposes their length
of service is inversely proportional to the time
of duration of the arc.
Outside Frame
The outside frame construction adopted for
the new locomotive (Fig. 5) is ver>' substantial.
The side frames are cut from solid rolled steel
plates; the end frames are built up with
structural steel channels, rolled slabs, and
wood bumpers protected by heavy face
plates.
The outside frame in this case differs from
the usual machine of this type, in that the
clearance between the rail head and lower
edge of side frame is ver\- high. Ordinarily
an outside frame machine would clear the rail
head at this point by three or four inches.
When the locomotive derailed, the frame
would settle down to about the level of the
rails and cover up all access to the wheels.
With this new construction, the high clear-
ance of the frame p«Tmits access to the lower
part of the wheel so that blocking or other
re-railing devices may be put into position,
and the locomotive gotten back on the track
practically as quickly as if the wheels were
outside of the frame.
Brahirvg Forward
' )Motori(\g Fbrward
^'Off
BrAKirvg Reverse \ ;
Motorirvg Reverse
,Fig. 3. Top of Electric Braking Controller Showing
Control Levers
Fig. 4. Arrangement of Restricted
Arc Chutes on Electric Br»k
mg Controller
A NEW TYPE OF GATHERING LOCOMOTIVE
449
There are, of course, some mines where side
clearance will not permit the extra width of
the outside frame type machine, but as a rule
the main objection to this type of construction
has been the difficulty of getting it back on the
track in cases of derailment. The high clear-
ance feature should remove this objection.
On the other hand, the outside frame con-
struction permits the use of a better journal
box, in that it is entirely enclosed at one end,
and the back end can be fitted with a dust
guard. With an inside frame both ends of the
box must be open and boxes must also be
made in two halves, which is not as good
construction either theoretically or practi-
cally. With the outside frame the greater
space between the side frames allows more
room for the equipment and permits the use
of a liberal amount of space for the motor-
man's cab.
Leaf Type Springs
Heretofore, practically all two-motor loco-
motives have been equipped with the round
wire coil type of journal spring, whereas
the heavier ' three-motor locomotives were
provided with leaf type springs to insure
smoother nmning and better distribution of
the weight on the drivers on rough and uneven
tracks.
As the result of operating experience gained
with the three-motor units, the new loco-
motive was provided with semi-elliptic leaf
type springs (Fig. 6) having an equalizing
used, than is the case with helical spring
design and, therefore, they are mechanically
stronger and less liable to breakage.
By using leaf springs and equalizers, the
two-motor four-wheel locomotives will accom-
modate themselves to inequalities in track
Fig. 6. Journal Leaf Springs with Equalizer Bar
levels, for the reason that any change in wheel
load is transmitted through the equalizing
levers to the other wheels, thereby practically
equalizing the weight on the drivers. Inci-
dentally, the equalizing lever greatly increases
the range of spring action and the tendency
toward derailment is thereby minimized.
Finally, the improved riding qualities of
the locomotive tend to reduce the wear and
tear on the track and roadbed.
Improved Cable Reel
This cable reel is an improved form of the
vertical axis motor driven type which has
been used successfully for a number of years.
No change has been made in the ball bearing
Fig. 5. Outside Frame Construction showing Liberal Space Available for
Equipment and Motorman
bar between the two journal springs on each
side.
Due to limited space in the overall dimen-
sions of mine locomotives, the leaf springs
can be designed with much greater margin,
approaching the elastic limits of material
motor, but the bearing mechanism of the reel
itself has been modified to secure greater
stability and better wearing qualities.
Instead of a large diameter of bearing made
up of a large number of small balls, the reel
now rotates on a heavy duty type combina-
450 May, 1920
GEXER.\L ELECTRIC REVIEW
Vol. XXIII, Xo. 5
tion thrust and step ball bearing (Fig. 7)
mounted at the center of the reel disk. The
double reduction train of gears is made up
entirely of forged steel gears and pinions, heat
treated.
be made without taking the axle out of the
frame.
The advantages of these demountable tires
will be fully appreciated by anyone who has
had to replace a shrunk on tire, as the com-
plete replacement of the new tire can be
effected by two men in about fifteen minutes
for each wheel.
With the co-operation of Mr. W. A.
Chandler, Electrical Engineer of the H. C.
Frick Coke Company, a 6-ton gathering
locomotive with an experimental braking
equipment of the tj-pe here described was
placed in ser\-ice at the company's coal mines
near Uniontown, Pennsylvania, and has been
in successful operation for the past two years,
during which time the necessan,- refinements
were worked out under actual service con-
ditions.
A 20-ton main haulage locomotive with
similar braking control has been handling
loaded trains on a 40U0-ft. line with four t<'
five per cent grades for about one year. It is.
Fig. 7. Under Side of Cable Reel showing
Motor-driven Gear Train
Demountable Tires
The construction of these tires (Fig. S) is
very simple and consists merely of two wedge-
shaped steel rings drawn together at suitable
intervals by bolts. In drawing these rings into
position the tire is forced to take its proper
alignment with respect to the wheel hub and
gauge line, and the wedging action of the
rings locks it securely in place.
It will be appreciated that the renewal of
this tire is a very much shorter job than in
the case of the ordinary shrunk on tire.
Socket wrenches are the only tools required,
and the change can be made in the locomotive
bam. With outside frame locomotives it is
only necessary to drop the axles, and with
inside frame locomotives the change can
Fig. 8. Drmountablc Tire* Assembled
and Disassembled
therefore, evident that the principles embod-
ied in the new gathering locomotive have fully
demonstrated their value in actual ser\Mco.
The new locomotive was designed by the
engineers of the General Electric Company.
Self-interest Will Solve the Problems Confronting
Electrical Development
Arranged for General Electric Review from an address before Schenectady Section A.I.E.E.
by A. Emory Wishon
Assistant General Manager San Joaquin Light and Power Corporation,
Pacific Coast Manager N.E.L.A.
In many respects the central stations of the country arc in the same position as the railroads; their rate
of income is regulated by commissions and the enormous increase in the cost of labor and materials has
reduced the return on the investment to a point where frequently it is difficult or impossible to attract
additional Capital for new development that is urgently needed. The solution of the difficulty lies in making
a clean breast of the situation to the public, and showing each community just what it is losing from the
inability of the central station to supply increased service. The central station that is fair and square in its
dealings has nothing to fear from taking the public into its confidence; and by convincing the individual of
the fact that whatever hinders the development of the central station will also keep money out of his pocket
the greatest obstacle facing the electrical industry will have been surmounted. — Editor.
The value to a public utility of preaching
the doctrine of Self-interest can be deduced
from the case of the railroads. A few years
ago we heard from time to time that the rail-
roads were in need of more revenue in order
to pay a fair return on the
capital invested and in this
way to attract new capital
for the ]3urpose of increas-
ing and replacing rolling
stock and improving trans-
portation facilities gener-
ally. We gave the matter
no thought; in fact, we
were not inclined to believe
that there was a real need
for additional revenue. We
preferred to believe that
the railroads were receiv-
ing ample return for their
services and straightway
forgot the matter. Later
on, when the war broke out
and our business was hurt
becati.se we could not get
shipments through, when
there was an embargo on
freight in all parts of the
country, and we could not
even get coal to run our
factories because of the
inability of the railroads to
haul it, we began to take
a personal interest in the
railroad situation; and lately we have actu-
ally become sympathetic with the argtiments
of the railroad managements for revised legis-
lation to enable them to earn a fair rate of
return. The average business man has lost
a great deal through the inability of the rail-
A. EMORY WISHON
roads to provide prompt and adequate freight
service, and today he realizes that his pros-
perity and that of his community is dependent
on the success of the carriers. He is now
an advocate of a sufficient increase in freight
and passenger rates, al-
though he knows that he
and his fellow citizens are
the ones who will have to
pay this increase. Self-
interest of the individual is
reflected from every angle in
the recent national legisla-
tion that has been enacted
for the benefit of the rail-
roads.
The electrical industry
today is equally as impor-
tant to the public as are
the railroads, and the
problems of the central
station — the source of elec-
tric energy — are very simi-
lar to the problems of the
railroads. In each case
their rates of return are
regulated by commissions.
Sufficient rates to return
the atithorized amount of
earning on the investment
will enable the power
company to finance and do
more development work ;
lack of rettirn will restrict
its expansion and accordingly the expansion
of every line of business connected with the
electrical industry.
The following statements are axiomatic:
(1) Central station development is the
barometer that indicates the degree of
452 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. ^
prosperity in the entire electrical industry,
because the central station is the source
of electric energy that is required by all the
common electric appliances, such as railway
and power motors, arc and incandescent
lamps, electric furnaces, flatirons and other
heating devices in the household, etc.
(2) If the electrical industry is to thrive
the central station must develop to its fullest
capacity. Central station expansion requires
financing, and to attract the necessary capital
a protected investment and a fair return
are necessary.
(3) A fair return is possible only through
unbiased and fearless legislation.
(4) Legislation should, and usually does,
represent the opinion of the voter.
An analysis of these statements leads to
the conclusion that the greatest problem that
faces the electrical industry today is to make
the public understand what electricity is
doing for the nation. This is a big under-
taking, but it is not so difficult as it appears
at first sight.
It is possible to create real personal and
active interest in the progress of electrical
development in any community by impress-
ing on the mind of the individual what the
world would be today without electric energy,
what it would mean to be without electric
light or power, without our trolley cars.
telephones and telegraphs, without the many
comforts and conveniences provided by
electricity in the home, and without pos-
sibilties. The electrical industry is that
wonderful activity which makes possible the
world of today, on which depends every line
of business for efficient and speedy pro-
duction, and on which will be based the
standards of living of the future.
However, these advantages conferred by
electricity are now largely taken as a matter
of fact by the public, and the ptiblic will not
be interested and will not understand the
problems that confront the electrical industry
until it is shown that these problems are also
the problems of the individual, because he
cannot do without the many things that
electricity is doing for him every day. It is
necessary to show every man separately what
stagnation in the electrical industry will cost
him in dollars and cents, or more correctly,
what he stands to gain by exerting his efforts
and influence to the end that electrical
development may progress to the fullest
possible degree.
It is a relatively simple matter to show that
any fonn of legislation, national or state.
that delays electrical development delays the
development and prosperity of the state and
of the nation, and any thinking man by means
of a few figures can show where any particular
business loses in dollars and cents when
hydroelectric development ceases. Prove
j'our case to the individual by proving that
the individual's pocketbook is hurt when
your business is hurt and you will have a
champion who will see that your business
will prosper.
When electrical development ceases the
prosperity of the dealer, contractor, jobber
and manufacturer ceases, and when electrical
development is encouraged their business is
furthered.
We will consider conditions that exist in
the West. Those who have studied the
industrial problems of the West know that
this section of the country- will not develop
ahead of its hydroelectric development.
Large tracts of western territory are arid
lands and depend upon hydroelectric power
for irrigation if agriculture is to be further
extended. If factories are to be built in the
West cheap power must be obtainable.
This western country- is yet a long way
from electric saturation and will be for many
years to come. The engineers and com-
mercial men of the central station companies
have plotted their anticipated load cur\-es
for several years into the future, basing their
estimates upon past experience. By super-
imposing these projected load curves a
composite cur\-e is obtained which shows
what increase in generating capacity can be
expected for this territor>' during the next
decade, if proper encouragement is given to
electrical development. The electrical invest-
ment to date is known and also the total
kilowatt capacity of equipment, and from
these figures the unit cost per kilowatt
installed is determined. We are thus enabled
to determine for any year in the future the
money that will be spent for electrical
development in the West.
Curves for the entire Pacific Coast will be
presented at the National Electric Light
Association's Convention at Pasadena in
May. These cur\-es will show that within the
next eight years approximately §500,000,000
of new capital will be expended in elec-
trical development in this section. From
the standard classification of accounting
established by the California Regulating
Commission for public utilities it is an easy
matter to determine the proportionate
amounts that will be invested in generators.
I
SELF-INTEREST WILL SOLVE PROBLEMS OF ELECTRICAL DEVELOPMENT 453
tlams, transfonners, copper, etc., during this
period if normal conditions continue. These
data are all available and it only remains to
put them together in graphic form to prove
to the individual and to the different interests
what this tremendous electrical development
in the West will mean to each one in dollars
and cents.
To show how this applies to an individual
we will take the case of the manufacturer's
representative. He is shown by definite
graphic proof that if conditions are encourag-
ing a certain load can be added to the Pacific
Coast central station systems in the next
year. We know the cost of construction
per kilowatt of capacity and quickly figure
out for the manufacturer's representative
just how many millions of dollars will be
required for added generating equipment,
dams, feeders, etc., to take care of this extra
load. Our segregated investment chart
shows, for instance, that eight per cent of
the money required for this addition to plant
will go into transformers. If the manu-
facturer's representative is wide awake and
on to his job he will know at once what part
of this transformer business should be his.
He will know what his net profits are to be in
the sale of those transformers, and he can
figure in dollars and cents just what it means
to him to have this development go through.
«o -
/
U^/r COST PCR KW OF
PEAX FOR 19:5 AND LAT^ff
/
HI -
/
139
120
iia
100
90
80
70
60
50
id
/
/
2'i -
.V
/
/
4^
/
y
2(165-
.^
f
A
/
.p
/
1365-
^
f\
y
108 -
^
/
/
9S -
217-
^
■^
>
/
^
/
^
^
?n
1
lib
19
1/
B
8
13
19
IS
4o
19
^l
19
11
19
?3
Fig. 1. Curve Showing Estimated Peak Loads of the San
Joaquin Light and Power Corporation
This is a direct appeal to the man through his
pocketbook, and he will be interested to the
extent that he will use his best influence to see
that conditions are favorable for the develop-
ment.
He does not stop there, however. He
wishes to convey to the manufacturer the
impression that he is on the job, and accord-
ingly communicates the glad message of Self-
interest to the sales manager, who will
straightway become interested in furthering
electrical development on the Pacific Coast.
/
na
J
f
1922
/
^/
1921
i^/
^/
f
i
7
1919
I
1919
.1
1917
1916
Fig.
10 15 » i5 30 56
Millions of Dollars
Total Capital Invested, Exclusive of Real Estate,
Supplies on Hand, etc., San Joaquin System
The manufacturer in turn spreads the message
to the steel mill, to the copper mill operator,
and to the railroad that will transport the
equipment, emphasizing the importance to
each one of encouraging electrical develop-
ment on the Pacific Coast.
That you may judge of the eft'ectiveness
of Self-interest in promoting favorable con-
ditions for electrical development, three
cur\'es are shown that have been compiled
by the San Joaquin Light & Power Corpora-
tion. Not forgetting the five hundred million
dollars we have said will be required for
electrical development in the West during
the next eight years, we will, however, con-
fine ourselves to the lesser requirements of the
San Joaquin Company, which are typical.
This company is at present working night and
day on a new 30,000-kw. hydroelectric plant
that will be in operation by September.
Also, five engineering crews are completing
the survey of a hydroelectric development
that will have an ultimate capacity of 160,000
kw. Conditions are the same the West over;
the public is clambering for electric ser\-ice
and the power companies are making every
effort to serve.
4o4 Mav, 1!)2()
GENERAL ELECTRIC REVI1-:\V
VoL XXIII, No. o
carpcnler to completely win him to this
hydroelectric development which make pos-
sible this home building program.
The results of efforts in the West in spread-
ing the doctrine of Self-interest demonstrate
beyond question that the solution of the
greatest problem that faces the electrical
Fig. 1 shows that during the ne.xt four
years the San Joaquin Company will require
$2(i, 0(10, 000 for hydroelectric development.
This statement is of monetary interest to a
great variety of businesses and individuals.
The Eastern bond house is interested in this
expenditure by the amount it means in
commissions on the underwriting and sale of
bonds; the electrical jobber is interested to the
extent that he will profit from the sale of
electrical supplies, not only supplies re-
quired for the development but those
which will be needed by the new indus-
tries which will make the development
necessary. The electrical manufacturer
is interested in this hydroelectric de-
velopment in the West because of the
profits he will derive from the manufac-
ture of apparatus; and viewed from
another angle, the various local indus-
tries are interested to the extent to
which they depend upon electric power
for operation and expansion.
Fig. 2 is a diagram that immediately-
excites Self-interest. The item that will
first command your attention is the item
that affects your business; it is invari-
ably so with ever>' man, and it is this
fact which proves that the theor>- of
interesting through Self-interest is sound.
Fig. 3 is a segregation by percentage
of home building costs. From over 2000
applications for power service made to
the San Joaquin Light & Power Com-
pany 1200 were applications for agricul-
tural service. Six hundred of these were
for the development of new lands re-
quiring homes, barns, fences, and all sorts
of fann equipment. Figures that have
been compiled show that for even,- home ^'^ ^ '-"'"'" s''°»""« scgrcBat.on p.g j. segregation of Homr building
m the country there are two homes in
town. Therefore, when electric ser\ice
is rendered on these farm a])i)lications
and the GOO new farm homes are built
there will also be 1200 town homes con-
structed, or in all ISOO additional homes will
have been built due directly to electrical de-
velopment. This represents a building con-
struction program involving .'§10,242,000. and
from our segregation diagram it is possible for
the lumberman, the brick manufacturer, the
lumber mill, the electrical jobber, and all the
several trades involved to figure just what
percentage of this business is his. With
I5H per cent of this investment in new
buildings going to carpenter work it is onh-
necessary to point out the figures to the
C2
lis
COti
06
■•J
16
''JM C^VSjMOfS PRE^IISES-^
TCLTPTOVf USEi [OUIP-^±^
CO'^CCIAL LA^PS a ECUIPiB
MUKiC^PAL Sr LIOHTIHCr-'^'m
ifMCTTn ■
ELECTtfiC SERVICES yS
7S
LIHE TFMSFOFMEPi 4 Ic
DEVICES ■ :;
cm
a
OS
MISC EQWPMEIfT B
SUBSTATIM EQUIP ■;
XlBSTATIOH BUILDI165 & B-
STRUCTUHES ■;
Hi
OVERHEM) SYSTEM l£
U)
fiCI.ES 4 FIXTURES
OOCI W5C EOUIPMLNT . .nr »"
3« iUBSTATIOM EQUIP % £
OJB SUBSTATION BUILOINOS . 3 >
75
oyLRHEAD srsrcH 1
3 ->
50
POLES & raruKES Bl a
MiSC EQUIPMENT B^ o
Vo'^-'prTcl^'"'^ 1 c.
rvRKACLS. BOILEJH M '^
HTDRO POnCKPUtlfT M£ <
EOUIPMEKT ■?; O
01
33
49
•♦6?
POWtR PLANT BUlLDINiS IC g
OENERAL srnCTURES ' Jj
no
1-
OAMS. nATCR D
CONDUITS & PENSTOCKS g
0:
a.
Mahtel -—
ail
TO CONTRACTOR
11
^
HARDWOOD rtOORS
PALNTI.Va
ELECTRICAL VVORK
1 PLfMBINS & 6A5
ROOFING
'f
oii
385
095
CARPENTER WORK
15 5
z
D
5
m
u
I
o
X
ll.
o
\~
in
o
\>
ce;
O
<
MILL WORK
9.0
1 PLASTTRIN6
BRICK WCffK
MANTEL
165
Z*
046
BWLDIN6 SITE
US
o
y-
<
HARDWOOD PLOORS
PAINT
ELECTRICAL WORK
PLUMBIN6 *
6AS FITTIN6
R00PIN6
NAILS & HARDWARE
MILL WORK
116
zss
zo
S87
1*
14
4»5
o
-i
LUMBER
140
1
LATH & PLASTER
BRICK
ZOM CRETE
of Capital Investment of San
Joaquin System. The item which
first attracts your attention is
the one that affects your busi-
ness, which proves the soundness
of the doctrine of Self-interest
Costs on Pacific Coast. This is
a Self interest chart which may
be used with unfailing success to
enlist the co-operation of every
business in the community
industPi' today will be an accomplished thing
if ever}' man in the ranks makes it his aim and
duty to imi)rcss upon the public what the
development of the electrical industry- means
to the individual in dollars and cents. When
a realization of the simple facts sinks home
the public will see to it that nothing interferes
to delay electrical development, that fair
and adequate legislation is enacted antl
supported to the end that the return on the
investment will be suHicient to attract the
necessary capital.
455
The Mariner: The First Electrically Operated
Trawler
By John Liston
I'l iiiiLATioN Bureau, General Eelectric Company
In previous articles in the Review we have described the electric propulsion equipments of the U. S.
cdllicr Jupiier and the battleship New Mexicv. With such large vessels the question of economy is- of prime
consideration, but for vessels smaller than, say lOOO tons displacement, it is doubtful whether this factor
alone would warrant the additional expense necessary for electric generators, motors and control equipment.
For such vessels it is the flexibility afforded by electric propulsion that strongly appeals to the marine en-
gineer. The electrically propelled vessel described in this article is of only 500 tons displacement. — Editor.
tained uniform rate of rotation for the
engines, positive control of the propeller
speed at all times, a high factor of safety by
means of three separate control station,
]jractically instantaneous reversal of the
propeller, and the use of electric motors for
driving auxiliaries such as pumps, com-
pressors, hoists and ventilating blowers.
The craft is of wooden construction, and
is rated at 500 tons with dimensions as
follows: length, over all, 150 ft.; beam, 24 ft.
The adoption of electric propulsion for the
beam trawler Mariner (Fig. 1) was the logical
result of the efficient and economical operation
secured with this system in numerous craft
of various kinds, both in Europe and America
during the past twelve years
In designing the equijjment for the Mariner.
the inherent flcxibilit_\- of the electrical method
of power application made it possible to
obtain high economy in fuel consumption,
especially under cruising conditions, sus-
Fig.
The Beam Trawler Mariner
456 Mav, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 5
THE MARINER: THE FIRST ELECTRICALLY OPERATED TRAWLER 457
Fig. 3. Engine Room (looking forward), showing Arrangement of Diesel Engines Driving the Main Generators
Fig. 4. Engine Room (looking aft), showing Main Generators with the Engine Flywheels Carried on the
Generator Shaft gearings
458 May, 1<I2()
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 5
•i in. ; mean draught, 1 1 ft. 9 in. Her cruisinj;
radius at 10 knots is (jOOO miles and, at three-
quarter speed, 9000 miles.
The propelling equipment (Fig. 2) com-
prises two eight-cylinder, four-cycle, 350
r.p.m. Diesel engines (Fig. 3), each direct-
connected (Fig. 4) to a 165-kw., 12.5-volt,
direct-current generator. The two self-
excited generators are normally connected in
series and supply current to a 400-h.p. 250-
volt, 200-r.p.m. motor (Fig. 5), which is direct-
coupled to the propeller shaft.
waterproofed, and the machines are so de-
signed as to prevent flashing in the presence of
moisture, due to either atmospheric conditions
or flooding of the engine room in rough seas.
In order to insure ample mechanical
strength for the electrical machinery, steel
castings were used for all rotating parts
which would be subjected to unusual strains,
or to shocks incident to operation during
stormy weather.
The 400-h.p. propeller motor is located
forward of the generating sets (see Fig. 2)
Fig. 5.
Forward End of Engine Room (looking forward), showing Main Generators and Propellor Motor
with Master Controller at Right
Two control stations are located in the
engine room — one provided with remote
control, and one arranged for emergency
manual operation; and a remote-control outfit
is also located in the pilot house.
Both the generators and the motor are
designed specifically for sea duty, and are
Ijrovided with non-corrodible fittings and
heat-resisting insulation throughout. The
bearings are a combination of waste-packed
and oil-ring type, with sjiecial jirovision
against the leakage of oil along the shafts,
when the machines are out of their nonnai
positions, due to the rolling and pitching of
the ship. Finally, armatures and fields are
and has a normal full-load speed range of
from KiO to 200 r.p.m. It is a compound-
wound machine and, when taking current
from both generators, it operates at 250 volts;
but, for slow cruising, one engine can he shut
down and the motor then receives current at
125 volts. Under these conditions it has a
speed range of from 70 to KiO r.p.m.
The proi)eller is 94 in. in diameter by (i.S in.
pitch and. at full-load rotation of 200 r.p.m..
gives a sjjeed of between 7 and 10 'a knots.
de|)ending ujion weather conditions. When
hauling the net the full horse ))ower of the
motor is developed at a propeller speed of
KiO r.p.m.
THE MARIXER: THE FH^ST ELECTRICALLY OPERATED TRAWLER 459
Fig. 6. Main Contrul Panel in Ent^mc Ruum and at Left
Motor-driven Bilge and Water-supply
Centrifugal Pumps
Engine-room control of all electrical cir-
cuits is secured by means of a main panel
board (Fig. 6), on which are mounted the
engine-room meters, generator field switches
and rheostats, switches and fuses for the pro-
]jelling and auxiliary motors, and an overload
rclav for the main hoist motor. The meters
arc mounted at the top of the panel and are
special instruments designed for shipboard
work being equipped with moisture-proof,
non-corrodible parts. The dais are black
with white markings, with radium ]5aint on the
needles and dial markings. A duplicate set of
these instrumenst is installed in the pilot house.
The starting rheostat resistance consists
of five boxes of grids (Fig. 7), which are
mounted on the starboard side of the engine
room. Just forward of these grids, the con-
trol contactors (Figs. S and !)) are located.
This group consists of the necessary current-
carrying contactors for starting, stopping and
reversing the motor, an overload relay and
motor-shunt field discharge resistance, and is
nonnally operated by means of one of two
master controllers — one located in the engine
room and the other in the pilot house.
During operation from cither of these
master controllers, the contactors are closed
Fig.
7. Engine Room i looking forward, starboard side), showing Bank of Resistance and Amount of
Working Space at the Side of the Engine
460 May, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 5
magnetically; but, if for any reason they
cannot be operated magnetically, handles
attached to cam shafts are provided which
may be operated manually to close the
contactors in the desired sequence. The
overload relay, in case of overload, opens the
circuit through the reversing contactor coils.
system is normally only a convenience, as com-
pared with the ordinary combination pilot
house and engine-room control; but, in enter-
ing and leaving slips in congested harbors, in
narrow and swift current waterways, and for
quick reversal or change of speed in emergen-
cies, its great practical value is obvious.
Fig. 8.
Arrangement of Main Control Contactors front view)
showing Arc Chutes and Flush Barriers
Fig. 10. Type cf Master Controller Used in
Pilot House and Engine Room
Fig. 9.
Main Control Contactors iback viewi. showing
Interlocking Levers for Hand Control
causing them to open the line circuit. The
handles for manual operation arc so inter-
locked that the reversing handle must be
operated before the accelerating handle; and,
therefore, the accelerating handle must l)e
turned ofT before the reversing handle can be
moved. This arrangement insures absolute
safety for the control system of the ship, even
in the very improbable event of failure of, or
injury to, the two remote control equipments.
One of the important advantages of electric
propulsion is that of remote control, which
permits the actual maneuvering of the shi]),
to he accomplished directly in the pilot liouse,
if desired, without the necessity for signals to
the engine room. At sea, this remote control
The type of master controller
installed on the Mariner (Figs. 10
and 11) consists of a cast-iron
frame, with a sheet-metal cover,
in which are mounted a main con-
trol cylinder and a reversing cylinder. The
construction of the frame and cover is such
as to make the controller practically water-
tight, the cover clamping against felt in a
groove in the frame. The control wiring is
taken out of the controller at the bottom
through the base.
There are two handles on the controller —
one main and one reversing. The main
handle rotates the main cylinder, which gives
17 operating i)ositions--one off-in^sition and
one overload relav reset position. The revers-
ing handle rotates the reversing cylinder and
has three ])ositions: ahead, olT antl astern.
These two handles are so interlocked that
the main handle cannot Ix- moved bevond the
THE MARINER: THE FH^ST ELECTRICALLY OPERATED TRAWLER 461
overload relay reset position unless the
reversing handle is in either the head or
astern position, and so that the reversing
handle cannot be moved unless the main
handle is in cither the off or reset position.
The rapidity with which the motor-driven
l)ropellcr can be reversed was demonstrated
Fig. n.
Arrangement of Wheel and Master Controller
in Pilot House
Fig. 12. Spring Thrust Bearing for Propellor Shaft
during the first trial trip when, with the pro-
peller rotating at from 193 to 19(1 r.p.m., it
was reversed from full speed ahead to full
speed astern in thirteen seconds: the actual
reversal of current in the motor being accom-
plished in two seconds.
Pilot-house control was used throughout
the test run, operating either one or both
generators with equal facility. The following
extracts from the report of the trial trip may
be of interest :
"On Wednesday morning, December 3,
1919, started from New London at 8:45 a.m.
and followed a course out around Fisher
Island, easterly between Point Judith and
Block Island, through Vineyard Sound and
around Pollock Rip Lightship; then northerly
to Highland Light at the end of Cape Cod;
and from there straight away for Gloucester,
which we reached at 5 :45 a.m. on December 4,
21 hours after leaving New London over a
course, estimated by the skipper of approxi-
mately 200 nautical miles."
"The best speed attained by the boat dur-
ing the trip was in the easterly part of Vine-
yard Sound where, with a favorable tide, it
reached approximately 1 1 3^2 knots. After
turning north along Cape Cod there was a
head wind with quite a rou^h sea, so that the
speed of the boat was considerably reduced."
"The Mariner ran very steadily and the
absence of vibration was very noticeable."
"No criticism could be made on the electri-
cal design of either the generators or motor,
as the machines operated over the entire
range of load without sparking or distress."
Fig. 13. Upper Housing Removedfrom Spring Thrust Bearing
Instead of the usual rigid multi-collar type
of thrust bearing, a self-oiling spring thrust
bearing (Figs. 12 and 13)ofthesingle-collar, self-
aligning type is used, located aft of the driving
motor and sustaining a thrust of T.oOO lb.
with the propeller revolving at 200 r.p.m.
462 May, l'J20
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 5
In addition to operating the propelling
equipment, electrical energy is used for
lighting and all auxiliary power purposes
and, when the main engines are shut down,
current is supplied by means of an independ-
ent 15-kw., 125-volt, oil-engine-driven gen-
erator (Fig. 15) installed in the forward end
of the main engine room on the port side.
The emergency air-compressor outfit is
driven by a direct-geared motor (Fig. 14),
and is ])rovided as an insurance against the
improbable loss of starting air for the engines.
Under these conditions it will be utilized to
fill the air-starting bottles, as the auxiliary
generating set can be started by hand.
The bilge and water-supply pumps are
small centrifugal units (Fig. (i), each driven
by a direct-coupled motor, and near the main
generators and propeller motor a small motor-
driven ventilating set is utilized to prevent
excessively high temperature in the engine
room.
The fishing operations are carried on by
means of a 6o-h.p., motor-driven, main
double-drum hoist, installed on the main
deck forward of the engine room, which
handles the haulage cables and ropes of the
net as they pass through the hoist brackets
fore and aft (Figs. 15 and Hi) on either side.
The unloading of the fish at the dock is
accomplished by means of a 5-h.p. motor-
driven whip hoist located near the forward
mast.
The Mariner is now regularly engaged in
commercial fishing. She was built by Arthur
Fig. 14. Motor-driven Air Compressor
D. Story, of Essex, Mass., for F. L. Davis, of
Gloucester, Mass. ; the engines were built and
installed by the New London Ship and Engine
Company, of Groton, Conn., and the com-
plete electrical-propelling equipment was sup-
plied by the General Electric Company, of
Schenectadv. N. V.
Fig. 15. Main Deck, showing Forward Net Drawing Tackle
Fig. 16- Hanillina Haul of Fish with MoloroperalcJ Hoist
THE MAKIXKK: THE FIRST ELECTRICALLY ()1M-:RATED 'I-RAWLER 40:i
The Electrically Propelled Trawler Mariner
464 May, 1920
GENERAL ELECTRIC- REVIEW
Vol. XXIII. \o. 5
QUESTION AND ANSWER SECTION
The purpose of this department of the Review is two-fold.
First, it enables all subscribers to avail themselves of the consulting service of a highly specialized
corps of engineering experts, or of such other authority as the problem may require. This service provides
for answers by mail with as little delay as possible of such questions as come within the scope of the Review.
Second, it publishes for the benefit of all Review readers questions and answers of general interest
and of educational value. When the original question deals with only one phase of an interesting subject,
the editor may feel warranted in discussing allied questions so as to provide a more complete treatment
of the whole subject.
To avoid the possibility of an incorrect or incomplete answer, the querist should be particularly careful to
include sufficient data to permit of an intelligent understanding of the situation. Address letters of inquiry to
the Editor, Question and Answer Section, General Electric Review, Schenectady, New York.
TRANSFORMERS: OPEN-DELTA
CONNECTIONS
(128) Where several substations are to be
tapped to the same three-phase line would
it not be feasible to use the open-delta
transformer connection in each substation?
If the open-delta connections are arranged
so that the open phase is different from
station to station, the system as a whole
operates as a closed delta with a rather long
connection between transformers. Although
this does not give as good an operating
condition as does a closed delta in each
substation, still it is one that has been
successfully used in many installations. The
objection to this connection is that the line
currents between substations will not be
balanced and the substations will carry only
86.6 per cent of their rated capacitv.
" E.C.S.
INDUCTION MOTOR: ROTOR DELTA OR Y
(693) Why are the rotor windings of slip-ring
induction motors usually Y connected?
The Y connection is employed because it
produces a secondar\' voltage 73 per cent
higher than that which would result from
using the delta connection. In medium size
motors relatively high voltage and low current
are desirable in the rotor, which are usually
bar wound, because these factors permit the
use of low-current capacity or small size
control apparatus.
For large size motors, however, the delta
connection of the rotor is used as commonly
as is the Y connection. This condition arises
from the fact that a Y connection for some
large rotors would produce too high a voltage,
one that would lower the factor of safety of
the rotor insulation and might be dangerous
to handle without specially designed control
apparatus. The adoption of the delta con-
nection for such rotors reduces the secondary-
circuit voltage in the ratio of 173 to 100 and
yet, at the same time, does not increase the
current to such a value as could not easily be
handled bv standard control apparatus.
AH. A.
ARRESTER: LOCATION OF FUSES
(659 ) Should a telephone or signal lightning
. arrester be installed with its fuse end
connected to the line or to the device to be
protected?
The position of the fuses depends upon the
service conditions of the particular system
under consideration. Fuses are used to guard
against three different conditions:
(1) To clear the instruments in case the
signal circuit becomes crossed with higher
voltage power or lighting circuits.
(2) To clear the signal circuit in case a
lightning arrester fails by grounding.
(3) To protect the signal or telephone
instruments in case of abnormal currents.
In order to meet the first two conditions, the
fuses should be on the line side of the lightning
arrester. In the case of the third condition,
the fuses could be on either the line or the
instrument side of the lightning arrester.
The objection to putting the fuse on the
line side is obvious, in that fuses must
necessaril_\- be of very small wire and would
frequently be blown by the lightning dis-
charges which they would have to carry.
It is therefore sometimes advisable to connect
the fuses on the instrument side of the arrester,
but this should not be done if there is any
possibility of the signal circuit ever becoming
crossed with higher voltage power circuits,
or if the arrester itself is unreliable or subject
to frequent grounds.
V.E.G.
TWO DOLLARS PER YEAR
TWENTY CENTS PER COPY
GENERAL ELECTPIC
REVIEW
VOL. XXIII, No. 6
Published by
General Eledric Company's Publication Bureau,
Schenectady. N. K.
JUNE, 1920
The Largest Hammer Head Crane in the World, Located at League Island Navy Yard, Philadelphia, Lifting a Load
of 1.010,000 Lb. From left to right the loads are, respectively: Switching Locomotive, 78,000 Lb.; Two
Loads of Steel Shapes, 416,000 Lb. each; Locomotive. 100.000 Lb. Total horse power of motors, 530
I See article, page 550)
For
Fractional H. P. Motors
Continuity of service uninterrupted
operation— is the key to maximum
production. Is it wise to limit the pro-
duction of an otherwise high-class ma-
chine by the use of even one part of
inferior quality? There have been fail-
ures among the hundreds of thousands
of high-speed, "tiORm^" equipped electrical
machines in service. But rare indeed
have been the cases where the failure was
the result of bearing trouble. Almost
invariably, the "NORmfl" Bearings have con-
tinued'on duty after other repairs were
made.
See that your Motors
are "NORmfl" Equipped
IRE Mwmm/^ (^^mF/^wf
ILs'Ej IglainKsl €n^
Ball, RoUei". Thrust and Combination Bearings
General Electric Review
A MONTHLY MAGAZINE FOR ENGINEERS
Associate Editors, B. M. EOFF and E. C. SANDERS
Manager. M. P. RICE Editor, JOHN R. HEWETT ,^ ^^^^^^ ^, Advertising, B. M. EOFF
Subscriplion Rates: United Slates and Mexico, $2.00 per year; Canada, $2.25 per year; Foreign, $2.50 per year; payable in
advance. Libiajy and Student Rates: United States and Mexico. $1.50 per year; Canada, $1.75 per year; Foreign, $2.00 per year;
payable in advance.
Remit by post-office or express money orders, bank checks, or drafts, made payable to the General Electric Review, Sche-
nectady, N. Y.
Entered as second-class matter. March 26, 1912. at the post office at Schenectady, N. Y.. under the Act of March, 1879.
Vol. XXIII. No. <i „, c;,„^;;/;^ix*;wf'^l/„,„y j^'nk, 1920
CONTENTS Page
Frontispiece: Airplane with SupercliarKcr in Flight over McCook Field, Dayton, Ohio . 466
Editorial: Relativity 467
Superchargers and Superchar^'in^ Engines 468
■ By Major George E. A. Hallett, U. S. A.
Maintaining Airplane Engine Power at Great Altitudes 474
By Lieut. -Col. V. E. Clark
The General Electric Turbo-Supercharger for Airplanes 476
By Sanford A. Moss
Relativity Theories in Physics 486
By Dr. Richard C. Tolman
The Production and Measurement of High Vacua — Part I 493
By Dr. Saul Dushman
Fundamental Phenomena in Electron Tubes Having Tungsten Cathodes — Part I . . 503
By Irving Langmuir
Electron Power Tubes and Some of Their Applications 514
By William C. White
Artificial Daylight for Merchandising and Industry 527
By G. H. Stickney
Enclosed Carbon Arc Lamps vs. Novalux Mazda Units 534
By H. E. Butler
350-ton Hammer Head Fitting Out Crane 550
By J. A. Jackson
Question and Answer Section ' 552
u
RELATIVITY
Science, as such, is distinguished from other
branches of human knowledfjc by its depend-
ence upon the use of the measuring stick,
balance, and penduUnn. Dynamics was born
as a branch of science when Gahlco used an
hour-glass and a yardstick to determine the
laws of falling bodies. Alchemy became
chemistry when Lavoisier weighed the oxygen
in mercuric oxide, and so we can continue
with illustrations throughout the whole his-
tory of science. All our so-called "laws" of
physics and chemistry are generalizations of
the quantitative results of innumerable experi-
ments. Progress in science has been achieved
by continued refinement of our methods of
measurements, and the approximate results of
yesterday are corrected by the more accurate
data of today. Under these conditions, the
" laws " of yesterday may be found to be inade-
quate to account for the new facts, and it may
be necessary to go so far as to revise views which
have hitherto been held as to the nature of the
phenomena under consideration. As Prof.
Lotka has expressed it in a recent article on
Relativity*: " If a new observation cannot by
any manner of means be made to fit into our
conception of the world, we may be forced to
change that conception."
The Theory of Relativity, of which Dr.
Tolman's paper in this issue is a splendidly
logical statement, is fundamentally an
attempt to reconcile with our ordinary
notions of dynamics as represented by New-
ton's laws of motion, certain experimental
facts which had previoush^ escaped observa-
tion because of the degree of accuracy
required in their determination.
As long as our experience dealt with
velocities small compared with that of light,
these laws were found to be apparently quite
adequate to correlate the observed results.
But, when we came to apply our ordinary
Newtonian laws to correlate the energy of
the extremely high velocity electrons emitted
by radioactive bodies with their mass and
velocity, we found these laws inadequate to
account for the experimental observation that
the mass of the electron increases with its
velocity, ski) < that it becomes infinitely great
for velocities approaching that of light.
vSimilarly in all our notions of energy, we
tacitly conceived as the vehicle of this energy,
mo\'ing bodies whether of atomic or ordinary
♦Alfred J. Lotka. "A New Conception of the Universe.
Einstein's Theory of Relativity, with Illustrative Examples. "
Harper's Magazine. March, lP2b, p. 477. This is a most interest-
ing article on this subject written in a popular manner and yet
thoroughly scientific. The reader who is interested in Relativity
will find it to be a splendid introduction to the present paper by
Dr. Tolman.
dimensions. But the ])henomena of radio-ac-
tivity have shown us the existence of stores of
energy in the atom it.self which are apparently
not kinetic in origin. Similarly Michelson
and Morley's experiment on the effect of the
earth's motion on the velocity of a light beam
led either to the hypothesis of Fitzgerald and
Lorentz that a moving body contracts or
else to the conclusion that our units of space
and time are different for moving bodies than
for bodies at rest.
The development of the consequences of
these facts by Einstein has led to a theory of
relativity which is extremely general in its
significance. As Dr. Tolman shows, Ein-
stein's theory is an extension of the simpler
theory of the relativity of uniform motion.
This simpler theory is an attempt to express
our laws of motion in such a general manner
as to be independent of the particular co-
ordinates with respect to which we ordinarily
define motion and the attempt succeeds as
far as systems in uniform relative motion are
concerned in space. An extension of this
view to systems having any type of relative
motion has led Einstein to a very generalized
theory in which gravitation itself appears
merely as a consequence of our usual limited
notions of time and space. That is, in the
proper system of co-ordinates, gravitational
effects disappear. This is the combined
significance of the condition for invariancy
and the equivalent hypothesis presented in
the latter part of Dr. Tolman's paper.
Such speculations will, of course, be
regarded by some as bordering on the meta-
physical. Nevertheless they are the logical
development of certain experimental facts
which our present refined methods of obser-
vation have discovered. Probably the con-
temporaries of Copernicus felt just as mysti-
fied about his theory of the universe as most
of us feel at present about Einstein's theory
of Relativity. Every new conception of the
universe has always had to contend with
a conservatism and inertia inherent in most
intellectual beings which tends to prevent the
rapid absorption of any new idea. Pragmati-
cally this is probably as it should be; for we
value new ideas onlv as they become useful in
explaining observed facts, and prophesying
new ones. Judged on this basis alone, the
Theory of Relativity represents an extension of
our ideas of theuniverse into a region which may
be even more incomprehensible than is the ordi-
nary notion of "infinity" as a mathematical
expression; but so are the facts for which it
attempts so successfully to account.
468 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 0
Superchargers and Supercharging Engines
By Major George E. A. H.\llett, U. S. A.
Chief of Powerplant Section, Engineering Division, Air Service, D.wton, Ohio
The "ceiling" or maximum attainable altitude of an airplane is limited by the engine output. As the
available engine power is reduced by the rarified atmosphere at high altitude, it is obvious that a raising of the
ceiling bv a considerable amount would require either the use of very much larger and heavier engines or some
means of supplying the present engines with fuel mixture at sea-level pressure. The latter method is the only
one worthy of studious consideration, and consequently much effort has been expended to develop a super-
charger or a supercharging engine that will meet the requirements successfully. In the following article, which
was delivered as a paper before the Society of Automotive Engineers, January 7 and 8, 1920, the author reviews
the work which has been done and outlines the possibilities of the future. — Editor.
The need for aeronautic engines that will
deliver the same power at 20,000 or even
30,000 ft. altitude as they develop at sea level
is very real and very great, in not only
military but also in commercial aviation.
Much success has already been attained with
supercharging devices in this country and a
certain amount of success in Europe. It must
be admitted that there have been some
failures also. It is the intention to -
otithne past developments in super-
charging in this article and to point out
the lines of attack which seem to be
meeting with most success.
Supercharging, as the term is gen-
erally used, means forcing a charge of
greater volume than that which is
normally drawn into the cylinders by
the suction of the pistons in conven-
tional internal-combustion engines.
When Supercharging is Needed
At 20,000 ft. altitude the atmospheric ".
pressure is roughly one half that at sea
level; hence about one half the weight
of charge is drawn into the engine and
less than one half the power is de-
veloped. At 25,000 ft. altitude less
than 25 per cent of sea-level power is
delivered. If at these altitudes air is
supplied to the carburetor at sea-level
pressure, or approximately 14.7 lb. per
sq. in. absolute, the power developed by
the engine becomes approximately the
same as when running at sea level.
The low atmospheric pressure and
density at great altitudes offer greatly
reduced resistance to high airplane speeds;
hence the same power that will drive a plane
at a speed of 120 m.]).h. at sea level will drive
it much faster at 20. ()()() ft., and still faster at
30,000 ft. altitude, and until apfyroxiniately the
same consumption of fuel per horse- power hour.
There is little to be gained by stu^ercharging
at sea level to increase the power of a given
size engine, because the clearance volume
must be made greater than normal to prevent
pre-ignition, with consequent decrease in the
expansion ratio and comparatively poor fuel
economy. The fact that the clearance volume
is increased removes the possibility of the
engine developing full power at great altitudes
unless a supercharging capacity greater than
anything heretofore considered feasible is
Fig. 1. Twelve-cylinder Liberty Motor with Supercharger of the Type
Shown in Fig. i
a\ailable. Supercharging, therefore, is most
useful in maintaining sea-level horse power in
engines ascending to or working at great
altitudes.
Superchargers
Superchargers usually take the fonn of a
mochanieal blower or pump an<l. of course,
require a driving gear of some kind. The
SUPERCHARGERS AND SUPERCHARGING ENGINES
469
types of blowers or compressors used to date
include the reciprocating. Root displacement,
and centrifugal types. The reciprocating type
was tried by the Royal Aircraft Factory earh'
in the war, on an air-cooled R. A. F. engine,
with practically no success. It seems that
this type of blower was found to be compara-
tively heavy and also unsuitable, due to the
pulsating pressure of the air delivered.
Rqczi ver
~^ Cuhnder
\ai\' ^ctrbunef-er
Fig. 2
Root Type Blowei so Arranged That Its Air Pulsations Synchronize
with the Suction Strokes of the Engine Pistons
Turbine Ejchausf
Discharge -
Bu-Pass '/aire
Air Compressor
Housing. _
Air Intakt
Airlmpeller
Member
Engine Exhaust
^ PipQ to Turbin*.
'Exhaust
Voire-.
Tntaks
■yafvs
- Intake
Pipe
Fig 3.
Air Discharge
to Carburefer
Centrifugal Type Blower Direct Connected to a Turbine Wheel That
is Impelled by the Engine Exhaust Gases
The Root type blower was tried by the
Royal Aircraft Factory with little or no
success. The trouble reported was "rough"
running of the engine on account of the pres-
sure pulsations in the air discharged by the
blower, which tended to overcharge some
cylinders and undercharge others, thus caus-
ing uneven impulses. It is reported that
mechanical troubles also developed with this
tvpe of blower. George W. Lewis, of the
National Advisory Committee for Aero-
nautics, is working on an improved Root type
blower, shown in Fig. 2. Here the pulsations
in the air discharged are synchronized with
the suction strokes of the engine. It will be
interesting to note how this develops.
The centrifugal type of blower was used
by Prof. Rateau, in France, early in the war.
He employed the exhaust gases of the
engine to drive a high-speed single-
stage turbine direct connected to the
centrifugal blower shown in Fig. 3.
Some success was had from the start,
but he encountered many mechanical
troubles. It is claimed in recent
reports that some fairly good results
are being obtained by the French.
The Royal Aircraft Factory experi-
mented in 19U) and 1917 with a gear-
driven centrifugal blower, but as soon
as an endeavor was made to run it at
speeds that would step up the pres-
sure to the 5 or (i lb. required, great
difficulties were encountered on ac-
count of the inertia and momentum
of the compressor rotor and the high-
speed end of the gear-train, which
resulted repeatedly in breakage of the
gears when the engine was acceler-
ated or decelerated. To eliminate this
trouble a friction clutch, designed to
slip under excess torque, was tried,
but only partial success was achieved,
and the clutch itself gave consider-
able trouble. Light, flexible vanes
were then tried on the compressor
impeller, but this expedient has not
proved successful to date. Similar
experiments were conducted by the
A. E. F. in France, but were concluded
by the signing of the armistice.
The United States Air Service
started work on the Rateau type
of turbo-compressor soon after we
entered the war. The work was done
under the supervision of E. H. Sher-
bondy, who worked in conjunction
with the Rateau-Bateau-Smoot Co.
which handled the Rateau patents in this
country, and designed a turbo-compressor
which seemingly embodied many improve-
ments over the Rateau type. Three of
these machines were built and given ground
tests on Liberty engines. The arrangement
of the engine and the supercharger is shown
in Fig. 1. Considerable trouble was encoun-
tered due to overheating of the exhaust-driven
470 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
turbine, and even the use of a special heat-
resisting metal in this part did not over-
come the trouble. Soon after Mr. Sherbondy
began work on the turbo-compressor, Dr. S.
A. Moss, of the turbine research depart-
ment of the General Electric Company,
was asked to carrj' on some work on the
same general type. He built one turbo-
compressor which was also a modification
of the Rateau type, but differed con-
siderably from Mr. Sherbondy 's machine.
This device was tested on a Liberty engine
at the summit of Pike's Peak and devel-
oped approximately sea-level horse power
there, at an altitude of 14,000 ft. It was
capable of making the engine pre-ignite
at that height.
After the armistice was signed all work
on the development of superchargers was
stopped. When the engineering division of
the Air Service took over McCook Field
and started to plan peace-time development,
the supercharger situation was carefully
considered. It was decided that it was
important to continue development work
along this line. It then became neces-
sary to decide whether work should be
continued on both the Sherbondy and
the Moss machines, and, if not, which
one should be developed. It was noted
that although Dr. Moss' machine was
comparatively crude, it contained some
inherent advantages over the Sherbondy
type, and no way was seen to overcome
the faults of the Sherbondy machine.
Therefore, the latter was dropped and the
General Electric Company was given a con-
tract to rebuild the old supercharger de-
signed by Dr. Moss. The new device is
now being tested in actual flight and is giv-
ing very interesting results. Figures on the
results obtained with the present Moss
supercharger are naturally confidential. The
indications are that the turbo-compressor
is very durable and probably will outlast
an aviation engine.
J. W. Smith, a designer and builder of air-
cooled radial engines, located in Philadelphia,
is known to have designed a turbo-compressor
for this type of engine. The B. F. Sturtevant
Co., at Boston, Mass., has at least partially
developed a belt-driven centrifugal com-
pressor for supercharging one of its aircraft
engines.
Carburetor Locations
There is still some question as to the best
location for the carburetor in relation to the
blower in supercharged engines. Apparently
all positions have been tried :
(1) It is possible to use the centrifugal
type of blower as a carburetor by placing a
fuel jet within its housing and allowing the
rotor to do the mixing. As the rotor usually
runs over 20,000 r.p.m., it will certainly mix
liquid fuel with air. This system would
require a manual fuel adjustment, such as is
used with the Gnome engine, for different
speeds. With this arrangement there would
be danger of an explosion in the blower in case
the engine back-fired, because the mixture in
the blower would be under pressure higher
than atmospheric.
(2) The carburetor can be placed on the
suction side of the blower. In this case the
evaporation of the fuel will assist in cooling
the charge during compression and the action
of the compressor will improve the mixing of
the fuel, but the danger from explosion
remains to be overcome.
(3) When the carburetor is placed in the
"normal" position and air is forced through
it, it becomes necessar\- to "balance" the
float-chamber with supercharger pressure.
This somewhat complicates the feeding of
fuel. Pressure gas-feed systems are " banned "
in military planes and in any case with a
pressure system the tanks would have to be
made comparatively heavy to withstand
the pressure which would be used at great
altitudes. Where gasolene pumps are used it
is necessary- to regulate their discharge pres-
sure as the plane ascends, because the fuel
must reach the float-chamber at a pressure
about 2}4 lb. higher than that at the super-
charger outlet. If the difference in fuel and
float -chamber pressures is not kept in con-
stant relation, the quality of the mixture fed
to the engine will vary on account of the
change in fuel level in the float-chamber.
The engineering division has developed a vcr>"
simi^le device that solves this problem effect-
ively and is entirely automatic.
It would naturally seem at first thought
that the extremely low temperatures always
found at great altitudes would make possible
the easy solution of cooling problems, but in
reality the low density of the air reduces its
heat conductivity and capacity for heat
absoq)tion to such a point that a supercharged
engine developing sea-level junver at 20.000 ft.
requires a little wore cooling surface than it
does when developing normal power at sea
level.
The Liberty engine and many others run
best with a water temperature of about 170
SUPERCHARGERS AND SUPERCHARGING ENGINES
471
deg. F. To maintain the cooling water at
this temperature in the reduced atmospheric
pressure at 25,000 ft. it is necessary to use
several pounds of air pressure in the radiator
to prevent the water from boiling away.
Very effective radiator shutters are needed
when the engine is throttled to make a descent
from altitudes of over 20,000 ft. to prevent
the water in the radiator from freezing before
wanner air is reached.
Contrary to expectations, the Moss turbo-
compressor now being tested at McCook Field
does not complicate the pilot's controls. On
a normal engine the pilot handles the throttle
and the altitude carburetor control which
thins down the mixture as he ascends. With
the turbo-compressor the altitude control
becomes unnecessary up to the altitude at
which the engine can no longer deliver sea-
level power but is used, as with a normal
engine, if the plane is driven higher.
With the Moss turbo-compressor, when
flying at low altitudes, the exhaust pressure
is allowed to "waste" through manually
operated "gates" in the exhaust pipes. As
the plane ascends the pilot closes these gates
a little at a time and after he reaches a great
altitude he can speed and retard the plane by
the use of these gates. He uses the throttle
only in case he wants to descend rapidly, when
he closes it. In our test flights we have pro-
vided the pilot with a sealed altimeter con-
nected only to the supercharger pressure, so
that it shows to what altitude this pressure
corresponds. When at great altitude the pilot
closes the exhaust gates until the pressure in
the carburetors causes the altimeter to show
sea-level pressure. This makes it unnecessary
for him to do any calculating. If he makes
the gauge read lower than sea level, the engine
will pre-ignite. We have already been able
to obtain sea-level pressure in the carburetors
at well over 20,000 ft. The exact height
cannot be mentioned at present.
With a normal engine the falling off in
power as the plane ascends does not cause as
much of a drop in propeller speed as might
be expected, because of the reduction in
density of the air in which the propeller is
working. Ovir best engines do not lose over
75 r.p.m. at 20,000 ft. When an engine is
supercharged so that the power remains
constant as the plane ascends, the propeller
tends to "race" at great altitudes. Therefore
it is necessary either to use a variable-pitch
propeller or to put on one that holds the
engine speed down too low for best perform-
ance near the ground, but also does not allow
the engine to race too much at great altitude.
In our present tests we are using an oversize
propeller and are getting surprisingly good re-
sults, but we also have variable-pitch propellers
about ready for test and should get much
better performance with them.
Supercharging Engines
As generally used; the term "supercharging
engines " refers to internal-combustion engines
in which compression in the crankcase or in
the lower end of the cylinders is used to force
an additional volume of air or mixture into
the working cylinders after completion oj their
normal suction stroke. Early in the war the
Army and the Navy each placed an order with
the Kessler Motor Co., Detroit, Mich., for
several experimental supercharging engines.
This type of engine, shown in Fig. 4, super-
charged each cylinder by the use of crankcase
pressure, as is possible in four-cycle engines.
Experiments were made using both air and
mixture in the crankcase. Considerable diffi-
culty was encountered in both the design and
construction of the engine and so far as the
engineering division has learned, no complete
tests have been run ; and in the small amount
of testing that has been done no very large
increase in power or brake mean effective
pressure has been shown officially. It is
believed that the frictional losses will prove to
be very high in this type of engine and that
the supercharging will be comparatively ■
limited. A similar engine which was tested
in this country did show very high frictional
loss, due partly to the work of operating the
valves which controlled the crankcase air.
An interesting problem in this type of
engine when using air in the crankcase is
whether a rich mixture should be fed through
the regular induction system and an effort
made to dilute it with the supercharged air, or
a normal mixture should be fed through the
induction system and an attempt made to
obtain perfect stratification and thus let the
supercharged air merely form a cool, elastic
and expanding cushion on the piston-head. It
is feared that in either case it will be difficult
to secure the desired results through a large
range of speeds and throttle positions.
There is an English make of supercharging
engine in which air is compressed under the
piston and by-passed through cylinder ports
at the bottom of every stroke (see Fig. 5),
supercharging, as in the Kessler engine, at
the end of the suction stroke and scavenging
at the end of the exhaust stroke. It is claimed
by the inventor that this scavenging makes
472 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
possible the use of higher compression and
greatly improves the fuel economy and brake
mean effective pressure. It is believed that
this engine will give rather limited super-
charging and it may prove difficult to control
the mixing or stratification of the air and
mixture at some speeds.
In an English rotarj- air-cooled engine the
pistons travel out to the cylinder-heads on the
scavenging stroke and the beginning of the
suction stroke and continue an extra distance
inward at the end of the suction stroke, thus
taking in a larger charge than that of a con-
ventional engine. The piston reaches only a
normal position at the end of the compression
stroke and continues an extra distance inward
at the end of the suction stroke, all by
means of an eccentric crankpin bear-
ing which is rotated on the crankpin
by gears of suitable ratio. This type
of engine must certainly give a very
limited amount of supercharging.
It is believed that supercharging
engines will necessarily give a rather
limited amount of supercharging. It
is also believed that considerable
difficulty will be encountered in
obtaining the desired stratification in
mixing conditions in the combustion
chamber through any wide range of
throttle positions. Also, some me-
chanical friction is added in this type
of engine and it must be borne in mind
that friction is particularly undesir-
able at great altitudes because it
remains nearly constant from the
ground up to great altitudes while the power
falls ofT rapidly; therefore, the mechanical
efficiency of the engine becomes very low.
The Root type of blower might be interest-
ing for supercharging jjurposes if the troubles
caused by the pulsating nature of its discharge
could be eliminated. It is hoped that Mr. Lewis'
efforts along this line will meet with success.
It is already frequent practice to build
aviation engines with compression so high
that the throttle cannot be fully opened on
the ground without injury to the engine.
In this way, perhaps, the same power is
obtained at oOOO ft. as can be obtained on the
ground. It has been suggested that this idea
be carried further and that an "oversize"
engine be Iniilt with much higher compression
so that the throttle cannot be opened fully
until a considerable altitude, such as l((,(l(r(l
or 15, ()()() ft., is reached. It has been stated
that such an engine could be made lighter,
in proportion to the cylinder sizes, tlian a
conventional engine, on account of the fact
that the throttle would never be opened near
the ground, but it is believed that when this
idea is investigated, it will be found that it is
the inertia forces quite as much as the explo-
sion forces that determine the necessary
strength in most high-speed airplane engine
parts and that therefore such an engine could
not be built light enough to make it practical.
In any case, it is doubtful whether this would
give a reallv good solution for flving at 25,000
or 30,000 ft.
It is possible that centrifugal compressors
can be operated satisfactorily by gears or by
a belt drive. It is known that some designers
are working on both of these problems.
Fig. 4.
Specially Designed Engine which is Supercharged by Pressure
Developed in the Crankcase
The turbo-compressor in which an exhaust-
driven turbine is used for driving the centrif-
ugal compressor seems to present one fairly
good way of accomplishing the desired pur-
pose. The turbo-compressor itself is very
simple, as there is only one moving part,
namely, the rotating element consisting of the
turbine wheel and compression impeller.
The bearin;;s of this rotating clement do not
seem to wear noticeably and the device
imposes very little drag on the engine when
not being used for supercharging. The turbo-
compressor is also an eflectivc exhaust muffler.
The Future of the Supercharger
It is believed that when the present typo
of turbo-compressor now being tested by the
engineering division has been more fully
developed, it can be built into an engine in a
form which will add less weight and less head-
resistance than the present machine, and
naturalh- when we kimw exactiv what addi-
SUPERCHARGERS AND SUPERCHARGING ENGINES
473
tional cooling surface is required at a given
height, it will not be difficult to build this
cooling surface into the airplane in such a
form that very little weight and head-resist-
ance will be added.
The uses of the supercharger for military
service can be divided into; first, for airplanes
in which it is desired to reach extreme altitude;
second, for airplanes in which it is desired
to increase the rate of climb and horizontal
speed and therefore maneuverability at alti-
tudes where it is intended to fight; and, third,
for airplanes which carry large loads, such as
bombers, which normally are handicapped bv
having a very low ceiling and whose entire
lnie+
Exhaust
Fig.
5. Cylinder of Specially Designed Engine which is Supercharged by
Pressure Developed Around the Piston and Beneath the Rings
usefulness would, if larger engines were
installed to pull them to a higher ceiling, be
lost on account of the larger amount of fuel
and other material that would have to be
carried, thus decreasing their radii of action.
In the first case it is believed that a special
supercharger can be built that will make
feasible much greater altitudes than any that
have been attained with the present General
Electric turbo-compressor; and it is considered
essential that we have airplanes capable of
reaching very great heights. In the second
case, it is pointed out that military machines
not fitted with supercharging engines, when
fighting at an altitude of 20,000 ft. or more,
are so near their ceiling that their rates of
climb, speed, and maneuverability are com-
paratively poor, but the use of a supercharger
seems to overcome this difficulty easily. When
a pilot climbs with a normal engine to 20,000
ft. and then levels ofif in horizontal flight, the
engine and propeller speed up perhaps 100
r.p.m. This, of course, enables the engine to
develop slightly more power. In the case of a
supercharged engine, especially with the
turbo-compressor type of supercharger, as
the engine speeds up in horizontal flight, the
temperature of the exhaust and the power
available from the exhaust increase, thus
building up the supercharging pressure and
giving considerably greater increased power
than with a normal engine.
The use of superchargers in com-
mercial airplanes of the future is
assured because superchargers will
make possible far more miles per hour
and more miles per gallon with a given
engine and airplane, and speed is the
main advantage of air over other kinds
of transportation. It is thought by
many qualified judges that, by flying
at a sufficient height with a super-
charged engine and a suitably de-
signed airplane, a speed of 200 m.p.h.
can be maintained.
In the heavy-load-carrying type of
plane which must necessarily cross
mountains or perhaps fly above
storms and clouds, the necessary
height can be reached with smaller,
cheaper, and more economical engines
if they are fitted with superchargers.
It is obvious that in really long cross-
country flights or trans-continental
with mail or passengers, the logical
is to fly at 25,000 or 30,000 ft. alti-
tude where the resistance to speed is low
and great speed can therefore be attained
provided the engine can deliver high power
economically, which it can do if equipped
with a supercharger.
As a graphic illustration of the advantage of
a supercharged engine, it is pointed out that
at 2.5,000 ft. altitude a supercharged 250-h.p.
engine will deliver as much power as a 1000-
h.p. engine without a supercharger; and of
course the former will weigh many hundred
pounds less, its fuel and tankage will weigh
very much less, the first cost will be much
lower and the structure of the airplane can be
made much lighter.
flights,
course
474 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
Maintaining Airplane Engine Power at
Great Altitudes
By Lieut. -Col. V. E. Cl.\rk
Engineering Division, Air Service, D.wton, Ohio
The following is a very brief abstract of Lieut. -Col. Clark's paper presented at the Aeronautic Meeting of
the Society of Automotive' Engineers, March 10, at New York City. The material omitted in this abstract is
principally aeronautical calculations of a rather highly technical nature. These furnish the data for plotting
the curves shown in Fig. 1. There seems to be no question but that the supercharger is initiating an era of
extraordinary development. In delivering the paper, Lieut. -Col. Clark specifically stated that his calculations
referred to the General Electric supercharger and that he computed that during Major Schroeder's recent
record altitude flight the plane was performing in accordance with the values stated in this article. — Editor.
In the summer of 1917, the Boiling Air-
plane Mission to Europe recommended in an
official report that our engineers direct
especial energy -toward the development of
means to maintain a high proportion of the
power of airplane engines at great altitudes.
The purpose of this article is to indicate the
possibilities and limitations of increasing
airplane speed by introducing means to main-
tain high engine power at great altitudes.
I have attacked the problem by selecting the
De Haviland Four as being an airplane typical
of present practice, and by endeavoring to
compute approximately the performances that
might be obtained at different altitudes with
various assumed ratios of actual engine power
at the altitude to the total weight of the air-
plane in e\-ery case.
Let us compare the speed of a present-day
airplane with that of a hypothetical airplane
in which is installed a means of maintaining
its power constant at all working altitudes.
Looking toward the future, it will be interest-
ing to assume that the total airplane weight is
the same in each case, 5000 lb., and the
engine develops 500 h.p. at sea level.
Case I. The engine power decreases with
an increase in altitude at the normal prescnt-
dav rate, with no novel means of maintaining
it."
Case II. Means are installed for main-
taining the engine power constant with
changes of altitude.
We will in the second case assume a con-
stant propeller eflicicncy of O.SO. From a
practical standpoint, the maintenance of such
an efficiency, constant at various speeds and
in different densities, is today impossible.
The development of the variable-pitch pro-
peller which, most fortunately, is contem-
porary with that of the supercharger, is lead-
ing in the desired direction, however. The
supercharger would, relatively, be of little
value without the variable-pitch propeller
which, set at a ven- low pitch, permits climb-
ing away from the ground, and, set at a ver\'
high pitch, shovild show good efficiency at
very high airplane speeds, in air of ver}^ low
densitv.
700O0 ■
6O00O ■
5OOO0 ■
.
.^^
L.
|!W
.se^
/
5S
r
/
^
e^.
/
/■
^'
Va
r
/
uo!
e>
/
.,/
f
/
p
>■
/
ft
N
J
f
Rt
/
^Ct
V
eo so looiro leo 150 ifo 1X160 no i«o ISO looemioixemx
SpeaamMPH.
Fig. 1. Stalling Speed and High-speed Curves of a De Haviland
Four Airplane with and without a Supercharger. Case I without
Supercharger; Case II with Supercharger i curve hypothetical)
Computation results in the cur\-cs shown in
Fig. 1 which present the stalling speed and a
comparison between the high speeds of the
planes in the two preceding cases at all
altitudes.
No suggestion will here be made as to the
basic principle of the device or means for
maintaining power with low density. Among
the soltitions suggested are a supercharging
device at each cylinder to increase the com-
pression and introduce more oxygen, a rotary
air compressor driven by an exhaust gas tur-
bine, or through gearing by a shaft, special
fuels and the combination of the special fuel
with a higher compression, etc. Designers
must consider the extreme low temperatures
encountered.
Incidentally, the air compressor for the
engine intake might also be used to maintain
MAINTAINING AIRPLANE ENGINE POWER AT GREAT ALTITUDES 475
good pressure and introduce extra oxygen in
a necessarily sealed compartment occupied
by the personnel.
If engineers should go through a few
numerical examples, following the method
shown (in the full text of this article) and
using the curves and noting results, they
might become interested in development along
this line.
In the general latitude of New York,
Chicago and San Francisco, suppose that we
could in certain seasons of the year, by rising
to an altitude of about 40,000 ft., encounter
a wind current having a velocity of 100 m.p.h.,
whose direction is such as to be "under the
tail." If we could maintain a speed through
the air of 200 m.p.h. at this altitude our
speed over the ground would be 300 m.p.h.
We could then, in flying time, go from Chicago
to New York in three hours and from San
Francisco to New York in nine hours.
Speed of travel or transportation makes for
saving in time which, from the practical com-
mercial standpoint, is tantamount to the
elimination of space. Bringing San Francisco
as near to New York as Pittsburg now is by
train, if it can be done, is a matter of tremen-
dous importance. We should, therefore, look
well into all means offering even the appear-
ances of feasibility which may be suggested
for helping toward this eventual accom-
plishment.
Le Pere Biplane after trial flight at McCook Field, Dayton. Ohio. The General Electric Super-
charger is shown mounted on the head end of the Liberty Motor
476 June. 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 0
The General Electric Turbo-supercharger for
Airplanes
By Dr. Saxford A. Moss
Turbine Research Dep.^rtment, Gexer.\l Electric Comp.\xy
Dr. Moss, who developed the General Electric turbo-supercharger, compares the device with others
intended for the same purpose and with supercharging engines. He then explains the design of the General
Electric supercharger, and narrates the very interesting history of its development which includes a series of
tests on Pike's Peak and also in flight. There is also included a very graphic description of the instruments
and method of calculation employed in accurately measuring airplane altitude. The author concludes his
article with a description of the performance of the supercharger. — Editor.
Introduction
An airplane flying at high altitude is in an
atmosphere of comparatively low density.
For instance, at 20.000 ft. altitude the density
is practically half that at sea level. This
means that a given volume contains half as
much actual air by weight. The cylinders
of an airplane engine are therefore charged
with an explosive mixture which has about
half the value of a charge at sea level. The
engine actuallv delivers about half of its sea-
level power at' 20,000 ft.
Both the low temperature and the de-
creased pressure at high altitude have effect
in fixing the high altitude density. Both the
decrease of temperature and the decrease of
weight of the charge affect the carburation at
high altitude. The fixed clearance volume
and the decreased initial pressure give a
decrease of compression pressure resulting in
a loss of efficiency. There is, therefore, a
combination of causes which gives as a net
result the decrease in engine power very
nearly proportional to the decrease in density.
At high altitude, the resistance of the air to
the motion of the airplane is decreased
directly in proportion to the decrease of
density. The power required for a given
airplane speed is therefore greatly reduced.
However, the engine power has been so
reduced that the usual net result is a con-
siderable decrease in airplane speed. When
the engine power is maintained at the sea-
level value, there is, however, a considerable
increase of speed at high altitude.
Filling the cylinders of an internal com-
bustion engine with a charge greater than that
which would normally occur, is called "super-
charging." Methods of doing this have
engaged the attention of a great many
experimenters.
The "gas turbine" is a prime mover in
which highly heated products of combustion
impinge directly on a turbine wheel. The
high thermal efficiency of the gas engine and
the rapid displacement of the reciprocating
engine by the steam turbine have caused a
great deal of effort to be spent upon some
combination of the two in the form of a "gas
turbine." Many inventors have proposed
various types of gas turbines and a number of
these have been developed to the point where
their operation is successfvd mechanically.
However, no type has yet shown sufticientiy
good efficiency to warrant commercial use.
The engineers of the General Electric Com-
pany have ver\- closely followed the various
gas turbine developments and have been
intimately in touch with the situation for
many years.
In 1903 the Company first began work on
the "centrifugal compressor." This is an
apparatus similar to the fan blower except
that the shape of the impeller blades and the
passages leading air to and from the impeller
are so arranged as to give efficiency ven-
much greater than that of the usual type of
fan blower, so that the apparatus forms a
satisfactory means for compressing air to
appreciable pressures. A line of single-stage
centrifugal compressors has been developed
for compressing air from one to five pounds
per square inch above atmosphere, to be
used for many industrial purposes; as well
as a line of multi-stage machines for com-
])ressing air and gas ujj to pressures of 30 lb.
per square inch above atmosphere.
The turbo-supercharger is a combination
of a gas turbine and a ccntriftigal compressor,
arranged as part of an airplane gasolene
engine. The hot products of combustion
from the engine exhaust are received upon the
turbine runner and furnish power whereby is
driven a centrifugal compressor mounted on
the same shaft, which compresses air for
supply to the carburetors. A more detailed
description is given later.
In the latter part of 1917 the National
Advisory Committee for Aeronautics re-
quested the co-ojH-ration of the General
THE GENERAL ELECTRIC TURBO-SUPERCHARGER FOR AIRPLANES 477
Electric Company in the development of the
turbo-supercharger in the United States.
Our experience with gas turbines and centrif-
ugal compressors led us to be greatly interested
and the work was pushed vigorously during
the war. An apparatus was constructed and
placed in operation on an airplane engine near
sea level. After a period of development, the
stage was reached where nothing more could
be done except at high altitude. However,
since the development was not sufficiently
advanced to warrant an airplane flight, the
entire testing apparatus was taken to the
summit of Pike's Peak. Here a further
period of development took place. The
apparatus was finally gotten into satisfactory
working order so that the airplane engine
developed the same power at the summit of
Pike's Peak as it originally had near sea level.
Arrangements had been started for installing
the apparatus on an airplane when the
Armistice intervened. Examination of the
results which had been obtained, by army
officials after the Armistice, led to a resump-
tion of the work and the apparatus was
finally installed on an airplane. A very good
showing was made from the first. The
increase of power at high altitude was such
as to give an entirely new set of conditions
from those under which the airplane originally
operated. This required various changes in
the entire airplane apparatus and develop-
ment was made of proper radiators, propellers,
gasolene systems, cooling systems, etc. This
work has been proceeding satisfactorily for
some time.
Development work on the turbo-super-
charger is also being carried on in France
independently of our work. So far as can
be seen from the published accounts of the
French work, our apparatus is on a larger
scale. We are supercharging a larger air-
plane motor and are carrying the super-
charging to higher altitudes. The mechanical
details of the French and General Electric
apparatus are quite different. The develop-
ment of a turbo-supercharger similar to that of
the French was started in this countr\- but
the design was modified considerably. Work
on this apparatus has not been carried to a
canclusion, however.
Our work was originally started at the
suggestion of Dr. W. F. Durand, then Chair-
man of the National Advisory Committee for
Aeronautics, who knew of our long experi-
ence with gas turbines and centrifugal com-
pressors. It has since been carried on under
the super\-ision at various times of Col. J. G.
Vincent, Col. T. H. Bane, Major H. C.
Marmon, Major G. E. A. Hallett and Major
R. W. Schroeder. Major Hallett has had
charge of the development since the Armi-
stice, and he has given considerable study
to the matter of superchargers in general.
The Turbo-supercharger Cycle
Fig. 3 of the foregoing article by Major
Hallett gives a detailed diagram of the prin-
ciples of the turbo-supercharger. The exhaust
of the airplane engine is received by an
exhaust manifold which leads it to a nozzle
chamber carrying nozzles which discharge it
onto the buckets of a turbine wheel. On the
same shaft as this turbine wheel is the
impeller of a centrifugal compressor. This
compresses air from the low pressure atmos-
phere to approximately normal sea-level
pressure and delivers it to an air discharge
conduit which supplies the carburetors.
The turbine nozzles are of such area as to
maintain within the exhaust manifold and
nozzle box a pressure approximately equal
to that at sea level. The difference between
this pressure and the altitude low pressure
gives a pressure drop for the exhaust gases
which furnishes the power that operates the
system.
Due to the respective temperatures, this
power input suffices to give the desired
compression and also to supply the inevitable
losses. However, in order to avoid back pres-
sure on the engine, above the normal sea-level
value, both turbine and compressor must be
designed with utmost attention to efficiency.
With an efficient arrangement, the engine
when at high altitude exhausts at normal sea-
level pressure and receives its air at the
carburetor at normal sea-level pressure.
Hence, normal sea-level power is delivered
at all altitudes up to the maximum for which
the supercharger is designed, so that the plane
speed will increase uniformly as the altitude
density decreases.
In order to reach this ideal there are vari-
ous auxiliary problems that have to be
solved; such as temperatvire rise of com-
pression, slight deficiency of oxygen at high
altitudes, effect of propeller on engine speed,
and various other effects. The work thus far
accomplished has demonstrated the validity
of the fundamental principles and has dis-
closed the problems of detail.
Mechanical Problems of Supercharging
The General Electric superchargers thus
far constructed have been designed to give
47S June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 6
sea-level absolute pressure at an altitude of
18,000 ft., which involves a compressor that
doubles the absolute pressure of the air. This
pressure ratio, with the quantity of air
involved, requires about 50 shaft horse power
input for the compressor. The design of a
complete power plant of this size to suit an
existing airplane engine, with such weight and
location as will not impair the flying char-
acteristics of the plane, has of course offered
many problems. The possibility of driving
the compressor of the supercharger by engine
power, instead of by the exhaust gases,
suggested itself. Indeed, obsen-ation of an
engine exhausting in the usual way into the
atmosphere, discharging flames through the
short red-hot spouts, with the almost incan-
descent exhaust valves in view, makes it seem
absurd to propose to pass these red-hot gases
through pipes with a pressure difference above
the surrounding atmosphere equal to its
absolute pressure, and more absurd still to
obtain power by discharging these red-hot
gases onto a turbine wheel rotating at 20,000
r.p.m. Nevertheless the turbo-supercharger
has made flight after flight with entirely
successful operation, while the mechanically-
driven supercharger has never endured in spite
of much effort. Much experience with the
operation of the gas turbine led the writer to
prefer its problems to those of the driving
mechanism of a supercharger operated from
the engine. The turbine involves merely the
addition to the compressor of a single extra
wheel, designed for the conditions, with
no extra bearings. The engine-driven scheme
involves a 50-h-p. transmission with multi-
plicity of gears, bearings, clutches, belts, and
the like. These offer more or less drag on the
engine when the supercharger is not in use at
low altitudes, and very serious problems of ac-
celeration when the supercharger is to be thrown
into action, since the engine will be then run-
ning at its full speed of about ISOO r.p.m.
It must be admitted that this is much the
simpler proposition, since a turbine wheel
has been designed which will endure.
The exhaust manifold and nozzle box have
proven to be a very efficient exhaust muffler
and conductor. Such a mufllcr and conductor
is needed in any event, and the design of
means for withstanding the increased pres-
sure difference of the turbo-supercharger has
been successfully accomplished.
Power for Turbo and Engine-driven Superchargers
An efficient turbo-supercharger theoreti-
cally deducts from the indicated horse power
of the airplane engine an amount correspond-
ing to the difference between sea-level
absolute pressure and altitude pressure. There
is this additional back pressure during the
exhaust stroke. The theoretical power avail-
able for driving the turbo-supercharger is
greater than this, however, owing to the fact
that there is available not only the energy
due to the direct pressure difference men-
tioned, but also the energ>- of perfect expan-
sion from the higher to the lower pressure.
If there were no turbo-supercharger the
engine would waste this energv- in sudden
pressure drop as the exhaust valves open.
The turbine can utilize this energA-. The
sum of these two amounts of available energy-,
multiplied by the efficiency of the turbine
wheel, gives the shaft power delivered to the
compressor.
For an engine-driven supercharger com-
pressor there is greater engine indicated
power due to a lower exhaust pressure.
However, the shaft power for the super-
charger compressor must be transmitted
through the engine connecting rod and crank
shaft, with losses, and then through the
supercharger driving mechanism with addi-
tional losses. The total shaft power thus
subtracted from the engine, multiplied by
the efficiencies of these two transmissions,
gives the shaft power delivered to the com-
pressor. This is the same as for the turbo-
supercharger. For a Liberty motor of about
401) h.p. and sea-level power at 18,000 ft.
altitude, this power is 50 h.p.
The comparison then is as follows: The
turbo-supercharger subtracts from the engine
indicated power, adds power of expansion
which would not otherwise be used, and has
turbine wheel losses. The engine-driven
supercharger puts this indicated power
through the engine (with some additional
loads on the pins and bearings) and has
engine and transmission losses.
With usual efficiency there is probably not
a great difference between the gross subtrac-
tion from engine ])ower in the two cases.
There is then the disadvantage of trans-
mitting the supercharger power through the
engine pins and bearings, as well as through
some mechanism l)etwe<.>n engine and su^xr-
chargcr, to be compared with the collection
of the hot gases under pressure (with muflling
advantages) and delivor>- to the turbine
wheel. As already mentioned, practical
success to date is in favor of the turbo-su[>er-
charger and the writer feels that this is really
due to its innate superiority.
THE GENERAL ELECTRIC TURBO-SUPERCHARGER FOR AIRPLANES 479
Engine-driven superchargers with positive-
pressure blowers have been proposed. These
have the additional disadvantage that with
the desirable pressure ratios of about two to
one there is an appreciable compression loss
due to the fact that the machine only dis-
places air and has no direct means for com-
pression.
It is to be noted that, although the power
required to drive the supercharger is sub-
tracted from the engine power, the remainder
at high altitude with an efficient super-
charger is equal to sea-level power. That is
to say, the supercharged engine delivers
power enough to drive the supercharger as
well as to deliver sea-level power to the
propeller. There is of course no way to
arrange for full power due to supercharging
with the additional power due to exhaust
at the low absolute pressure of high altitude
and without expenditure of power for super-
charging. Without a supercharger the engine
has the advantage of a very low exhaust
pressure, but the explosive charge is so small
that the gross power has the well-known low
value at high altitude.
Supercharging Engines
Supercharging engines of various kinds,
in which the engine crank case or the engine
cylinders themselves are arranged for addi-
tional compression, have been discussed by
Major Hallett, and shown to give excessive
weight and complication as compared with a
turbo-supercharger.
A very simple form of supercharging has
frequently been used wherein an engine of
large displacement, but with very high com-
pression pistons has been fitted to a com-
paratively small plane. In such a case, the
throttle could not be opened wide near sea
level because the compression would be exces-
sive and serious pre-ignition would result; to
say nothing of the damaging effect on the
engine by delivery of the full power cor-
responding to the displacement with sea-
level charge. At altitude, however, a full
charge at the altitude density is taken, and on
account of the high compression pistons this
is compressed to a proper amount for good
operation. , ^pme high altitude flights have
been made in this way with a single seat
plane and engine with a displacement cor-
responding to 400-h.p. at sea level. The power
at high altitude was possibly 100-h.p. A 100-
h.p. engine with a turbo-supercharger would
give the same power at altitude and weigh very
much less.
Since such an engine has normal com-
pression pressure at high altitude, the power
will be very nearly proportional to the density
of the charge. There will be no loss of
efficiency due to decrease of compression
pressure. The altitude power will then
varv directly with the cylinder displacement
and inversely with relative density at altitude.
Major Hallett points out that with such an
engine the weight is nearly proportional to
the displacement. Hence such an engine
will weigh nearly twice as much as a super-
charged engine for 18,000 ft. altitude con-
ditions, and nearly four times as much for
35,000 ft. conditions where the density is
one fourth that at sea level. There is some
deduction from these figures due to the fact
that the weight will not go up quite as fast
as the displacement and because the super-
charger's weight is not negligible. However,
the situation in the main is as represented.
Engines have also been proposed with
crank-case compression, either with individual
connections or with a receiver. With a four-
stroke cycle, two crank-case ends supercharge
a single cylinder. However, with the mini-
mum crank-case clearance thus far suggested,
the maximum compression pressure possible
is not sufficient to give supercharging at an
appreciable altitude.
Design of General Electric Superchargers
The machines used thus far have been
designed to give sea-level pressure at 18,000
ft. altitude, which corresponds to a pressure
ratio of about two. The rated speed for
these conditions is 20,000 r.p.m. Sea-level
pressure has readily been obtained up to-
22,000 ft. altitude. The control is entirely
by hand operation of waste gates, which
permits of free escape of some of the exhaust^
gases.
The entire apparatus, exclusive of exhaust,
manifold and air discharge conduit, weighs
about 100 lb. The exhaust manifold and air
conduits have nearly the same weight as.
equivalent parts with no supercharger.
The turbine and compressor wheel have
diameters somewhat less than a foot. The
present design has been hampered by necessity
for accommodation to existing engines and
planes. It is proposed, however, to con-
struct apparatus in which engine and super-
charger are integral, with all parts arranged
for the full possibilities of the combination.
The essential features of the design are
various arrangements of ducts for cooling the
several parts, means for accommodating the
480 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
temperature expansions, means for handling
the temperatures which exist, and design of
both turbine and compressor to give utmost
efficiency.
History of General Electric Supercharger
The combination of airplane, propeller,
engine, radiator, cooling system, and super-
charger are so intimately associated that no
adequate tests can be made without the
complete system in operation at full speed
at altitude. Altitude chambers exist for
tests of engines alone, but none are arranged
for inclusion of the propeller. What tests
were possible were first run with steam with
the supercharger alone at the Lynn Works of
the General Electric Company. Additional
tests were run with the
supercharger and Liberty
motor on dynamometer
stands at McCook Field,
Dayton, Ohio, the Experi-
mental Station of the Engi-
neering Division of the
Air Sen.-ice. These tests
were necessarily made with
nearly sea-level initial pres-
sure. Even slight super-
charging under such condi-
tions involved increase of
compression pressure, and
this instantly caused pre-
ignition. Both sets of tests
gave means for perfecting
the mechanical operation
of the supercharger, but
gave no information as to
increase of engine power
under altitude conditions.
So far as could be seen everything was operat-
ing in accordance with scheduled expectations,
but there was not sufficient assurance to war-
rant an airplane flight.
During the initial development of the
Liberty motor a testing expedition had been
sent to the summit of Pike's Peak, and it was
decided to repeat this performance with the
supercharger. Fig. 1 shows the motor truck
that was prepared for the expedition. The
Liberty motor carrying the supercharger
was mounted on a cradle dynamometer, with
scales and all arrangements for accurate
measurement of power, gasolene consump-
tion and the like. In fact, a complete testing
laboratory was provided. The motor truck-
was shipped by rail to Colorado Springs, and
then proceeded by its own power to Pike's
Peak summit on the "Pike's Peak Auto
Highway." This is a well constructed, very
tortuous mountain road twenty-eight miles
long.
Pike's Peak Summit has an altitude of
14,109 ft. It is the highest point in the
United States easily reached by road. The
summit is a slightly rounded rocky flat
about 100 yds. in diameter. On it are two
stone houses, one at the terminus of a cog rail-
road and the other about one hundred yards
distant at the terminus of the auto highway.
The motor truck was set up near the latter.
Fig. 2 shows the nature of Pike's Peak
summit. Fig. 3 shows the way the test car
was left after each day's work. Fig. 4 shows
its condition on many of the mornings. There
were, however, man\- pleasant days when the
Fig. I.
Motor Truck Carrying Liberty Motor and Complete Equipment Tor Testint the
General Electric Supercharger in Rarified Atn\ospherc at Pike's Peak
testing work could be carried on with facility.
Fig. r> shows the rear of the test car on a
pleasant day.
The testing work at the summit lasted
through September and half of October, 191.S.
The u-sual difficulties with experimental work
were, of course, encountered with the addition
of many delays, due to the cold and snow, and
distance from repair shops. Minor changes
were ma(fe in a little shack at the summit,
but all the machine work and changes of
appreciable magnitude were made at Colorado
Springs. The apparatus was finally arranged
to give good mechanical operation, and a
number of tests were run showing the ytCT-
formance of the engine with the supercharger
opened u]) to the maximum limit possible.
The supercharger was ilesigned for operation
at l.S, ()()() ft. with some margin. It was
THE GENERAL ELECTRIC TURBO-SUPERCHARGER FOR AIRPLANES 481
'-''km*
i%
s<
S 2
5 H
482 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 6
possible at the existing altitude of 14, ()()()
feet not only to supercharge so as to give full
sea-level power, but also to overcharge so as to
cause the engine to pre-ignite.
It was agreed that results of the Pike's Peak
tests warranted the immediate installation of
Fig. 6 Major R. W. Schroeder, Lieut. G. W. Elsey, and the
Author (right to left)
the supercharger on an airplane, and arrange-
ments for doing this were in progress when the
Armistice caused a cessation of the work.
After the Armistice, careful re-examination
of the situation resulted in resumption of the
work in the early part of 19 HI. Various re-
arrangements were made in view of the
experience gained at Pike's Peak and the
apparatus was finally installed on an airplane.
After a number of tests on the ground, flight
tests were made.
It soon developed that a ver\- appreciable
increase of power was easily obtained when
the supercharger was opened up. The whole
airplane installation was not properly arranged
to take advantage of this power, however, and
changes were necessar\- in the radiator, cool-
ing system, propeller system, gasolene tank,
and pump system, etc. Changes in these
parts have been made from time to time,
and this work is still in progress. As the
work proceeds more and more power is
developed by the engine. Changes have also
been made in the supercharger itself.
Many remarkable flight tests have been
made. In fact, during the early work a
flight record of some kind was broken at ever\-
flight. Appreciable progress has already been
made, but the full capacities of the apparatus
have not yet been reached, and further
improvements of performance are to be
expected.
Fig. 7 shows the airplane installation, and
Fig. () shows Major R. \V. Schroeder, who has
made all of the flight tests to date, together
with Lieut. George W. Elsey, who has made
all of the flight observations to date. The
aviators are of course clothed for the intense
cold of high altitudes and carr>- the parachutes
that are now regularly used by the U. S. Air
Service in experimental work.
Measurement of Altitude
The altitude of an airplane is measured by
an altimeter such as is shown in Fig. S. This
is essentially an aneroid barometer. It com-
prises a chamber almost wholly exhausted of
air, on one side of which is a flexible metal
diaphragm. As the atmospheric pressure
Fig. 7.
Supercharger Equipped Airplane. The extra long propeller is used to hold the engine tpeed down to normal
in the rarified atmosphere of high altitude
THE GENERAL ELECTRIC TURBO-SUPERCHARGER FOR AIRPLANES 4S3
presses on this diaphragm to a greater or
lesser extent, the diaphragm moves in or out.
This motion actuates a train of mechanism,
ending in a needle moving over a scale. The
temperature of the instrument itself must, of
course, have no effect on the readings. Tem-
perature compensation is arranged for bv leav-
ing a certain amount of air in the\-acuum cham-
ber and also by use of metal, in one of the levers,
which has an appreciable coefficient of expan-
sion. This temperature compensation is never
quite exact, however, and a slight correction to
the indications must be made, to take account
of the actual temperature of the parts of the
instrument at the time of an observation.
The reading of the instrument with tem-
perature compensation taken into account
gives the absolute pressure at the altitude
in question and it is from this absolute pres-
sure that the altitude is computed. Knowing
the absolute pressure at the field from which
the flight is made, as given by the barometer,
the absolute pressure at altitude as given by
the altimeter reading, and the temperature
of the column of air between these two points
at a number of heights, the difference in eleva-
tion can be computed by appropriate for-
mula. There exist tables of average values
of temperatures at various altitudes to enable
this computation to be made approximately
for an average case. However, where an
actual altitude record is involved, the actual
temperatures at various altitudes during the
ascent must be observed and mserted in the
formula. The determination of the altitude
in a record flight is therefore a matter of
some complexity. It has been very carefully
done in the case of the supercharger flights at
McCook Field.
The instrument in Fig. S is an indicating
instrument. The instruments actually used
for the final computation of altitude records
are recording instruments called "baro-
graphs," which operate on the same principle.
Fig. 10 shows the autograjahic record of such
an instrument. After a flight the recording
instruments used are removed from the plane
and placed under the bell-jar of an air pump,
connected with a mercury column, while
the clock which causes the rotation of the
record paper is still running. Autographic
records are thus obtained at a number of
known values of absolute pressure, as shown
by the mercury column. This gives an
accurate calibration and establishes the
absolute pressure at the maximum altitude
attained. During record flights three inde-
pendent barographs are used for certainty.
Fig. 9 shows observations of temperatures
at high altitudes for a great many of the
supercharger flights. From the actual values
of these temperatures for a given flight and
the barograph record mentioned, the maxi-
mmn altitude is computed.
19 ■>,
Iff '-
Fig. 8. The Instrument which Indicates to the
Aviator the Height at which He is Flying
The amount of supercharging is measured
by a recording barograph of the same kind,
which is not exposed to atmospheric pressure,
however, but is enclosed in a sealed chamber
connected by a pipe line to the air conduit
at the carburetor inlets. By means of the
known temperatures the altitudes correspond-
ing to this record are known, so that there is
given a record of the equivalent altitude of
the engine. This is practically sea level as is
shown by the lower curve in Fig. 10.
4S4 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 6
-■40 -50 -20 -10 0 10 20 30 do » £0 TO 80
Atmospheric Tcmp«rature°F
Fig. 9. Curves of Temperature at High Altitudes. The plotted
points were made during supercharger airplane flights
The upper curve in Fig. 10 gives readings
of a Verturi-meter-Pitot-tube arrangement,
which gives the air speed. These readings
are cahbrated by an actual flight near the
ground over a measured course of three miles
with the use of stop watches.
By these methods very accurate knowledge
has been obtained of the performance of the
supercharger under many conditons.
Supercharger Performances
The supercharger which has been used to
date was primarilv designed for high speeds
at altitudes of IS^OOO to 22.000 ft. The Le
Pere plane on which the installation was made
had a ceiling of about 20,000 feet with two
men. and a speed at this altitude of 70 miles
per hour. With the supercharger in use, a
speed of about 140 miles an hour has been
attained at 22.000 ft. As already pointed
out, this has been attained with various parts
of the plane installation in a partially de-
veloped state. Theoretical computations have
been made showing that much higher speeds
at high altitudes are to be expected. The
progress of the flight tests to date indicates
that the theoretical expectations will be full\-
realized.
The making of high altitude records has
been ver}" attractive and the supercharger has.
of course, been used for this purpose as well
as for the speed courses mentioned. Success-
ively higher altitudes have been reached as
experience has been gained regarding the
manipulation of oxygen, gasolene, and other
details. The highest altitude reached with
two men was on October 4. 1919, with Major
R. W. Schroeder and Lieut. George W. Elsey.
The maximum indicated altitude was 32.335
ft. Various computations from ver\- corn-
Fig. 10. Sample Barograph Curve Record of an Airplane Flight. From »uch ■ record, and the temperature* •« fthown
in Fig. 9. the true altitudes are calculated
THE GENERAL ELECTRIC TURBO-SUPERCHARGER FOR AIRPLANES 485
plete observations, give the actual height
above the ground as 31,800 ft. Complete
details of these computations, as officially
verified, are given in Flyini^ for January, 1920.
This figure is about one mile higher than the
nearest two-man altitude record without a
supercharger.
On February 27th, Major Schroeder made a
flight alone, attaining an actual height above
the ground finally computed as 36,130 ft.
((i.S.5 miles). The lowest temperature reached
was minus 67 deg. F. At the maximum
altitude. Major Schrocder's oxygen apparatus
failed and he became unconscious and lost
control of the plane. The recording instru-
ments, of course, continued to work and these
show that there was an almost vertical fall
of about five miles in two minutes fan average
speed of fall of 150 m.p.h.). Observers in
Dayton saw the plane spinning around as it
fell. Major Schroeder became semi-conscious
as he neared the earth and, at an altitude of
about 3000 ft., he succeeded, in a half-dazed
semi-automatic way, in righting the plane and
making a good landing in his own field, again
becoming unconscious. He was taken to a
hospital in a serious condition, but has since
almost completely recovered. The super-
charger, engine, and plane were in perfect
working order after the flight.
At the maximum altitude attained, record-
ing instruments showed that the plane was
still climbing at the rate of about 125 ft. per
minute and it was estimated that an altitude
of 40,000 ft. would have been attained if the
oxygen apparatus had not failed.
General Electric Supercharger Equipped Le Pere Biplane photographed from another
plane while test flight at McCook Field, Dayton Ohio
486 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 6
Relativity Theories in Physics
By Dr. Richard C. Tolman
Fixed Nitrogen Research Laboratory
The results observed very recentlj' during the eclipse of the sun were so startlingly in accord with Ein-
stein's theon' of relativity that the subject at once received universal attention and a great deal of discussion
arose in the daily papers over the exact significance of the nfew theory. The paper by Dr. Tolman is a concise
and extremely interesting exposition of Einstein's theory in relation to the older theories of relativity, and is
written bv one of the best authorities on the subject. — Editor.
In the following paper we shall first present
a description of the general nature of relativity
theories and then, by way of illustration, shall
give brief and hence necessarily incomplete
accounts of three relativity theories which
have actually been used in physics. The first
of these will be the theon,' of similitude, (or
theory of the relativity of size) ; the second.
Einstein's original theory of the relativity of
uniform motion; and the third, Einstein's
general theor\- for the relativity of all types
of motion, with its applications to gravitation.
The Nature of Relativity Theories
The general idea of relativity arises from the
fact that all our quantitative judgments are
in the nature of comparisons. To make a
quantitative judgment, we compare the phe-
nomenon under consideration as to size, as to
position, as to velocity, or what not, with some
standard reference sj'stem. Our quantitative
judgment is thus in the nature of a relation
between the phenomenon under considera-
tion and the standard reference system.
Thus I can speak of the length of a table
relative to the length of a standard meter-
stick, or relative to the length of a foot rule,
or relative to the length of any other chosen
standard. I can speak of the position of the
planet Mercury relative to a system of co-
ordinates having the earth at their origin, or
relative to a system of co-ordinates having the
sun at their origin. I can speak of the velocity
of a man in a railroad train relative to the car
in which he is located, or relative to a station
platform past which the train is moving. To
speak of absolute length, absolute position,
or absolute velocity, would be meaningless.
All quantitative judgments are thus relative to
the more or less arbitrarily chosen standard
system of reference.
It must be particularly noticed that this idea
of the relativity of quantitative judgments
is such as to make the nature of these judg-
ments depend not only on the properties of
the phenomenon to be judged, but also on the
particular choice of reference system which is
made. In general, a change in reference sys-
tem will be accompanied by an appropriate
change in the quantitative judgment.
For example, suppose I am interested in
giving a quantitative description of a circle.
If I take as my reference system a set of
Cartesian co-ordinates having the center of the
circle at its origin, the mathematical equation
which gives such a quantitative description of
the circle will have the familiar form
where a is the radius of the circle measured
in the particular units of length employed.
If now, however, I change my reference
system by choosing a new and shorter stand-
ard of length 1 m times as long as the original,
the equation describing the circle will be trans-
formed into
x'-+y- = m-a'-
Or if, on the other hand, I change my refer-
ence system to a new set of Cartesian co-
ordinates parallel to the first, but having its
origin not at the center of the circle, the equa-
tion for the circle will assume the form
{x-xoy-+(y-yor- = a'-
where xo and yo are the co-ordinates giving
the position of the center of the circle.
As still a further type of change of reference
system, I might change from Cartesian to
polar co-ordinates, and my equation for the
circle would then assume the simple form
r = a
provided the center of the circle lies at the
origin of co-ordinates.
Attention should be paid to the fact that
the change in quantitative statements which
accompanies a change in reference system
may be one either of form, or merely of the
numerical values entering into the quantita-
tive statement. Thus, when I change from
Cartesian to polar co-ordinates, I have a
change in the form of the quantitative state-
ment. By changing, however, from one set
of Cartesian co-ordinates to another set of
Cartesian co-ordinates, which diflfers from the
RELATIVITY THEORIES IN PHYSICS
487
first only in the magnitude chosen as unit
length, I obtain a change merely in the numer-
ical values entering into the description, but
not a change in the form of the description.
Of course, the mere fact that the form and
numerical content of the equations of physics
are dependent on choice of reference system,
is not, itself, sufficient to permit the drawing
of definite conclusions as to the nature of
physical phenomena. In order to obtain such
conclusions, we must know how the equations
of physics are dependent upon the choice of
reference system. This information is usually
most succinctly expressed by a statement as
to those things which remain invariant (i.e.,
are not changed), when the transformation to
the new reference system is made. In fact,
any relativity theory can be most conveniently
founded on a statement as to the type of
change in reference system which is to be
considered and a statement as to the invari-
ants for this transformation. On the basis of
these two statements, it will then be possible
to build up the whole theory of relativity for
the particular branch of investigation under
consideration.
In carrying out such an application of
relativity methods, we are of course at liberty
to consider any change in reference system
that we may desire. The gist of the problem
lies in determining what shall be invariant
when the transformation to the new reference
system is made. The decision as to this is
usually presented in the form of a postulate,
which presents our preconceived ideas as to
to those things which will not be affected by
the change in reference system contemplated.
Theory of Similitude
Let us now consider as a simple example of
the application of relativity methods, the
theory of similitude,' or perhaps, as it might
better be called, the theory of the relativity
of size.
The fundamental idea of the theory of
similitude is that there ought to be no signifi-
cance in the choice of any particular length
(1) See Tolman. Phys. Rev.. S. 244, 1914; 4. 14.5. 1914; 6.
219, 1915; S. 8. 1916; S, 2.'i7. 1917. Buckingham, ibid.. 4. 345.
1914. Nordstrom. Finska Vetenskaps Soc. Forh.. S7, 1914-15;
Afd. A. No. 22. Ishiwara. Science Report of Tohoku Imp. Univ. ,
.5. .33, 1916. Ehrenfest-Afanassjewa. Phys. Rev.. 8. 1. 1916.
Bridgman, ibid., «, 423, 1916. Karrcr, ibid.. S. 290, 1917. Davis
Science 50. 3.38. 1919.
The theory in question was originally called the theory of
similitude since the underlying postulate on which it may be
founded was first stated in the form:
The fundamental entities out of which the physical universe
is constructed are of such a nature that from them a miniature
universe could be constructed exactly similar in every respect to
the present universe.
In the present paper we take a form of statement for the
fundamental postulate which shows more clearly the relation
between this and other relativity theories.
as the standard length, in terms of which all
other measurements should be made. Since
the length of an object is, in any case, merely
a relative matter, and since it is meaningless
to speak of absolute lengths, it would seem as
if the general laws of physics describing
classes of phenomena ought to be entirely
independent of the choice of standard length,
although, of course, the numerical values
entering into the description of any particular
phenomenon will depend on this choice.
This general idea can be expressed more
definitely by the following postulate upon
which the theory of similitude may be founded :
A change is possible in the magnitudes of the
standards for the measurement of the different
quantities of physics, including any desired
change in the standard of length, which will
leave all the general equations of physics abso-
lutely invariant, both as to form and numerical
content.
By the term " general " equations of physics
we are to understand those equations which
describe classes of phenomena, rather than
equations which mereh' describe one partic-
■ ular phenomenon. Thus, for example, the
equation, C=tD, giving the relation be-
tween the circvimference and diameter of any
circle would be a "general" equation and
would be absolutely invariant both as to
form and as to the numerical value of the
quantity ir, for any change in the standard of
length. A statement, however, as to the
diameter of some one particular circle, such
as D = 24, would be a "special" equation,
and would, of course, not be invariant in
numerical content if we changed our stand-
ard of length.
In order to satisfy our postulate, we must
be able to find a set of equations by which we
can transform quantities of length, and such
other quantities as may be necessary, to a
new set of standards of different magnitude,
and yet leave all the general equations of
physics absolutely invariant, both as to form
and numerical content. Moreover, this set
of transformation equations must correspond
to any desired change in the standard of
length.
As a matter of fact, it has been possible to
find such a set of transformation equations.
For the five fundamental kinds of quantity —
length, time, mass, quantity of electricity, and
entropy — the transformation equations have
the form:
l' = x' t'^xt
m
S' = S (1)
488 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 6
Since x may be any desired number, it is seen
that these transformation equations cor-
respond to any desired change in the standard
of length and, by trial, it can be shown that
the substitution of these equations will, as a
matter of fact, leave all the general equations
of physics absolutely invariant.
Having obtained these equations for the
transformation of the five fundamental kinds
of quantity, it is easy to obtain transforma-
tion equations for any desired kind cf quantity
merely making use of the definition of the
derived quantities in terms of the funda-
mental quantities. Thus we can write down
the following further transformation equations
for velocity, energy, frequency, and force, etc. ;
X X- ■ Xi
To illustrate the usefulness of the theory
of similitude, we may use these transformation
equations to derive a general relation con-
necting energy and frequency. This was
done, as a matter of fact, by Dr. Karrer' of
the Fixed Nitrogen Research Laboratory, at
a time when it was important to supplement
our inexact empirical knowledge of the relation
between energy and frequency by theoretical
investigations.
Let us suppose that our experimental
investigations have indicated that there must
be some general relation between the energ>-
given up, or absorbed by an oscillating
svstem, and the frequency of oscillation.
Expressing this fact, let us write the equation
E = <i>{v) (3)
where <l> is the unknown function, the form
of which we wish to determine. In accord-
ance with our postulate, 4> must be entirely
invariant when we change to our new stand-
ards of reference. Hence we can obviously
also write
E' = <i>{v')
where <^ has the same form as above. Sub-
stituting our transformation equations (2)
for energv and frequencv, we obtain
or combining with (3)
E = .i>{v)=x<t>(^\ (4)
^ loc cit.
* In another place. I hope to show the bearing of the theory
of simihtude on Einstein's solution of the problem of gravitation.
' For an English account of Einstein's theory of the
relativity of uniform motion, see Cunningham. "The Principle
of Relativity. " Cambridge University Press. 1914; Silberstein.
"The Theory of Relativity." Macmillian. 1914. Tolman. "The
Theory of the Relativity of Motion." University of California
Press. Berkeley. 1917.
It will be seen by inspection that the only
solution for this functional equation is
E = hv (5)
where /z is a constant. By this simple process
we have thus derived the fundamental equa-
tion of the quantum theory.
As another illustration of the usefulness of
the theorv^ of similitude, we may note that
Newton's equation for the gravitational at-
traction between bodies is not invariant when
we substitute the transformation equations
given above. As a matter of fact, the equation
l^-
transforms into
f = x-k-
r-
(7)
when we substitute the transformation equa-
tions. This alone should make us suspect
that Newton's equation of gravitation is not
one of the general equations of physics. It
may give correct numerical results, but its
failure to conform with the requirements of
the theory of similitude indicates that some
more fundamental treatment of gravitation
is demanded and this, as a matter of fact,
has been provided by Einstein's general
relativity theory, which will be described in
the last section of this paper.'
Einstein's Theory of the Relativity of Uniform
Motion
Einstein's first work on the theory of the
relativity of motion- was based on the
general idea that co-ordinate systems in uni-
foriii relative motion must be entirely equiva-
lent to each other. Since there is no such
thing as absolute velocity, it would seem
as if one co-ordinate system should be just as
good as another, moving relative to the first
with some uniform velocity, and that the equa-
tions of physics ought to be expressible in
such a form as to show their independence
of the choice of reference system. In par-
ticular, it would seem as if the description of
the simplest of all kinematical occurrences:
namely, the spreading out of a light dis-
turbance in free space, should be absolutely
invariant for all co-ordinate systems in
uniform relative motion. This idea may be
stated in the form of the following definite
postulate;
The general laws of physics are expressible
in equations which are invariant, when we
change from one set of space-time co-orJtnatcs
to another set moving relative to the first with
RELATIVITY THEORIES IN PHYSICS
489
uniform velocity; and, in particular, the equa-
tion which describes the way light spreads out
in free space, is completely invariant in form
and numerical content for such a change in
reference system.
The latter part of this statement is some-
times called the second postulate of relativity
and is stated in the form: The velocity of
light in free space appears the same to all
observers, regardless of the motion of the
source of light and the observer.
This statement, which was originally taken
by Einstein more or less as an unproved
postulate, has, as a matter of fact, received
very satisfactory experimental proof. Thus
the Michelson-Morley experiment, which
compares the velocity of light perpendicular
and parallel to the earth's motion around the
sun, may be regarded as showing that the
velocity of light is unaffected by a simulta-
neous motion of source of light and observer
through any suppositious ether, and the
recent work of Majorana' on the velocity
of light reflected from a moving mirror, and
more recently from an original source set in
motion, show that the velocity of light is
independent of the relative motion of source
and observer.
The revolutionary nature of this postulate
must not be overlooked, and no attempt to
conceal it can be tolerated. Suppose, for
example, that 5 is a source of light and A and
B two moving systems. A is moving towards
the source S, and B away from it. Observers
on the systems mark off equal distances aa'
and bb' along the path of the light, and deter-
mine the time taken for light to pass from a to
a' and b to b' respectively. Contrary to what
seem the simple conclusions of common sense,
the postulate requires that
•
S
1-
b'
the time taken for the light to pass from a to a'
shall measure the same as the time for the light
to go from b to b'. Such a consideration
makes the path obvious by which the theory
of relativity has been led to strange con-
clusions as to the intercomparison of measure-
ments of length and time made in systems
moving relative to each other.
If our postulate is true, it is evident that
we must completely remodel our so-called
"common sense" ideas as to the nature of
space and time, which, however, have been
1 Phil. Mag. S6 163 (1918) Phil. Mag. 37 145 (1919).
built up through a long ancestral experience
which involved such small relative velocities
as to make the difference between the correct
and the "common sense" ideas of space and
time negligible.
Returning now to the statement of the
fundamental postulate which we gave above,
we can proceed to the development of the
original Einstein theory of the relativity of
uniform motion.
Using Cartesian co-ordinates, we may
obviously write the following equation as a
description of the way in which a light dis-
turbance in free space spreads out
(sy^TiyK^)'
(s)
where c is the velocity of light. Multiplying
through by dt''- and transposing, we obtain the
equation in the form
dx''->rdy-'+dz"--c''dt- = 0 (9)
In accordance with our postulate this
equation must transform identically into
itself, when we change to a new system of co-
ordinates in uniform motion relative to the
first. Our problem is to find a set of trans-
formation equations which will obey this
requirement, and which will reduce, at low
relative velocities to the form required by
our "common sense" ideas of space and time,
since these are known to be adequate when we
deal with velocities small compared with that
of light.
As a matter of fact, if we consider two
systems of space-time co-ordinates, S and S',
such that S' is moving past .S in the X direc-
tion with the velocity V, it can be shown that
the desired transformation equations have the
form
X - 17
y' =y
-V^lc^
t' =
t-X/Vc^
It will be seen by trial that these equations are
such as to transform the equation
dx'^-\-dy'''+dz'^-cW = 0
identically into
dx'^ + dy'+ds'-c^dt'^O
and that they fulfill the further requirement of
reducing to the familiar Galilean form
x' = x— Vt
y'=y
z' =z
1' =t
490 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
when the relative velocity of the systems V is
small compared with the velocity of light c.
Referring again to our fundamental postu-
late, Einstein's theory requires not only that
the equation for the spreading out of light in
free space must be absolutely invariant for
the transformation in question, but that all
the laws of physics must be expressible in
equations which are invariant for this same
transformation. This latter requirement has
been of great importance in the modem
development of theoretical physics, by pro-
viding important information as to what must
be the nature of the general equations of
physics.
These new investigations in theoretical
physics have shown that it is possible to
retain Hamilton's principle, as the funda-
mental law from which the equations in the
most varied branches of physics can be
derived. This has been done by showing that
the quantity Hdt occurring in Hamilton's
fundamental equation
€'
8lHdt = 0
(11)
assumes for ever\^ branch of physics the same
form
c
-u'dt
(12)
where the symbol - indicates that we are
to sum up the quantity in question for all
parts of the system in question', Eo is the
energy of each separate portion of the system
as measured by an observ'cr at rest with
respect to that particular portion of the
system, and u is the velocity of that portion
of the system.
With the help of this expression, we may
now write as the fundamental equation for
every branch of physics
6f-I^V^^u'dt = 0 (13)
To show as a matter of fact that this funda-
mental equation is invariant for the trans-
formations under consideration, we may note
that bv introducing the substitution
* For a continuous system the summation will have to be
made by a process of integration.
' For an English account of Einstein's general relativity theory,
see Eddington. Report on the Relativity Theory of Gravitation,
London. 1918.
For a discussion of some of the philosaphica! implications of
relativity, see Wilson. "Space. Time, and Gravitation," THt
Scientific Monthly. 10. 217 (1920).
' Owing to the limited nature of the invariance. prescribed
by the above postulate. Wilson, Astrophys. J. iS, 244 (191~).
would prefer to call this postulate a co-variance principle rather
than a relativity principle.
we can transform equation (13) into
iC—I -V - (dx= +d-f +dz' —c'- dC) = 0
(14)
This makes it evident that the same trans-
formation equations which leave
dx'-+dy'-+dz^--c'-df-
invariant will also leave Hamilton's equation
unchanged in form. Since this equation can
be made the starting point for ever\' branch
of physics we have thus shown the general
applicability of Einstein's theory of the rela-
tivity of uniform motion.
Einstein's Theory of Gravity and General Relativity
Einstein's original relativity theon.- concerns
itself solely with the consideration of systems
in uniform relative motion. It is obvious that
we wish to be able to employ reference systems
having any type of motion relative to each
other. Thus, accelerated systems of space-time
co-ordinates, rotating systems, or systems mov-
ing in any desired manner, ought all to be
utilizable for the description of physical
phenomena. As a matter of fact, Einstein's
theor>' of general relativity shows that all
possible co-ordinate systems are equally justi-
fiable.
As a basis for Einstein's general relativity
theorj'- we may take the postulate :
The laws of physics can be expressed in a set
of equations which are invariant in form,
although not necessarily in numerical content
for any possible transformation of space-time
co-ordinates.
Since this principle requires invariance
merely in form, and not in numerical content'
it might seem to be of little importance since
the well-known theories of curvilinear or
generalized co-ordinates have already pro-
vided general methods for expressing equa-
tions in such a way that their form is inde-
pendent of the choice of reference system. We
shall see, however, in what follows, that
Einstein is able to relate the change in the
numerical content of equations, accompanying
a change in reference system, to the change
in gravitational field which is also found to
accompany changes in reference system.
Einstein assumes that the general nature of
the relation bctwen gravitational field and
reference system is such that by using a
system of co-ordinates which has the natural
acceleration of gravity, the equations of
physics with certain restrictions will assume
the same form, as in a space free from gravita-
tional action. It is by a combination of the
above postulate with this further principle
RELATIVITY THEORIES IN PHYSICS
491
that Einstein is led to his important con-
clusions.
Returning now to the requirements of our
fundamental postulate, Einstein takes as his
general equation for the way in which hght
spreads out when referred to any set of
space-time co-ordinates, (or in space having
any gravitational field) instead of the simple
equation
the more general equation
gndxi-+g,2dxidx2i+gi3dxidx3+ . . . +gudx4- = 0
or
J gij dxi dxj = 0
(15)
and takes, instead of Hamilton's princii^le in
its earlier form
'/-
E,
■Z V -(dx^+dy'^+dz^-
'C' df^) = 0
the more general equation
5/z
'^—\' giidxi'^+gi2dxidx2 + g\3dx\dx3-\- ...+g„dx-i=0
or bC~I—\/igijdxidxj = 0 (16)
J ' C 1
In these equations, the four generalized co-
ordinates Xi X2 Xi and X4 replace the previous
co-ordinats x y z and t.
Furthermore, it will be noticed, as is
required by our fundamental postulate, that
these equations will be transformed into new
equations of exactly the same form by any
possible transformation of co-ordinates. Thus
if we put
Xi=Xi {Xi
Xi = X2 {xi
X3=X3 (x{
Xi = Xi {Xi
Xi Xi Xi)
Xi Xz Xi)
Xi Xi Xi)
Xi' X,' x:)
where the four functional relations may be
anything at all, we may then write
, hx\ , , bxi , , . hx\ , , bx\ , ,
dxi = ^,dx:' + ^—,dxi + ■r~-,dx3 + Y~idxi
oxi oxi 0x3 6x4
dxi
5X2
5xi
-,dx,' +
dxi = T-^dxi + . . .
OXi
dxi = -T-^,dxi+ ■ ■ ■
OXi
and substituting into
I gij dxi dxj
1
it will be found that we obtain an expression
of exactly the same form
I g'ijdx'idx'j
1
thus showing us that the fundamental
equations of physics, (15) and (16), are
invariant in jorm for any transformation
of co-ordinates. It must be noted, however,
that in general the numerical content of this
expression will not be invariant and the
quantities g'a will have different numerical
values from the quantities gij.
Our next problem is to determine the values
of these quantities gij. This has been done by
Einstein by obtaining an inter-relation be-
tween these quantities and gravitation, his
fundamental idea being that the presence and
magnitude of a gravitational field is entirely
dependent on the particular choice of co-
ordinates made. According to this idea, any
observer finds a gravitational field at any
point in space, only because he is using a set
of co-ordinates which does not have the
natural acceleration due to gravity at the
point in question. In fact, Einstein's specific
assumption is that it is always possible at any
point in space and time, for a limited region
surrounding that point, to choose a set of
co-ordinates such that the equations of
physics will all have the simple form found
for them in the original simple theory of
uniform relative motion. In other words,
at any point in space and time it is always
possible to choose a set of space-time co-
ordinates such that the expression
gndxi^+giidxidXi-\-gi3dxidx3+ . . . +gndxc
will reduce to the simpler form
dx'-+dy-+dz''-c''-dt'-
These co-ordinates will be called the "natu-
ral" co-ordinates for the point in question. It
must be noted, however, that Einstein's
assumption is, that this particular choice of
co-ordinates can be made only for a limited
region in space and a limited duration in time.
These ideas as to the inter-relation of
gravity and choice of co-ordinate system, and
the possibility of transforming away a
gravitational field for a limited region in space
and time, by a proper choice of co-ordinates,
may be illustrated by considering the phe-
nomena inside of a freely falling elevator. It
is evident that an observer inside this elevator,
using meter-sticks and clocks which have the
same downward acceleration as everything
else inside the elevator, would obtain for
the phenomena inside his limited region, the
same laws as would be found by an observer
in free space completely removed from any
gravitating bodies. The observer inside the
elevator would find no evidence of any attrac-
tion due to gravity. The fioor of the elevator
would exert no upward pressure on his feet.
Bodies would have no tendency to move down-
ward with reference to the elevator itself and.
492 June, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII. No. G
if thrown across the elevator, would move in
straight lines instead of parabolas, referred
to the walls of the elevator.
In order now to obtain definite conclusions
as to the values that are to be assigned to the
quantities gij which occur in our general
equations of physics, we may set in definite
form the additional hypothesis used by Ein-
stein as follows:
At any point in space and time it is always
possible to choose a set of space-time co-ordinates
with reference to which the quantities gij, and
their first differential coefficients will assume the
simple values which they have in free space.
This principle has been called by Einstein
the equivalence hypothesis, since it requires
that at least as far as first order diiferentials.
a gravitational field of force shall be identical
with the field of force which can be generated
in free space solely by a choice of co-ordinates.
In order to derive from this equivalence
hypothesis, definite equations connecting the
quantities ga with Xi Xi X3 and X4, Einstein
has found it necessary to develop a very
elaborate mathematical theory, which is
beyond the scope of this paper. As a final
result, however, using polar co-ordinates, we
obtain the following expression for the
quantity - gij dxi dxj, in the neighborhood
1
of a central attracting body such as the sun.
1—-—- )c°-dt' (17)
c-r J
drf_
2*1
In this expression r is the radius from the sun
to the point in question, m is the mass of the
sun, and k is the gravitational constant.
Substituting this expression into our general
equation (15) for the spreading out of a light
disturbance, we obtain of course a description
of the way light will move in the gravitational
field of a central attracting body of mass m.
and Einstein has predicted on this basis that a
ray of light grazing the sun's limb will be bent
inward lay the gravitational field so as to give
a total deflection of 1.7')". This is the pre-
diction which was recently tested by British
astronomers by taking photographs of the
sun and the stars surrounding it during the
eclipse of May 29, 1919, and comparing the
relative positions of the stars with those
obtained when the sun was not present. As
a matter of fact, the rays of light from stars
were bent in towards the sun by an amount
almost exactly that predicted.
Substituting the expression (17) for the
value of 2' g,> dxi dxj in the neighborhood of
t
the sun into (Hi), we shall obtain a modified
form of Hamilton's principle which will per-
mit us to determine the motion of a particle
in the gravitational field surrounding the sun.
A computation on this basis has been carried
out by Einstein, which shows that the path
of a planet around the sun should not be quite
a stationan,- ellipse, but that the major axis
of the ellipse should gradually rotate. For the
solar planets the only case in which this effect
is large enough to be determined is for the
orbit of Mercury, and Einstein has calculated
that the long axis of Mercur\''s ellipse should
rotate 43" per century, which just removes
the previous unexplained anomalies in the
orbit of this planet.
These two confirmations of Einstein's
theory are certainly ver\- compelling. Ein-
stein has made a further prediction that the
frequency of vibration of an atom should
depend on the gravitational potential at the
point where it is located, and he has calculated
that there should be a measurable deflection,
towards the red, of lines in the specta which
originate from the strong gravitational field
at the surface of the sun. This prediction has
failed to receive confirmation, although it has
been carefully looked for by St. John at the Mt.
Wilson Obser\-atory. Further investigation as
to the theory and as to the experimental test
of this part of Einstein's work must certainly
be undertaken.
In conclusion, I wish to express my own
feeling that in Einstein's latest theory, allow-
ing for further modifications which will un-
doubtedly be introduced, he has made a
contribution of fundamental importance for
theoretical physics. He has shown that all
the laws of ]jhysics can be expressed in equa-
tions which arc completely invariant in form
for all possible transformations of space-time
co-ordinates. Although the equations of
physics are not invariant in numerical content
for these transfonnations. nevertheless, he has
shown that the changes in numerical content
are to be simi^ly accounted for by their rela-
tion to the gravitational field, found by the
particular observer in question.
Einstein has thus solved the two age-long
problems of the relativity of all motion and
of the uniformity of gravitation. All systems
of space-time co-ordinates are equally justi-
fiable for use in the description of physical
phenomena without reference to their state
of motion, and all bodies at a given point in
space, regardless of the material of which they
are composed, will experience exactly the same
gravitational acceleration, since this gravita-
tional acceleration is due to the particular
choice of space-time co-ordinates made.
493
The Production and Measurement of High Vacua
Part I
By Dr. Saul Dushman
Research Laboratory, General Electric Company
The marvelous development during the past few years in the application of hot-cathode devices to the
field of wireless telephony and telegraphy has been largely due to the great progress in the art of producing
and maintaining extremely high vacua. Simultaneously, because of the knowledge concerning the structure
of the atom revealed by investigations at very low gas pressures, added interest has been directed to the whole
subject of the production of high vacua. This article, which is the first of a series by Dr. Dushman, discusses
the fundamental principles of the kinetic theory of gases which are of importance in connection with the sub-
sequent discussions of methods for the production and measurement of high vacua. — Editor.
INTRODUCTION
"Nature abhors a vacuum." This state-
ment represents the sum total of the knowl-
edge possessed by the ancients of a field of
scientific investigation which within the past
decade has yielded results of extreme im-
portance. In 1643, Torricelli, a pupil of
Galileo, showed that nature abhors a vacuum
to a limited extent and the discoverer of the
fact that the atmosphere exerts a pressure
equivalent to that of a mercury column 32
inches in height, is remembered by the desig-
nation "Torricellian vacuum" for the space
above the mercury in the barometric tube.
No doubt Torricelli imagined that this
space is a "perfect void." We now know,
however, that in this space there is mercury
vapor at a pressure corresponding to about
two or three millionths of an atmosphere and
also traces of water vapor and air whose
pressure may often amount to one or more
millionths of an atmosphere.
In 1654 Otto von Guericke invented the
first mechanical air-pump which was sub-
sequently improved by Boyle, Hawksbee,
Smeaton and others. During the two hun-
dred years or so that followed, the interest in
low pressure phenomena was more or less
academic and often that of the dilettante.
The paths of glory laid out by Newton,
Laplace and Maxwell in mathematical phys-
ics, and by Priestly, Lavoisier and Faraday
in experimental science, were so enticing
that little or no enthusiasm could be aroused
in investigations of "empty space." How-
ever, with the development of the carbon
filament lamp on the one hand, and the
discovery by Geissler and others of curious
electrical phenomena in gases at low pres-
sures, there began a series of investigations
in this field which have not only increased
enormously our knowledge of the technique
for the production of lower and lower pres-
sures, but have also led to results which have
profoundly affected our views of the nature
of matter and energy.
When Crookes first observed the phenomena
of cathode rays, he thought that he had dis-
covered a fourth, or radiant state of matter.
A further investigation of this subject by
J. J. Thomson led him, as is well known, to
the conclusion that in the conduction of
electricity through gases at low pressures,
the negative current, or so-called cathode
rays, is carried by extremely small corpuscles
or electrons, whose mass is about one two-
thousandths of that of a hydrogen atom,
while the charge is exactly the same as that
carried by a hydrogen ion in electrolysis, but
opposite, of course, in sign. These electrons
are the principal carriers of the current in all
cases of conduction in gases at low pressures.
It was also obser\'ed that electrons are
emitted from metals under the influence of
light, and Richardson showed that electrons
are emitted from incandescent metals. The
conclusion was therefore drawn that electrons
are present in the atoms of all elements — a
conclusion which was very soon corroborated
by obsen-ations on the radio-active elements.
With the discovery by Roentgen of X-rays,
the study of so-called vacuum tube phe-
nomena entered upon a new phase which has
led not only to increased knowledge of the
structure of matter and the nature of X-rays,
but also to vast improvements in both the
devices for the production of these rays and
their application to medical diagnosis and
therapy.
The mutual effects of purely scientific
discovery and technical achievement have
at no other time been better illustrated than
in the history of the development of the hot
cathode high vacuum devices which play
such an important role at the present time
in both the application of X-rays and of
wireless telephony. The history of this
development has been so interwoven with
494 June, l«-20
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
the progress achieved during the past decade
in the field of high vacua that a few remarks
on this subject may not be out of place in
this connection.
It has already been mentioned that elec-
trons are emitted from the surface of incan-
descent metals. A careful study of the varia-
tion in the number of electrons emitted per
unit area with change in temperature led
Richardson to the theor\^ that the electrons
are emitted from the metal by a process quite
similar to that of ordinary evaporation. The
mathematical relations are the same in both
cases and, as in the case of ordinary molecular
evaporation, it is also possible to calculate
the heat of evaporation of the electrons for
different kinds of surfaces.
This view of the existence of an electron
emission per ipse was opposed by a large
number of investigators who maintained that
the observed emission of electrons is a second-
ary effect due to chemical reactions at the
surface, between the metal and the residual
amount of gas present in the vessel. There
was some excuse for this view, as Richard-
son's experiments were not carried out at
very low pressures. The conclusion was
therefore, quite prevalent that in a "per-
fect vacuum" the electron emission would
disappear.
A similar view was held with regard to
the photo-electric effect, in which case elec-
trons are emitted by the action of ultra-violet
and ordinary visible radiation.
In order to throw some light on these
problems. Dr. Langmuir carried out a series
of experiments on electron emission in which
special care was taken to obtain extremely
low pressures. The results of this investiga-
tion showed that not only does the electron
emission persist even in the best obtainable
vacuum, but that the rate of emission at
any given temperature is a specific property
of the metal. It was found that the j^ower
of emitting electrons is also greatly decreased
by slight traces of different gases, even at
ver\' low pressures. However, if the vacuum
is sufficiently good this electron emission is
quite reproriuceable and constant, so that
further improvement in degree of vacuum
causes no increase in emission. It was also
observed that at these low pressures the elec-
tron emission exhibits space charges effect,
that is, the mutual repulsion between the elec-
trons emitted from the hot surface limits the
further emission of electrons, and the elec-
tron current to the anode is then de])en(k'nt
u]wn the anode voltage. Such an effect
could arise only under such conditions that
the number of positive ions formed by collis-
ions between electrons and gas molecules is
extremely small, in other words, at ver\-
low gas pressures. This accounts for the
fact that this phenomenon was not obser\-ed
by previous investigators.
These discoveries immediately paved the
way for the development of the hot cathode
X-ray tube by Dr. Coolidge and also led to
the development of other hot cathode devices,
such as the kenotron, pliotron and dynatron.
whose application in wireless telephony and
telegraph\- has been of immense importance.
At the same time the necessity of producing
and maintaining high vacua in these devices
has led to a vast amount of improvement in
methods of exhaust.
While the phenomena of electrical con-
duction in gases at xery low pressures have
thus ser\-ed to arouse a great deal of interest
in the subject of high vacua, a number of
investigations in other fields of physics and
chemistry have also led to greater interest
in the same field. The work of Knudsen.
Smoluchowsky, Gaede and others on the
application of the kinetic theon* of gases to
low pressures, and the striking results ob-
tained by Langmuir on the mechanism of
chemical reactions at low pressures have led
to new views upon the nature of chemical
and physical forces between atoms and we
can look forward, as a result of these investi-
gations, to solving some of the most vexing
problems in both physics and chemistry by
a study of the phenomena in gases at vcr>-
low pressures.
Of necessity, as the technique of high
vacuum production has improved, methods
have been developed for measuring these
extremely low gas pressures. A great deal
of literature has been published during recent
years on this whole subject, and a great
deal of information has been graduallx"
acquired in different laboratories about the
actual technique of producing and measuring
these pressures. In view of the imixirtant
results to be expected from further investiga-
tions of low pressure I'henonicna it has been
thought worth while to describe in a scries of
articles not only the methods available at
present for the production and measurement
of high vacua, but also to a lesser extent the
more important results which have been
obtained by the different investigators who
have studied the i)hysical and chemical
|)henomena exhibited in gases at very low
pressures.
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
495
KINETIC THEORY OF GASES.
APPLICATION TO GASES AT
LOW PRESSURES
Laws of Boyle and Gay-Lussac
The state of a gas is ordinarily defined bv
means of the volume which is occupied by a
given mass under definite conditions of tem-
perature and pressure. The three laws of
Boyle, Gay-Lussac and Avogadro may be
combined in the form of the well-known
relation :
PV = vRT (1)
where P and V denote the pressure and
volume respectively, T denotes the absolute
temperature (degrees Centigrade-f 273), v is
the number of mols (mass in grams divided
by the molecular weight) and /? is a constant
for al! gases.
The value of this constant is derived from
the experimentally determined value of the
volume of one mol of an ideal gas at given
values of P and T. As standard pressure we
shall consider that of 1 niegahar. By defini-
tion, this is equal to 10^ dynes per cm.^, and
corresponds very closely to a pressure of
7.50 mm. of mercury at 0 deg. C, lat. 45 deg.,
and sea level.
For r = 273.1 and P=\ megabar, V =
22,708 cm.' per mol.
Hence, i? = 83. 15 X 10*^ ergs per degree abs.
Denoting the weight of gas by m, and its
molecular weight by M, equation (1) may
therefore be written in the form.
P l' = 83.15X10' wi^
(la)
where P is measured in bars (dynes per cm.-)
and V in cm.^
Now the pressures which we ordinarily deal
with in high vacuum phenomena range from
1 to 10~^ bar and even less. It is evident that
at these pressures the volume of even a very
small amount of any gas may be quite con-
siderable. Thus, by applying the above
equation to the case of hydrogen (M = 2.016),
we find that the volume occupied by 1 milli-
gram of this gas at a pressure of 1 bar and
20 deg. C. (room temperature), is 1.209X10'
cm.', while at standard pressure the volume
is only 12.09 cm.'
Kinetic Theory of Gases
For a proijcr understanding of phemonena
in gases, more especially at low pressures, it
is essential to consider these phenomena from
the point of view of the kinetic theory of
gases. At the present time we can, as a
matter of fact, regard this theory as much
more than a mere hypothesis. The evidence
of the actual existence of atoms and mole-
cules is so conclusive that very few would
care to believe to the contrary. On the other
hand, the theory has enabled us to interpret
and prophesy so many facts about gases
that one naturally uses this point of view in
discussing any phenomena in gases.
The kinetic theory of matter, and more
especially that of gases, rests essentially upon
two fundamental assumptions. The first
of these postulates is that matter is made up
of extremely small particles or molecules,
and that the molecules of the same chemical
substance are exactly alike as regards size,
shape, mass, and so forth. The second
postulate is that the molecules of a gas are in
constant motion, and this motion is intimately
related to the temperature. In fact, the
temperature of a gas is a manifestation of
the amount of molecular motion. In the
case of solids, at least those that are crys-
talline, it has been shown by the investiga-
tions of Bragg and others that the atoms
which constitute the molecules when the
substance is in the gaseous state are arranged
in definite space-arrangements, and in this
case the effect of temperature increase con-
sists in increasing the kinetic energy of vibra-
tion of the atoms about their mean positions
of equilibrium. But in the case of gases the
effect of increased temperature is evidenced
by increased translational kinetic energ}^ of
the molecules, and a relatively simple cal-
culation based on these assumptions leads to
the relation.
MG^_2,
(2)
where G denotes the so-called mean velocity of
the molecules at the absolute temperature T.
Substituting for R the value already given,
this equation may be written in the form,
G=.
3RT
\ M
15,800
T
yJM
cm. sec.
(2a)
Table I gives the values of the mean velocity
at 0 deg. C, and 20 deg. C, for some of the
more common gases.
It follows directly from this equation that
at constant temperature the rates of fiow of
different gases through a narrow opening
must vary inversely as the square roots of
the molecular weights. This conclusion is of
importance in connection with exhaust prob-
lems since it indicates that heavier gases
must be more difficult to pump out than
lighter ones.
49() June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. G
Maxwell's Law of Distribution of Velocities
It is evident that even if all the molecules
in a given volume actually possessed the
same velocity at any initial instant, the con-
stantly occurring collisions would disturb
this equal distribution of velocities and a
-
/^
N
1
/
1
'c,
V'
1
i
\
1
\
j
"
3
\
>■,
/
A \
1
/
1-9
V
\
1
^
*
0
^
!
*
^'
,«■
\
1 1
/
1^
1-
^l
/ ;
Si
\i 1
1
s;
*
-S ^
/
.\^
T
<
4-^ T
.
^
^-^X
1
J
1
[
1
^-
in
^
b\
1
1
X
It ''I
'r4-UJJ
LU
LJ
Fi^. 1
non-uniform distribution would soon be
established. By applying the laws of proba-
bility. Maxwell showed that it is possible to
calculate the law according to which the
velocities of the molecules would be dis-
tributed at any temperature.* The curve
shown in Fig. 1 represents graphically the
distribution of velocities at any temperature,
in terms of the most probable velocity, whose
value is taken as unity. The significance of
this curve can be understood better by means
of the results tabulated in Table II. Under
* For a further discussion of the Distribution Law and other
aspects of the kinetic theory of gases the reader may be referred
to the following books and articles:
J. H. Jeans. The Dynamical Theory of Gases, 1916.
Meyer. Kmetic Theory of Oases.
K. Jellinek. Lehrbuch der PhysikaHschen Chemie.. I. 1, 1914.
W. C. McC. Lewis. Kinetic Theory of Oases. 1915.
S. Dushman. The Kinetic Theory of Gases. General Elec-
tric Review. Vol. US. I9I.i.
t K. Jellinek. loc. cit.. p. 222.
Ax is given the range of velocities and under
Ay the fraction of the total number of
molecules which have velocities corresponding
to this range. Thus 16.1 per cent of all the
molecules have velocities which range between
0.9 and 1.1 times the most probable velocity
at any temperature. Similarly it follows that
68. 4 per cent of the molecules have velocities
ranging between O.o and 1..3 times the most
probable velocity, while only 3.1 per cent
have velocities that exceed 2..i times the
most probable velocity.
TABLE lit
Ai
Ay
J- V
^v
0 -0.1
0.001
1.3-1.5
0.112
0.1-0.3
.021
1..5-1.7
.078
0.3-0.5
.063
1.7-1.9
.0.58
0.5-0.7
.112
1.9-2.1
.034
0.7-0.9
.149
2.1-2.5
.030
0.9-1.1
.161
2..5-3.0
.008
1.1-1.3
.1.50
0.5-1.5
.684
0 -2.5
.969
As shown in Fig. 1. the most probable
velocity (which may be denoted by M') is
different from the mean velocity. G. and the
relation between these two values of the
velocity is given by the
which can be readily
equation to the cun-c
tribution law:
following equation,
deduced from the
for Maxwell's dis-
W = ':^ c = J~~ = 1 2.900,
(.
Ci
In addition to these values of the velocity,
it is important, in connection with a large
class of applications of the kinetic theory of
gases, to know the arithmetical or average
velocity of the molecules at any temperature.
TABLE I t
MEAN VELOCITY X 10"* CM. SEC."*
Avrraae Velocity
at 2(f C. ■
Gas
A*
At OP C. j .^t 20° C.
n.
2.016
1 .838
1.904
1.755 XU>» cm. sec."'
0,
32.00
0.4613
0.4778
0.440
N,
28.02
.4928
.5106
.471
Air
28.96
.4849
.5023
.463
Hg
200.6
.1842
.1908
.176
CO2
44.0
.3933
.4076
.376
N2O
18.016
.6148
.6368
.587
A
39.88
.4133
.4282
.395
NH3
17.02
.6328
.6554
.604
CO
28.00
.4933
.510.)
.471
} These data are taken from the author's paper on the Kinetic Theory of Gases. loc. cit.
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
497
This is usually denoted by Si and
calculated bv means of the relation,
n = ^|S
'Air G =
\ Mtv
= 14,500
T
\M
be
(4)
The values of the average velocity at room
temperature for some of the more common
gases are given in the last column of Tabic I.
Number of Molecules Per Unit Volume
According to Avogadro's law the number of
molecules per gram-molecular weight of any
gas ought to be the same. The problem of
accurately determining the value of this con-
stant, which we shall denote by A'o, has
naturally been the object of a large number
of investigations, and a number of different
methods have been used in order to deter-
mine it. The phenomena of Brownian move-
ment, the accurate determination of the
charge on an electron, counting the number
of alpha particles expelled from a gram of
radium, and finally the study of the laws of
black body radiation — all these methods have
led to approximately the same value for A'o.
According to Millikan, whose determination
is undoubtedly the most accurate we have,
this constant has the value of 6.062 XIO"^
From this value it is readily calculated that
the number of molecules per cubic centimeter
of an ideal gas at a pressure of 10" bars and
0 deg. C, is 2.67X10'«.
Let us now attempt to interpret this
magnitude. The highest vacua attainable at
present range around 1G~* bar. Even at
this extremely low pressure, which would
ordinarily be regarded as a "perfect vacuum,"
the number of molecules per cm.^, at 0 deg.
C, is still 2,670,000.000, a number which is,
roughly speaking, of the same order of mag-
nitude as the total population of the earth.
Rate at Which Molecules Strike a Surface
It was shown by Meyer that the number of
molecules of a gas at rest as a whole that
strike unit area per unit time is equal to
34 M fi. where n denotes the number of
molecules per unit volume, and S2 denotes
the average velocity.
Substituting for n and 9. the values pre-
viously given, we obtain the relation,
N P 18
l/4«o=l/4^-
= 2.6.53X101
Rf
M
(5)
For air at 20 deg. C, and 10'' bars, the
number of molecules striking 1 cm.- per
second, is 2.S8><^(F.
Equation (5) may also be expressed in
terms of the mass (w) of gas that strikes 1
cm.- per second.
Let p denote the density of the gas.
Then,
iti= 1 4 w )H fi= 1 A p9.
MP ^ ... _ . P
0 = 43.74X10-" , (6)
ART
For air at 20 deg. C, and lO" bars,
w= 13.8 gm./ cm." sec.
As has been shown by Langmuir, equa-
tions (5) and (6) are extremely useful in the
consideration of rates of evaporation of metal
in vacua, and also in the study of the kinetics
of chemical reactions at low pressures.*
Mean Fiee Path of Molecules
While the individual gas molecules in a
gas at rest possess very high velocities, as
shown above, it is a matter of ordinary
observation that gases diffuse into each other
very slowly. This is explained on the kinetic
point of view by assuming that the molecules
do not travel continuously in straight lines,
but undergo frequent collisions. The use
of the term "collision" naturally leads to the
notion of <ree path. This may be defined as
the distance traversed by a molecule between
successive collisions. Since, manifestly, the
magnitude of this distance is a function of
the velocities of the molecules, we are further
led to use the expression "mean free path"
(denoted by L), which is defined as the
average distance traversed by all the mole-
cules between successive collisions.
Simple considerations show that the value
of L must vary inversely as the total cross-
sectional area of the molecules per unit
volume. Taking into account Maxwell's
distribution law and the fact that the mole-
cules exert attractive forces on each other,
it can be shown that L is given by the relation,
L =
1.402
yi irn d-
(>4)
(7)
* Phys. Rev., g. 329, 1913, also Jour. Am. Chem. Soc, 37, 1139.
1915.
where d denotes the molecular diameter, and
C is a constant for each gas (Sutherland's
498 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 6
constant) which is a function of the attrac-
tive forces between the molecules.*
The value of the molecular diameter may
be derived, as shown by van der Waals, from
the critical temperature and pressure of the
gas. This value of d can then be used to
calculate L by means of the above equation.
It is, however, more usual to calculate L from
the coefficient of viscosity, or heat conduc-
tivity of the gas. for it is evident that
whether it be transference of momentum
from one layer to another, as in viscosity, or
transference of increased kinetic energy, of
the molecules, as in heat conductivity, the
rate of transference must depend upon the
number of collisions which each molecule
experiences as it passes from point to point.
It should be observed in this connection that
in the case of air especially, and of most other
gases, the value of the coefficient of viscosity
(denoted by t]) has been determined with a
high degree of accuracy.
The general relation between L and rj is of
the form.
r] = k p G L,
(S)
where p denotes the density, and G the root-
mean square velocity. For approximate
calculations the value of k is ordinarily taken
as J-3. Boltzmann and Meyer have both
derived different values of k by taking into
account Maxwell's distribution law. Accord-
ing to the latter_thc relation is.
7j = 0.3097pO/. (9)
while Boltzmann derived the relation,
77 = 0.3502 pfi/,. (10)
From the above equations it follows that
the magnitude of I. varies inversely as the
pressure, that is. the lower the pressure, the
greater the value of the mean free path.
Table III. taken from the writer's i)aper on
the kinetic theory of gases, gives the values
of L at room temperature and standard
pressure, for different gases. Under these
conditions, the length of the free path for
most gases is about 10"'' cm. But at 1 bar
the value of L is as much as 10 cm., so that
* See the writer's articles on '"The Kinetic Theory of Gases."
General Electric Review, ts. 1042-48. IDl.'S. for a more
detailed discussion of the derivation of this equation and of the
following equations for the calculation of L. from the coefficient
of viscosity. Heydweiller, Ann. Phys.. .J2, 127:1. 101.'J. uses the
constant 1.319 instead of 1.402. as above, and combines this
with equation (9) below, for the calculation of the molecular
diameter.
t Dynamical Theory of Ga.ses, 191(i. p. IMl.
to. Sackur. Ann. Phys., iO. 97. 191.!.
% A. Heydweiller. Ann. Phys.. it. 127:i. Ull.t.
the molecules travel considerable distance
without suffering any collisions. We shall
show, in a subsequent section, how this con-
clusion is in splendid agreement with the
phenomena observed at low pressures.
TABLE III
L X 10« cm..
ColUsion-
Gas
at 20» C.
frcquency
and 10* bare
aiL X ia-«
Air
9.376
4940
H,
17.-14
10060
He
27.45
4.545
N,
9.287
.5072
o.
9.931
4432
A
9.879
3998
CO,
6.148
6115
CO
9.232
5101
NH,
6.60
9152
From the values of L and 12. the value of
the collision-frequency, j may be denved.
This number thus expresses the number of
collisions per molecule per second. The
values for some of the gases, at room tem-
perature and standard pressure, are given in
the last column of Table III.
Molecular Diameters
Using the above values of /.-. it is i)ossible,
from equation (7) to calculate the molecular
diameters for different gases. Owing, how-
ever, to the fact that different investigators
have used different values of the constant in
the numerator of equation (7), there is no
exact agreement with respect to the values of
d thus derived. Table IV gives a summary
of the values obtained by different methods
of calculation. Under I are given the values
calculated by the writer by means of equa-
tions (7) and (10). using the values of C as
deduced b\- Sutherland. Column II gives
the values assigned by Jeans as the mean
values derived from three different methods
of calculation. t Sackur has also attempted
to deduce the value of d by several different
methods +. and concludes that the most
probable values are those given in column
III. Ileydweiller's values. § obtained by
using equation (7) with the substitution of
l..'{19 for the constant 1.402. and combining
this with equation (9). are given in the last
column.
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
499
TABLE IV
MOLECULAR DIAMETERS (d X 10« cm.)
Gas
1 II
III
IV
Ho
2.403 2.68
1.90
2.176
He
1.905 2.16
1.77
N,
3.146
3.76
2.40
(h
2.975
3.62
2.30
A
2.876
3.64
2.68
CO
3.190
3.78
2.50
CO2
3.335
4.54
2.76
cu
5.36
3.30
3.693
Br,
3.74
I.
4.52
H2O
2.27
2.26
NHa
2.967
General Considerations Regarding Gases at Low
Pressures
As has already been stated, the pressures
which interest us in the study of high vacuum
phenomena usually range below 1 bar. At
these pressures the mean free paths of the
molecules are at least of the same order of
magnitude as the dimensions of the vessels
used in experimental work. Thus, at 1 bar
the mean free paths for most gases are about
10 cm. (Table III). It therefore follows that
the majority of the molecules travel in straight
lines as far as the dimensions of the vessels
will allow, and the number of inter-molecular
collisions per second becomes relatively small
as compared with the rate at which the
molecules strike the walls. The following
considerations will probably serve to explain
the significance of this statement more fully.
Consider a cube, whose volume is D^ cubic
cm., and let n denote the number of molecules
per cm.^. The number of collisions between
gas molecules, per second is
C--
The total number of molecules striking
the walls of the cube in each second is
0
A = Oi?= n —
4
Hence,
That is, the ratio
A L
CO
C D-
between the rate at which the molecules
strike the walls and the rate at which they
collide with each other is given by the ratio
between the lengths of the mean free path
* I. Langmuir. Am. Inst. Electr. Eng. Trans., also. Phvs. Rev.,
St 329 191,3
t S! Duihman. Phys. Rev., o. 223, 1915.
J L. Dunoyer. Les Idees Modemes sur la Constitution de la
Matiere. p. 2\o, 1913. This article contains a very interesting
discussion of low pressure phenomena, especially of Knudsen's
Work (see p. 500).
and of the side of the cube. It can be readily
shown that no matter what the shape of the
vessel, the ratio -=; is proportional to that
of y^, where D is the distance between the
walls. Thus, if D is of the order of magnitude
of 10 cm., A is greater than C when L is
greater than 10 cm., that is, when the pres-
sure is lower than 1 bar (approximately).
Consequently we should expect to find that
at pressures of 1 bar and lower, the molecules
travel in straight lines toward the walls of
the containing vessel.
A very common illustration of this fact is
the production of sharp shadows in vacuum
type incandescent lamps. As has been shown
by the investigations of Langmuir and
Mackay*, the blackening of ordinary tung-
sten lamps is due to the evaporation of metal
from the filament. The pressure in this type
of lamp under operating conditions is less
than 0.01 barf, so that the mean free path of
the tungsten atoms is of the order of several
hundred centimeters. Consequently, colli-
sions between these atoms and molecules of
residual gas are very rare, and the tungsten
atoms travel directly to the sides of the bulb,
where they are immediately condensed. By
interposing some object between the filament
and the walls, very sharp shadows can be
produced, if the vacuum is good. On the
other hand, the shadows are very much
blurred if there is present in the bulb a pres-
sure of even several bars of some inert gas
like argon. Similar phenomena are observed
in the evaporation of other metals, like
mercury and sodium. +
Laws of Molecular Flow
It is evident from the above considerations
that at very low pressures the rate of flow
of gases through tubes or narrow apertures
must be limited solely by the frequency with
which the molecules strike the walls of the
tube or aperture and may thus be thrown
back in the direction of incidence. At
higher pressures the rate of flow of gases
through narrow tubes is governed by Poiseu-
ille's law. If Qi denotes the amount of gas
(measured in terms of P. I') which flows
per second through a tube of diameter D
and length /, and 77 denotes the coefficient of
viscosity, Poisouille's law may be expressed
bv means of the equation,
_DHP.-P.)P (11)
500 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 6
where P is the pressure at which Q is measured
and Pi — Pi denotes the difference in pressure
at the two ends of the tube. At very low
pressures this relation is no longer valid, and
for a reason which is self-evident. At ordi-
nary pressures the rate of flow of gases must
be limited by the frequency of collisions
between molecules, hence the necessity for
introducing the coefficient of viscosity in the
formula for the rate of flow. At very low
pressures, however, where the length of L,
the mean free path, is much greater than that
of D, it is meaningless to speak of a coeffi-
cient of viscosity and it is therefore neces-
sary to discard the hydrodynamical equations
upon which Poiseuille's relation is based, in
order to arrive at a more accurate relation
for the rate of flow of gases through tubes.
A similar difference has been observed for
the laws of heat flow in gases at low and high
pressures. For pointing out the manner of
attacking both these problems and deducing
a number of relations which are applicable to
gases at low pressures, we are indebted to the
theoretical and experimental investigations
of M. Smoluchowsky, M. Knudscn, and W.
Gaede, who, since 190<S, have pubHshcd a large
number of papers dealing with this subject.
The term "molecular flow" was suggested
by Knudsen to designate the condition of
gases flowing through tubes at such low
pressures that collisions between the mole-
cules are infrequent as compared with col-
lisions at the walls. As has been shown above.
at these pressures L is much greater than D
and the ratio y increases with decrease in
pressure, so that any molecule striking the
inner surface of the tube at any point is
repelled all the way across the tube until
it strikes the opj^ositc wall. Knudsen now
assumes that any plane surface, no matter
how smooth it may a])pcar, consists in reality
of teethlike projections which are probably
due to one or more atoms being irregularly
piled up above the surrounding atoms; that
is, these projections are of molecular dimen-
sions, and they are irregularly distributed
over the surface. Consequently, "a gas
molecule on striking the surface is repelled in
a direction which is totally indejjendent of
the direction of incidence, and the distri-
bution of directions of an infinitely large
number of molecules after reflection from a
surface follows Lambert's cosine law for the
reflection of light from a glowing body."
« Ann^^r Phys.. 2S, 7.5, 1908. and 2S. 999, 1908. Also XI. L.
Dunoyer, loc. cit.
Introducing Maxwell's distribution law and
Meyer's equation for the number of mole-
cules in a gas at rest that strike unit area,
Knudsen arrives at the following relations
for the case in which the diameter of the tube
or aperture is infinitcsimally small as com-
pared with the length of the mean free path.*
In the case of a circular tube of diameter
D, and length /, the quantity of gas, 0«, which
flows through per second, with a difference of
pressure P2 — P1, is given by the equation.
where
(32 =
H'i =
P2-P1
&l
2.394 /
V2ir D' D^
(12)
(13)
and pi denotes the density at 1 bar pressure
and the temperature of the tube.
From the gas laws it follows that,
Pi =
M
83.15X10' T
It will be observed that equation (12) is
analogous to Ohm's law, so that we may
speak of the term ll'i \/'p[ as the resistance
to flow of the tube at the temperature T for
a given gas. For different gases, the value
of the resistance varies as the square root of
M.
For the case of a circular opening in a thin
plate, equation (12) is still valid, but the
value of Il's, the "resistance" is given by
the equation.
W - l27r_3.184
(14)
where .4 is the area of the opening, and D
its diameter.
Ilencc, where wc have a tube of diameter
11 and length / connecting two vessels at low
l)rcssures. the total resistance to flow of this
tubing for a gas of unit density, is
2.394 / 3.1X4
£>»
D'
(15)
By means of equations (12) and (15) it is
possible to calculate the quantity of gas that
can flow through any given tube or opening
at low |)ressurcs. The value of 0 is obtained
in terms of P W that is, the volume in cm.'
at a given jiressure P, in bars. As an illus-
tration of the application of the above equa-
tion. Table V gives the volumes (in cubic cm.)
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
501
TABLE V
/
D
w
<3j (Air)
Qj (Hyd.)
1
10
1
10
1 .
1
0.1
0.1
5.578
27.124
2712.4
24258.4
5204.
1070.
10.70
1.196
19710.
4053.
40.53
3.60
of air or hydrogen (at 1 bar pressure) that
would flow through different sizes of tubing
for a difference of pressure of 1 bar, and room
temperature. For air at 293 deg. abs, and 1
bar pressure, p = 1.189x10"^ and for hydrogen,
under the same conditions, p = S.271 X 1()~".
From equation (15) and the data in Table V
it is evident that for long tubes of very small
diameter (capillaries) the end correction is
negligible. The values of Qo for air and
hydrogen may then be derived from the data
in the table for / = 10 and D = 0.1 by apply-
ing equation (13).
These examples illustrate the effect of
narrow tubes on the rate of exhaust at low
pressures, and it is therefore absoluteh'
essential that in experiments at low pressures
where maximum speed of exhaust is desired,
the connecting tubing should be as large in
diameter as practicable, and also as short as
possible.
Laws of Flow at Higher Pressures
The equations given above are strictly
accurate only at such low pressures that
J- is infinitesimally small. Actually it
has been found by Knudsen that the equa-
tions are accurate to within 5 per cent even
at pressures where -j- = OA. For air at
room temperature and 1 megabar, the value
of L is 9.4X10'^ cm., and at a pressure P,
9 4
L = '-^. So that in case of a tube 1 cm.
in diameter, the equation for molecular flow
would be accurate to within 5 per cent for
all pressures below about 3.76 bars.
It is of interest in this connection to discuss
briefly the manner in which the rate of flow of
gas through a tube varies at higher pressures.
If we denote the ratio, 7^5 — ^^-r- bv F,
it is evident from the above discussion that
for very low pressures this ratio is constant
and independent of the pressure. As, how-
ever, the pressure is increased the value of F
is observed to decrease at first until it reaches
a minimum value which is about 0.95 of its
value at very low pressures. As the pressure
is increased still further, F increases and the
rate of increase with pressure is given by
Poiseuille's law. From experiments over a
large range of pressures with different gases,
Knudsen has derived the following semi-
empirical relation which is found to hold at
all pressures:
(1+CiP)
F = aP-\-b
where.
Z)<
12s -ql
{l+c,P)
(Poiseuille's constant)
(16)
/
— ,-— ^ (Coefficient of molecular flow)
H \/pi
\/plD .
Ci — — and C-)
1.24VP1 D
For ordinary pressures this equation as-
sumes the form already given for Poiseuille's
law, equation (11), while at very low pres-
sures it becomes identical with equation (12).
In order to illustrate the application of
equation (16) and also show the effect of
pressure on the rate of flow of gases it is of
interest to calculate by means of this equation
the value of F at different pressures for air
flowing through a tube 10 cm. long and 1 cm.
in diameter, at room temperature. In Table
VI, F expresses the volume in cubic cm.,
measured at 1 bar pressure and room tem-
perature that flows through the tube for a
difference of pressure of 1 bar at the ends and
an average pressure of P bars.
TABLE VI
n =135-6. 6=1070
£1 =0.19033
c: =0.2360
P (bars)
F (equation 16)
F-1070
106
13.56 X
10«
13.56 X 10«
100
2227
1157
50
1555
485
20
1160
90
10
.10.58.1
-11.9
5
1025.7
—44.3
4
1023.6
-^6.4
3
1025.2
—44.8
1
1043.6
—26.4
0.1
1065.4
- 4.6
0.01
1069.6
— 0.4
These results have been plotted in Fig. 2.
It is seen that the minimum value in F occurs
at about 4.5 bars. Even at this pressure the
dift'erence between the value calculated by
means of equation (16) and that calculated
502 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
by applying the simple equation (12) com-
bined with equation (15) is less than 5 per
cent of the value, F = 1070, calculated by the
last mentioned method. Table VI also shows
that the resistance of tubes is very much
greater at extremely low pressures than at
ordinary pressures.
1072 ■
1068 [
ioe4^
/060-
1056 -
J052-
/048-
1014
1040
/036
i03Z-
/028-
I0Z4-
5
Fig. 2
Thermal Molecular Flow
In experiments on gases at low pressures
it is often the practice to keep different parts
of a system at different temperatures. A
usual case is where an appendix or trap con-
nected with the vessel to be exhausted, is
kept immersed in liquid air, while the pressure
in the system is read bv some form of sensitive
gauge.
If the pressure is so low that ^ for
the connecting tube is very high, gas is
observ^ed to flow jrom the- colder to the hotter
parts until a sufficient pressure is developed
to check it. The condition of equilibrium of
pressures in the two parts of the system is
then given by the relation,*
As the pressure increases, the amount of
flow from the colder to the hotter parts
gradually decreases to zero and then reverses,
so that at ordinary pressures the condition of
equilibrium is
P, = P, or ^ =
|P2
Equation (17) is immediately applicable to
the case mentioned above where a trap con-
nected to an exhaust system is immersed in
liquid air (7' = .SS approx.). The pressure in
the latteristhen-vl;j^ = 0.55 of that in the rest
of the system.
• KnudKpn. Ann. Phys.. SI. 20.5 and 33. H.^i. 1910.
G. D. West. Proc. Phys. Soc. (Lond.l :»1. 27.S (1919). This
paper givi's a critical discussion of the laws of thermal trans-
piration over the whole range of pressure.";.
(To bf Cvnliriufd\
♦
503
Fundamental Phenomena in Electron Tubes
Having Tungsten Cathodes
Part I
By Irving Langmuir
Research Laboratory, General Electric Company
As a result of the pressing need for electron tubes during trie war, the development of these devices has
proceeded rapidly in the last few years. The phenomena upon which the operation of these tubes depends
are so diflferent from those of most electrical devices that confusion has frequently arisen as to the interpre-
tation of their various characteristics. The author discusses the fundamental factors, such as the electron
emission from the filament, the effect of space charge, the disturbances caused by the current passing through
the filament, etc., and endeavors to clear up many of the mysterious effects that have been observed. This
article was read as a paper last November at a symposium on electron tubes held in Chicago by the American
Physical Society. — Editor.
In a paper* published in 1913, it was shown
that in a two-electrode thermionic device
having only a negligible gas residue, and
operating with relatively large currents, the
characteristics consists essentially of two
parts.
In one part of the characteristic the
current is practically independent of the
applied voltage, but increases rapidly if the
tempjrature of the filament is raised. This
part of the characteristic we will refer to as
the "saturation region." The current is
primarily determined by the electron emission
from the cathode.
In the second part of the characteristic the
current increases with the applied voltage,
usually about in proportion to the three-
halves power of the voltage, but the current
is practically not affected by a change in
filament temperature. This part of the
characteristic we will refer to as the "space
charge region." Under these conditions the
current is limited primarily by the electro-
static field of the electrons in the space
between the electrodes.
In discussing the fundamental phenomena
in electron tubes, we must keep clearly in
mind the distinct nature of each of the two
factors just mentioned. It will therefore be
desirable to discuss these factors separately
and later consider how they may co-operate
to determine the characteristics of a given
device.
* Langmuir: Phys. Rev. 2, 450 (19L3). Two other papers
giving some new data and a clearer discussion of the theory were
soon afterwards published in the Physikalische Zeitschrift
Vol. 1.5. pages 348 and .516 (1914). A review of the history of
these theories and a discussion of their application to electron
tubes for use in radio work were published the following year.
General Electric Review. 18, May. 1915. and Proc. Inst.
Radio Engs. 3. 261 (1915).
t The temperature of filaments are expressed on the absolute
or Kelvin scale as denoted by the symbol °K. The method of
determining the temperatures from the characteristics of the
filaments have been published: Langmuir. Phys. Rev. 7, .'i02.
(1916) and Gener.\l Electric Review 19. 208 (1916).
The fundamental phenomena underlying
the two different parts of the characteristics
are (1) the electron emission from the cathode,
and (2) the space charge between the elec-
trodes.
Electron Emission. When a metal is heated
to high temperature in an extremely high
vacuum, electrons are emitted from its
surface. These electrons are emitted with
certain initial velocities, depending on the
temperature of the heated metal or cathode.
It has been shown by Richardson and others
that these initial velocities depend only on
the temperature of the cathode and not on the
material of which it is constituted. All the
electrons emitted do not have the same
velocities — the velocities of the individual
electrons are distributed around an average
value according to the laws of probability.
In this particular case the distribution of
velocities is expressed by a law derived by
Maxwell and usually known as Maxwell's
Distribution Law. The average kinetic
energy of the emitted electrons has been
found to have the same value as that of gas
molecules in a gas at the same temperature
as the cathode, and the distribution of the
individual velocities around this mean value
is also the same.
Although the actual average velocity of
emitted electrons is very high when expressed
in ordinary' units, such as miles per second,
the effects produced by these velocities are
strikingly small. Because of the very large
electric charges on the electrons and their
small masses, these initial velocities do not
enable the electrons to move against anything
but small retarding potentials. For example,
the average kinetic energy of the electrons
emitted from a cathode heated to 2400° K.f
is onlv sufficient to allow the electrons to
504 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
move against a retarding potential of 0.31
volts; with a filament at 1200° K. the average
velocity corresponds to 0.15 volts. To
illustrate the meaning of Maxwell's Distri-
bution Law applied to the case of a filament
at 2400° K. the following figures are given:
90 per cent of the emitted electrons are capa-
ble of moving against 0.022 volts; 75 per cent
can move against 0.059; 50 per cent against
0.143; 25 per cent against 0.29; 10 per cent
against 0.48; 1 per cent against 0.95; while
only one out of a thousand can move against
1.42, one out of a million against 2. So and only
one out of a billion against 4.27 volts. The
higher the temperature of the filament the
greater the voltages against which the
electrons can move; in fact, this voltage
increases directly in proportion to the absolute
temperature of the filament.
The number of electrons emitted from a
given cathode in high vacuum depends to a
very marked extent on the nature of the
material constituting the cathode, on the
condition of its surface and on the tem-
perature. The way in which the electron
emission varies with the temperature is
usually given with satisfactory accuracy by
an equation which was derived in 1901 by
Richardson. This equation is
i = as/Te-r W
Here i is the current emitted per unit area
from the cathode. In other words, i is
proportional to the number of electrons
emitted per unit area per second; T is the
absolute temperature of the cathode, a and b
are constants depending on the nature and
condition of the surface of the cathode, and
e is the base of the natural system of loga-
rithms, which is 2.7 IN.
In deriving this equation it was assumed
that the number of electrons per unit volume
in the metal remains substantially constant
while the temperature has been raised, and
that the external electric field produced by a
positive anode is without effect in drawing
electrons out of the metal. In order that the
current obtained in any actual device may
correspond with the above equation it is
necessary that the conditions be so chosen
that all of the electrons which are emitted
are able to flow to the anode. If this is the
case and if the external field does not cause
an increase in electron emission, then the
current passing through the tube is inde-
pendent of the voltage applied to the anode
and the current is said to be saturated and is
usuallv referred to as the saturation current.
For the case of a filament of pure tungsten
in a very high vacuum the electron emission
is given bv the equation :
52500
i = 23.6X10Vre — T- (1)
where / is expressed in amperes per sq. cm.
From this equation the values of i given in
Table I are calculated.
TABLE I
ELECTRON EMISSION FROM *
PURE TUNGSTEN
Absolute
Amperes
Absolute
Amperes
Temp.
per Sq. Cm.
Temp.
per Sq. Cm.
1500
0.58 XIQ-'
2300
0.1377
1600
5.42X10-«
2400
0.365
17(1(1
37.8 X10-*
2500
0.891
1800
214 X10-«
2600
2.044
1900
0.00103
2700
4.35
2000
0.0042
2800
8.33
2100
0.0151
2900
17.1
2200
0.0483
3000
31.7
Although it is frequently possible under
experimental conditions to obtain saturation
currents which remain constant over wide
ranges of voltage, there are two factors which,
even in the complete absence of gas effects,
may cause the current to increase with the
voltage over that part of the volt-ampere
cur\'e where ordinarily the saturation current
is to be expected. The first of these effects
has been experimentally found by several
observers and particularly by Schottky (Phys.
Zcit. 15, 872, 1914). Schottky has given both
theoretical and experimental reasons for
believing that with potential gradients of the
order of magnitude of a million volts per
centimeter, the electron emission from a
metal can be vcr>- greatly increased because
of an actual pulling of the electrons out of
the metal by these fields. Although fields of
this order of magnitude can be rather easily
obtained experimentally, most practical
devices utilizing thermionic currents do not
have electric fields around their cathodes
sufficient materially to increase the current
in this way.
The second effect is caused by hetero-
geneity in the surface of the cathode.
When a metal which gives a high electron
emission is in electric contact, with another
metal giving smaller electron emission, there
is a contact difference of potential between
these metals, and there is an electric field
produced in the space between the surfaces of
these metals. This contact difference of poten-
tial has been the subject of discussion for
nearlv a hundred vears.but within more recent
ELECTRON TUBES HAVING TUNGSTEN CATHODES
505
' years, through the work of Richardson and
others, has assumed increased importance.*
According to this theory a small surface of
very high electron emission (which we will
refer to as an active surface) will have a
positive potential with respect to surrounding
areas having lower electron emission (inactive
surface). If the active surfaces are small in
extent compared to the inactive ones the
electric field close to the surface of the
cathode will be largely determined by the
negative field of the inactive areas. Since the
electrons escaping from the active areas must
pass through these negative fields, the
effective electron emission may be greatly
cut down unless an external field is applied
sufficient to counteract or neutralize the
negative field produced by the inactive areas.
A mathematical analysisf shows that when
the sizes of the active and inactive areas are of
molecular dimensions or rather are of the
dimensions which one might expect by a
random distribution of active molecules over
the surface, very large external fields are
necessary in order to get saturation current.
When to this effect of active and inactive
areas we add the geometrical surface irregu-
larities such as small elevations and depres-
sions due to crystalline structure, etc., it is
apparent that conditions should be expected
to arise in which the volt-ampere character-
istics may increase with the applied voltage
in a complicated way over the range usually
called the saturation region.
Among the numerous experiments which
have confirmed these theoretical conclusions.
I will mention only one in detail. A thermionic
device having a tungsten filament containing
a trace of thorium and made up in such a way
that a particularly high vacuum is inaintained.
can be treated so that the filament acts in one
case like a pure tungsten filament, and in a
second case like a pure thorium filament (as
far as electron emission is concerned), while
in a third case a fraction of the surface of the
cathode is covered with thorium atoms so
that the surface is not homogeneous. These
changes in the condition of the cathode can
be brought about at will merely by heating
the cathode at a series of different tem-
peratures in the highest vacuum. For example,
if the filament is heated for a short time to
1900° K., thorium diffuses from the inside
of the filament to the surface and gradually
completely covers the surface of the filament
* For a general discussion of contact potentials and for
references to the earlier literature see Langmuir, Trans. .\mer.
Electrochem. Soc. 29. 125 (1916).
t This work will probably be published within the ne.\t year.
t Annalen der Physik JiT. 573 (1915).
with a layer of thorium. On the other hand,
when the filament is heated for a few minutes
at 2S00 or 2900° K., all the thorium distills
off the filament, leaving a surface of pure
tungsten. If, however, the thorium be dis-
tilled from the filament at a lower tem-
perature or for a shorter time, it is possible to
leave the surface covered partly with tungsten
and partly with thorium.
When the surface is entirely covered with
thorium the electron emission at a given
temperature is many thousands of times
greater than that from the pure tungsten
surface. By lowering the temperature of the
thorium covered filament it is possible to get
the same emission in the two cases. It is
then found that under both of these conditions
a very definite saturation current is obtained.
That is, there is a wide range of voltage over
which the current remains practically constant.
On the other hand, if the surface of the
cathode is made heterogeneous by having
both tungsten and thorium present on the
surface it is found that no well defined
saturation current is obtained, but the
current gradually increases as the voltage is
raised. The volt-ampere characteristic is
very markedly different from that which is
found in either of the two previous cases.
This test is best made by adjusting the
filament temperature so that the current
that flows with an anode voltage of 200 is the
same in each case. These experiments can
only be made in an exceptionally high
vacuum because even slight traces of gases
such as water vapor, or oxygen entirely
destroy the activity of a surface of thorium
and slight traces of positive ionization
produce a disintegration of the surface to
such an extent that the minute traces of
thorium are removed from the surface. It is
clear from these experiments that the failure
to reach saturation is no indication whatever
of the condition of the vacuum within the
tube. With one and the same tube, without
change in vacuum conditions, the volt-ampere
characteristics can either be made flat, giving
a good saturation current, or be made to
curve continually upward even at high
voltages so as to give little indication of a
definite saturation value. Other experiments
seem in a general way to indicate that
Wehnelt cathodes as ordinarily made are far
from homogeneous, and that even in the
highest vacuum it is difficult to get a well
defined saturation current.
This effect also explains the results obtained
by Schlichtert in which he found that with a
506 June. 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
contaminated platinum surface no definite
saturation was obtained at a hundred volts,
while with a clean surface saturation was
reached at five volts or less.
In determining the saturation current in a
particular device it is usually necessan- to
take great care that the temperature of the
cathode remains constant. The electron
emission ordinarily increases so extremely
rapidh- with the temperature that a ver\^
few degrees change in temperature of the
cathode will cause a relatively large change
in the electron emission, and therefore in the
saturation current. In cases where fairly
large currents are made to pass through a
thermionic device and fairly high voltages
are used, the heat generated at the anode and
radiated from it may be sufficient to cause
changes in the temperature of the cathode,
and in this way, unless extremely sensitive
methods of determining and checking the
cathode temperature are employed, it may
happen that this cause produces an apparent
increase in the saturation current when the
voltages applied to the anode are increased.
An efTect of this kind is apt to be particularly
important in those cases where the cathode
operates at a low temperature, as for example
where a Wehnclt cathode is used. Either by
careful control of the cathode temperature or
by making measurements of the electron emis-
sion by momentary application of voltage to
the anode, it is possible to distinguish between
an apparent increase in saturation current
due to this heating effect and a real increase
in electron emission due to heterogeneity
of the filament surface or to other causes.
Very minute traces of impurities in the
filament or on its surface may cause great
changes in electron emission, and as the
condition of the surface may change during
the heating of the filament it may happen
* This equation was independently derived by the writer
(Phys. Rev. g. 450 (191.3]) and applied to the case of conduction
bv electrons.
Lilicnfeld (Phys. Rev. S. .364. |iyi411 claims to have found
the law that the current increased with the 3 2 power of the
voltage in some of his work published in 1910. A careful study of
Lilicnfeld 's data shows, however, that in his experiments the
current did not even appro.ximately vary with the 3.2 power of
the voltage. The original data upon which Lilicnfeld liases his
claim are those given on page 69S of his 1910 article (.\nnalen
der Physik. Vol. 32). It there appears that no current flowed until
the voltage between the sounding electrodes was 102 volts and
when the voltage increased from 102 to 116 volts there was a
17-fold increase in current.
The current thus increased with the 23nd power of the
voltage instead of with the 3 2 power. At higher voltages the
rate of increase became gradually less, but over the range in which
Lilicnfeld claims to have found the relation, the 3-2 power law
is not even approximately fulfilled. What Lilicnfeld did find w.as
the purely empirical relation that, beginning from 102 volts, in
his device the current increased in proportion to the 3 4 power
of the quantity 1'' — l'..* where V is the difference of potential
between two sounding electrodes and l',. is lO.i volts. At the
higher voltages before 1' became very larfje compared to V*.
Lilienfeld found tliat even the above empirical relation did not
hold.
that the saturation current changes markedly
with the time so that irregularities in the
volt-ampere characteristics may result.
Space Charge. When a positive potential is
placed on an anode, in proximity to a heated
cathode in a high vacuum, and the filament is
heated to a low temperature so that it emits
relatively only few electrons, the velocity of the
electrons increases steadily as they move
between cathode and anode. However, with a
given voltage on the anode the velocity which
the electrons acquire is perfectly finite so that
the electrons take a certain time to pass across
this space. When the temperature of the
cathode is raised and the electron emission
increases, the number of electrons in the
space between cathode and anode at any
given time increases at first in proportion to
the electron emission. Now these electrons
in the space tend to repel those which are
leaving the cathode, and it is clear that as we
increase the total number of electrons carr\-ing
the current, a point must ultimately be
reached at which the repulsive force caused
by the electrons in the space will be sufficient
to neutralize the attractive force caused by
the positive potential on the anode. Under
these conditions the current still flows to the
anode because of the initial velocities of the
electrons, but if the current should then
increase enough so that the repulsion of the
electrons in the space is able to exceed the
attractive force due to the anode sufficiently
to counteract the effect of the initial velocities,
then any additional electrons emitted by the
cathode will be forced to return to the
cathode. The current flowing from cathode
to anode will thus reach a definite limit at
any given voltage on the anode. In other
words, the space between cathode and anode
has a finite current-carrying capacity for a
given anode voltage. No mere increase in
electron emission from the cathode can
cause the current through the device to
increase beyond that set by this limitation
to the current.
The quantitative theor>- of the volt-ampere
characteristics of an electric discharge in
which the current is carried by ions of only
one sign was worked out by Qhild {Physical
Revieu\ 32, 4!i2 |liUl]). For the case of a
discharge occurring between two parallel
plane electrodes Child obtained the equation:*
Ott > m X-
Here i is the current flowing between the
electrodes per .square centimeter of surface;
ELECTRON TUBES HAVING TUNGSTEN CATHODES
507
e is the charge on the ion; w is the mass of
the ion; \' is the difference of potential
between the anode and cathode; and x is the
distance between these electrodes. Child
derived and used this equation in connection
with a discussion of the maximum current
that could be carried by positive ions in a
gaseous discharge.
When equation (2) is applied to the case of
a discharge carried by electrons only, the value
p
of is very much larger than in the case of
m
discharge carried by positive ions. If we take
the value of this ratio as found for the elec-
trons and substitute in the equation and
adopt as our units the volt, ampere and
centimeter the equation becomes
1/3/2
i = 2.33X10-«^^ (3)
In this equation, i represents the current-
carrying capacity of the space between the
electrodes in amperes per square centimeter
when the difference of potential V is applied
to the electrodes, and the distance between
the electrodes is A" centimeters.
For the entirely analogous case of a pure
electron discharge from a straight cylindrical
filament to a co-axial cylindrical anode, I
have derived the equation
9 \m r
where i is the current per centimeter of
length along the axis and r is the radius of the
cylindrical anode.
If the units are expressed in volts, amperes
and centimeters this equation becomes
y3/2
1 = 14.65X10-" (5)
r
This gives the maximum current-carrying
capacity of the space between the cathode and
anode in amperes per centimeter of length for
the case of a small heated wire in the axis of a
cylindrical anode of radius r centimeters,
having a positive potential of Y volts with
respect to the cathode.
Assumptions Underlying the "Space Charge Equa-
tions"
In deriving the above equations. Child,
Langmuir, Schottky, and presumably Arnold,
made two fundamental assumptions; first,
that the initial velocities of the electrons had a
negligible effect under the conditions to which
the equations were to be applied, and secondly,
that the temperature of the cathode and
therefore its electron emission were so high
that a further increase in temperature would
not cause an increase in the current. This
second assumption is equivalent to assuming
that the filament is at such a temperature
that it emits a surplus of electrons. It is
necessary to make some such assumption
in order to bring into the mathematical
equations the fact that we wish to consider
the space charge limitation of current instead
of that due to the electron emission from
the cathode.
We have already seen that the average
initial velocity of the electrons emitted from
the filament at 2400° K. is only sufficient
to cause electrons to move against a negative
potential of 0.31 volts. Since the space charge
equation is ordinarily used in connection
with thermionic devices, in which voltages of
from 10 volts up to many thousands of volts
may be used, it is clear that the initial
velocities are very small compared to those
produced by the applied voltages. In general,
therefore, it would seem that we are justified
in neglecting these initial velocities. Of
course, what actually hapjjens is that, when
the electron emission from the cathode exceeds
the current-carrying capacity of the space,
some of the electrons begin to return to the
cathode. When this occurs there is at the
surface of the cathode an opposing electric
field, and the potential in the space at a short
distance from the cathode surface is negative
with respect to the cathode itself. The
electrons which return to the cathode are
naturally those which were emitted from it
with the lowest velocities. There is thus
directly in front of the cathode a place where
the potential is a minimum and it is at this
place that the potential gradient is zero.
In the derivation of the space charge equa-
tions the assumption of negligible initial
velocities introduces an error which is roughly
proportional to the ratio of the few tenths of
a volt to the applied anode potential. The
assumption that the potential gradient at the
surface of the cathode is zero is strictly only
permissible when we neglect the initial
velocities. If the initial velocities were
actually zero, then there is no question that
the potential gradient at the surface of the
cathode would be zero when the current is
limited by the current-carrying capacity of
the space. If we assume that this zero
potential gradient does not exist at the
surface of the cathode, but exists at the point
of minimum potential at a distance of a
thousandth of an inch or so from the cathode,
then the assumption is entirely permissible.
508 June, 1920
GEXER.\L ELECTRIC REVIEW
Vol. XXIII. Xo. 6
It is also apparent that this assumption of
zero potential gradient taken together with
an assumed zero initial velocity, is merely
equivalent to stating that a surplus of
electrons is emitted at the cathode, and it is
essential to make this assumption if we wish
to have the resulting equation give us the
maximum current-carr\'ing capacity of the
space.
Of course, there are various other tacit
'assumptions which are made in the deriva-
tion of the space charge equations (2) to (5),
such as, for example, that the whole of the
cathode is at one potential, that the electrons
do not lose energy by collisions with gas mole-
cules, that there is no magnetic field which
interferes with the free motion of the elec-
trons, etc.
The space charge equation (3) giving the
current between parallel plane electrodes is
not of a form which can be readily tested
experimentally because it is difficult to work
with a plane cathode surface of large area.*
The fact, however, that this equation indi-
cates that the current when limited by space
charge should increase in proportion to the
three-halves power of the voltage is the most
significant feature.
The equation (5) dealing with the case of a
straight cathode wire in the axis of a cylindri-
cal anode is one which can be tested experi-
mentally with high accuracy. A large number
of experiments by Dushmant and by Schottlcy
have shown that this equation holds with a
verv' satis''actory degree of accuracy, the
experiments being made of course under
conditions of very high vacuum. This
agreement is not only accurate in regard to
the current increasing in proportion to the
three-halves power of the voltage over a wide
range, but also in regard to the actual
numerical values of the current obtained at a
given voltage.
By a general method of reasoning published
in 1913, J it was concluded that the three-
halves power relation derived directly for the
case of the parallel plane electrodes and the
cylindrical electrodes should also hold for
electrodes of any shape, provided that every
part of the cathode surface is heated to a
temperature high enough so that a surplus of
* Germerhausen (Physik. Zeitsch. /S.^ICM |l91.il) applied
Equation Ci) to electron current between .1 plane Wehnelt
cathode and a parallel anode and obtained Rood agreement.
t Dushman, Phvs. Rev. i. 121 (1914).
Schottky. Physik. Zeitsch. IS. 624 (1914).
t Langmuir: Phys. Rev. .'. 4o9 (1913).
II See footnote on page 87S of Schottky's subsequent paper
(Physik. Zeitsch. IS. 872, |1914|).
electrons is emitted, and that the boundaries
of the space through which the discharge
takes place are either at the potential of the
cathode or at the potential of the anode. In
the argument given in support of this relation
it was not made clear that the rigorous deriva-
tion depends upon the assumption of negligible
initial velocities and also upon the fact that the
paths which the electrons take under these con-
ditions remain the same if the potentials of the
bounding surfaces are all increased in the same
ratio. Schottky (Phys. Zeitsch. 15. 520 [1914])
criticised this conclusion and maintained that
the 3/2 power law should hold only for those
cases in which the electrons travel in straight
lines from the cathode to the anode. As a
result of correspondence, however. Schottky
recognized that the paths of the electrons
remain unchanged when the voltage is raised,
and therefore admitted'! the general validity
of the 3 2 power law for all cases where the
effects of initial velocities were negligible.
By far the best indication, however, that
the three-halves power relation does hold for
electrodes of even complicated shapes is the
experimental proof that it holds over wide
ranges of voltages in actual thermionic devices
in which the filaments and even the anodes
are in the form of wires twisted into a great
number of irregular shapes. In fact, the
experiments seem to show that except under
unusual conditions the relation holds ver\-
nearly as well as for the case of concentric
cylinders. This experimental proof justifies
the assumptions made in the derivation of the
general equation. In devices with three
electrodes it is usually found that the three-
halves power relation holds satisfactorily
either when both cold electrodes arc connected
together as anode or when one is used as
anode and the other is connected to the
cathode. In this latter case, however, the
range over which the relation holds is usually
more restricted. The reason for this will be
discussed below.
Of course, it is only to be expected that the
space charge equations will apply only when
the electron emission from the cathode is so
great that a surplus of electrons is emitted
from every part of the cathode surface. Even
under these conditions there are various
factors which may cause deviation from the
three-halves power law or which may cause
the currcnt-carrx'ing capacity of the space to
vary, even without change in the anode
potential.
We will discuss these various factors.
ELECTRON TUBES HAYING TUNGSTEN CATHODES
509
Effects Due to Initial Velocities of the Electrons
I have already discussed how with the
parallel plane electrodes the initial velocities
of the electrons produce a small deviation
from the three-halves power relation when
the voltage applied to the anode is not very
large compared to the voltages of a few tenths
of a volt corresponding to the initial velocities.
Of course when voltages below two or three
volts are used on the anode these deviations
from the three-halves power law become
relatively large. In the case of coaxial
cylindrical electrodes the electric field around
the cathode is particularly strong close to the
surface of the cathode, and the effects due to
initial velocities are thus less important than
in the corresponding case of the parallel plane
electrodes. Still with anode voltages of less
than five or ten volts, the initial velocities
cause quite perceptible, and at very low
voltages, relatively large deviations. In the
general case of electrodes of any shape
whatever it is not always true that the initial
velocities have the same relatively small
effect as in the two simple cases of the plane
and cylindrical electrodes. Wherever the
design of the apparatus is such that there is a
strong electric field close to the surface of the
cathode the deviations from the three-halves
power relation are not greatly different from
those in the case of the concentric cylindrical
electrodes. Thus, the cathode instead of being
straight can be bent into complicated shapes,
such as a rather open helix or a V, U or W
shape. The anode also may have various
shapes of this kind, and yet in all these cases
the three-halves power ratio may be found
to hold over a very wide range of voltages
and even hold fairly accurately at voltages as
low as five volts. That is, the current at five
volts may differ only by a small percentage
from the current calculated by extrapolating
from the volt-ampere curve at higher voltages
according to the three-halves power law.
In those cases, however, where the appara-
tus is so designed that the electric field close
to the surface of the cathode is made abnor-
mally small by the presence of an auxiliary
electrode or the walls of the vessel, the effect
of the initial velocities of the electrons is
exaggerated so that marked deviations from
the three-halves power law may hold at
comparatively high voltages. Such an effect,
for example, is obser\'ed when a standard
Coolidgc X-ray tube, having a focusing shield
nearly surrounding the cathode, is operated
at potentials of only a few hundred volts,
that is, under conditions far removed from
those at which it ordinarily operates. The
presence of this focusing screen or shield,
which is connected to one end of the cathode
filament, makes the electric field close to the
surface of the cathode very small indeed
compared with what it would be if the shield
were not present. Under these conditions the
number of electrons which escape from the
cathode and pass to the anode depends very
greatly on the initial velocities of the electrons.
Thus, with voltages on the anode much less
than those needed to give saturation, the
current flowing through the tube is found to
depend to a considerable degree on the
temperature of the cathode; in other words,
on the initial velocities. Furthermore, the
current is found to increase with the voltage
considerably more slowly than corresponds to
the three-halves jjower law even at voltages
of several hundred volts. However, as the
voltage is raised the slope of the volt-ampere
characteristic gradually approaches closer to
that corresponding to the three-halves power
law, so that at 2000 volts there is reasonably
good agreement with this law.
Another case in which the effect of initial
velocities may be abnormally increased is that
in which the grid of a pliotron is connected
to the negative end of the filament, while
various positive potentials are applied to the
plate or anode. The shielding action of the
grid has an effect similar to that of the
focusing shield in the Coolidgc tube, although
the effect is usually not nearly so marked.
Under normal conditions it is usually found
that the current flowing to the anode follows
the three-halves power law quite satisfactorily
under these conditions over a wide range of
voltage, but that as the anode voltage is
lowered deviations from this law begin to
occur at, for example, 20 to 50 volts on the
anode, instead of five or ten volts as in the
case where the grid and anode are connected
together.
A simple analogy may make clearer the
reason for the effect of initial velocities being
so greatly exaggerated when the field around
the cathode is made small. Suppose, for
example, a man stands on the ground holding
a handful of sand and throws this up into the
air while the wind is blowing. The sand will
evidently be carried away by the wind. This
condition is pather analogous to the case of
electrons emitted from a filament surrounded
by an electric field which tends to draw the
electrons away. The wind corresponds of
course to the electric field. Under these
conditions, it is clear that the particular
510 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 6
velocity with which the man throws the sand
upward is not important so long as the
velocity is sufficiently great for the sand to
leave his hand. Consider now the case of a
man standing at the bottom of a well, with
the wind blowing over the ground around the
well : If the man throws the sand up as before,
it will all fall to the bottom of the well
because it is protected from the action of the
wind. If, however, he throws it upward with
sufficient velocity so that some of it rises
above the level of the ground, it will then be
carried away by the wind. The amount that
is thus carried will depend entirely on the
velocity with which the sand is thrown
upward. This is analogous to the case of the
Coolidge X-ray tube in which electrons are
emitted at the bottom of a depression which
protects the surface of the cathode from the
action of the external electric field.
Experiments have shown that if a tube is
made up exactly like a standard Coolidge
X-ray tube, except that the focusing shield
is omitted, the three-halves power law then
applies with verj' satisfactory' accuracy down
to comparatively low voltages.
Effects Due to the Charging Up of the Walls of the
Vessel
In cases where the anode does not nearly
surround the cathode some of the electrons
emitted from the cathode are able to pass to
the surface of the glass until a negative charge
develops on the glass sufficient to prevent any
further flow of electrons to the glass. In
other words, the glass will become slightly
negatively charged with respect to the
cathode. The actual potential which the
glass may reach in this way depends mainly
on how good the insulating conditions are,
and, as we shall see later, depends to an
extraordinary degree on the presence of even
relatively minute amounts of gas ionization.
According to Maxwell's Distribution Law,
there is no definite upper limit to the velocities
with which the electrons are emitted from
the hot cathode, but, as has been shown, only
about one out of a million of the electrons is
able to move against 2.85 volts, and only one
out of a billion against 4.27 volts. Such
extremely small currents as those correspond-
ing to this last figure would not ordinarily
be able to make up for the electric leakage
from the surface of the glass. Under usual
conditions, therefore (and this is in general
accord with experiment), the glass walls, and.
in fact, any well insulated electrode placed
within the vessel, charges up to potentials
in the neighborhood of about two volts. But
under exceptional conditions of high vacuum,
etc., this negative charge may be a couple of
volts greater.
Where the anode nearly surrounds the
cathode the slight negative potentials on the
glass have, of course, relatively little effect
on the electric field between the electrodes,
and therefore are not important in their effect
on the current-carrsing capacity of the space.
In those cases, however, where the anode is
relatively small, or is placed at a considerable
distance from the cathode so that a relati\-ely
large surface of glass is exposed, the effect
of the charges on the glass frequently becomes
of the utmost importance in modifying and
increasing the effect of the space charge, and
in weakening the field around the cathode.
In this way it may happen that the effects
of the initial velocities may be exaggerated,
just as in the case of the presence of an
auxiliary electrode close to the cathode. In
the case where the anode and cathode both
consist of small L'-shaped filaments placed
at some distance apart in a rather large glass
bulb, the effect of the charging-up of the walls
may become so great as entirely to stop the
current that flows between the two electrodes
even when several hundred volts are applied
to the anode. Such effects as th's can be
observed when the greatest effort is made to
obtain a high vacuum and the entire bulb is
completely immersed in liquid air so as to
improve the vacuum and decrease the electric
conductivity of the glass surfaces. Under
slightly poorer vacuum conditions, where
effects of this kind may occur, but are less
marked, it may happen that no current will
pass from cathode to anode until a certain
critical voltage is applied to the anode. Above
this point the volt -ampere characteristics
correspond accurately under ordinary condi-
tions to the three-halves power law. These
effects can be eliminated by rendering the
glass walls electrically conducting, as for
example, by distilling tungsten, copper or
other metallic substance on to them from the
filament and by connecting this conducting
surface to the anode or cathode in such a way
as to prevent the accumulation of negative
charges. ^
In properly constructed thermionic devices
effects of this kind usually do not occur at all
under normal conditions. These effects have,
however, misled a great many scientific
investigators and have undoubtedly led Pring
and Parker in their work on thermionic
emission from carbon, and Hallwachs in his
ELECTRON TUBES HAVING TUNGSTEN CATHODES
511
work on photo-electric emission, to conclude
that in a very high vacuum the thermionic
emission and photo-electric emission vanish.
Of course, subsequent work has entirely
disproved these conclusions.
Effects Due to the Current Used to Heat the Cathode
Under the usual operating conditions
ordinary thermionic devices have a cathode
which is in the form of a filament and is heated
by the passage of current through it. There is
thus a potential drop along the filament and
the current through the filament produces a
magnetic field surrounding the wire. There
are thus three ways in which the current
flowing through the filament may influence
the volt-ampere characteristics and cause
deviations from the three-halves power law :
(A) The potential drop along the wire
makes the potential diff'erence between the
anode and cathode vary for different parts of
the cathode surface.
(B) If the different portions of the cathode
are in close pro.ximity to each other, two
parts having marked difference of potential
produce an effect exactly like that caused
by the grid of a pliotron.
(C) The magnetic field causes changes in
the paths of the electrons.
Let us discuss these effects separately :
(A) Potential Drop Along Cathode. As an
extreme case where the potential drop along
the wire influences the characteristics of the
discharge, let us consider a long straight
filament placed in the axis of an equally long
cylindrical anode. We will assume that the
voltage drop along the wire is 100 volts. If,
now, we give the anode a potential of ten volts
(positive) with respect to the negative end of
the filament, it is clear (neglecting initial
velocities) that current- will flow only from
that part of the cathode which is negative
with respect to the anode. The current will
thus flow only from a section of the cathode
located near the negative end of the filament,
which is one tenth of the length of the
cathode. If, on the other hand, the anode is
charged to plus 20 volts with respect to the
negative end of the filament, then current
will flow from two tenths of the length of the
cathode while the other eight tenths will be
entirely inactive as far as contributing to the
total current is concerned. For each small
section of the cathode the three-halves power
law will apply with reasonable accuracy, but
the total effective length of the filament also
increases in proportion to the applied voltage.
From this it is readily seen (and this is
confirmed by more rigorous mathematical
analysis) that as long as the potential of the
anode does not exceed that of the positive end
of the cathode, the total current increases in
proportion to the five-halves power, that is,
the two and one-half power of the voltage.
With anode voltages higher than that of the
positive end of the filament this relation will
no longer hold, but the slope of the volt-
ampere characteristic will gradually approach
more and more closely to that corresponding
to the three-halves power law as the voltage
is raised beyond this point. It is thus seen
that the maximum effect of the voltage drop
along a straight filament with the anode at a
constant distance from it is not greater than
that corresponding to an increase of the
exponent from 3/2 to 5/2.
It is clear that in studying the character-
istics of a device in which there is a large
voltage drop along the filament, the voltage
applied to the anode should be measured
from the negative end of the filament, because
most of the electron emission, and in some
cases all of the electron emission, comes from
this end of the filament.
(B) Grid-like Action of One Part of the
Filament on Another. With special construc-
tion of the cathode by which the two ends of
the cathode are brought close together, or
where the cathode is made as a double helix,
this grid-like action may become fairly
marked. It is still more pronounced where one
end of the filament is supported by a frame-
work in proximity to the filament or to the
path that the electrons take between the
cathode and anode. This framework corre-
sponds exactly to a grid connected to one end
of the filament. The current-carrying capacity
of the space between cathode and anode
depends upon the direction in which the
current is made to pass through the cathode,
the voltage of the anode with respect to the
negative end of the filament being kept
constant. Thus, if the framework is positive
with respect to the rest of the filament, the
current-carrying capacity of the space will be
greater than if the framework is negative
with respect to the rest of the filament. In
general the 3/2 power relation will hold more
accurately if the framework is connected to
the negative end of the filament than if
it is connected to the positive end. Further-
more, with a given potential on the anode
(always with respect to the negative end of
the filament) the current-carrying capacity
of the space will vary with the current flowing
through the cathode, because the potential
512 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. G
on the positive end of the filament is thus
made to change and this alters its effect as a
grid. In case the anode is maintained at a
given potential with respect to the positive
end of the filament (which is not customarity
done) this may readily lead to a decrease in
the current-carr\-ing capacity of the space
when the filament temperature is increased.
This effect is simply due to the negative end
of the filament becoming more strongly
negative when the filament temperature is
raised.
(C) Effect of Magnetic Field. Richardson*
has pointed out that the magnetic field pro-
duced by the current flowing through the
cathode tends to interfere with the free
motion of the electrons and tends in fact to
cause them to return to the cathode. This
effect is particularly marked when the
cathode is such that it requires large heating
currents of many amperes, and is also more
marked when the electrons are moving under
the influence of a weak electric field, that is,
when the anode voltages are low.
For the case of a straight filament carrying
one ampere the electric field needed to draw
the electrons through this magnetic field is
only 0.2 volts. The effect is thus of the same
order of magnitude as that due to the initial
velocities of the electrons. With filaments,
however, which take 10 or 20 amperes to heat
them, this effect would be quite serious except
when high anode voltages are used.
Elimination of Effects Caused by Current Passing
Through the Cathode
All of the effects caused by the current
passing through the cathode can be entirely
eliminated by making measurements of the
volt-ampere characteristics during short inter-
vals of time during which the current flowing
through the cathode is interrupted. If these
time inter\'als are made of the order of a
hundredth of a second or less, the filament
temperature remains practically constant.
This may be accomplished by use of a rotating
commutator. This method was apparently
first used for measurement of thermionic
currents by O. von Baeyerf and was subse-
quentlv used bv Schottkv (Annalcn der
Physik 44. 1021 ['ini4]). This simple method
of determining the characteristics of a
thermionic device eliminates all the rather
complicated secondary effects produced by
(A) the potential drop along the cathode;
(B) the grid-like action of one part of the
» The Emission of Electricity from Hot Bodies" (1916), by
O. W. Richardson, p. 67.
t Physik. Zeitsch 10. 168 (1909).
cathode on another; and (C) the magnetic
field caused by the current flowing through
the cathode.
Effects Caused by Lack of Uniformity in the
Temperature of the Cathode
The filament used as cathode is cooled at
its ends by the leads which carry the current
to it. and it is also cooled by any supports used
to hold the filament in place. Although the
central part of the filament may be heated to
such a high temperature that it emits more
electrons than can be carried through the
space with a given voltage on the anode, the
ends of the filaments are always cooled to such
a low temperature that they emit practically
no electrons. A mathematical and experi-
mental analysis has shown that under ordi-
nar\^ conditions the cooling effect at the two
ends of a tungsten filament lowers the total
electron emission from the filament by an
amount that corresponds roughly with the
electron emission from a length of filament
having a voltage drop of 1.4 volts. Thus, if
the voltage used to heat the cathode is 14
volts, the cooling effect of the leads normally
decreases the total electron emission by about
10 per cent below that which would be
obtained if the entire length of filament were
heated to the same temjierature as that which
exists at its middle. With a filament, however,
requiring only 4.2 volts the cooling effect of
the leads would reduce the total electron
emission by about one third. These results
are given simply to indicate the order of
magnitude of the effect produced by the
cooling of the leads.
With filaments which take three volts or
less the relati\e decrease in electron emission
for a given heating current is much greater
than would be calculated by the above simple
rule.
In the derivation of the space charge
equations, particularly that derived for the
case of electrodes of any shape, it was assumed
that every part of the cathode surface was
heated to a temperature at which a surplus of
electrons was produced. Now, in all cases in
which we are dealing with actual filaments
there is a short length of filament near the
ends on which the electron emission is so
small that there is not a surplus of electrons.
In fact, over this region of the filament the
current flowing from it will Ix" saturated.
This causes a slight deviation from the 3 2
power law, but with filaments requiring three
or four volts or more this particular cause of
deviation is an extremelv small effect. This is
ELECTRON TUBES HAVING TUNGSTEN CATHODES
513
proved by observation of the actual character-
istics of electron discharge devices having
filaments of this character. It is found that
the volt-ampere characteristics do follow the
3 2 power law over a wide range of voltage.
Of course the higher the voltage used in the
filament the smaller this effect becomes, but
the evidence seems to indicate that even for
the filaments having the lowest drop used in
practice this particular effect is negligible in
causing departure from the 3/2 power law.
The cooling eft'ect of the leads, however, has
a more marked effect on the characteristics
in causing a change in the current-carrying
capacity of the space because of the change
in the area of the heated part of the filament
when difl^crent filament temperatures are used.
Thus the higher the filament temperature the
longer the section of the filament which will
be heated hot enough to emit a surplus of
electrons. With filaments in which the
voltage drop is only three or four volts and
operating at normal temperature a marked
increase in filament temperature will increase
the length of the heated region sufficiently to
cause several per cent increase in the current-
carrying capacity of the space, simply
because it increases the cross-section of the
space through which the electrons pass. With
filaments in which the voltage drop is larger,
however, this effect becomes proportionally
less.
This effect of the cooling of the leads
sometimes causes the apparent current-
carrying capacity of a given device to depend
on the cathode temperature. This should not
be mistaken for an effect due to the initial
-velocities. The effect due to initial velocities
is noticed particularly with low anode volt-
ages, whereas that due to ths cooling of the
leads is effective at all anode voltages practi-
cally equally. The volt-ampere curve in this
latter case still corresponds to the 3/2 power
law, but the coefficient K in the equation
i = KV^I'^ increases as the filament tem-
perature -is raised simply because the length
of the heated region changes with the tem-
l)erature of the filament.
Effect of External Magnetic Field
If a thermionic device is placed in a
magnetic field, as, for example, by bringing
an electro-magnet or permanent magnet close
to the bulb, the paths of the electrons are
changed in a way similar in principle to that
caused by the magnetic field due to current
flowing through the cathode. With ordinary
thermionic devices of the type I am consider-
ing an external magnetic field usually
decreases the current-carrying capacity of the
space for a given anode voltage. The lower
the voltage on the anode the more marked
this effect becomes. As a result of this an
external magnetic field usually causes the
current to increase with the applied voltage
on the anode more rapidly than when the
magnetic field is absent. Thus, in the ordinary
device the 3/2 power law holds accurately
over a wide range when there is no field, but
a strong magnetic field has the effect of
making the current increase more rapidly
than the 3/2 power law requires. An effect
of this kind is ordinarily only noticeable when
rather strong external fields are applied, and
weak fields like that of the earth's magnetic
field are negligible in their effects.
{To be concluded in July issue)
514 June. 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 6
Electron Power Tubes and Some of Their
Applications
By William C. White
Research L.\bor.\tory, Gexer.^l Electric Company
The interest and activity in radio transmission is on the increase and this fact has focused attention on
the three-element electron tube, otherwise the pliotron. In this article there is given a very thorough expo-
sition of the factors of design and construction that determine the output of three-element power tubes. This
is followed bv a description of a typical pliotron power tube, after which are discussed the properties of oscil-
lating circuits for these tubes. The article is concluded with descriptions and illustrations of radio transmit-
ting sets. — Editor.
There is a considerable field of application
for three-element electron tubes as oscillators
or amplifiers to give outputs of several
hundred to several thousand watts.
For this purpose it is possible to utilize
a ver\- large number of small tubes operating
in parallel, but for reasons of expense, com-
plexity, liability to breakdown of units, and
space required, such a solution of the problem
is impracticable and becomes increasingly so
the greater the power output required.
In radio transmitting apparatus particularly ,
there is a field for continuous wave high
frequency outputs of about one kilowatt.
Excluding amateur installations, the greater
proportion of radio transmitting sets are of
the so-called medium power type giving out-
puts of high frequency energy of 250 to 2500
watts into the antenna. With the usual
size of antenna employed and the logical
organization of wave lengths, operation in the
wave length range of 300 to 2000 meters
(1,000,000 to 1.50,000 cycles) is usually
desired.
At the present time spark sets, mostly of
the quenched gap type, are used for this class
of transmitting equipment.
The adoption of continuous wave trans-
mission to supersede damped wave trans-
mission has brought about wonderful im-
provements in the case of very low power
and ven.- high power installations.
Spark sets of the medium power class
mentioned above cannot be replaced b\- the
type of apparatus used on the high power
continuous wave sets.
Neither the high frequency alternator,
the Poulsen arc nor the so-called timed spark
systems are practical for this particular class
of stations, principally because of their
inability to operate at the short wave lengths
required and also due to lack of flexibility of
wave length in the case of the alternator and
poorer characteristics of the arc at low power.
Also in the case of radio telephony, voice
control is a comparatively simple matter
with three-element electron tubes.
Therefore it is logical to believe that this
class of medium power transmitting equip-
ment is a particularly premising field for
the three-element vacuum tube.
There are also several other fields of applica-
tion, some of which will be described later.
Since the three-element power tube itself
is the basis of such equipments, its character-
istics and the features required for power
output will be first considered.
THE THREE-ELEMENT POWER TUBE
There are a number of factors in the design
and construction of a three-element vacuum
tube which may limit its output.
(1) Dissipation of energy in the form of
heat at the anode so great that deterioration
of the vacuum results, or certain of the parts
lose their mechanical strength, or even melt.
(2) Insufficient electron emission, resulting
in a definite limitation of plate current and
therefore limiting the input energy to the
tube.
(."1) Insufficient exhaust treatment.
(4) Insufficient dielectric strength in the
materials holding the electrodes in place and
in the lead-in wires or terminals.
(5) Insufficient mechanical strength of the
electrodes or their parts, so that the high
anode voltage causes a displacement and
probable short-circuiting, due to the mechan-
ical force of the electrical field.
((>) Improper geometrical design or con-
struction so that the electrical constants of
the tube are incorrectly proportioned to the
conditions of operation. This allows the
factors which cause the first five limitations
to be at values above the possible minimum
These factors will now be taken up more
in detail, together with comments n-lating
thereto.
ELECTRON POWER TUBES AND SOME OF THEIR APPLICATIONS
oL-J
(/) Dissipation of Energy from the Anode
Of the input to the plate circuit of the tube a
portion appears as output and ])ractically the
entire remainder is lost as heat at the plate.
In the types of oscillating circuits used at
the present time the output approximates
50 per cent of the input, actual values ranging
from 25 per cent up. Therefore, in the design
of the anode and bulb provision must be made
to dissipate more energy than is produced.
The heat energy leaves the anode or plate
mostly by radiation, but a certain proportion
is carried away by conduction through the
plate supports. The necessary plate area
acting as anode is more or less determined
by the desired electrical characteristics, but
its actual area, to facilitate radiation of heat,
may be increased by vanes or other forms of
attached surfaces. It is good practice to
make the plate out of the same piece of
metal as the vanes, as it is surprising how
poor the heat conductivity between two
pieces of metal may be that are in intimate
contact by a process such as riveting. This
effect is often plainly apparent in plates that
are running at a red heat by the difference
in color of parts supposedly in intimate
contact.
In the choice of material for the plates of
power tubes the use of tungsten or molyb-
denum is very desirable, as it gives the follow-
ing advantages of energy dissipation for a
given area:
(1) These metals, in the form of vacuum
tube plates, can be freed from gas to a greater
extent than any other metal.
(2) They retain good mechanical strength
at a bright red heat.
(3) The rate of evaporation of the metal
which would cause blackening of the bulb is
very small, even at a bright red heat.
(4) On account of their very high melting
points they can stand up under very high
inputs of energy over a short period of time.
This is important in safeguarding the tube
from excessive inputs with no output when,
from some accidental cause, a high voltage
power tube stops oscillating and the plate
current is only limited by the electron
emission.
The writer has noted this latter defect in
many foreign power tubes. Their rated
operating condition is so near the absolute
limit which the tube will stand that any
change in conditions is liable to spoil the
vacuum, or a mishap stopping oscillations
almost instantly causes destruction of the
tube by excess dissipation of energy. It
certainly is not good engineering practice to
so rate vacuum tubes. This defect in tubes
of foreign design is usuilly accentuated by
the fact that the design is of rather low
impedance for the voltage employed, thus
making the plate current at full rated plate
voltage greatly in excess of the normal load
current in the oscillating condition.
When a tube fails to function properly and
the characteristic blue glow of excessive
ionization appears, this gas usually comes
from the metal of the plates, the glass, or
from other metal portions of the tube, such
as plate supports, grid frame, or supporting
sleeves.
Extravagant claims are often made as to
the output of a vacuum tube. If the tube is
used in the usual types of circuits, so that
the efficiency is of the order of 50 per cent,
an examination of the size and material of
the anode gives very good evidence as to the
power capabilities of a tube, assuming it is
satisfactory in all other respects.
A little experience with tubes of various
designs gives a person a very good idea of the
approximate amount of energy which can
be liberated from any form of anode of par-
ticular material without either rendering it
too weak mechanically, or too hot so that
excessive evaporation takes place.
{£) Insufficient Electron Emission
When a tube is delivering alternating cur-
rent energy, either as an amplifier or oscillator,
a certain average direct current is supplied to
the plate from the direct current source.
Electrons, of course, cannot be "stored
up " and used even a short period of time after
their emission. Therefore, there must be a
constant emission of them sufficient to meet
the maximum demand portion of the plate
current cycle.
With a pure sine wave form of plate cur-
rent an emission corresponding to double the
average plate current is required to give the
peak value. Also during the period of the
cycle when the plate current is maximum the
grid voltage and therefore the grid current
is at a positive maximum. A direct current
meter placed in the grid circuit of an oscillat-
ing tube supplying energy under proper
adjustments shows a current of roughly 10
per cent of the average plate current. There-
fore the peak value of electron current to the
grid is several times this value, running up
as high in certain cases as the average plate
current itself.
It is desirable to have a certain excess
electron emission so that the tube will deliver
516 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
practically full energy over the range in
variation of filament current that is liable to
occur in actual supph' circuits or through the
limits of variation of a filament current-
regulating device.
Also slight differences in the dimension of
filaments for different tubes of the same type
necessitates usually a further allowance of
excess emission.
These factors taken together usually require
that the emission at rated filament current
be three to five times the a\'erage current
while oscillating, as measured on a direct
current instrument.
There is also an efl'ect which is negligible
on receiving and other low power tubes, but
which is an important factor on high voltage
power tubes. This is the effect on filament
temperature, and therefore upon the electron
emission, of the electron current of the plate
circuit, adding to or subtracting from the
filament current.
If the filament is operated from a direct
current source through a series resistance in
the circuit the electron current will add to
the filament current if the negative of the
plate voltage source is connected to the nega-
tive filament terminal and it will subtract
from it if connected to the positive filament
terminal. This effect will not be uniform over
the entire length of the filament but will be
variable, being a maximum or minimum at
the end of the filament to which the negative
terminal of the plate source is connected.
When it is remembered that the average
plate current of a tube usually has a value
between 2 and 7 per cent of the filament
current and that a ."J per cent increase or
decrease in the heating current of a tungsten
filament respectively halves or doubles the
filament life, the importance of allowing for
this is apparent.
This filament heating effect of the plate
current also has another important aspect.
If for any reason a high voltage power tube
stops oscillating, the plate current will usually
rise to a value limited only by the filament
emission. If the return or negative of the
plate source is connected to the negative end
of the filament this abnormal current flow
will increase the temperature and therefore
the emission which in turn increases the plate
current, this effect being accumulative, often
V destroying the filament in a few seconds.
Therefore, on a high \-oltagc power tube
with direct current filament excitation it
is advisable to connect the negative terminal
of the plate voltage source to the positive
end of the filament.
If possible, alternating current should be
used for filament excitation with the regulat-
ing rheostat in the power side of the trans-
former circuit and the return of the grid
and plate circuit made to a center tap on the
winding supplying the filament.
This connection assures minimum dis-
turbance in the plate and grid circuits from
the frequency of the filament source.
The plate circuit current in this case divides
evenly between the two filament legs. Also
the direct electron current and the alter-
nating filament current add at a 90 deg. dis-
placement to give the combined heating cur-
rent so that the additive effect is much smaller.
For instance, a one-ampere electron current
added to a four-ampere filament current
would give a five-ampere heating current at
the hottest part of the filament if it were d-c,
and only 4.13 amp. if a-c.
For a tungsten filament there is a definite
relation between the electron emission cur-
rent, filament heating watts and temperature.
A curve is shown (Fig. 1) giving the plot
between milliamperes emission current per
'
1
1
1
r"
1
>!
1
1
t
k"
1
»
f
1
I
? m
t
J. >i3
I
1
r
/
1 ^
/
/
Q
/
3
'
•*'
—\
ZHxi zioo Z3O0 i*oo Z500 uao
Fig. 1. Variation of the " Efficiency " of Electron Emission
with Temperature for Tungsten Filament Cathode
watt of filament heating energy and absolute
temperature of the filament. This relation
holds for any diameter or length of filament,
but its values are subject to some modifica-
tion for particular cases owing to the cooling
effect of lead wires and filament supports and
ELECTRON POWER TUBES AND SOME OF THEIR APPLICATIONS 51^
also variations in emissivity factor for dif-
ferent samples of tungsten. This curve is
plotted from Dr. Langmuir's published data.*
In types of high power tubes in which the
plates are of tungsten or molybdenum and
are located close to the plate and operated
at a bright red heat, the filament temperature
will be higher than when the plates are cold.
Therefore in the operation of such a tube the
filament current may be reduced after the
plates have come up to their normal operating
temperature.
Although in vacuum tube circuits it is
advisable to include a voltmeter or ammeter
in the circuit of the filament or filaments,
such an instrument should not be wholly
relied upon for filament adjustment.
The best practice is to operate tubes at the
lowest filament temperature consistent with
satisfactory operation. In this way maximum
tube life will be obtained.
{3} Insufficient Exhaust Treatment
As is now well known, it is not only neces-
sary to reduce the gas pressure in the biilb to
a minimum, but it is even more important to
free the internal parts from gas so that the
pressure of gas in the bulb remains low through-
out the life of the tube.
The exhausting process for a vacuum tube
increases in diffic'ulty the higher the power of
the tube and the higher the voltage at which
it is to operate.
This condition arises at higher powers and
voltages, owing to the fact that the positive
ionization eff^ects are greater and the tem-
perature of the parts higher.
Keeping the glass walls of the tube cool by
artificial means will help to better the vacuum,
because it not only prevents the glass from
liberating gas, but may actually enable it to
absorb some gas that might be liberated from
other parts of the tube.
(4) Insufficient Dielectric Strength in the
Materials Holding the Electrodes and in the
Lead-in Wires or Terminals
In a three-element oscillating tube the
maximum voltage occurs between the grid
and plate and may easily reach a value three
times the normal operating plate voltage.
This is due to the fact that with a pure
inductance in the plate circuit the current
may vary between zero and twice normal
value each cycle, and therefore the voltage
between filament and grid may vary between
zero and twice normal. At the same time
*"The Characteristics of Tungsten Filaments as Functions of
Temperature," by Irving Langmuir, G. E. Review, Vol. 19, No.
3, March '16.
there is a ISO deg. relation between grid
and plate voltage; therefore, with a tube of
low amplification constant the grid voltage
may easily reach the value of average plate
voltage.
It will therefore be seen that with a tube
operating at a plate voltage of several
thousands, the dielectric strain may be
considerable. Owing to the temperature at
which power tubes operate this factor is made
more serious.
A high vacuum is the best insulator under
these conditions and air at atmospheric
pressure also is a good insulator.
Glass, however, is necessarily used for sup-
ports. The dielectric strength of glass
decreases rapidly with increase of tem-
perature. This is true of all grades of glass
but the effect is much more marked in some
grades than in others.
Hot glass is conductive and acts like an
electrolytic solution. Bubbles of gas form
at the negative electrode and if this electrode
is one of the seal-in w-ires leakage of air soon
results.
It is interesting to note that in a tube in
which the leads are brought through a pinch
seal the electrolysis is much more serious
with the tube oscillating than with the tube
operating non-oseillating with the same plate
voltage and energy loss in the tube.
With the best grades of glass at a tem-
perature of about 400 deg. C. the dielectric
strength at high frequency is less than a layer
of air of equal thickness.
Therefore, in the design of a high power
tube it is necessary to have the electrical
path between electrodes through the glass as
long as possible and located in one of the
cooler parts of the bulb.
Under certain conditions of improper
adjustment of the oscillating circuit the
voltage between grid and filament may rise
very high. This necessitates careful insula-
tion between these leads both in the tube and
in the base.
Related to this cjucstion of dielectric
strength between electrodes is the question
of high frequency dielectric losses in the
material employed, such as the glass of the
tube and the insulating materials of the bases.
This factor becomes very important because
often these dielectrics are subjected to an
intense electric field of high frequency.
If the materials used have a high dielectric
loss heat will be generated at the points of
loss, adding to the liability of breakdown and
also decreasing the efficiency of oscillation.
518 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 6
(5) Insufficient Mechanical Strength to With-
stand the Mechanical Force Due to the
Electric Fields
The filament and grid mesh are most
subject to this strain, because of their com-
paratively small size. The most usual effect
of this strain is a contact between filament and
grid or grid and plate.
On several occasions the writer has seen
double helix spiral filaments pulled into
almost two parallel strands by this force.
Also, if the plates are operated by alternating
current of a commercial frequency, some of
the grid strands may have mechanical
resonance to this frequency and vibrate to
destruction.
When operating an audible frequenc}'
oscillator it is not uncommon to have the
tube emit a distinctly audible tone of the
frequency generated.
(6) Improper Geometrical Design or Con-
struction
If the value of the amplification constant
is too low, an excessive grid excitation voltage
is required for power oscillation. This makes
more difficult the problems of mechanical
and dielectric strengths.
If the value of plate impedance is too high
for the voltage used or for the output desired,
the grid current will be excessive while
oscillating because the grid must be carried
to a high positive value to obtain the neces-
sary maximum of plate current.
In this case also greater emission will be
required to supply the added grid current.
This question of projxT proportioning of
the electrodes to obtain the best electrical
constants is not w-ithin the scope of this
paper. These two examples were stated
to give an idea of some of the factors in-
volved.
As in most cases of design, the final choice
of constants is a compromise which, it is
believed, will give the best results under the
conditions of service.
In the case of a power vacuum tube, cer-
tain practical considerations usually decide
the electrical design rather than the choice
of the electrical constants for absolute maxi-
mum output.
In the case of a tungsten filament type of
tube the plate voltage is chosen as high as
possible consistent with the operating con-
ditions and procurability of the voltage
source.
* "The Vacuum Tube as a Generator of AllernatinR Current
Power." by J. H. Morecroft and H. T. Frii?. Presented before
the Philadelphia meeting of the A.I.E.E., October 10. 1919.
The desired power output being decided
upon, and knowing the probable efficiency.
the input direct current is then determined.
It is then best to design the elements of the
tube so that at full plate voltage, with the
filament at a maximum temperature, the
plate current will not exceed two or three
times normal oscillating value when the grid
is at zero potential.
Although a design made in this way will
show a higher impedance and therefore a
somewhat lower output than the use of a
lower impedance, the ease of handling and
safety to the tube during telegraphic opera-
tion and particularly while making adjust-
ments more than compensate, it is believed,
for the loss in maximum output.
For efficient operation and use of the tube
as a modulator, as large a value of amplifica-
tion constant as possible is advisable; this
choice, of course, assuming a value of im-
pedance as chosen above.
In the foregoing it is understood that by
impedance is meant the value obtained from
the slope of the plate-voltage plate-current
cur\'e. This, of course, is not the apparent
resistance of the tube found by dividing the
plate voltage by plate current for a given
value of grid voltage.
It is possible that future developments in
the vacuum tube art might make maximum
output a better criterion than ease and safety
of operation, the latter being gained by
auxiliary and protective devices.
The question of the possible efficiency of
the vacuum tube as a generator of oscillations
is dependent to some extent on the proper
tube design, but much more largely on the
circuit employed. This question has been
quite thoroughly investigated and reported
upon in a very interesting paper recently
presented.*
DESCRIPTION OF PLIOTRON POWER TUBE
As an example of a three-element power
tube, the Type P pliotron will be described.
This tube is rated at 2.'i() watts output with
!.")()() volts on the plate. The filament con-
sumes about SO watts. The plate current at
full load is approximately ;?0() milliampcrcs.
This tube is shown in Fig. 2. a view of the grid
filament and plate elements being also included.
The bulb and glass parts arc constructed
of Pyrex. a special strong heat resisting glass.
The globular part is .1 in. in diameter and
there are two arms extending from opposite
sides, making, when based, a total overall
length of approximately M^^ in
ELECTRON POWER TUBES AND SOME OF THEIR APPLICATIONS
519
The cathode arm is 2i'2 in. and the anode
arm II/2 in. in diameter. The net weight is
approximately 25 ounces. For shipment the
tubes are crated individually.
There are three terminals in the base at
the cathode end. The center blade is the
grid terminal, the two pins being the filament
terminals. The anode terminal is the cap
at the anode end.
The filament is of ductile tungsten wire
W-shaped in form. The range of filament
current used is 3.5 to 4.0 amp.
The U-shaped grid frame and the wire
forming the grid mesh are of tungsten and
are freed from gas by the exhaust treatment.
The anode or plate are of tungsten sup-
ported by molybdenum rods. The plates
and these rods are, of course, thoroughly
freed from gas during the exhaust treatment.
Under normal operating conditions the
plates run at a dull red heat. During exhaust
treatment they are brought up to a brilliant
red heat by electron bombardment dissipat-
ing nearly 2 kw. of energy. An average
characteristic curve plotted between grid
Fig. 2. Large Type of Pliotron Tube. Assembled and
Disassembled
voltage and plate current for this type of
tube is given in Fig. 3.
It is interesting to estimate the future
development in size and output of the three-
element power tube.
In the opinion of the writer, the physical
dimensions of the glass bulb pliotron are
somewhere near their limits. This is because
large glass constructions are expensive to
build, difficult to safely ship, and a very slight
mishap such as a small crack or leak prac-
tically destroys the total value of the tulDC.
It is believed that the much-to-be-dcsired
increase in power output per tube unit will be
obtained in one or more of the following ways:
1
/
1
/
f
Plate I/OHO(!I> 1500
F/lorrtentaC S6Amp
/
/
f
/
t
?
/
^
/
f
$ /so
%
/
f
/
4&
f
Ij 100
S
/
/
1
/
f
/
so
J
/
/
/
/
X
-60 -40 -20 O *20
Grid l^/togemth Respect to Neqotiye Filament End
Fig. 3. Typical Characteristic Curve for the Type of
Pliotron of Fig. 2
(1) Increase of plate voltage for the glass
tubes up to the neighborhood of 100,000
volts, or whatever limit is set by the feasi-
Vjility of the production and use of such a high
voltage. By the use of these high voltages
in special circuits, which seem possible of
development, a very high efficiency is prob-
able so that possibly 100 kw. may be gen-
erated by a pair of tubes each not much larger
except for terminal arms than the Type P
tube described.
(2) By the use of a hermetically sealed
metal tube.
(3) By the use of a metal or glass tube or
combination in which the vacuum is main-
tained by a continual or intermittent method
of exhaustion.
In any case the use of as high a voltage as
practical is desirable.
The advantages of increased voltage are
due to the fact that for a given output the
currents are smaller, thus simplifying the
problems of emission and space charge. The
outputs and efficiencies can be made higher
520 Tune, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. G
because more voltage is available for the
output circuit and a smaller proportion is lost
in the tube.
OSCILLATING CIRCUITS FOR POWER
TUBES
A great deal has been published on this
subject and the principles involved are pretty
widely understood. However, such publica-
tions have dealt mostly with receiving cir-
cuits in which the requirements are somewhat
different.
Some general considerations from the
viewpoint of power tube operation will be
given and these points explained in connection
with a typical circuit.
In practically all the forms of oscillating
circuits used at the present time three funda-
mental adjustments are necessary in order to
get full output at the best efficiency and
desired frequency. In some forms of circuits
the variation of one factor will change another,
making the complete adjustment a complex
matter.
These three fundamental adjustments are:
(1) Variation of inductance or capacity or
both in the resonance portion of the circuit
to obtain the frequency desired.
(2) Adjustment of the ratio of transforma-
tion between the plate circuit and the load
circuit.
(3) Adjustment of the voltage value, phase
relation and normal d-c. potential of the
energy fed back to the grid for self-excitation.
The first and probably the third of these
adjustments are well known and need no
further explanation; but the second one, the
adjustment of the ratio of transformation
between the plate circuit and the load circuit,
is not as well understood as it should be.
Since power tubes arc used mostly for
radio transmission, an antenna as a load will
be considered.
For energy calculations the antenna ma>-
be replaced by a resonant circuit having
concentrated inductance and capacity. In
an artificial circuit of this sort the energy
radiating property of the antenna is replaced
by a resistance and it is customary to con-
sider a resistance in series with the inductance
and capacity rather than a higher resistance
shunted across the two.
For instance, ten amperes flowing in an
antenna circuit of 0.001 microfarad and one
millihenry at the resonant frequency of
1,59,000 cycles (ISSn meters wave length), will
give an energy dissipation of 1000 watts if
the series resistance is ten ohms. An equix'a-
Icnt dissipation in energy would take place
if the series resistance of ten ohms were re-
placed by a resistance of 100,000 ohms
shunted across the capacity.
For ever\' condition of operation of a
three-element tube as an amplifier or oscilla-
tor, there exists a particular value of plate
impedance and the use of this value of
resistance in the plate circuit will give a
maximum output of energy. This is a well
known fact and has been brought out by
many writers.
Now supposing the normal operating imped-
ance of the tube to be 5000 ohms, it is verj-
apparent that the voltage generated by the
tube is not at all suitable for direct excitation
of the antenna. Therefore, a transformer or
its equivalent must be placed between the
plate circuit and the antenna circuit.
In the case mentioned the ratio of trans-
formation between impedances of 5000 and
100,000 would be the square root of this
ratio or approximately 1 to 4.5.
The actual turn ratio would be consider-
ably higher than this, owing to the fact that
in most radio frequency transformers the
voltage ratio falls below the turn ratio.
An equivalent to a step-up transformer is
a variable coupling, but in this case the plate
circuit must be made resonant by the addi-
tion of a capacity shunted across the plate
inductance.
In order to get full output from a tube it is
neccssar\' to have only resistance effective
in the plate circuit. Any excess value of
inductive or capacity reactance means a
heavier current for the same energy delivered
(the so-called wattless component) and this
component gives an added loss in passing
through the necessary impedance of the fila-
ment to plate path in the tube.
In any circuit built to energize an antenna
at a number of quite widely different wave
lengths, this ratio must be varied for each
wave length in order to get best results,
because the effective antenna resistances will
vary. The tube impedance under fixed
operating conditions remains quite constant
except through very extreme ranges of fri'-
quency where the capacity effects between
electrodes becomes appreciable.
One form of typical oscillating circuit quite
widely used for energizing a radio trans-
mitting antenna is shown in Fig. 4.
The first variation mentioned, that of
change of inductance or capacity to get the
desired wave length, is accomplished in this
case bv variations of the inductance of /-.4.
ELECTRON POWER TUBES AND SOME OF THEIR APPLICATIONS
521
The inductance Lp and L^ have a fixed
close couphng and the second adjustment,
that of variation in ratio, is accomphshed by
variation by taps in the number of turns of
Lp. In most cases the ratio of Lp to La is
a step-up one; but if the tube impedance is
unusually high and the antenna resistance
(expressed in the usual way as equivalent
series ohms) rather higher than the average,
the ratio in the direction as above stated may
be slightly step-down.
If a direct current ammeter is placed in
one of the leads from the direct current source
Eb, so as to indicate the average plate cur-
rent lb and an ammeter inserted in the antenna
it is very easy to set the transformer ratio at
the proper value.
Starting with a large number of turns on
Lp, other adjustments being properly made,
both the value of plate current h and an-
tenna current h will be low.
As the number of effective turns on Lp is
decreased the values of h and la rise. Since
Eb is held constant while antenna voltage
increases with increase of la, the rate of
Fig. 4. Typical Oscillating Circuit for Energizing a Radio
Transmitting Antenna
increase of la will be less for a constant value
of efficiency.
In most cases of high voltage power tubes,
the adjustment is made for best efficiency
considering the maximum safe load on either
the tube or the source of supply at Eb- In
either case the value of Lp is adjusted until
the ratio of , is a maximum. If the number
of effective turns of Lp is decreased beyond
the best point there will be an increase of /;,
with little or no increase in la. Finally, if
the number of effective turns of Lp is unduly
decreased, the oscillations will usually stop
and lb take up a value dependent upon the
static characteristics of the tube.
Of course as in the case of transformer
design at commercial frequencies it is not
only the ratio of turns between Lp and La that
is important, but also the actual number of
turns used on the coil Lp. Too few turns
will give an excessive exciting current relative
to normal load current and too many turns
an excessive impedance to the load current.
Owing to the fact that air with its uniform
magnetic permeability is usually used for
the magnetic circuit of the coils the permis-
sible range of design variation is increased
over that of commercial frequency trans-
formers.
If a variable coupling between Lp and La is
employed rather than variable turn ratio, a
condenser must be placed across Lp to form
a resonant plate circuit if full output is
desired. If this condenser or an equivalent
is not used there will be an excessive inductive
reactance in the plate circuit.
The third adjustment, that of grid excita-
tion voltage, is accomplished by variation of
Lg, its coupling to La and the value of Rg (see
Fig. -t).
In the actual arrangement of a circuit of
this type incorporated in a finished piece of
apparatus, there are several features which
are of interest.
In radio telegraphy the telegraphic dots
and dashes require rapid make and break of
the energy to the antenna. This is usually
accomplished by placing a key in series with
the high grid resistance Rg of Fig. 4. In this
way it is only necessary to make and break
the relatively low current and low potential
of the grid circuit. Opening the circuit at
this point allows the condenser Q to charge
up so as to give the grid a high negative value
of potential, thus cutting off the flow of plate
current to the tube which in turn reduces the
input and therefore the output to zero.
Now, the condenser Cg has usually a small
value of capacity of the order of a few hundred
micro-microfarads. Therefore, the rate of its
charging up and the resultant rate of decrease
of the plate current h is extremely rapid.
522 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
much more rapid than the opening of a pair
of contacts because there is no arcing or
sparking.
In most power tube work the source Ei, is
a high voltage, direct current generator. In
such a machine the number of turns in an
armature coil is large and therefore the arma-
ture inductance is high. If the flow of cur-
rent through this inductance is suddenly
interrupted there naturally follows a con-
siderable step-up of voltage across the arma-
ture terminals.
Therefore, there is an abnormal instan-
taneous voltage strain liable to cause a break-
down, which breakdown of insulation usually
occurs between some point on the armature
coil near the plus end and the core of the
armature, because the negative armature
terminal and frame are at, or near, ground
potential.
As an example, a certain more or less stand-
ardized form of high voltage direct current
generator rated at 1. ")()() volts. 2 kw., 1800
r.p.m., had an armature inductance of approxi-
mately one henry. The full load current was
1.33 amperes and it is readily seen what a
high surge of voltage was caused when this
current was brought to zero in a very small
fraction of a second.
It is therefore necessary to put some form
of protective device across the terminals of
Eb which will pass current when the voltage
arises above a certain predetermined value.
Aluminum cell lightning arrester equip-
ment* has proved to be an excellent form of
protective device and it has been found
advisable to use them when the energy input
exceeds 500 watts at 500 volts or over. For
the lower powers a condenser of as large a
value as practical should he shunted across the
source Eb-
When a circuit is incorporated in a finished
piece of apparatus in which the wiring is more
or less com])licated by the requirements of
operation, it is often noticed that the adjust-
ments, particularly that of the grid excitation,
are more critical than in a similar but simple
experimental circuit. This can be usually
accounted for by the capacity effects between
adjacent wires and inductive clTects between
wires of different circuits which are naturally
closer spaced and of greater number than on
an experimental layout.
In operating tubes of the larger type in
parallel it is usually advisable to include a
fuse and ammeter in the plate circuit of each
tube. The ammeter allows the location of a
defective tube to be noted and the fuse throws
out of circuit a tube drawing excessiv^e current.
In the parallel operation of tubes, trouble
is often experienced with sudden values of
excessive plate current. One cause of this is
the oscillation of the tubes at a very high
frequency, the value of which is determined
by the capacity between the electrodes and
the inductance of the leads to the grid and
plate. In order to avoid these occasional
abnormal plate currents it has been found
advisable to insert in the grid circuit of a
few of the tubes a very small inductance of
only a few microhenries value. This pre-
caution prevents the operation of the bank of
tubes at this ultra frequency. Such a coil
is inserted in the grid circuit as near the grid
terminal of the one or two tubes as possible.
Where a considerable number of power
tubes are operated in parallel the total plate
impedance may reach a rather low value of
the order of a few hundred ohms. Also at
very high frequencies as small an inductance
*'* The Construction and Maintenance of Aluminum Cell
Arrcesters." by R. T. Wagner. Gknkrai. Elkctrk Rkvikw.
Vol. 16. January. Ifll.i.
Fig. 5. Pliotron Radio Transmitting Set In the Open
Position Shown AH Circuits are Disconnected
as 10 microhenries is closely comparable to
the tube impedance, and therefore under
these two conditions of high frequency
( 1,000.000 cycles or more) and low total plate
impedance it is important that the total
plate im])edance be as far as possible localized
ELECTRON POWER TUBES AND SOME OF THEIR APPLICATIONS
523
in the tubes and output transformer. Ten or
fifteen feet of plate circuit wiring may have a
very appreciable effect under these con-
ditions.
For this reason it is usually advisable to
bridge the d-c. power leads by a capacity as
near the tubes and output transformer as
practical; this capacity of course to have an
impedance (at the frequency used) low in
comparison with the tubes.
RADIO TRANSMITTING SETS
Brief descriptions and illustrations of pieces
of apparatus used in the application of
pliotron power tubes to various fields, fol-
low.
Description of Power Tube
In Fig. o is shown a radio transmitting set
arranged for radio telephony or either con-
Fig. 6.
Another Multiple Pliotron Set Arranged for Only
Three Wave Lengths
tinuous wave or modulated wave telegraphy.
Six Type P pliotrons are used, and for con-
tinuous wave telegraphy a maximum antenna
radiation of 12 to 15 amperes is obtainable in
a six-ohm antenna.
A plate voltage of 2000 is used, obtained
from a double commutator direct current
generator of 3J^2 kw. capacity. Because of
this high voltage the set is completely
screened in. To gain access to the set it is
mounted on wheels running on rails in the
supporting frame and when pulled forward
to the open position as shown in Fig. 5 all
circuits to the set are disconnected. This is
following the modern engineering practice
of removable switchboard panels of the truck
type.
The particular set illustrated is equipped
for five wave lengths, any one of which may
be instantly put into use by a shift of the dial
lever switch to the right of the panel.
It will be noted that there are four
adjustments for each wave length. These
adjustments correspond to those previously
outlined.
(1) Adjustment by change of antenna
inductance to the exact wave length desired.
(2) Adjustment of transformer ratio be-
tween the plate circuit of the tubes and the
antenna circuit.
(3) Adjustment of grid excitation voltage
by:
(a) Coupling of grid coil.
(6) Amount of inductance in grid circuit.
Means are also provided for conveniently
adjusting the value of grid resistance (Rg of
Fig. 4) which controls the value of normal
operating negative grid potential. A dial
switch is provided for this purpose which can
be operated from the front of the panel.
On the upper one of the three panel sections
are mounted various controls and instru-
ments. The plate current ammeters for
individual tubes are located in the oblong
protective cases. These protective cases are
used because these instruments are located
in the positive or "high side" of the plate
circuit. The other instruments are for plate
voltage, total plate current, filament voltage
and antenna radiation current.
The white circular disk to the right side
of the panel is a wavemeter condenser for
determining the wave length of the trans-
mitted energy. The transmitter for telephone
operation and the key for telegraphic opera-
tion are shown in operating position in Fig. 5.
This view also shows two of the individual
tube plate fu.ses mentioned in a previous
paragraph.
Fig. 6 shows another multiple pliotron
tube set, of simpler design, as it is equipped
for only three wave lengths and has no
52-i June, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, No. 6
adjustments from the front of the panel for
best operating conditions.
This view shows the method of tube
mounting in a spring suspended cradle so
that the set may be operated under conditions
of mechanical shock or vibration.
Fig. 7. Pliotron Panel for 50.000 Cycles
This view also shows three auxiliary pieces
of apparatus, the aluminum cell protective
device in the upper box, dry batteries for
transmitter excitation and normal grid volt-
age in the lower box, and the transformer for
filament lighting from an a-c. source with a
center tap connection to minimize the a-c.
potential effect.
Equipment for Production ot High Frequency
Energy
Two views of pliotron panel equipment for
supplving oO.UOO-cycle energy arc shown in
Fig- ■'^■
Instruments are provided for indicating
filament voltage, plate voltage, plate current,
grid current and high frequency output
current.
Means are provided for filament regulation.
a slight variation of output frequency, and
for an adjustment of grid excitation voltage
and ratio of transformation between tube and
load.
The dial switch for this latter adjustment of
transformer ratio can be seen on the left side
of the set in Fig. 7.
In this same view, at the bottom of the
left side, a spark gap is shown. This is for
protection of the resonance condenser of the
load against over-voltage.
If a large number of pliotron tubes are to
be operated in parallel, it is best to arrange
the design so that there are unit panels each
controlling a certain number of tubes. By
combining these panels an equipment of any
size can be provided for.
These unit panels contain only the pliotron
tubes and their individual auxiliary pieces
of control apparatus, the inductance capaci-
ties and generating equipment being separate
units.
Such a unit panel for six large tubes is
shown in Figs. S and 9.
Fig. 8.
Pliotron Unit Panels Containing Only Pliotrons
and Individual Control Apparatus
Referring first to Fig. S, each tube position
is equipped with the following devices:
(1) Plate ammeter, under a protective
cover.
(2) Individual filament rheostat, the con-
trol knob Ix-ing located below and
ELECTRON POWER TUBES AND SOME OF THEIR APPLICATIONS 525
slightly to the left of each plate am-
meter.
(3) A plug switch for throwing in or out
of circuit each filament.
A similar plug having the contacts
connected to an ammeter through
flexible leads is provided.
By this means the filament
current of any tube may be
individually adjusted
to the desired value. ^^^^^Bi
This ammeter is ^^^^t 'jjfiujjiBaA*
located in the center of the
panel and the plug is in its
holder near the lower right-
hand corner of the panel.
Referring now to Fig. 9:
(4) Individual plate fuses are pro-
vided. The white porcelain
fuse blocks showing plainly on
the back of the panel.
A typical power control panel for use in
connection with the above unit is shown in
Fig. 10.
From left to right the instruments are:
Filament volts, filament amperes, plate
volts (in hecto-volt scale divisions) and plate
Fig 11. Thirty-tube Pliotron Panel
Fig. 9. Rear of Panel Shown in Fig. 8
All filaments, grids and plates are each
connected in parallel and brought out to
terminal posts mounted at the rear of the
framework. Such units are used for either
amplifier or oscillator equipment.
Fig.
10. Power Control Panel for Use with Pliotron
Units Shown in Figs. 8 and 9
current in amperes. ^The left switch is for
the field circuit of the filament generator and
the right switch for the separately excited
field of the high voltage d-c. plate source
generator.
526 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
A high voltage plug switch is used in the
plate circuit to absolutely disconnect the
high voltage source to insure safety when
handling the circuits.
The lower left-hand rheostat knob is for
filament voltage adjustment and the right-
hand rheostat knob is for plate voltage
adjustment. In general, therefore, filament
control is on the left and plate voltage control
on the right.
On the rear of the panel is located the
rheostats, plug switch mechanism, fuses,
voltmeter resistances and terminals.
Power Tube Amplifier Equipment
A panel equipment of this type containing
30 pliotron power tubes is shown in Figs.
11 and 12. The tube panels are made up on
the unit plan previously described.
In Fig. 11 the power panel is the right-hand
one and contains instruments and control
apparatus for filament and plate sources of
supply.
In this equipment forced cooling is used
because of the large number of tubes in close
proximity. The blower and air ducts are
plainly shown in Fig. 12.
This equipment operates normally with a
direct current plate voltage of 2300, the total
plate electron current averaging about two
amperes.
Fig, 12 Rear View of 30 lube Pliotron Panel Shown in Fig 11
527
Artificial Daylight for Merchandising and Industry
By G. H. Stickney
Engineer Lighting Service Department, Edison Lamp Works, General Electric Company
The great difficulty in approximating daylight by artificial means is owing mainly to the indefiniteness
of the term dayUght. Daylight, as we know it, is a combination of direct sunlight and reflected skylight and
varies widely, depending on the degree of cloudiness, state of atmosphere, the angle at which sunlight enters
the atmosphere, etc. The closeness with which artificial light should approach average daylight varies for
different purposes, and the requirements are arranged in three groups by the author. The needs of the silk
dyer are most exacting and are closely followed by the requirements of the woolen industry. Other industries
invol'ving color matching come in the second classification; while the color matching needs of the merchant are
the least exacting and form the third class. The quality of artificial lighting demanded by the first class is
produced by expensive and inefficient color matching units. A more efficient light source of lower color accuracy
is provided by the daylight Mazda, and in conclusion the author mentions a number of specific applications
where modified light by means of this lamp can be employed to advantage. — Editor.
Introductory
In view of the lack of a general understand-
ing of the features of lighting for the purpose
of inspection or selection of colored materials,
it seems desirable to present a general review
of the subject.
For ordinary purposes the color of artificial
light is not highly important so long as it is
pleasing and does not depart too far from that
of daylight. In fact, a yellow tone in light is
often desirable, on account of artistically
pleasing qualities. On the other hand, there
are certain applications in connection with the
manufacture, inspection and sale of colored
materials where it is highly important that
they be viewed under an illumination that is
much closer to daylight in color than is the
illumination from ordinary lamps.
Since the appearance of colored objects
varies more or less with the color of the light
falling on them, it is important, especially for
artistic articles which are seen under daylight,
that color determinations and selections be
made under a light of daylight quality. This
is especially the case with garments in which
different materials, such as woolen cloth, silk
linings, braids and buttons, are combined, as
the apparent colors are not always affected
similarly or in the same degree by a change
of light. In such a case, parts which har-
monize under one light may clash under
another. It should be noted also in this con-
nection, when such articles are likely to
receive their principal use in the evening,
that the colors should be inspected under the
predominating artificial light.
Color of Natural Light
Although daylight is universal, and almost
as intimately experienced as gravity, few
realize how complex it is in its composition,
nor the extent to which it is subject to varia-
tion in intensity and color.
Light emitted by the sun traverses the
92, ()()(), ()()() miles of space with presumably no
apparent color change. However, on entering
the earth's atmosjihere it becomes modified.
Small particles of water, vapor, clouds and dust
in the atmosphere tend to deduct, especially
the short waves or blue rays, from the direct
sunlight. Part of this light is scattered and
received as skylight, so that we receive
a combination of direct filtered sunlight and
skylight. It is evident, therefore, that the
character of daylight, as we know it, depends
to some extent upon the state of cloudiness,
the angle at which sunlight enters the atmos-
phere, etc. In interior lighting other factors
enter: for example, the position of the sun
with reference to the window exposure, and
the color of nearby buildings which may be
reflecting light into a room. Fortunately
some of these factors tend to compensate for
each other, and the variation is ustially less
than might be expected.
While the variation of daylight colors is
sufficiently large to render accurate deter-
minations difficult, it mtist be remembered
that they are small compared with the
difference between average daylight and
unmodified light of practically all artificial
illuminants.
For accurate color matching purposes,
experts have always preferred the light from
the north sky — i.e., that from which direct
sunlight is always absent. The apparent
advantage of this is that it is subject to less
variation than any other natural light. Such
light contains more blue than average day-
light, and undoubtedly the latter would have
been preferred if it were obtainable as a fixed
standard. Artificial lighting can be and has
been produced which is more accurate as a
color standard, but stich lighting is expensive
and therefore practicable only where the value
of the accuracy is great or the areas to be
lighted small.
Demands for White Light
Observations of the practical use of day-
light lead to the conclusion that there is a
wide range of demands as to accuracy of
528 June, U)20
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. G
color matching. That the silk dyer needs an
accurate standard is evidenced by the pains
taken, even at considerable expense, to work
always under unobstructed north light.
The woolen industrv' apparently has a
slightly less exacting demand, and the cotton
industr\^ less yet, although compared with
most industries these (i. e., textile manu-
facturing), along with the manufacture of
celluloid, ivor\-, and a few other things, may
be classed in a separate group with exacting
demands.
Nearly all other color industries find day-
light from any direction acceptable, and ven.-
so that it is evident that store managers have
not recognized any such demand for color
accuracy as have the manufacturers. Thus, a
third grade is formed.
In artificial lighting, the production of day-
light color is usual!}- secured at the expense of
efficiency, the sacrifice depending upon the
degree of accuracy required. There is, there-
fore, a demand for several compromises,
depending upon the relative importance
given to these two elements, viz., cost of light
and color accuracy.
Since there is today a relatively small
demand for a highh- accurate light and a much
220
200
180
t)0
^ 140
W
120
0
■x> 100
Id
"o 60
1 —
A~.
\
s
^
\
\
/
,/
t
N
\^
■'"x
/'
f^
t
h>
3_
.Y
'
-— -
^'
H
y
1='^
:^.:.
<^
/
j:_
—
—
—
—
y
y^
y'
/'
\
^
40
20
y^
^y
>
>•
>
40
45
Fig. 1.
.50 .55 60 .65
Wave- Length. \.n A
.70
.75
Spcctrophotometric Curves of Typical Color Modifying Glass
^Energy Intensity for Various Wave lengths
A— Blue skv
B — Average Jaylight (black body at 5000° k.)
C — Mazda lamp at 19 lumens per wait (black body at 2850° k.)
D — Daylight Mazda lamp
E — .Accurate color matching type unit
F — Typical daylight enclosing globe
few take pains to eliminate colored light
reflected from buildings. These fall in a
second grade as to color accuracy.
When we come to the sale of all these goods,
we find a much lower standard of accuracy
acceptable. Many of the finest dr>--goods
stores are in the business centers, where the
surrounding buildings must necessarily modify
the light to a considerable extent and subject
it to variation through the more or less direct
reflection of sunlight from adjacent structures.
Still further is the light modified h\ window
shades and hangings, as well as wall finishes,
mahoganv furniture and other woodwork.
larger demand for a more efficient light, even
though less accurate, it is evident that
ordinan.- demands can readily be taken
care of by the modified light from the
incandescent lamp, especially since the
Mazda C lamp has made a much whiter
light .source available. Particularly for the
general lighting of large interiors, such as
salesrooms, efficiency and attractive appear-
ance, as well as color accuracy, are of great
importance, and the more accurate unit
used for localized lighting of small areas
will not so adequately meet the require-
ments.
ARTIFICIAL DAYLIGHT FOR MERCHANDISING AND INDUSTRY
529
Method of Modifying Artificial Light
In modifying the light of an illuminant
for color matching purposes, the best method
at present available is that of passing the light
through glass so colored as to absorb part of
those radiations which are in excess, so as to
produce approximately the same balance of
light rays as exists in daylight. This means
that the intensity and therefore the efficiency
is reduced. While a considerable modification
can be made with relatively little loss, further
correction involves much larger sacrifices. It
is no simple matter to produce glass suitable
for this purpose, since ordinary glasses absorb
too much of one color and not enough of
another.
It is not so difficult to produce a light which
appears white or displays a few colors correctly.
This does not seem to be understood by
many who are manufacturing and selling glass
for this purpose, so that it is necessary to use
considerable discretion in the selection of
color screens. The only safe method of deter-
mining the correctness of the glass filter is to
supplement careful spectrophotometric tests
by the practical demonstration of the light on
a large number of colored samples selected
throughout the range of colors.
Accurate Color Matching Units
There are two or three makes of color
matching units which employ Mazda C lamps
with colored glass screens very accurately
chosen to modify the light in such a manner
as to produce a mean between north skylight
%
ft
^
V
■-
\
-J
c
V
S3
\
■s
■"
-•
s
\
"v
J^
~
P. .40 45 SO 55 fcO .65 70 75
V/ave- LeriolK \n >
V B G Y O R
.TranSTn\ss\on curves of lypical color
Tnodxfyxng glass
^erccT\\ \ransmvss\on for various wave IcnOihs)
A- Glass plate accurate type un\t
B- Glass of DaylioKt Mazda lamp "bulb
C- Typical DayUgKl enctosmg globe.
Fig. 2. Transmission Curves of Typical
Color Matching Glass
and average daylight. Such a unit is rela-
tively expensive and inefficient, but this form
has proved satisfactory where the most exact-
ing color requirements are encoimtered. In
general, these accurate color matching units
♦ " Effect of Color of Walls and Ceilings on Resultant
Illumination," by A. L. Powell. G-E Review, March. 1020.
consist of a metal reflector arranged to con-
centrate the light downward through a blue
green glass filter plate. They are available
in sizes from 150 to 500 watt. Such fixtures
are ordinarily employed to light a table top,
or an area of 6 to S sq. ft. on which colored
Fig. 3. Accurate Color Matching Units for Industrial Purposes
Light from a Mazda C Lamp is Reflected Downward
Through a Glass Color Screen. The unit on
the right, of the angle type, produces high
illumination on vertical surfaces
material is inspected. They have been em-
ployed in commercial dye-houses, for cotton
grading, color printing, paint and ink mixing,
and many other industrial purposes. They
have also'been used in lighting color booths in
silk and dress goods departments of dry-
goods stores, haberdasheries and tailor shops.
As a rule such units are not sufficiently eco-
nomical for general lighting of stores. Some of
the fixtures are so designed that either arti-
ficial daylight or ordinary incandescent light-
ing can be obtained by the mere throw of a
switch, enabling the goods to be examined
under both conditions of use.
The Daylight Mazda Lamp
As before pointed out, there is a very large
demand for a more efficient, although less
accurate degree, of color modification. This
is obtained with the Daylight Mazda lamp.
This lamp is provided with a scientifically
determined blue glass bulb, while the fila-
ment is run at a higher temperature, giving
a whiter light than produced by the regular
Mazda C lamp. Daylight Mazda lamps are
made in sizes from 75 to 500 watt, and, while
the efficiencies of the various sizes are slightly
different, they correspond approximately to
those of the larger Mazda B lamps. The
lamps, therefore, are applicable to general
lighting.
As pointed out in the article on the "Effect
of Color of Walls and Ceilings on Resultant
Illumination,"* where Daylight Mazda lamps
are used in semi-indirect or other ornamental
glassware, it is important that the glass and
reflecting surfaces, such as the walls and
ceilings, be white; yellow tinted glass or room
530 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
finish tends to counteract the efifect of the bulb
thereby lessening the special advantage of the
light for color purposes.
In the case of the Da>-Hght Mazda lamp,
in spite of the popular opinion to the contra^y^
the blue bulb does not add anything to the
color of the light, but on the other hand sub-
tracts a certain percentage of the rays which
are predominant in the unmodified light.
It is evident from this that the efficiency
of the Daylight Mazda lamp is necessarily
lower than that of the regular Mazda C lamp.
This fact must always be borne in mind when
installing Daylight Mazda lamps, since it
will be necessary- to use approximately 35 per
publicity and injure other similar products
which have considerable merit.
This situation is ameliorated, however,
b}' the fact that many merchants who think
they desire color matching illumination really
are better off with a light of yellowish tint
which has the advantage of producing a more
cheerful appearance in the store.
In conclusion, it may be noted that, while
there seems to be a general impression that
an exact duplication of daylight color is
needed in lighting, there is in reality a wide
diversity of requirements, relatively few of
which include a high degree of accuracy.
Even where colored materials are handled and
Fig. 4.
A Counter Type Color Identification Unit Showing the Application in Merchandising
Greatly Simplifying the Selection of Colored Materials
cent more wattage than when clear bulb
Mazda C lamjjs are installed to obtain the
same ilhunination.
Color Modifying Globes
Besides the units already described, there
is on the market a considerable variety of
enclosing glassware sold for color matching
purposes. Some of these equipments have
considerable merit. The majority have little
advantage beyond an api)arent whitening of
the light. Some, however, while having a
considerable absorption, actually lessen the
color matching value of the light of the
Mazda C lamp. Unfortunately, there are
manufacturers of such glassware who claim
in their advertising "perfect color matching
effect." Such claims tend to discredit all such
sold, the common illuminants wnthout color
modification meet the large majority of cases.
Specific Applications of Modified Light
.S^)rt•,v.- -Daylight Mazda lamps find a
wide field of api^lication in store lighting.
This type of illuminant will never supplant
the regular Mazda C lamp for general illu-
mination ; for the public, as a whole, prefer the
somewhat wanner hue of the latter, as it
makes the store look cheerful and inviting.
Prominent merchants have stated that, in
over !)() percent of their sales, color matching
is not an element. They feel it desirable to
sell goods under the conditions most favorable
to their best appearance. Many of the
finest garments will be worn at night and arc
designed to be most attractive under average
ARTIFICIAL DAYLIGHT FOR MERCHANDISING AND INDUSTRY
531
night illumination such as furnished by the
regular Mazda lamps. Nevertheless, there are
certain goods which should be lighted with a
nearer approach to daylight than furnished
with clear bulb lamps. In this class fall those
which would be worn largely out of doors.
Daylight Mazda lamps used for general
store illumination give a distinctive appear-
ance to the store which has a distinct advertis-
ing value. In fact, some stores so lighted
make it a feature in their newspaper and other
advertising calling attention to the "Day-
light" store. Even though Daylight Mazda
lamps are not used for general illumination
throughout, there are certain departments
which will require such a light to display the
goods to the best advantage, for example,
men's clothing (particularly blues and blacks),
linens, which appear pure white rather than
slightly yellowish, furs, jewelry, silks and shoes.-
Many of the most progressive stores in the
country have supplemented the general
illimiination with the accurate type of color
Fig. 5. A Color Identification Unit Placed
Above the Triplicate Mirror in a Clothing
Establishment. Confidence in pur-
chasing garments is much greater
when artificial daylight is
available
matching units over the counters and in other
parts of the store where color matching is an
important element. Larger types of these
same units are employed over the triplicate
mirrors in the clothing department with very
satisfactory results.
These units provide local lighting of the
high intensity suitable for the critical exam-
ination of colored fabrics. It is not necessary
to enclose such devices in a booth, for the
amount of light directly beneath the fixture
is so much greater than the illumination
necessary for the store as a whole that the
mixture of color of light does not affect the
result. Where accurate color comparison
units are installed the customer can have
absolute confidence in his judgment. These
save a great deal of the clerk's and cus-
tomers' time by avoiding the necessity of
carrying merchandise to the doorway or
window for inspection, and are a decided
economy in store operation.
Show' Windows. — Daylight Mazda lamps
in the show windows cause them to stand out
prominently in comparison with other forms
of illumination. While they are not neces-
sary as standard show window equipment,
every store of any appreciable size should have
a complete set for at least one window, so that
when displays requiring such a quality of
light are in place they may be employed.
Laundries. — Spots and stains on goods are
usually of a brownish or yellowish tint. If the
light source is rich in yellow, then these
blemishes tend to fade into the white back-
ground and are hard to detect. Daylight
Mazda lamps in steam laundries over the
folding and inspection tables enable the
operator to catch many an improperly
cleaned piece and thus keep up the standard
quality of the work.
Textile Mills. — Without doubt, most of the
processes here are carried on without regard to
color. After the warp is made up and the
shuttles filled with thread of the proper
color, operations proceed almost automati-
cally, but there is the liability of the mixing
of color which may throw out an entire piece.
Daylight Mazda lamps are being installed
over looms in a number of instances to avoid
this difficulty. It must be borne in mind
that the color of light produced by these lamps
is not accurate enough for color matching as
the term is used in the dye-house, and the
more accurate units should be recommended
for such work. Color matching devices are
also very desirable in the final inspection
departments, bleacheries and show-rooms.
Concentrating Plants. — Iron and zinc streaks
must be recognized as they are found on the
concentrating tables. The zinc is of a rusty
gray color and the iron of rusty red brown
color. In a mill lighted by ordinary types of
lamps it is difficult to tell these colors apart,
532 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
yet it is necessary to separate most of the iron
ore from the zinc. Some of the mines in the West
have installed Daylight Mazda lamps for local-
ized lighting over the tables for this purpose.
Chemical Laboratories and Sugar Refineries.
— Daylight Mazda lamps have given quite
satisfactors- ser\'ice in illuminating the centrif-
ugal machines in iise in the laboratory. They
assist in the discriminations of color necessarv-
for the grading and matching of cane sugar.
Some prominent chemical manufacturers are
employing Daylight Mazda lamps for watch-
ing changes of color in their testing depart-
ments, particularly in connection m-ith titrat-
ing. There are also nimierous places about
Printing. — It is quite difficult to detect the
yellow half-tone from the white background
when illuminated by regular Mazda lamps,
and other colors are not shown in their true
relation. Some of the large lithographing
companies are using Daylight Mazda lamps
to light the receiving end of the presses and
also in the proof room and artist's workshop.
The accurate type of color matching units,
however, are particularly adaptable to the
final inspection and for the absolute assur-
ance of satisfacton,- night work. By the
proper use of modified light, overcast and short
days put no check on the art or press work in
the lithographing plant.
Fig. 6. A Color Identification Unit in Which a Large Mazda C Lamp Housed in the Center Produce*
Accurate Color Matching Light After Passing Through a Large Circular Color Screen at
the Bottom. Smalt Mazda B lamps at either side provide a convenient means
of comparing artihcial lighting color effects with daylight
the chemical plant where verA- definite de-
mands for constant north sky color quality
exist, and with the devices mentioned this
can be satisfactorily met.
Photographic Supplies. — Some manufactur-
ers of photographic materials are employing
Daylight Mazda lamps for examining the
quality and color of prints; correct shades of
blues, blacks and sepias can be deicnnined
far more readily with this tyjie of illuminant.
Daylight Mazda lami)s are also used for micro-
scopic and lower power photomicography anti
in connection with the spectrograph for de-
termining color sensitiveness of photographic
emulsions. Where a high degree of accuracy
is required the accurate type of color modify-
ing device is essential.
Cigar Fact jries. — Cigars are graded accord-
ing to shade, and minor differences in color
must be detected in ins])ecting and sorting.
The Daylight Mazda lamp assists in this work.
A high quality of product can be obtained in
the factory lighted with these lamps. This
factor alone offsets the slight additional cost
of operati(m.
Miscellaneous fnJustries. — Other fields in
which modified light has proved useful are oil
refineries, where it is necessary to determine
the difTerence between grades of oil; in fruit
packing houses, where oranges and lemons ar«
sorted according to color as well as to size; in
paper mills, where the sample room is so
illuminated; in the jewelry trade, for the criti-
cal examination <if stones such as diamonds
ARTIFICIAL DAYLIGHT FOR MERCHANDISING AND INDUSTRY
.533
and pearls ; in metal working, for the selection
of brass by color; and in miscellaneous places
about flour mills, rubber goods and garment
factories, button factories, potteries, paint
factories, etc
Medical. — The Daylight Mazda lamp has
proved quite a boom to the medical profession,
in the chemical laboratories and assisting in
microscopic examination. For diagnosis of
skin disease and retina examinations, promi-
nent specialists have employed this form
of illuminant. In the operating room, the
various tissues are revealed more accurately
when examined under the light of the Dayli,ght
lamp — as, for example, when operating on a
jaundiced patient whose tissues are yellow and
whose blood gives all the tissues of the body
a yellow tint a yellow light would be unsatis-
factorv'. During operations for gall stones,
yellow bile ducts, red arteries and blue veins
must be distinguished one from the other.
One of the first uses for which the Daylight
Mazda lamp was placed was the examination
of X-ray negatives. A suitable light for this
purpose is necessar^^ for. after the plate is
developed, it is necessary to inspect this ver\^
carefully to determine the ailment or discover
the fracture. The negative is in general
illviminated by ven,' diffused light from the
rear, and experience has proved that the
whiter the light the greater the ease of the
examination. Diffused light for this purpose
is obtained b}- using what is practically
indirect ilkunination. A box or frame to hold
the negatives is painted flat white on the
interior surface. Lamps are concealed from
view and equipped with reflectors to direct the
light on this white background. From here it
is reflected to an opalescent ground glass plate
covering the opening or mouth of the box.
In dental work the Daylight Mazda lamps
used for general illumination of the office or in
the concentrating spot lamp assist materially in
detectingdecayedspotsand diseased conditions.
Many accurate color matching units are used
by the dental supply companies for the match-
ing, grading and sorting of artificial teeth.
Art Galleries and Museums. — Daylight
Mazda lamps are used in many instances
with splendid results for illuminating paint-
ings. The artist paints his pictures under
natural light; he places the colors on his
canvas with particular relation to each other.
Each small area of the picture is blended with
the next and viewed as a whole. If the color
of light is such as to materially modify the
relation between these various areas of the
picture, then his object is defeated. Portions
may be intensified: others dulled. The better
the painting, the greater the demand for suit-
able lighting. Many of the art exhibits
throughout the country- are visited at night
bv the general public, and for this reason the
question of the correct artificial illumination
is of much importance.
Hotels. — The field of apj^lication of Day-
light Mazda lamps is far broader than one can
imagine and it would be out of the question to
attempt to enumerate all the various applica-
tions for which they find use in the hotel.
If the sample room is fitted with Daylight
Mazda lamps, then the critical examination
of goods on display is facilitated. In the linen
department they enable the help to readily
detect spots on tablecloths and napkins.
Over the cigar counter they present the dis-
plav in more nearly its true value.
BIBLIOGRAPHY
The following list indicates some of the
leading articles on the general question of
artificial daylight and kindred subjects which
have appeared in the technical magazines
during the past few years :
"Color Values of Light from Electric Lamps,"
G. H. Stickney, Trans. I. E. S., VoL 5, p. 431.
•• Daylight and Artificial Light," E. L. Nicols,
Trans. 'I. E. S., Vol 3, p. 201.
" Daylight Efficiency of Artificial Illuminants," H.
E. Ives", Bulletin Bureau of Standards, Vol. 6, p. 231.
"The Subtractive Production of Artificial Day-
light," H. E. Ives and M. Luckiesh, Electrical World,
Mav -t, 1911.
" The Relation Between the Color of the Illuminant
and the Illuminated Object," H. E. Ives, Trans. I.
£.5., Vol.7, p. 62.
"Color of Illuminants," L. A. Jones, Trans. I. E.
.S., Vol. 9, p. 687.
"Artificial Davlight: Its Production and Uses," M.
Luckiesh and F. E. Cady, Trans. I. E. S., Vol.9, p. 839.
" Development of Daylight Glass," E. J. Brady,
Trans., I. E. S., Vol. 9. p. 937.
"Some Data on Artificial Daylight Units," C. H.
Sharp, Trans. I. E. S., Vol. 10, p. 219.
" AGaseous Conductor Lamp for Color Matching,"
D. McFarlan Moore, Trans. /. E. S., Vol. 11, p. 192.
"Colored Glass in Illuminating Engineering,"
H. P. Gage, Trans. I. E. S., Vol. 11, p. 10.50.
"Commercial Uses of Filtered Light," C. E.
Clewell, Electrical Review, April 7 and 14, 1917.
" Mazda C-2 Lamps and Their Applications,"
A. S. Turner, Jr., National Electrical Contractor,
October, 1917.
" The Color of Artificial Light in the Medical Pro-
fession," A. L. Powell, Medical Times, October, 1917.
" Mazda C-2 Lamp and Its Applications," A. L.
Powell, Electrical Engineering, December, 1917.
"Artificial Daylight in the Industries," M.
Luckiesh, Electrical World, September 28, 1918.
"A Color Symposium," Trans I. E. S., Vol. 13,
p. 1.
"Symposium on Camouflage." Trans. I. E. S.,
Vol. 14, p. 216.
"The Daylight Mazda Lamp in the Psychological
Laboratory," G. J. Rich, American Journal Psy-
chology, July, 1919.
534 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, \o. 6
Enclosed Carbon Arc Lamps vs. Novalux Mazda
Units
By H. E. Butler
Illuminating Engineering Laboratory, General Electric Company
During the period of the war and during the coal strike, street lighting was curtailed in this country in the
interests of economy. Two very important truths have been revealed as the result of this experience: first, that
reduced street lighting is accompanied by an increase in crime and accidents; second, that an appreciable waste
of power exists in street lighting systems because of the use of the enclosed carbon arc lamp. The author takes
up this second matter and by detailed analysis shows the fallacy of the policy of keeping in ser\-ice obsolete
forms of lighting units. The author's conclusions are accepted by the most progressive central stations in the
country, but there still appears to exist an opportunity for education among others in the industry. — Editor.
A few years ago it was estimated that
approximately a half million series enclosed
carbon arc lamps were in use for street lighting.
This number has been materially reduced
during the past few years, but there are
still in use throughout the country' many
of these obsolete lamps. It is for the purpose
of illustrating the inefficiency and poor
economy of these lamps that the following
data have been prepared.
The enclosed carbon arc lamp was devel-
oped to overcome the disadvantages of the
open arc lamp; viz., short life of carbons and
poor distribution of light on the street surface.
However, the long life of the carbon trims
and the steady and relatively uniform dis-
tribution of light from the enclosed lamp were
obtained at a sacrifice of efficiency. These
characteristics are now available in the
luminous arc lamp and Novalux fixtures but
at an efficiency very much greater than existed
in the earlier lamps. The luminous arc lamp
is a more efficient producer of light ; however,
as this article is to cover data only on incandes-
cent lamps and obsolete enclosed carbon arcs
the luminous arc lamp will not be discussed.
In Figs. 1, 2, 3 and 4, the enclosed carbon
arc lamp is compared with the 250, 400, 600
and lOOO-c-p. series Mazda lamps respectively.
These charts indicate the relatively low
efficiency of the carbon arc as compared with
the incandescent equipments. Another fea-
ture which shows the Mazda lamp to advan-
tage is the rated life per trim for the carbon
arc as compared with the rated life of the
Mazda lamp. It is well, however, to remem-
ber that, while the life of a series incan-
descent lamp is rated at 1350 hours, it is not
economical to allow the lamp to remain
in service without attention until its life is
spent. Periodic visits should be made for the
purpose of cleaning the equipment; otherwise
a serious loss of light will result from the
accumulation of dirt on the glassware. It
will be further noted from these charts that
a wider range of lamp sizes and a greater
choice in types of light distribution are avail-
able with the Mazda lamp than with the
carbon arc lamp. The illumination data in
this article comprise a picture of the unit, the
candle-power distribution curves, and the
calculated illumination curves; the latter
give the average and minimum foot-candles
and the uniformity of illumination for various
size lamps and spacings. The uniformity fac-
tor is the ratio of minimum to maximum foot-
candles.
From these cur\-es, the following are
some of the questions that may readily be
answered :
With a specified spacing and equipment,
what will be the average and minimum
illumination and the uniformity on the
street ?
Taking equal average or minimum intensity
of illumination as a basis, what will be the
spacing required for the various units with
their dilTcrent equipments?
From this and the wattage, what is the
power consumption per linear foot?
Taking eqtial uniformity of illumination as
a basis, what will be the sjjacing required?
The candle-power distribution of the en-
closed carbon arc lamps are shown in Fig. 5.
together with cun-es giving the average mini-
mum illumination and the ratio of the mini-
mum to the maximum foot-candles for various
spacings on the street. The alternating-
current series enclosed carbon arc lamp is less
efficient than the direct-current enclosed
carbon arc lamp because the alternate
cooling of the electrodes in the former
lamj) causes an additional loss of heat and
consequently a lower temperature of the
carbon points. It is interesting to compare
the photometric cur\-e in Fig. 5 with those of
the incandescent equipments in Figs. 7, S, and
9. The angle of maximum candle-power in
ENCLOSED CARBON ARC LAMPS VS. NOVALUX MAZDA UNITS
535
^ -
I;
»-*Trp
^r
-co^
XX^r
c^ciM' a£/sM ot/rrr*
^zr^^r
^nt^ZCO CO,
njrv<gSr£M
iXJOM^-iS
X^l-
■n^ivasTSAi'
^AtitX Cl^^^^
rijjvtsaf'^'v
-.M^i-C ^tx
roi^G sr^/f
waay^f^
lii \ ...Jm
Fig. 1. Obsolete Series Enclosed Carbon Arcs Compared with Novalux Unit Equipped with 250-c.p. Series Incandescent Lamp
^
iff
,0-0
J
"C+SJ^ '
BSfejF
-CD-
oa^e> c«.*MB oc e*^£.o3^e
-CS:^
7LJVG JTVV
nCD-^
TuiVISa TC'-y'
^„u« cj' agJ ^^M*i:er^ 7£a
yonnssT^ft/
uja n^.s /.xio ^^»
ru^aare-zv
^^.^M^i^ma
Tl*V(SJ7-ifA'
T-4**'*^ i^anMS *e*i«.i^ 'BWK*'
II \ jjB
Fig. 2. Obsolete Series Knclosed Carbon Arcs Compared with Novaluz Unit Equipped with 400-c.p. Series Incandescent Lamp
The lo-amp. 400-c-p. Mazda series lamp is operated from an auto-transformer
1
536 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
I-
4"'
■* s
■<J
t
li
— d^-
, 4f>*x.c^^es^
"^fj^S^^^
T^
c eivc^am^f
di^Mlc zi.rA/? cyuropa.
I
1^1 I
Fig 3. Obsolete Series Enclosed Carbon Arcs Compared with Novalux Unit Equipped with 600-c.p. Series Incandescent Lamp
The 20-amp. 6()0-c-p. Mazda series lamp is operated from an auto-transfornier
^3
-.3
■t
t
-^^
Tj:-
III ...jfl
Fig. 4. Obsolete Series Enclosed Carbon Arcs Compared with Novalux Unit Bquipped with 1000-c.p. Series Incandescent Lamp
The 20-amp. 10(H>-c-p. Mazda series lamp is operated from an atiu^transformer
ENCLOSED CARBON ARC LAMPS VS. NOVALUX MAZDA UNITS
537
Fig. 5a. Enclosed Carbon Arc Lamp with Light Opal Glass
Inner Globe, Clear Glass Outer Globe, and Street Reflector
the incandescent units is brought closer to the
horizontal, and therefore those rays which
have to travel the greatest distance possess
the greatest intensity. This results in greater
Fig. 5b. Initial Distribution of Candle-power in a Vertical
Plane of the Unit Shown in Fig. 5a
Curves .-1. B, and C correspond to the lamps named in Fig. 5g
uniformity of illumination on the surface of
the street, as may be seen by the curves of
the ratio of minimum to maximum illumina-
tion. On residential streets, parkways, and
O.J
«1
I
K
0
?
In
r
O./S
\
\ o./
5
K
0
1
i O.OS
%
n
1
0,4
ll -
1
0
I
o/
,
,
^
—
—
—
— ■
—
—
—
—
—
—
.
—
—
—
—
—
—
—
—
—
—
—
^
—
—
—
—
—
—
—
—
,
1
—
—
—
—
—
—
—
—
—
L
1 —
—
—
—
—
1
—
—
—
—
—
—
—
—
1_
—
,
I
4-
' —
—
—
—
—
— ■
—
—
\
-L
—
— ,
—
—
1
1
\
~\
— '
— 1
—
—
—
—
—
—
—
_
1 —
,
—
—
—
—
—
—
—
—
\^
A
\
1
\
\
—
—
—
—
1 — 1
—
—
—
—
\
1
\
.^
__
^
;
,^
^
k,
1
,.-'
1
' —
' —
—
' —
■ —
—
'
' —
—
\
\
—
i-
—
—
—
—
,
^
\
c
\
V
—
V
—
—
—
—
' —
' —
<j
\
>e
\
—
_\
^
ac
ik
t£l
at
-Ei
.
\
V
\
y
—
—
V
—
—
' —
—
' —
1 —
\
^
»~
L^^
\
—
—
A
,
,
1
i
N
^\
^
V
f—
■^^
^~
S
""■
^
^~
^~
H
^
'^
is
"^
~
—
—
' —
_\
V
u
—
—
—
—
s
—
^
—
—
tl
**
,^
~-
r
\
s
—
V
a
s
—
—
—
—
s
—
—
•-
^ ,
-^
-
—
—
s^
_i
_
L_
^='
—
■ — •
—
—
—
—
—
—
'
joo
zoo <9oo
Fig. 5c. Calculated Illumination Values on Street Surface Along Center Line of Street
Lamps on one side of street only, on 4-tt. bracket arm. Height, 35 ft. Width of street, 50 tt.
A: 6.6-amp. d-c. series enclosed carbon arc lamp B: 6.6-amp. a-c. series enclosed carbon arc lamp
C: 7.5-amp. a-c. series enclosed carbon arc lamp
53S June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
<
1^
<?
i
^
1
\^
^
/^^ ?^
)wc\i
-V
r
^
4
^
■^^Stf
~i>^
-^^^"■■^
1 *S5**«
^«»
^
Wf
^'fe
K
,^
-^
5^
kH
t$j
\XW5
if<x.
/^~-~
y^
^
yTy/yj/
\\^>c\
\, , /
^ --^*V_^
"A )
NySCy/
^
s
1
^
\y\y
o^
-J_ /ooo __
4^
>\y^/
Fig. 6a. Novalux Pendent Unit with Diffusing Glass Globe
Reflector Used with 250-. 400-. 600- or 1000-c.p.
Mazda Series Lamp
boulevards where it is possible to space the
units close, diffusing glassware will give
satisfactory illumination and eliminate the
glare which is so often experienced with other
equipments, and also more ornamental and
pleasing effects will be obtained. For such
Fig. 6b. Initial Distribution of Candle-power in a Vertical
Plane of the Unit Shown in Fig. 6a.
Cunes E.f.GandH correspond to curv-es .4. B.C.and DinFig.6c
ser%'ice a distribution similar to that shown
by Fig. 6 is most suitable. It is not possible,
however, to make general statements covering
the use of illuminating glassware, as each
problem requires a carefuJ study of existing
conditions.
^ Soo
.^oo
■300
eoo
/c»o coo
Fig 6c. Calculated Illumination Values on Street Surface Along Center Line of Street
Lamps on one side of street only, on 4-ft. bracket arm. Height. 2."> (t. Width o( street. 60 ft.
.1. B, C. D: Diffusing Glass Globe and Steel Reflector
1000-c-p., 20-amp. Mazda Series Lamp, PS-40 bulb B: 600-c-p.. 20-amp. Mazda Series Lamp. SP-40 bulb
400-c-p., 15-amp. Mazda Siiries Lamp, PS-40 bulb
2o0-c-p.. 6.6-amp. Mazda Series Lamp. SP-.i.'i bulb
ENCLOSED CARBON ARC LAMPS VS. NOVALUX MAZDA UNITS
539
Fig. 7a. Novalux Pendent Unit with Bowl Re-
fractor and Reflector Used with 400-, 600-.
or lOOO'C.p. Mazda Series Lamp
Fig. 7b. Initial Distribution of Candle-power in a Vertical Plane of the Unit
Shown in Fig. 7c
Curves A, B and C correspond to the lamps named in Fig. 7c
The general order of illumination intensities
for utilitarian street lighting where pendent
type lamps are appropriate is as follows:
Average Hor. 111. in
Foot-candles
Important side streets 0.10 to 0.25
Residential streets 0.01 to 0.05
Suburban roads 0.005 to 0.01
The most useful basis of comparing
illuminants is, perhaps, the total lumens or
light flux delivered by lamps as it is usually
possible by selecting proper refractors, reflec-
tors, and glassware to distribute the light in
the manner most suitable for the particular
conditions at hand. These lumen values are
Fig. 7c. Calculated Illumination Values on Street Surface Along Center Line of Street
Lamps on one side of street only, on 4-ft. bracket arm. Height. 2.5 ft. Width of street, 60 ft.
A, B, C: Prismatic Glass Bowl Refractor and Steel Reflector
.•I.- 1000-c-p., 20-amp. Mazda Series Lamp, PS-40 bulb B: 600-c-p., 20-amp. Mazda Series Lamp. PS-40 bulb
C: 400-c-p., 15-amp. Mazda Series Lamp, PS-40 bulb
540 June. 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
Fig. 8a. Novalux Pendent Unit
with Dome Refractor and Stip-
pled Glass Globe* used with
250-. 400-. 600- or 1000-c.p.
Ma2da Series Lamp
given in Fig. 10 for the following lamps which
are considered in this article:
ENCLOSED CARBON ARC LAMPS
C.6-anip. a-c. enclosed carbon arc, light
opal inner, clear outer, standard reflector.
7.5-amp. a-c. enclosed carbon arc, light
opal inner, clear outer, standard reflector.
Fig, 8b Initial Distribution of Candle-power in a Vertical Plane of the Unit Shown in Fig. 8a
Curves .4 . B. C and D correspond to the lamps named in Fig. 8c
G.G-amp. d-c. enclosed carbon arc, light
opal inner, clear outer, standard reflector.
NOVALUX UNITS WITH MAZDA LAMPS
Form tt pendent Novalux unit with
250, 400, 600, and lOOO-c-p. Mazda series
lamps, diffusing gkibe with and without
reflector.
^-yw»ct5 oA/ o/\t£ ^/o£ or- s-neccr
o/\/ -^rr o/e*^cfecr ^^je*^ /t£/^/^t zsn . v^iDm cv ^tif^et ^o ft.
1
I,
—
~
—
—
—
—
—
-^
oj
—
— h
—
—
-1
—
o.s
—
^
1
'-^-
—
—
—
— .
— .
_
— ,
8
1
1
1
1
1
o^
—
—
—
~
—
—
—
—
—
—
— '
—
—
—
—
—
—
—
-
e-
—
^
—
-
^
—
—
-
—
-
—
-
—
-
—
—
—
-
^
—
^
—
—
-
-
—
~
-
-
-
-
_J
-
-
-
—
—
^
y_
—
—
—
-
zi
.
=
-1
1
-
-
-
-
=
-
u
k
\
I
_J
—
-
-
—
r
—
—
\
-
=
Zj
-
-
1 ^^
1
^
g
-
—
-^
-^
—
—
j\
-
-
-
—
^
—
—
B
N
V
—
—
-V
^
^
-
=
-
=
-
\
•*
ia
•c^
3
*A
r/S
■«<!r.
—
^
^
%
-S
^
^
^
—
—
—
-S
-^
t-
-^
-
-
-
-
1
-
\
L
-
—
— ,
S
—
s;
li
:s
"S
—
I
^
-
-
-
-
-
I
-
-^
\
y
-
-
—
^
s
>^
<»h
_
J
-\
<-
r
-
-
-
-
-
T
\
-
-
N
\
-
-
-
^
^
v>
5
■^
zj
^
_
—i
K
A
A.
\
A
^
-
-
-
«u
-
-
^
^
—
—
.^
-
^
::;:;
-
-S
2[
\
^
5
v
s
^
-
-
-
S
S-
m
a.
—
—
E
—
—
—
-
!!>
*:
^55
».
^
r
-
-
-
—
-
wo
JOO
too
JOO
lOO
too
lOCt
Fig. 6c. Calculated Illumination Values on Street Surface Along Center Line of Street
Lamps on one side of street only, on 4-ft. bracket arm. Hcij^ht. 'J.'j ft. Width uf street. 60 ft
A, B. C. D: Prismatic Glass Dome Retractor and Stippled Glass Globe*
A and E: 1000-c-p.. 20-amp. Mazda Series Lamp. PS-IO bulb B and C. 600-c-p.. 20-amp. Maida Series Lamp. PS-W bulb
C and C. 400-c-p., 15-amp. Mazda Series Lamp. PS-40 bulb Wand H. 250-c-p., 6.6-amp. Maida Series Lamp. PS-35 bulb
*Tbe illumination on the street will be practically the same for a rippled glass globe.
ENCLOSED CARBON ARC LAMPS 15. NOVALUX MAZDA UNITS
Fig. 9a. Novalux Pendent Unit with Prismatic Refractor and
Reflector Used with 250-, 400-, 600- or 1000-cp.
Mazda Series Lamp
Form 6 pendent Novalux unit with 250,
400, 600, and lOOO-c-p. Mazda series lamps,
equipped with bowl refractor and reflector.
Form 6 pendent Novalux unit with 250,
400, 600, and lOOO-c-p. Mazda series lamps,
equipped with Holophane prismatic dome
refractor and stippled glass globe.
Fig. 9b. Initial Distribution of Candle-power in a Vertical ] ;
Plane of the Unit Shown in Fig. 9a
Curves A, B,C and D correspond to the lamps named in Fig. 9c
Form 6 pendent Novalux unit with 250,
400, 600, and 1000-c-p. Mazda series lamps,
equipped with Holophane prismatic band
refractor and reflector.
The comparative figures of lumens reveal
the inferiority of the enclosed carbon arc.
The best that can be obtained from the en-
JCoo
.jra
p
^
__
_^
-_
M-
«t-
—
—
—
H-'
,
0
\
^
^ -
\
v.
s
\,
\
s
\
,
,
,
—
/00 /too J0o y*oo i^oo
Fig. 9c. Calculated IHuiiiination Values on Street Surface Along Center Line of Street
Lamps on one side of street only, on 4-ft. bracket arm. Height, 25 ft. Width of street. 60 ft.
A.B.C, D: Prismatic Glass Band Refractor and Steel Reflector
JOOO-c-p., 20-amp. Mazda Series Lamp. PS-40 bulb B: ,60a-c-p.. 20-amp. Mazda Series Lamp. SP-40 bulb
.400-c-p., 15-amp. Mazda Series Lamp, PS-40 bulb
D: 250-c-p.,6.6-amp. Mazda Series Lamp. SP-35 bulb
542 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
closed carbon arc lamps (direct-current series
type) is 7.4 lumens per watt as compared with
10 to 14 lumens per watt for the incandescent
units, depending upon the size of lamp and
equipment. A general idea of the relative
appearance of an incandescent and an
enclosed arc installation may be had from
Figs. 11 and 12. These are actual photo-
graphs and illustrate very forcibly the non-
unifonn illumination on the street surface
from the arc installation as compared with
the Novalux system.
Table I shows the electrical data relating
to the several systems considered in this
article and also the saving in power and the
relative power cost data of the carbon arc and
incandescent systems. An examination of
these data, together with the photometric
data previously referred to, will convince the
most skeptical that it is poor economy to
continue operating the enclosed carbon arc
lamp in lieu of the more modem incandescent
equipments. Due to the advancement in the
efficiency of street illuminants and equip-
ments, it is possible to maintain higher
standards of illumination than was possible
with the series enclosed carbon arc lamp
for the same expenditure of money. W.
D'A. Ryan, Director of the Illuminating
Engineering Laboratory, has advocated higher
TABLE I
ELECTRICAL DATA
Type of Unit
Line Amperes
Lamp Amperes
Volts at Lamp Terminals
Watts at Lamp Terminals
Line Loss
Efficiency of Constant-current Transformer
Efficiency of Brush Arc Generator Driven
by Synchronous Motor
Combined Efficiency
Watts Supplied at Switchboard 468
Hours Lamps Burn Each Year 4000
Kw-hr. Consumed per Lamp per Year . . .
Total Lumens per Lamp:
With light opal inner plobe, clear outer
globe, and street reflector
With light Carrara globe and reflector. .
With bowl type Holophane prismatic
refractor and reflector
With Holophane dome refractor and
stippled glass globe*
With Holophane band refractor and
reflector
6.6 Ampere
7.5 Ampere
6.6 Ampere
Form 6
Form 6
Form 6
A-C. Series
A-C. Series
D-C. Series
Novalux
Novalux
Novalux
Enclosed
Enclosed
Enclosed
250 C-P.
400 C-P.
600 C-P.
6.6
7.5
6.6
6.6
6.6
6.6
6.6
7.5
6.6
6.6
15
20
77
77
(0
23.4
37.1
51
425
480
495
155
234
340
5%
5%
5%
5%
5%
5%
96%
96%
86%
96%
96%
96%
91.2%
91.2%
81.7%
91.2%
91.2%
91.2%
468
527
605
170
257
373
4000
4000
4000
4000
4000
4000
1872
2108
2420
680
1028
1492
1810
2170
3645
1725
2750
4070
1670
2820
4225
1913
3060
4580
1780
3040
4510
* The illumination on the street will be practically the same for a rippled glass globe.
SAVING IN KILOWATT- HOURS AND MONEY PER LAMP PER YEAR BY REPLACING
ENCLOSED CARBON ARC LAMP WITH NOVALUX UNIT AND MAZDA SERIES LAMP
(.Power at 1.5 Cents per Kw-hr.)
6.6 Amp. A-C. Series Enclosed Carbon Arc
7.5 Amp. A-C. Series Enclosed Carbon Arc
6.6 Amp. D-c. Scries Enclosed Carbon Arc
rORM 6 NOVALUX
250 c-p.
400 c-p.
Kw-hr.
1192
1428
1740
Money
$17.88
21.42
26.10
Kw-hr.
844
1080
1392
Money
$12.66
16.20
20.88
600
c-p.
Kw-hr,
Money
380
616
928
$ 5.70 ■
9.24
13.92
The 1000-c-p. lamp is not oonnxjeped' i« this ta-ble, as its size confines it very larL'clv to hiiih intinsitv White
Way lighting'-wher«-local'<XJn<Jition9.pr«clude instilling the luminous arc lanii
ENCLOSED CARBON ARC LAMPS VS. NOVALUX MAZDA UNITS
543
standards of illumination for many years and
has shown the many advantages in their use.
His recommendations of higher standards of
street illumination have been accepted by some
of the largest cities in this country and in
foreign countries and they have installed
street lighting installations which have more
than doubled the old standards of intensities.*
The simplicity, flexibility, and efficiency for
the incandescent system extends beyond the
lighting unit itself to the station equipment
re^/VDEfJT W^IT WITH
LIGHT C^/TYT-iff'i OLOSe
N»B7..ZO'ST££i. /f£FL£CT0/f
l=^/^0£fvT U/VIT HlTH
ST'>'f*Ljeo aiAsa otoae
l^£.'^D£fjT OH IT MTH
so area, ^e^t-ccnj^
vKS
Fig. 10. Light Flux Values of A-C. and D-C. Series Enclosed
Carbon Arcs and Novalux Series Units for Street Lighting
where we find a variety of apparatus suited to
all possible conditions of service. The illustra-
tions shown in Figs. 13a and 13b indicate the
general structure of two types of constant-cur-
rent transformers; and Figs. 13c, 13d, and 13e
a complete line of auxiliary transformers.
Fig. 13a shows the station type air-cooled
constant-current transformer design in sizes
from 3 to SO kw. to operate series street-
lighting circuits. These transformers are
standard for 2300 volts, 60 cycles primary,
6.6 amp. secondary and for one and two cir-
dilits. They may be designed for any circuit
» 'Intensive Street Lighting." by W. D'Arcy Ryan. General
Electric Review, May, '1920, page 362. ;■■•■.■
or commercial frequency. This type of trans-
former is recommended in places where the
streets that are to be lighted are economically
close to the station or substation. Otherwise,
the pole type con,stant-current transformer,
as shown in Fig. 13b, is recommended. The
features of the station type constant-current
transformers are as follows:
1 . Constant current within one per cent of
nonr.al from full load to short circuit, regard-
less of fluctuations in primary voltage, lamp
failures, grounds, or short circuits.
2. Automatic regulation; no change in
taps for variations in load; no adjustments
necessary.
3. Instantaneous regulation; balancing
mechanism supported on ball bearings.
4. Maximum insulation between all parts.
5. High efficiency and power-factor.
6. Ventilated, air-cooled, impregnated wind-
ings.
7. All parts visible and easy to keep clean.
Fig. 13b illustrates the use of the pole type
constant-current transformer. This trans-
former adds another important link to the chain
of constant-current transforming devices.
The demand is urgent and the field is wide.
Series street lighting systems require con-
stant current, and constant-current trans-
formers have always required a substation
with control panels and an attendant, there-
fore, it has been difficult to provide street
lighting for smaller towns and villages where
the revenue derived would not be sufficient
to warrant the installation of a substation
and attendant.
Larger cities also have experienced difficulty
in solving the demand for higher intensities
and more units in their suburbs. The growth
of these outlying districts has been so rapid
that it has been almost impossible to keep
pace. When it becomes impracticable to run
circuits from the control station because of
the distance and the copper required, it is not
always advisable to erect a substation, but
if it is, the growth is usually so rapid that
there is an interval before the substation can
be erected when the lighting service is apt
to be inefficient or ineffectual.
The type of transformer shown in Fig. 13b
has been designed for such service. It is
entirely automatic and positive in action.
It does not require a substation or an attend-
ant, and it can be controlled by an oil time
switch. These features are combined with
as close current regulation through as wide a
range as offered by the best station type con-
stant-current transformer. The^current froi;a
54-4 June, 1920
GENERAL ELECTRIC REVIEW
Vol XXIII, No. 6
full load to no load is maintained within one
per cent of normal. This feature alone
practically guarantees the life of the Mazda
lamps operating on a circuit controlled by such
a transformer. The efficiency is the same as
for the station type transformer and the
ing and acts instantaneously to check surges
on the line which would tend to shorten the
life of the lamps. The moving secondary
coil with its high repulsion gives almost per-
fect regulation from full load to dead short
circuit. It protects the lamps not only from
Fig. 11. Night View Showing Street Illuminated with Obsolete Enclosed Carbon Arc Lamps
Fig. 12.
Night View Showing Street Illuminated with Pendent Type Novalui Units Equipped
with Refractors
power-factor is 20 per cent higher than for
any previous design of pole-type regulating
transformer.
For the operation of Mazda C lamps, this
transformer is ideal. The high internal re-
actance serves to protect the lamps at start-
changes in current due to changes in second-
ary load, but also from fluctuations in primary
voltage.
The construction of this transformer con-
tains no untried features, but simply combines
various features incorporated in several
ENCLOSED CARBON ARC LAMPS VS. NOVALUX MAZDA UNITS
545
C^va»teCtjrre^K
Fig. 13a.
Illustrating the Station Type Constant Current Transformer and
Its Application
different types of transformers which have
been in production for a long time. The core
is the standard, three-legged construction
with coils surrounding the center leg. The
primary coil is fixed at the bottom of the core,
and above is the floating secondary coil.
The balancing mechanism, however, has been
modified so that an exact line up of the coils
is not necessary for satisfactory regulation.
In order to give protection to the lamps
at starting, the minimum react-
ance of the transformer is made
fairly high. This results in a motion
of the coil of only a few inches,
while at the same time the repul-
sion of the coils is high, thus
giving excellent regulation. This
transformer may be tipped 10 deg.
from the vertical in any direction
without affecting the regulation.
This is much more than the trans-
former would be called upon to
stand in actual service. The coils
are liberally designed so that temper-
atures come within A.I.E.E. require-
ments. A single adjusting lever
permits adjustment of the secondary
current to the desired value. The
final current adjustment is made in
the factory and no further change in
adjustment should be necessary.
After being installed this trans-
former requires no more attention than one
of the constant-potential type. It is used
in sizes from 1 to 20 kw., 6.()-amp. secondary,
and for any commercial frequency.
Fig. 13c illustrates the series transformer
designed for the purpose of operating series
circuits of low-voltage in conjunction with
the main series circuit. This is accomplished
by using a series transformer having a one
to one ratio, the secondary being well insu-
T/Af£- SMTCH
-« — « — #-»--•
Fig. 13b. lUusCrating the Pole Type Constant Cuixent Tranaformer and Its Application
546 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
lated from the primary. The primary winding
is connected in series with the main series
circuit so that, under all conditions of load
on the secondary, the primary carries the full
current of the main circuit which is main-
tained at its normal value by a constant-
current regulating device. Series transformers
of such construction are made in many sizes
from 0.04 to 10 kw.
With this type of transformer, the instal-
lation can be arranged so that all street lamps
will come on at the same time and the
i- Primori/Cof/
-^SL Transformer
Secondary Coil
Main Series Circuit
ih-Uh
O-
a-
h ^
Protect'm Device -
T-f
f
/^ /^ /^ r^Auxi/iaru Series Ciruit
Fig. 13c. Illustrating Low Voltage Series Circuit with Constant Current Transformer and Its Application
SBffl£S C/ZTCUIT
0
15 OK to /IMr LAMP
Fig. 13d. Illustrating tie Series Auto Transformer and Its Application
ENCLOSED CARBON ARC LAMPS VS. NOVALUX MAZDA UNITS
547
expense of switching on by hand can be
eliminated. The transformer permits the use
of the series lamp which is more efficient
than the multiple lamp. It is used for
supplying current to one or more lamps
connected in series and located where the
high potential of the ordinary constant-
current series circuit would be objectionable.
These transformers are operated on
loaded series systems and, consequently, if the
secondaries become open-circuited, are sub-
jected to sinusoidal excitation which gives a
high distorted voltage. With the larger sizes
of transformers where this open circuit
voltage may become dangerous to the
of the air gap which, being protected, is very
uniform. When the gap breaks, the metal
flows and fills the hole so as to form a short
circuit across the transformer. When the
handle containing the film is inserted in the
holding clips, the protective device circuit is
opened and the film left in position to operate
in case the system is open-circuited. The
protective devices are mounted in steel cases
adapted to either pole or subway use in
accordance with the arrangement of the
transformer with which they are used.
As these transformers are designed to
operate from a circuit where the current is
held constant, the field for them necessarily
ISO/^ZO AMP LAMP
-^. °^-:^^/o-^-. c^.^/^^^ -^'iJ'-^' ■& .<^
Fig. 13e. Illustrating the Single Lamp Series Transformer and Its Application
insulation or to operators, film protective
devices are used. Film cutouts are not
employed with the smaller sizes of trans-
formers as these will operate without over-
heating or breakdown on open circuit.
The protective devices are designed with
clips to short circuit the secondary sj^stem
when the handle is removed. A second pair
of clips in the handle holds the films which,
on account of the relatively high voltages
encountered, are of a new type. The films
for higher voltages consist of plates of soft
metal cemented to the two sides of a fiber
disk through which a hole is pierced. The
thickness of the fiber determines the strength
lies in the vicinity of constant-current series
circuits.
Certain classes of lighting require lower
potential than that existing on series arc or
incandescent circuits and, to provide for this,
companies would be compelled to run multiple
circuits from the Central Station, often at
a considerable expense, if it were not for
this , transformer. Some of this low-voltage
lighting is supplementary to the regular street
lighting system and, filling the same function,
it is desirable to control it simultaneously
with the street lights. This transformer
aflfords the ideal method, ; for this control
as the low-voltage circuit is turned on and
548 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 6
off with the closing or opening of the main
constant-current transformer circuit.
Places where this transformer can be used
to advantage are as follows:
1. Isolated side streets or allej-s where it
is desired to install series incandescent
mounted on the pole shown in Fig. 1 1 and is
recommended where high voltage going
through the pole is not objectionable or the
appropriation is limited so that a safer
transformer cannot be used such as is shown
in Fig. 13e. When the auto-transformers are
Fig.
14. Comparative Floor Area and Efficiencies of the Brush Arc Generator vs, the Constant
Current Transformer System Based on Approximately Equal Capacity
lamps and where the only a\'ailable circuit is
the alternating-current series circuit.
2. In places where high potential is imprac-
ticable; e.g., where the line would be installed
on telephone poles, or where a few small units
in a building are required and a multiple cir-
cuit is not available.
3. On bridges where it is necessan.' to
eliminate high potential.
4. For underground circuits leading to
ornamental poles.
These transformers are designed in sizes
from 0.4 to 1.0 kw. and for any current or
frequency, having a ratio of the primary
amperes to the secondary amperes of 1:1.
The secondar\- is highly insulated from the
primar\'.
Fig. 13d illustrates the use of the series
auto-transformers designed for operating high
current Mazda series lamps such as the
15-amp. 400-c-p., 20-amp. (iOO-c-p. and
20-amp. lOOO-c-p. lamps. It is designed
primarily to operate these lamps at high
current density to obtain the advantage in
efficiency over straight current lamps. They
are made for both pendent and ornamental
Novalux units.
There is one winding and taps come cut
for the lamp, as shown in Fig. 13c, therefore
the circuit to the lamp is not insulated from
the high potential series circuit.
For aerial work there are no particular
objections, and for ornamental lighting the
auto-transformer is placed in the casing
used, the underground highly insulated cables
are carried up the post to give proper pro-
tection against grounds.
Fig. 13e illustrates the use of the single-
lamp series transformer also designed for
Fig. 15
Series TransfcHmer Mounted in Baae of Pole on Strap
Iron Support Embedded in Concrete
operating high current Mazda series lamps.
It is built in capacities to take care of
400, 600 and 1000-c-p. lamps. Standard
primar\- windings are for (>.(> amp. and stand-
ard secondaries for l."> amp. 400 c-p., or 20
amp. 600 and 1000 c-p.
ENCLOSED CARBON ARC LAMPS VS. NOVALUX MAZDA UNITS
549
They are entirely enclosed in steel casinj;
and are weatherproof. Two types are used;
subway and aerial. Leads on the subway
type are brought out through galvanized iron
wiping sleeves so that the lead sheath can be
readily wiped on. For aerial use the leads
are brought out through porcelain bushings.
These have the following advantages:
L High efficiency series lamps can be used
where high potential is impracticable.
2. They protect the lamp from surges in
the line.
per cent of all the line trouble which occurs
between the pole and the lamp.
When lamp wattage varies between 8 per
cent above and 20 per cent below normal, the
secondary current will not vary more than
1 per cent with normal current and fre-
quency.
An interesting comparison of the floor space
required for Brush arc machines operated by
synchronous motors, which supply current to
direct-current enclosed carbon arcs, as against
the constant-current transformer equipment
Fig. 16. Series Transformer Buried in Ground
3. They are a valuable adjunct to "Safety
First" in ornamental street lighting, due to
the fact that the secondary is highly insulated
from the primary- and permits the use of high
efficiency series lamps in business districts
where ordinances prohibit high tension wire
above street surface.
4. They save the expense of high-voltage
conductors, heavy insulation and high tension
cutouts, which materially assists in liquidat-
ing the difference between the first cost of
auto-transformers and series transformers, the
latter being naturally somewhat higher priced.
Furthermore, this low voltages eliminates 75
(the latter to supply current to the incandes-
cent lamp), is shown in Fig. 14. The Brush
machine may be found in many of the cities
that are still using the carbon arc lamp.
It will be obvious to any central station
manager, after reviewing these data, that the
time has arrived to replace the enclosed car-
bon arc lamp with more efficient equipment.
Their continued use can be attributed to no
other cause than backwardness and failure
to realize the opportunity for increased effi-
ciency, increased capacity, and better service
offered by the later developments in street
lighting equipment.
550 June, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 6
350-ton Hammer Head Fitting Out Crane
By J. A. Jackson'
Power .\nd Mining Engineering Dep.\rtment, Gener.\l Electric Comp.\ny
The massiveness of the armament and machinery of our modern battleships is well indicated by the capac-
ity of the huge crane described in this article. It seems scarcely probable that the maximum capacity of this
crane, 350 long-tons, will ever be required for a single load, but we are progressing at such a rapid rate that
another two or three years may produce a need for a crane having a capacity greatly in excess of the one de-
scribed here. — Editor.
There was recently put into sen'ice at the
League Island Navy Yard, Philadelphia, Pa.,
the largest hammer head type fitting out
crane which has ever been built. The crane
is approximately 250 ft. high overall and can
handle a 350-long-ton load at 115 ft. radius,
and a 50-long-ton-load at a 190 ft. radius.
The maximum lift of the main hook from
the deck of the pier on which the crane is
located is 145 ft., but the drum holds suffi-
cient cable so that the hook can be lowered
25 ft. below the deck of the pier, making a
total lift for the main hook of 170 ft. The
50-ton hook has a maximum lift of ISO ft.
The total length of the swinging jib is
300 ft. of which 100 ft. is on one side of the
center line and contains the machinery' house
and counterweight, while the other 200 ft.
is on the opposite side of the center line and
contains the runways for the trolley carriages.
The machinery house has a floor space of
approximately 70 ft. by 50 ft. and is ser\'ed
by an overhead travelling crane of 35 tons
capacity and approximately a 50-ft. span.
The runway for this crane is extended outside
the machinery house so that the crane can
run out through a motor-operated sliding door
and lower its hook to the pier deck when
necessary to lift anything from the ground to
the machinery house.
The machinery house contains all the
machinery and electrical equipment with the
exception of the slewing motor, drum con-
trollers, and the master switches. The slew-
ing motor is located in a irtotor house about
50 ft. from the ground at the lowest part of
the revolving member of the crane. All mas-
ter switches and drum controllers are located
in an operator's cab, so placed as to have a
clear view of the crane hooks at all times.
The revolving member, which with its
live load weighs approximately 5,500.000 lb.,
is supported at the top of the stationary mem-
ber on a roller bearing, which bearing can l)e
readily inspected by raising the entire revolv-
ing member by means of hydraulic jacks per-
manently located on the crane. The revolv-
ing member is carried down to apjiroximately
50 ft. of the ground, where it terminates in
a large ring running on rollers located on the
stationary part of the crane. A rack is also
fastened to this ring, which rack is connected
to the slewing motor through a suitable train
of gearing.
There are two entirely independent main
hoists, each having a capacity of 175 tons
and each main hoist has its own independent
trolley. The hoisting engines, however, for
both the two hoists and the two trolleys, are
arranged so that they can be coupled together
mechanically so as to operate as a single unit.
When thus operating, the two master switches
for the hoist motors have their shafts coupled
together so that they operate in unison from
one handle. The two drum controllers for
the two trolley motors are similarly coupled
together.
The drum for each main hoist has right
and left hand groo\-es, from which two ropes
lead off through suitable guiding sheaves to
the trolley wagon. At this point each rojje is
reeved through four sheaves, giving an eight
part line between the trolley wagon and the
hook. After passing through the sheaves, the
two ropes are taken to an equalizer sheave to
insure equal division of the load on the two
ropes.
There arc four gear reductions between the
drum and the motor pinion, and an actual
test with a 350-ton load showed an efficiency
of 52 per cent from the hook to the motor
pinion. The crane was designed for a hook
speed of 2.5 ft. per min. when hoisting full
load, but an actual test showed a speed of
2.62 ft. per min.
Each main hoist is also equipped with a
gear change which increases the hook speed
from 2.5 ft. ])cr min. to 10 ft. per min. for hand-
ling lighter loads at higher sjieeds. Each main
hoist dnmi weighs approximately 37 tons.
All tlic trolley wagons are rope driven from
hoisting engines located in the machinery
house, and in the case of the two main trolleys,
the hoisting engines have gear changes which
give a speed of 12 ft. per min. in low gear and
100 ft. per min. in high gear. Actual tests.
350-TON HAMMER HEAD FITTING OUT CRANE
551-
however, with full load on the hook, give a
speed of about 19.3 ft. per min. in low gear.
The auxiliary' hoist is geared for a hook
speed of 12 ft. per min. with a 50-ton load
and its trolley is geared for 100 ft. per min.
The auxiliary trolley runs on separate tracks
from the main hoist trolleys so that its oper-
ation is entirely independent of them.
The slewing motor is geared to give one
revolution of the crane jib in 12 minutes, but
an actual test showed one revolution in
approximately 9 minutes with a 125 per cent
load on the hook.
One MDS-107 mill-type motor is used on
each of the main hoists, the auxiliary hoist
and on the slewing motion, while each of the
three trolleys is equipped with an MDS-104
motor. All motors are series wound. Tests
showed that all motions are conser\'atively
motored, especially the slewing motion. It
was considered advisable, however, to use
the same size motor on the slewing motion as
on the hoist motions, on account of the spare
part situation.
The control equipment for all hoist motions
consists of a standard contactor panel slightly
modified to obtain additional points of con-
trol, which were required by the government
specifications.
Each hoist master switch gives six-hand
points in each direction and all lowering
points give both power and dynamic braking,
depending on the load requirements.
Extra heavy duty resistors were necessary
to meet the severe dynamic braking condi-
tions encountered when a load is lowered
throughout the maximum lift as required by
government tests.
The lowering speed with full load on the
hook was approximately 3 ft. per min.; thus
it required about 57 min. to lower through the
maximum lift. Each hoist equipment has a
geared type limit switch to prevent over-
travel.
The slewing motion has straight reversible
magnetic control with a five-point master
switch. The panel contains plugging relays
and the resistor is laid out to give safe plug-
ging-
All three trolley motions arc controlled by
drum type manual controllers with vertical
operating handles. Resistors are laid out for
plugging.
Track type limit switches limit the travel
of all trolley motions. The protective panel
is a standard panel with two gravity reset
overload relays in each motor circuit. The
protective panel also contains contactors on
which the trolley motion limit switches
operate.
Series wound solenoid brakes are used on all
motors, those on the slewing and trolley
motions being set for very low retarding
torque values.
The operator's cab is equipped with com-
plete electric bell signaling devices to signal
to the ground or to the machinery house, and
it is also equipped with indicators to show
the exact position of each of the trolleys on
the runway.
The crane is equipped with a passenger
elevator which starts at the bottom of the
revolving member and runs to both the
operating platform and on up to an obser-
vation tower at the highest part of the
crane.
The crane was subjected to a very severe
series of acceptance tests by the government,
which included the handling of a 25 per cent
overload throughout the maximum lift with
the trolley run out to the maximum radius.
It passed all tests successfully and the cover
illustration of this issue of the Review shows
the crane at the completion of the test.
552 June, 19-20
GENERAL ELECTRIC REVIEW
VoL XXIII. No. 6
QUESTION AND ANSWER SECTION
The purpose of this department of the Review is two-fold.
First, it enables all subscribers to avail themselves of the consulting service of a highly specialized
corps of engineering experts, or of such other authority aj the problem may require. This service provides
for answers by mail with as little delay as possible of such questions as come within the scope of the Review.
Second, it publishes for the benefit of all Review readers questions and answers of general interest
and of educational value. When the original question deals with only one phase of an interesting subject,
the editor may feel warranted in discussing allied questions so as to provide a more complete treatment
of the whole subject.
To avoid the possibility of an incorrect or incomplete answer, the querist should be particularly careful to
include sufficient data to permit of an intelligent understanding of the situation. Address letters of inquiry to
the Editor, Question and Answer Section, General Electric Review, Schenectady, New York.
GROUNDS; THREE-WIRE THREE-PHASE
DELTA
(215) Ordinarily, 220/110-volt three-wire
systems have their neutral grounded.
However, the grounding of a three-phase
delta three-wire secondary distribution
system, as illustrated in Fig. 1, would
cause a circulating current through the
ground from each phase to the other two,
due to the 110-volt delta potential. Should
such a system be grounded; if so, how
should the grounds be connected?
The National Electric Code states that
circuits should not be grounded in such a
manner that the ground connection will carry
CALCULATION: INDUCTANCE, CAPACITY
AND RESISTANCE IN SERIES AND
IN PARALLEL
(216) Please solve by the method of complex
quantities the problem given in Q. & A.
No. 190.
The question referred to asked: "What is
the total impedance of the circuit illustrated
in Fig. 1? Explain the method of calculation."
As in the two methods of solution previously
published,* the following solution by the
complex quantity method will be divided into
distinct parts: (2) the impedance of the
parallel group, {2) the impedance of the series
group, and the combination of these two.
r- 220 *1
W-iio f 110
216) Fig. I
an appreciable amount of current. Obviously,
the ground connections shown in Fig. 1 (or
even only two) could not be used for they
would short circuit half the voltage of the
delta. There have been a few cases known
where the middle point of one phase only was
grounded, dependence being placed upon this
to furnish sufficient ground connection for
the other phases.
If the transformers are single-phase and
each phase is separated from the other two
on the secondary side, the middle points of
the individual phases could of course be
grounded.
F.R.F.
(i) Impedance of the parallel group.
This is composed of three impedances in
parallel ;
Zi = ri->rjxi
in which :
r, = 4 .t,=0
rj = U x, = 2
Hence
Z, = 4;Z, = /2;Z.= -;t)
QUESTIONS AND ANSWERS
553
The corresponding admittances are
Z, 4' - Zo
.. 1 1 .1
Zz 76 6
;^
?•
In complex quantities, the joint admittance
of a number of parallel-connected admittances
is equal to the sum of the individual admit-
tances, thus
and the impedance of the parallel group is
^4 V(Ir+(|)V +^3 V(i)=+(iW
1 144 1 144
Z„ = -X^+;3X^=1.44+;1.92.
* These appeared in the December, 1916, Review, p. 1135,
but are repeated in the following for the sake of completeness,
and for the benefit of those who were not subscribers at that
time. — Editor.
One method is based on graphics; this is by far the simpler
mode. The other employs only arithmetic; this method furnishes
a more exact answer than does the former.
Since there are two distinct combinations of the resistance,
magnetic reactance, and capacity reactance in the circuit, the
problem will be divided first into two parts, viz., (a) the imped-
ance of the parallel group and (b) the impedance of the series
group. The combination of these two impedances will then be
the total impedance of the circuit.
The impedance of a parallel circuit is equal to the reciprocal
of the vector sum of the reciprocals of the ohmic values of the
sub-circuits. The impedance of a series circuit is equal to the
vector sum of the ohmic values of all the jjarts of the circuit.
9'
t t
0.167
//
J
/
0.5
X
/ 6
b'
U — o.zs — -
{2} Impedance of the series group:
Z6 = r+/.v = 3+/(2-l)=3+/
Combination of the two groups:
The joint impedance of a number of series-
connected impedances is equal to the sum of
the individual impedances, thus the total
impedance of the system is
Z = Z«+Zfc = 3+1.44+y(l + 1.92)
= 4.44+y2.92
and the absolute value is
Z = v/(4.44)=-j- (2.92)2 = 5.314
The complex quantity method furnishes a
more exact solution than does either the
graphical or the arithmetical method.
L.G.
Arithmetical Method
This method employs the general formula:
' 'Impedance equals the square root of the sum of the
resistance squared and the arithmetical difference between the
magnetic and capacity reactances squared."
laj For the parallel group the reciprocal of the ohmic
values of the sub-circuits are used, the reciprocal of the resultant
giving the impedance. This coincides mathematically with the
graphical method described in Method I (c).
\(K)^+(H-KP
= 2.4 ohms.
(b) For the series group the general formula is applied
directly
Impedance = \/32+(2-l)2
= .3. 1 6 ohms.
Impedance =
h
■"L 3 -J
/
1.
?:i^^
^^^
e
d
(216) Fig. 2
'216i Fig. 3
12161 Fig. 4
Graphical Method
(Assume the vector direction for resistance to be hori-
zontally to the right, for magnetic reactance to be vertically
upward, and for capacity reactance to be vertically downward.)
(a) For the parallel group lay out the vectors, as shown in
Fig. 2, equal to the reciprocal of the ohmic values of the con-
ductors, i.e., for resistance draw a'b' }\ or 0.25 of a unit to the
right, for magnetic reactance h'g' }•> or 0.5 upward, and for
capacity reactance gf y& or 0.167 downward. The vector, a'f,
that joins this last point to the first will be at some angle d to the
horizontal and will scale 0.417 in length. The impedance of the
parallel group will therefore be ^ -"- _ or 2.4 ohms.
0.417
(b) For the series group lay out the vectors, as shown in
Fig. 3, directly equal to the ohmic values of the conductors, i.e..
for resistance draw/rf 3 units to the right, for magnetic reactance
dh 2 upward, and for capacity reactance he 1 downward. The
resultant vector, fe. will be at some angle <^ to the horizontal and
will scale 3.16 ohms, which will be the impedance of the series
group of the circuit.
Combination of the two groups:
Lay out a vector, af. 2.4 units in length at the angle 8 to the
horizontal, see Fig. 4. From the end of this line lay out a vector,
A. 3.16 units in length at the angle 4> to the horizontal. The line
bridging these two vectors from end to end. ae, will represent the
total impedance of the circuit, the value of which will V)e found to
be 5.32 ohms by scaling the length.
Combination of the two groups:
Since the influence of the parallel group on the power-factor
is not the same as that of the series group, the respective resultant
ohmic values of the two groups must be added in accordance
with the difference in phase angle, in order to obtain the total
impedance.
This is best accomplished by squaring the arithmetical sum
of the two resistance components of the two groups, adding to
this the square of the arithmetical sum of the resultant magnetic
or capacity reactance of the two groups, and determining the
square root of the whole. In the symbols of Fig. 4 this is
ae = \/iab ■^bc)'-\-{cd +d€)^
ah =2.4 X^' (from Fig. 2) =2.4 X-^^ =1.44
bc=fd (from Fig. 3) =3
a =hf = 2AX-rf, (from Fig. 2) =
rif =
0.417
-1 (from Fig. 3) =1
Impedance, therefore, equals
= v'(l-44+3)! + (1.92 + l)2
= .5.32 ohms.
ECS.
18
GENERAL ELECTRIC REVIEW
JUNE. 1920
Where to Get G-E Service —
Quick service is best obtained from the nearest G-E
sales office, distributing jobber, or foreign representative
For Business in the United States
G-E Sales Office C-E Dutribatiat Jobber
Alabama. Blrmlnehatn MaiUiews EIk Supply Co.
ArkBQSaj. LItUe RocK
California. \m3 AiuEelest Paclflc Statra Electric Co.
California. Oofclandl Pacific Statra Eleclrlc Co.
California, San FYandacoJt Pacinc sute* Elw-trlc Co
Colorado. Dcnvert Tbe Hendrle A BoltboII MIC A
Sup. Co.
Connecticut. Hartford . .
Connecticut, New Haven
Connecticut. Waterbury^ New Eofland Eog. Co.
District of Cotumbla. Wublng-
ton National Elec'l Supply Co.
Florida. Jacksonville Florida Elec. Supply Co.
Florida. Tampa; Florida Elee. Supply Co.
Georela, Atlantajt Carter Eleeirlc Company
Georsla. Savannahl Carter Electric Company
Illinois, Chicago it Central Electric rompaoy
Commontrealib Ediaoo Co.
Indiana. Fort Wayne -
Indiana, Indianapolis Indianapolis Flee Supply Co.
Indiana. South Rendt South (tend Wc- irlc Co.
Iowa, Dea Moines MWI-West Electric Co
Kentucky, Louisville Belknap Hardware A Manufac-
turlns Co . Inc
Louisiana, New Orleuu Gulf Stales Electric Co.. loo.
Maryland, Baltimore Southern Electric C^.
Massachusetts. Hostont Petting eU- Andrews Co.
Massachusetts. Sprlncneld
C-E Sale
Off
New 1 ■ ^
New Vort Cit
It
Work* at Schenectady. N Y. Cnundan*. 340
seres Floor space. 5.800.000 square (fcw.
Other manxifacfunng plants of the General
Electric Company are located at Harnaon. N J.
Newark. N J . Lvnn. Mass. Pitttfield. Maam..
Ene. Pa . Cleveland. O.. and Fori (Varne. /n«t
C-E DUtributifiB Jobber
-Havens F.lcciric <~o , Inc.
Robertson-Cataract Elee.Oo.
E B. I^tham A CoiApani
Royal Eastern Elec'l Sup.
Klbley-Pluoaa Eler Corp.
Co.
. n'beeler-Grees Elee'l Sup. Co.
' Mohawk Elee'l Sup. Co!
.Frank C. Teal Companr
Massachusetts, u orcester
Mlcblsan. Deuolt
Michigan, Grand Rapid!
Minnesota. Uuluth Northwestern Electric Equip-
ment Company
Minnesota. Mlnnespolist Peerless Electric Co-
Minnesota. St Fault Northwestern fc:iee. Equip. Co.
Missouri, Jopllnt
Missouri, Kansas CItyt The B-R Electric Co
MlMMurL St. Ixiulst Wwco Supply Company
Montana, Buttet Butte Electric supply Co.
Nebraska. Omaha Mid-West Electric Co
New Jersey. Newark: Trl-fliy Elecwic Co , Inc.
I No G-E Office
New York. NIacara Falls...
New York. Rochester
New York, ^heoectady
New York. Syracnae . .
North Carolina. Charlotte
Ohio, ClnclnnaUt Tbe F D. Lawrence Else. Ce.
Ohio, Cleveland Republic Elw-trtc Co
Ohio, Columbus The E->ner A Honkltu Co.
Ohio. Dayton Tbe Wm Hall EIrrtric Co.
Ohio, Toledo W. G. Nagrt Eketrte Co-
Ohlo. YounestowTi
Oklahoma. oklah»nu Cltyt .'«oulbwest O-E Co.
Orecon. Portlan<|t PacMc States ElceUlc Co.
Pennsylvania. K.ric .
Pennsylvania. Pbtladelphlatt.PhltadHphta Eleriric Compsar
Suppl>- rvpsrtment
Pennsylvania, Plttsbunht Unloo Electric i^'ompany
Rhode Island , Prov Idence .
South CaroUaa, Columbia: Prrry-Mann Elee. Co . loc
Teooessee. Chattanooca James Supply Compaoy
Tennessee, KnoKvllle
Tennessee. Memphis Eleeirlc Supply Compsar
TiiiiiiiMi I " mill (III
Texas. Dallast Southwest C^E Co.
Texas, El Pasot Suuthwcat C.-E Co.
Tevas. Hnusliiot . , S.iulh«rst (i-E Co
Ctah, Salt lAkeCityt t'apttal Electric Company
Washlnston, Spokane. .
Wa9hlnet'>n. Tacoma
West Vlrclnla. ChsricstoO
Wlaoonaln. MHwaufcre
Fur Hawaiian business ad'lress Cation. Nclll A Cowpeny. LM-.
lilt: ■ "
tUarrhouse (Servke Shop.
Diatribulors for tbe Cooeral EJoclric Company Oulsids of ih* United Slates
INTERNATIONAL GENERAL ELECTRIC COMPANY. INC.
120 Broadway. New York. N Y. 83 Cannon Street, London Scbearclady. N. Y
Foraim Offices and ReprsMntalivos
Artentlna: General Electric. .^^ A.. Muenos Aires.
Australia: Australian General Electric Co., Ltd . Sydney and
Melbourne.
BelKlum and Colonies: -Soclete d'Elecirlclte el de MecanlQue
Procodes Thomson-Houston A Carels tl-oclcte Anooyme.
Brussels.
Bolivia: International Machinery Co , I.a Pa» and Oruro.
Rrsilt: Grnrral Electric, S A.. Rio de Janeiro and Sao Paulo.
Canada: Canadlun (General Electric Co.. Ltd.. Ontario.
Chile: International Machinery Company, Santiago. Anto-
fasania and Vali>aralso.
China: Anderson. MF><-r A Company, Ltd . Sbanxhal
International General Electric Company, Inc (General repre-
sentatives of the tar East excluding China and Japan )Sban<bal.
Colombia: W es^elhocft A I'oor, Darranqullla and Bogota.
Cuba: General Electric Company of Cuba. Havana
Dutch East Indies: International General Electric Co,, Inc..
Soerabala. Java
Ecuador: <'arliis Cordovet, Guayaquil and Quito.
Egypt: British Thomson-Houston Co.. Ltd., Cairo.
France and Colonies: Compagnle FrsocaUe Thomson-Houiton.
Paris.
Great Britain and Ireland: British Thomson-Houaton Co..
Ltd , Rugby: Internallonal CcneralElcctrk Co.. iDC London.
Greece and Colonies: Compagnle Erancatse Thomson- II oustoo.
Paris
India: Rrltl'h ThoTn^on-Ilniulon Co.. Ltd.. Calcutta and
Bombay: International General ElecUtc Company, loc..
Calcutta
Italv and Colonies: Franco TcmI Sorleta Anoiilna. Milan.
Japan: Shlbaura Engineering Works. Tokyo: Tok>o Eleetiie
Co . Ltd , Kawasaki
Meiico: Mexican General Electric Co.. City of Mexko sad
GuailaU]:ira
New Zealand National Elecirlcal* F.nglneerlngCo., Aucklaad.
Dunetlln. Chrlstchurch. and Wrlllncion.
Paraguay: i'.ciicrsl Electric. S A., Buenos Aires.. ArgvatUw.
Peru: W*. R. Grace A Co., Lima
Philippine Islands: PaclOr Commrretal Company. Manila.
Porto Rico: International General Electric Co . Inc.. San Juan.
Ru-<sla: Waeobshichala Elertrlchcskata Kompania. Peutxrnd
and Vladivostok.
South .\rrlra: South African General Electric Co..Lld.,Jo>
hrinni'sburg and Capetown.
Uruguay: lieiieral Electric, S. A.. Montevideo.
vruciuela: Wesselhoen A PooT. Caracas.
SElectdc
SchenSitady.NY. ^^a** O KO.
aey
Sales Olficcs in
all large cities
95-294
Sav you sa2i' it advertised in the Gkner.m, Kleitrk' Review
TWO DOLLARS PER YEAR
TWENTY CENTS PER COPY
GENERAL ELECTFIC
REVIEW
VOL. XXIII, No. 7
Published by
General Electric Company's Publication Bureau,
Schenectady. N. Y.
JULY, 1920
AUTOMATIC EXHAUSTING AND SEALING MACHINE FOR INCANDESCENT LAMP MANUFACTURE
For
Fractional H. P. Motors
IT is easy to claim precision. It is a harder
matter to maintain it, especially in quan-
tity production. That "NORmfl" Precision is a
tact, and not a mere claim, is proved bv the
daily performance ot hundreds ot thousands
ot high grade, high speed electrical machines
in which "NORfflfl" IJearings are standard. It
is high precision unitormlv maintained which
explains "NORmfl" silence and serviceability at
high speeds.
See that your Motors
are "NORmfl" Equipped
Ball, Rol lei; Thrust and CcmbinatiQn Bearings
General Electric Review
A MONTHLY MAGAZINE FOR ENGINEERS
.. „ „,^T. ,-,. , „^„.„^,„ Associate Editors. B. M. EOFF and E. C. SANDERS
Manager. M. P. RICE EdUor. JOHN R. HEW ETT ,„ ^^^^^^ ^f Advertising, B. M. EOFF
Subscription Rates: United States and Mexico, $2.00 per year; Canada, $2.25 per year; Foreign, $2.50 per year; payable in
advance. Library and Student Rates: United States and Mexico, $1.50 per year; Canada. $1.75 per year; Foreign, $2.00 per year;
payable in advance.
Remit by post-office or express money orders, bank checks, or drafts, made payable to the General Electric Review, Sche-
nectady, N. Y.
Entered as second-class matter, March 26, 1912, at the post office at Schenectady, N. Y., under the Act of March, 1879.
Vol. XXIII. No. 7 „ya„^r^E^:^^i^cLpa,.y -U-LV. l^^^O
CONTENTS Page
Frontispiece : 30,000-kw. Curtis Turbo-generator Set, Philadelphia Electric Company . 556
Editorials ; Temperatures in Large Alternating-current Generators 557
Exciters and Excitation Systems 558
Speed and Power-factor Control of Large Induction Motors .... 559
Temperatures in Large Alternating-current Generators 560
By W. J. Foster
Exciters and Systems of Excitation 566
By H. R. SUMMERHAYES
Gaseous Conduction Light from Low-voltage Circuits 577
By D. McFarlan Moore
Fundamental Phenomena in Electron Tubes Having Tungsten Cathodes — Part II . . 589
By Irving Langmuir
The Safety Car 597
By W. D. Bearce
The Production and Measurement of High Vacua — Part II 605
By Dr. Saul Dushman
Two Years' Service of Battleship New Mexico 615
Electric Power in the Oil Fields as a Central Station Load 616
By W. G. Taylor
Theory of Speed and Power-factor Control of Large Induction Motors by Neutralized Poly-
phase Alternating-current Commutator Machines 630
By John I. Hull
— o
.5 «
Z '-/^
> S
o £
§0.
II
T E
-Q
ut £
O
i
5
H
i
TEMPERATURES IN LARGE ALTERNATING-CURRENT GENERATORS
Temperature is a matter of so great
importance in rotating electric machinery
that all standardization of the American
Institute of Electrical Engineers is now
based upon it. The approved ratings of
machines are determined by considerations
of the temperature rises as related to the
materials employed. Hence it has come
about that designers of electric machines
have given great attention to the ventilation
problem. This is most commendable in
itself, and has resulted in a decided advance
in the construction of machines.
It often happens that wrong ideas exist
regarding machines when established rules
are applied. For instance, there is a prev-
alent notion that any electric machine is
underrated if the temperatures when operat-
ing at the rating given to it by the manu-
facturer are lower than those designated by
the rules of the A.I.E.E. for the type of
insulation used. Temperature rises are often
less than those permissible for the reason that
purchasers of machines also insist upon
certain other desirable characteristics.
The temperature of an electric machine,
or any electrical apparatus, is the resultant
of two factors; first, the quantity of energy
in the fonn of heat losses that are attendant
upon the operation of the apparatus, and,
second, the effectiveness of the dissipation
of this heat energy. It therefore happens
that of two machines of the same rating,
operating at the same load under the same
conditions, the one showing higher tempera-
tures may be the more efficient. This is
undoubtedly contrary to the general notion
regarding the matter. It is simply necessary
to reflect that a machine that is constructed
as a good blower and passes through itself a
large quantity of air, can remove from itself
a much greater amount of energy in the form
of heat than can one that is sluggish in its
air circulation.
Another idea that is generally prevalent
is that the facts taken as the basis of standard-
ization are established beyond dispute. In
the case of electric machines, it is generally
believed that the temperatures specified in
the rules for different classes of insulation
represent the dividing point between what
is safe and what is dangerous, and that it is
perfectly safe to operate below the specified
temperature and unsafe to operate above it;
whereas, as a matter of fact, there is no hard
and fast division point and, consequently,
it is often sensible to operate at temperatures
quite a little below those allowable. It is
the life of the machine, the freedom from
renewal of parts and from repairs in general,
that should be considered in connection with
the original cost and the efficiency at which
the machine operates.
As yet no accurate method of determining
the real internal temperatures of insulated
windings has been developed for the regular
commercial operation of machines. These
internal temperatures could be derived with
a high degree of accuracy from measure-
ments on the outside of the insulation and
from laboratory determinations of tem-
jDerature drop through insulations made up
of various materials, were it not for the
impossibility of determining the actual quan-
tity of heat to be removed. It is quite
generally known that in most electric con-
ductors there are other losses than that
represented by the flow of current against
resistance. Most alternating-current gen-
erators, even of the largest size, can be
designed with these parasitic losses so small
in quantity as to make the internal tem-
peratures so low that the deterioration of
insulation will be slow. This fact has been
demonstrated in many instances by high-
voltage generators, having only varnished
cloth insulation, operating from 15 to 20
years without a single breakdown.
558 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 7
EXCITERS AND EXCITATION SYSTEMS
At the annual convention of the American
Institute of Electrical Engineers held at
White Sulphur Springs, June 29th to July
2nd, a session under the auspices of the Power
Stations Committee was devoted to papers
on exciters and systems of excitation.
A complete analysis of the factors determin-
ing the selection and general design of exciter
systems was presented by J. T. Barron and
A. E. Bauhan, both of whom are connected
with large operating companies. A shorter
paper was presented by Messrs. Parker and
Meyer, of the Detroit Edison Company,
which discusses briefly the advantages and
disadvantages of various excitation schemes
and outlines the essential requirements from
a broad point of view as power house design-
ing engineers.
A paper by H. R. Summerhayes, of the
Engineering Department of the General
Electric Company, gives a broad discussion
of the advantages and disadvantages of
various types of exciters and of exciter drive,
and refers to past practice and to the trend
of present practice in the selection of exciters.
Mr. Summerhayes' paper will be found in
this issue.
Papers by Messrs. Cox and Michener, of the
Southern California Edison Company, describe
the excitation arrangement used in existing
and proposed plants of that Company, and
J. D. Ross, of Seattle, Wash., gave data
on the exciter practice in a number of hydro-
electric plants in the Northwest.
A paper by Messrs. Boddie and Moon, of
the Westinghouse Company, calls attention
to a nimiber of characteristics of design of
exciters to be used with automatic regulators,
and characteristics affecting parallel opera-
tion; also the advantages of shunt exciters
as compared to compound.
Thus, the field of exciter systems was
co\ered by a nimiber of i)a])ers from manu-
facturers, from operators, and from engineers
having to do with the design of new stations.
It was noticeable that all of the authors
attempting classification of exciter systems
divided them into two general classes: first,
the common bus system with exciters operat-
ing in parallel, and second, separate exciters
or individual exciters for each generator not
operating in parallel. This division was
independent of the method of drive. The
general conclusion reached from the opinions
expressed in the papers and brought out in
the discussion was that for plants where the
speed is not too low or too high to obtain a
good design of exciter, the direct-connected
individual exciter provides a most reliable
form of excitation at the lowest cost. There
are some engineers, however, especially those
connected with large city central stations,
who still prefer the common bus exciter
system on account of the fact that a storage
batten,- can be operated on this bus at all
times, ready to take up the excitation load in
case of exciter trouble. The general opinion,
however, and the trend of present practice
is toward the individual direct-connected
exciter, which has the advantage that its
circuit to the generator field is short and
simple and not liable to trouble, that its
method of drive is exceedingly reliable and
efficient, and that trouble on the exciter
affects only one generator. Proof of the
reliability of direct-connected exciters for
hydro-electric generators is given in the deci-
sion of the Southern California Edison Com-
pany, as described in the paper by Cox and
Michener, to supply in their latest plant one
direct-connected exciter without any operat-
ing alternate exciter, although a spare direct-
connected exciter ready to be mounted will
be kept in stock. The use of direct-connected
individual exciters with large steam turbines
was also favored by Messrs. Parker and Meyer
on account of the reliability of such units,
as proved in actual experience.
^Ir. Summerhayes points out that in plan-
ning power stations, engineers sometimes call
for direct-connected exciters on steam tur-
bines large enough to excite more than one
turbine. This practice is undesirable, since
such a large exciter, if over-hung, may require
a shaft extension so long, and the weight of
the exciter may become so great, that it
interferes with the operating balance of the
main unit. The exciter drive should not lie
permitted to introduce any uncertainty into
the operation of the main turbine unit and
for this reason the size of exciters should be
limited to those which may be safely over-
hung on the turbine shaft. When common
bus excitation is used this bus should not
be used for the supply of auxiliaries, or for
working electrically-operated switches in a
plant where continuity of service is essential,
since troubles on the auxiliaries or on the bus
may affect the excitation, and vice versa —
because in case of short-circuit on the alter-
nator high voltage may be induced in the field
circuits which may affect the control circuits.
EDITORIAL
559
SPEED AND POWER-FACTOR CONTROL OF LARGE
INDUCTION MOTORS
In the early days of electric power trans-
mission the motor problem and the generator
problem were the same, and the universal
answer to both was the d-c. commutator
machine, whose voltage control as generator
and whose speed control as motor could be
economically attained with a field rheostat.
But the transmission problem demanding
small currents and their accompanying high
voltages, and the generation and utilization
problem demanding low voltage, forced the
development and almost universal use of
the constant-voltage constant-frequency poly-
phase a-c. system. Then came the polyphase
induction motor — rugged, simple, efficient,
with high starting torque (especially when
used with external resistance and wound rotor)
and low starting current, well nigh perfect
except for two characteristics, viz. : for its
excitation the line has to furnish magnetizing
current ; and it has but one economical speed,
and more serious still, one stable speed, syn-
chronism, which it approximates either loaded
or light.
The.se limitations have not overcome its
advantages, and therefore the induction
motor has had a marvelous growth in size
and numbers. But the recent attention that
has been given the indirect costs of poor
power-factor has brought one of these dis-
advantages to the foreground, while the
urgent need of varying the speed of many
large induction motor units (especially in
rolling mills) which are supplied with a-c.
power has kept the problem of speed control
of large induction motors in a prominent
position pending a satisfactory solution.
For the solution of the latter numerous
schemes have been advanced, but so far only
two liave assumed any real commercial
importance, aside from the arrangements
which may be classed as multi-speed motors
where two or more motors of different speeds
are combined in one mechanical structure.
One of the schemes involves the conversion
of the slip ring energy of the induction motor
into mechanical power by means of a rotary
and d-c. motor, and is somewhat loosely
called the " Kraemer" system. The other
involves the conversion of the slip ring energy
into mechanical power by a polyphase a-c.
commutator machine, and is commonly
called the "Scherbius" svstem.
Thus, in order to regulate the speed
economically, recourse is had again to the
commutator. Why' In the synchronous
machine, there is a relation of proportionaUty
between speed and voltage; but speed and
frequency are also in fixed proportion, so
that to vary the speed by merely changing
the field strength is not possible, and the
attempt to do it results in magnetization or
de-magnetization by wattless current from
the line. In induction machines, the speed
can be varied independently of primary
voltage and frequency if voltages are intro-
duced into the secondary at slip frequency.
Secondary resistance drop can thus be used
to reduce the speed, but this gives poor
efficiency and unstable speed.
Only in commutator machines can be had
frequencies capable of changing without a
change in speed, and vice versa; and by placing
the commutator machines in the secondary
circuit of the induction motor to be regulated,
the commutator machines can be made of
reduced capacities, and if desired, can be
removed from the main motor, as con-
trasted to placing the commutator on the
main driving machine. This of course vastly
simplifies the design problem, cheapens the
drive, and increases its dependability. In
the Kraemer system the rotary converter
speed is of course determined by the slip
frequency, but in this case the desired varia-
tion between slip frequency and speed is
between slip frequency and the speed of the
d-c. regulating motor. As so far developed
commercially, the Kraemer or "rotary
converter" system regulates the speed only
below synchronism, so that the motor may
be operated without the auxiliaries only at
its top speed. This also characterizes the
Scherbius system as built in Europe and
as first built in the United States. John I.
Hull, who is the author of an article on the
subject in this issue, has however developed
this system to a point where it is possible to
regulate the induction motor above as well
as below its synchronous speed, so that the
normal speed of the motor lies in a much
used part of the speed range, permitting the
regulating set to be shut down and its wear
and losses avoided during a great number
of operations. Clearly this affords many
operating advantages, while decreasing the
size of the auxiliaries and increasing the
overall efficiency.
560 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo.
Temperatures in Large Alternating-current
Generators
By W. J. Foster
Alterxatixg-currext Exgixeerixg Departmext, Gexeral Electric Company
When solving the temperature problems that arise in the design of large alternating-current generators,
the first factor to be considered is the relationship between the space occupied by the object and the total
heat losses to be dissipated. Other factors are those of the heat density and thermal conductivity of the various
materials and the construction employed. In the following article, which was delivered as a paper at the
annual Convention of the A. I. E. E. June 29-July 2, the author calls particular attention to certain advantages
to be gained by reversing the present usual practice of ventilating large hydro-electric generators by taking
in air directly from the generator room and piping it out of doors or to some point in the building remote
from the generators. — Editor.
What shall be considered a large generator;
5000, 10,000, 15,000. or 20,000 kv-a. ? Shall it
be the rating alone that is considered, or shall
we take such factors as speed into account ?
Undoubtedly, a large proportion of the
general public think of a large machine as one
that occupies a large space compared with
other machines used for the same purpose.
They judge largeness by physical dimensions
alone. At the same time, it is safe to say the
more intelligent of the general public think of
size in terms of the work that can be done.
Probably a machine of 10,000 kw. or more is
regarded by them as a large machine. To the
engineer, largeness involves the difficulties
inherent in design and construction. A
1000-kv-a., 10,000-cycle alternator is a large
one; a 5000-kv-a. generator of 3(500 r.p.m. is
large. Considered strictly with reference to
the temperature problem, the engineer would
hardly consider a 20,000 kv-a. generator a
large one if the periodicity and potential were
those in regular commercial use and the speed
were 100 r.p.m. or thereabouts. However, for
the purpose of this article we will consider a
20,000-kv-a. machine as large.
There are two principal factors in the tem-
perature problem in even,- case; first, the
total losses or the amount of heat energy to
to be disposed of and its concentration; and,
second, the means that can be provided for
dissipating the heat in such a manner as not
to cause damage to any part of the machine.
The problem may be attacked along the lines
of reduction of losses or of devising such con-
structions that the heat may be more eflfec-
tively dissipated.
In a rotating dynamo-electric machine
three sources of heat are always involved;
first, hysteretic lo.sses in the magnetic ma-
terial ; second, the resistance to flow of current
losses in the windings; and, third, the fric-
tional losses in the bearings and the windage.
The first two are electrical in their nature;
the last is mechanical.
Combined with hysteresis losses in the
magnetic material are more or less eddy cur-
rent losses. The total losses in the magnetic
material are dependent upon such factors as
the degree of lamination employed, the
character of the insidation between lamina-
tions, the amount of pressure employed in
clamping the cores, as well as the character
of the steel employed. In like manner, the
resistance losses in the copper are frequently
accompanied by eddy current losses, the
amount of which is dependent upon such
factors as the stranding of the conductor, the
pitch of the winding, and the arrangement of
the turns. The windage losses, or the losses
that result from either the fan action of
the rotating parts or from the disk action or
rubbing of a re\-oh-ing body on the surround-
ing air, are dependent upon the peripheral
speed and the details of design of the parts
that are producing fan action.
In considering the temperattu-e problem,
the first and most fundamental consideration
is the relation of the space occupied by the
object to the total heat losses to be dissipated.
A 20,000-kv-a., 100-r.p.m. machine compared
with one of the same output at ISOO r. p.m.,
has its losses generated in a space eight times
as great in terms of cubical space occupied, or
appro.ximately two tintes in the projected
area occupied. We may well think of the
temperature problem in terms of heat losses
to space occupied. Below a certain value of
this constant it is absurd to use ventilation
housings, no matter how great the rating of
the machine, as such housings have the effect
of preventing the natural means of heat dissi-
pation; viz., convection and radiation, and
ventilation housings in such cases result in-
higher temperatures, unless forced draft is
provided, which results in a decrease in the
TEMPERATURES IN LARGE ALTERNATING-CURRENT GENERATORS 5G1
efficiency of the unit and can be justified
only on the score of reduction of noise or
some similar reason.
Second in importance to the space factor
comes the heat density factor. By this we
mean the quantity of heat energy passing
through a unit area of material.
The third factor is the thermal conductiv-
ity of the various materials; a factor which
depends not only upon the thermal properties
per se of the materials but also upon the man-
ner in which the materials are put together.
Classificatioft of Machines with Respect to Ventila-
tion
Attempts have been made to standardize
various classes of machines with respect to
ventilation, but the writer thinks it safe to
say that nothing yet has been suggested
which appeals to engineers in general as
entirely satisfactory. Possibly it is desirable
to have a large number of classes of machines
to fill in the gap between the extremes of the
lowest speed small capacity machine that
requires no special provision and may be said
to depend upon natural ventilation alone,
and the highest speed machine that requires
the most careful artificial ventilation. It is
difficult to classify the types that have al-
ready been developed to fill in this gap as they
blend into one another.
The points to be kept in mind in the design
and construction of machines in general, not
particularly those standing at the extreme
ends, are: first, the obtaining of a supply of
cooling air from a region well removed from
the space into which the outlet air is dis-
charged; second, the placing of barriers or
bafflers to assist the flow of air and to prevent
re-circulation; third, the providing of ample
cross section in all parts of the paths of flow
of cooling air and the avoidance of sharp con-
trasts in the cross-sectional area of the paths,
especially the avoidance at any point of
greatly reduced cross-section that would in-
troduce great resistance; and, fourth, the
avoidance of "churning of air," or internal
circulations, which are often hard to prevent
by reason of the irregular shapes of the dif-
ferent parts of the machine.
Closed and Semi-closed Ventilated Machines
In almost all large generators, whether
hydraulic or steam turbine, it is necessary
either to pipe air to the machine or away
from it, or both to and away from it. Prob-
ably the most common practice is to pipe air
to the machines, allowing it to escape through
the stator frame into the dynamo room, the
escape often being arranged so as to be up-
wards, which is preferable on account of the
greater comfort to the operators, the reduc-
tion in noise and the slight reduction in tem-
perature obtained by the lower temperature
of the air immediately around the machine.
The writer wishes to call attention to cer-
tain advantages that would result in reversing
the common practice of the present time, in
the ventilation of large generators in hydraulic
units, and to take the air in directly from the
room and pipe it away either to some point in
the building removed from the machine or to
out of doors. The advantages of this arrange-
ment have already appealed strongly to the
operators of some of the largest hydraulic
generators, and such a system is now in use
in a few plants. It is a much simpler matter
to draw air into the rotor direct from the
room at the two ends of the generator than to
provide the necessary space for the air con-
duits and the housings required either at the
one end or the two ends, which almost in-
variably involve greater distance between
bearings and, consequently, an increase in
both the diameter and length of the shaft and
corresponding parts. A great advantage of
the scheme of piping air away is the more
comfortable temperature of the dynamo room
in hot weather. It is never necessary to be
in an atmosphere of higher temperature than
that existing out of doors, whereas, in case of
the more common practice, the air surround-
ing the machine has its temperature raised
several degrees above that of out of doors,
due to the heat that has been added to it
when passing through the machine.
Water Cooling
Water is an ideal agent for cooling purposes.
At first thought it seems strange that it has
not been made greater use of in removing
heat from large machines. A small quan-
tity of water, on account of its high specific
capacity, would suffice to remove heat from
a large generator, but the difficulty is in ar-
ranging jackets that will prove safe and
can be located in close enough proximity
to the parts in which the losses are generated
to remove such losses without a considerable
drop in temperature through the intervening
walls of material. It is apparent at once that
water-cooling is much better adapted to the
stationary than to the revolving parts.
While it is possible to arrange for a flow of
water through the revolving parts, it is
probably not possible to so arrange the flow of
562 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo.
water as to cool the surfaces where most
of the heat is generated. Hence, a system of
water-cooling would be dependent upon the
joint action of air-cooling and would require
a design that would re-circulate the air sur-
rounding the rotor in such manner as to most
I
!
I
1
1
1
t
1
1
^
-+-
Sffiutnaj
1
'
;
,
1
^1
{
1
\
1
1
1
^^
■
1
1
1
1
Vj
'^
1
1
'
;
s^
■n^
1
J
1
1
^
^
"■^
■^
ei
^
1
_^
_^
!r^
==-^
W&5£ci-
"**
1
="
^
'^
—
' — '
in Fieldi
II
^
— 1
— 1—
.4..
_L
^
M*
1 — '
-^
_H
^-
i-
TT
—
—
-—
==
1^ ,
-
^
— ^
^.
_^
id
i 1 1 1 1 _l_l
-d
=3=
^^
-|-
-p
-^~
~r
H
=r
--^
T
-^
-i
~—
1
.
;
1
1
1
1
;
'
1
1
1
_
,**-*
1
,
■
■—
' (
-+-
,
„j^
■^
1
-
-r-
,
1
1
1
r ' I 1 1 ■ 1 ;
,
1
zoo
ISO
100
so
6000 8000 lOOOO 12000 I4O00 16000 ISOOO
7000 9000 IIOOO 13000 ISOOO 17000
Peiiflhtrol Helocit4f • Feet ptr Minutf
Fig. 1. Losses and Efificiencies as AflFected by Choice of Diameter
for a 20,000-kv-a.. 11.000-volt. Three-phase. 360-r.p.m.,
60-cycle,0.8-power-factor. Two-bearing Generator
effectively carry the heat from the surfaces of
the rotor to the surfaces of the water piping
in the stator. Another objection is the danger
of injuring a machine in case the water cir-
culating system becomes leaky. Still another
is the danger of too great an accumulation of
dampness due to condensation, at certain
times, of the moisture in the air on the water-
cooled parts. It is doubtful whether water-
cooling can ever be a competitor of air-cooling
in dynamo electric machines.
Peripheral Speed of Rotor
It is possible that the electrical advantage
of higher peripheral speed in almost all de-
signs of large generators is not fully appre-
ciated by many designers themselves. The
losses in both the iron and the copper are
almost universally less at the higher periph-
eral speed in any practical problem. As-
suming the same characteristics electrically
in all respects, such as saturation cur\^e,
armature reaction per unit pitch, etc.. the
losses in the armature teeth are less at the
higher peripheral speed. They are exactly
inversely as the peripheral speed, while the
core losses proper remain practically constant.
Copper losses in both armature and field are
less at the higher peripheral speed, unless
the speed is carried to an absurd limit. The
windage losses will be increased. Hence,
considered electrically, the most efficient
design will be that where the windage losses
begin to increase so rapidly as to offset the
combined reduced losses of core and windings.
For the purpose of illustration, the cur\-es in
Fig. 1 have been worked out for a 20,000-kv-a.,
60-cycle, 360-r.p.m., 3-phase. 11,000-volt
generator at peripheral speeds var\4ng from
(iOOO to 18,000 feet per minute.
The variation in the segregated losses, as
affected by different rotative speeds, is illus-
trated by the curves added in Fig. 2 which
show such losses for 20,000-kv-a. generators
throughout the range 100 to 600 r.p.m. It
should be understood that these generators
are designed with identical electrical char-
acteristics and have the same temperature
rises; viz., those corresponding to the A.I.E.E.
Standard for Class "A" insulation. They
are what the writer considers normal in de-
sign for the output at the several speeds.
■^100
■
;
1
l""
I— _
__
^
t^
.*— rt — 1
i «
^^
—
-
-^—
£ffK
"^r
;
~
■"
■
^
%
1
s
<
•■
E
Tffll
"
■1
"
"
rtr
"J
"
5
~
4"
-
""
1 1
'/M/ '
\400
k
■
*
1 1
^
hH
h
«■
■*
'
1
•^J^ti
J^
-
.-
[_ ; ti ■ t "
A-^
rf
•«■
^300
— "
—
-
— r~
^ftJ
~*~
"■
~
~"
1
1
_. a'^^i'
-
'
-
\
-
"
■ "-
■■
1
^
^
^
Fr/ttipfi
M^
^
VN
ki
rr
1
1 1 1 1
li
W
u
w
J
W
#
M
n
RT
n
n
Rtr otut ions ptrAfint/t*
Fig. 2. Losses and Efficiencies as Related to Rotative Speed of
a 20.000-kv-a,. 10.000-volt, Three-phase, 60-cycle,
0.8-power-faclor Generator for Speeds from
too to 600 r.pm.
The lOO-r.p.m. generator has a peripheral
velocity of about 7.")00 ft. per niin.; the tiOO-
r.p.m.. l.>,00() ft. per min. The first has its
losses generated in a space of approximately
2200 cubic feet, the last in approximately
,)50 cubic feet. The total losses of the first
TEMPERATURES IN LARGE ALTERNATING-CURRENT GENERATORS 563
arc (i20 k\v., or 270 watts per cubic foot of
space occupied; of the last, 495 kw., or 900
watts per cubic foot of space occupied.
Hence, the ventilation problem is quite dif-
ferent in the two cases. The first might be of
the open type, drawing its ventilating air
from the room and returning it directly to
the room; the last must be enclosed, prefer-
ably totally enclosed.
Ventilation Ducts in Armature Cores
The common practice for ventilating anria-
tures is to provide at short intervals in the lam-
inated core, a narrow passage, usually % in.
or 3^ in., for the air to be driven through
radially by the fan action of the rotor, or in
special cases by an external fan. The flow of
air is in any special case dependent upon the
details of construction, such as the character
of the space blocks, how these are located
with respect to the coils in the slot ; the niceties
introduced at the entrance from the airgap
in the way of treatment of the retaining
wedges of the windings, the exact location of
the end of the spacers, etc.
Good results are usually obtained by having
a ventilating duct about every two inches.
Theoretical Temperature /?i\se
t^ SO
zs
3000
Zzooo
^
■«,
ifi
.
•*
^
—
•-
/
\
^
V
*N
^
~^
—
—
>
^
^
l/filnrif-ii
of A/ra
to ^>r n
■"s
^
^
/ct
iv.
inimii
Fig. 3. Ventilation of Stator Core Having Ducts Equally
Spaced and the Air Admitted at the Two
Ends of the Air Gap
the most efficient spacing being dependent
upon such factors as the radial depth of the
core, the length of the core, and the pressure
of the cooling air. In long cores the spac-
ing may be graded and the sections of core at
the middle made smaller than at the ends.
There are two reasons for this arrangement;
first, some of the heat at the ends travels to
the head of the core where the cooling con-
ditions are usually good and, second, the
ventilating air gathers up heat as it passes in
from the head towards the middle and hence
gv 3-0
IS
3000
«, zooo
1000
Theoretical Temperature Rise
of IVinding in Slots
>
rn
kP
~
~
~~
~
~
r
..
-
~
=
r:
"
-
'"■
■"
"
~
Theortiicol Temperature R/ss
s
^
of Venti/atinQ Air
^
>
^
-
-
^
"
""
^
-'
^
^
■"
^
ve/outf^ of Air at
^
entrance to Air Oucts
L
LI
DfflDOQ
D
f^Qtor (Re¥o/¥ing riefdj
DDDQDDliiraDD
Fig. 4. Ventilation of Stator Core Having Ducts Unequally
Spaced^and the Air Admitted at the Two
Ends of the Air Gap
is not as good a cooling medium when it enters
the ventilation duct as the air in the ducts
nearer the head. When fans are mounted at
the two heads of the rotor and end housings
are placed on the stater so as to establish a
good air pressure, the pressure in the ventilat-
ing ducts increases from the head to the
middle. Hence, the quantity of air passing
through is greatest in the duct at the middle,
decreasing toward the heads. The curve in
Fig. 3, entitled "Velocity of air at entrance
to air duct," was plotted from air pressure
readings made on a large turbo-generator
with equally spaced ventilating ducts. The
other curves were determined from a con-
sideration of the heat dissipation problem.
In like manner. Fig. 4 shows curves for a later
turbo-generator with stator core sectionalized
in such a manner as to equalize the tempera-
tures throughout the length of the core.
Giving Direction to Cooling Air
It is quite wonderful what improvements
are sometimes accomplished in cooling ma-
chines by very simple expedients. Sometimes
it is advisable to arrange a machine so that
it is obliged to take all of the cooling air in
564 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 7
at one end and to discharge it at the other.
This, as a rule, is especially helpful to the rotor.
But, in general, machines arranged with
a radial system of air ducts through the core
should draw the air in equally from both ends.
Often a machine that seems at a casual glance
HO
~\
-
_
m
-
r
1
r
'TFt^
—
-
—
^
L
1
--
-
X
^' ---
/
-
+-
(
'
^0
[
:
*
-
_
-f- Armo
tare PVind/ng
..7.
—
^
—
7^
~
--
^
= = = -- +
,
1
^-
1
I
jaw
'
/
.^
!
1
i
— ■
/
' /■
1
/
/
1
\ 1
^
'
//
^40
~'t7~
-T
1
(• V,
■
77
I
T
~1
n\
1
1
zo
// J.^
t2-4+-
exate^taAfOf
wo/vo/ts —
»
„-^-^ — ' — ' 1 .1,
\'SZ\ ': L
' \ \ , ' : 1 I 1 . 1 1 1 1 i 1
Ff -rt n i 1 1 1 1 1 1 M 1 1 1 1 1
as 10 IS it) zs it iS
Time-Hourz
Fig. 5. Time Required to Reach Constant Temperature in an
18,750-kva., 11,000-volt, Three-phase, 60-cycle. Cylindri-
cal-rotor Generator Operating at Overload
as a Synchronous Condenser
to be sjTnmetrical as to the two ends proves
to be a surprise in taking practically all its
air from one end. In such case, little, if any,
air passes outward through the radial air
vents; in fact, sometimes the air will pass
inwards in some of the vents. The remedy
is usually ver>' simple; any little barrier in-
terposed in the path of the axial flow will
restore the desired circulation and often re-
duce the temperature several degrees.
The poles themselves act as fan blades on
the rotors of many salient pole generators,
and no fans or fins for additional fan effect
are required. It may not be generally known
that even in cases where carefully designed
fans similar to those used in large cylindrical
rotor turbo-generators are employed, the
poles themselves contribute more to the
blower action than the fans. The problem of
ventilation in salient pole machines is more
complicated than in cylindrical rotor ma-
chines, where the blower action is more
largely due to fans designed for the purpose.
Heat Flow
The most efficient ventilation of a large
electric generator requires circulation of the
cooling air in such manner as to bring it in
contact with large surfaces of the solid mate-
rials in which heat is being generated and close
to the sources of the heat generation.
The heat resistance of the various materials
entering into the construction, such as copper.
magnetic steels, and various insulating mate-
rials, is quite well known.
An analysis of heat flow in a 30,000-lc\--a.
generator from the inside of an armature
coil at the middle point of the core to the
ambient cooling air is as given in Table I
for the following four designs:
(a) 4000-volt mica-insulated coils to with-
stand A.I.E.E. high-potential tests.
(b) 11,000-volt mica-insulated coils to
withstand A.I.E.E. high-potential tests.
(c) 11,000-volt mica-insulated coils with
copper density same as in 4000-volt
design.
(d) 11,000-volt mica-insulated coils to
withstand high-potential test of three
times normal, instead of two times plus
1000 volts (A.I.E.E. Standard), with
coils of same external dimensions as
those of (b), so as to be assembled in
the same slots.
TABLE I
TEMPERATURE DROP IN DECREES C.
Coil Design
(»)
(b)
(c) 1 (d)
Drop through insulation.
Drop through core
Drop at surface
Drop in cooling air
21
6
16
15
30
4
11
15
48 67
6 5
16 14
15 15
Total drop
58
60
85 101
Heat Storage
Heat storage must be reckoned with when
ratings for intermittent loads are to be given
to a generator; but for continuous load ser\-ice,
as in nearly all large commercial generators,
the heat capacity properties are chiefly of
scientific interest, except when the duration
of heat runs in acceptance tests is under con-
sideration. Fig. 5 shows cur\-es of time re-
quired to reach constant temperature, in the
case of an lS,750-lc\--a. turbo-generator at
overload corresponding to about 20,000 kv-a.,
which may be taken as typical of the mod-
em large cylindrical rotor generator. This
set of curves represents three runs under
widely different conditions. The cur\-es,
"Field winding" and Armature winding,"
were determined in the same run. The
curve. " Uncxcited" shows the rate of temper-
ature rise when the heat is generated by
windage alone, as measured by detectors
embedded in the armature slots. The cur\-e.
"Excited to normal volts," shows the rate of
rise measured in the same manner, when the
heat is that of core losses on open circuit in ad-
dition to the windage. Fig. 6 gives cur^•es
TEMPERATURES IN LARGE ALTERNATING-CURRENT GENERATORS 565
of temperature rise in a single run on a salient
pole generator. Comparing the two sets of
curves, it is interesting to note the quicker
rise in the field winding of the salient pole
machine, where nearly all the heat passes
directly into the cooling air from the surface
of the bare copper, over that of the cylindrical
rotor machine where the field winding of each
pole consists of several coils embedded in slots
in the magnetic material.
High vs. Low Temperature Generators
Undoubtedly any machine, electric genera-
tor or strictly mechanical machine, such as
the steam turbine, would be better off if it
could always maintain the same temperature
in all its parts. A rise of temperature beyond
certain limits, repeated often enough, results
in deterioration in most electric generators.
This is due primarily to an effect of heat that
is mechanical in its nature; viz., a change in
size. Much can be done to minimize the dele-
terious effects of change in size of the various
parts by introducing constructions in detail
parts that automatically adjust for changing
size. But it is extremely difficult in certain
parts to protect materials of quite frail me-
chanical nature, like many insulations, from
the effects of change in compression or what
is more serious, shght movements of different
degree in different places. Looked at in this
way, it is desirable to have a generator of low
temperature rise. But it is not always con-
venient or possible to build low temperature
generators if machines are to be produced
equal in capacity to prime movers. Further-
more, high temperature machines are justified
in cases where increased efficiency or lower
cost will more than offset the shorter life.
With reference to relative efficiency and
cost, it is quite apparent that the high tempera-
ture machine has the advantage in the case
of a generator whose insulations are suitable
for the higher temperature, and whose effi-
ciency is still rising with increase of load, as in
most alternating-current generators, and whose
cost is only slightly increased by enlarging the
shaft and other mechanical parts involved.
It may be laid down, as a rule, that the
high temperature machine costs less, but it
must not be taken for granted that its effi-
ciency is better. In fact, most of the lower
speed machines of low temperature rise have
better efficiency than corresponding machines
of high temperature, unless the designer has
been grossly careless in taking care of the
ventilation of the former.
It is possible to design most machines for
low temperature rise without great additional
cost, and to have decidedly better efficiency.
This condition obtains especially in connec-
tion with large salient pole generators of low
speed and low or medium potentials. As a
rule, machines of this class, designed for 50
deg. C. rise, permit of decided increases in
(,0
r
.
.
— .
—I
— r
— r
■— r
— T
— r
— r
— T
— r
— Tl
-
-
^
-
-
-
-
-
-
-
-
^
^
r/e/<f t^/n<fmg
~
H
"
~
y
/
-
-
-
_
r
-
-
7
/L
-
r-
~
-
C-
-
__
^
=
=
r=
40
t
A
-
;
z
^
;
5
E
Z
-
rt
Armoture Vrndma
-MM III
^
-
—
=s
f
^
-
-
-
^
-
P
f^
>
Armature Core |— |
30
■
/'
^
.-
"
/
f
^
i
/
y^
f
/
ZO
/
/
/
V
r
f/
I
f
1
10
u
f
f
'
O 05 10 15 20 ZS 30 35 40
Time -Hours
Fig. 6. Time Required to Reach Constant Temperature in a
12,500 kv-a, 22,000-volt, 50-cycle Synchronous
Condenser Having a Salient Pole Rotor
the amount of copper in both armature and
field without any change except slightly
larger slots in the armature. In addition, a
higher grade of magnetic steel may be used
than that called for by temperature con-
siderations. Often one per cent in efficiency
at full load may be gained at an increase in
cost of 10 to 15 per cent. In other cases as
much as Yi per cent efficiency can be gained.
The resulting generators may have only 35
or 40 deg. C. rise at rated load.
Possibly the author is on dangerous ground
in discussing the advantages of generators
that do not conform to the Standardization
Rules of the A.I.E.E. However, it is not
for a moment his intention to reflect in the
slightest on the standards that have been set,
but he wishes to point out gains to the user
to be had by following along more conserva-
tive lines in certain cases. Again, he realizes
the possibility of trouble to himself and his
ilk from urgent requests that may be in store
from buyers of generators, when generators
are under consideration that cannot economi-
cally be built for temperatures below the
Standards of the A.I.E.E. in the high and
low temperature classes. It is well to add
that designing certain sizes for low temper-
atures, where only slight gain or no gain
whatever in efficiency results, often involves
hardship and results in certain risks being
taken that are wholly unjustified.
3(36 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No.
Exciters and Systems of Excitation
By H. R. SUMMERHAYES
Engineering Department, General Electric Company
The continuity of service rendered by a generating station is directly affected by the reliability of its
excitation system. Therefore, in the selection of a system, first cost and economy of operation are of lesser
importance. In the following article, which was read as a paper at the annual convention of the A.I.E.E.
June 29 to July 2, systems of excitation are grouped and discussed as common excitation plants and as
individual exciters. In addition to a comparison of these two systems and a comparison of shunt, compound,
and commutating-pole exciters, there is included a discussion of such related factors as: method of exciter
drive, voltage, rheostats, field switches, batteries, voltage regulators, and station auxiliaries. — Editor.
In laying out the excitation system for the
generators of a central power station the
primary requirement is reliability; that is,
continuity of serv'ice. First cost and economy
in operation are secondary, but, nevertheless,
must be given consideration.
To meet the first requirement :
(1) The exciters should be machines of
good design and liberal size.
(2) The method of drive should be reliable.
(3) All electrical connections and %viring
should be as short and simple as pos-
sible, and located and supported so as
to be safe from external injury.
(4) The method of control should be
simple and reliable, and the operation
convenient.
(,)) Resen-e capacity should be supplied
and reserve driving source.
The systems of excitation which have been
used or proposed may be divided into two
general classes:
(1) Common excitation jilant (exciters
operating in parallel on a bus supply-
ing excitation to all generators).
(2) Individual exciters (not operating in
parallel).
The first system was for many years the
standard American practice for both steam
and hydro-electric plants, excepting in some
small plants where belted individual exciters
were commotily used.
Euroi)ean practice, on the other hand, has
shown a preference for individual exciters,
and in recent years American practice has
tended toward their use, for reasons which
will be discussed.
One reason for the American preference for
a common excitation plant may have been
the use of large alternators driven by low-
speed Corliss engines, on which it was rela-
tively expensive, in cost and floor space, to
arrange for direct-connected exciters.
At the same period, European plants were
installing high-speed vertical engines, for
which the exciters on account of the high
speed were of small dimensions and weight
and could readily be overhung on extended
^hafts.
When steam turbines came into general
use, manufacturers were somewhat unwilling
to lengthen their shafts and to complicate
their problems of balance, expansion, etc.,
for the purpose of adding direct-connected
exciters, and for vertical shaft turbines there
was the further objection that the exciter
would be in an inaccessible location. There
was also the conservatism of power plant
engineers and the general appreciation of the
reliability of excitation afforded by ha\nng a
battery floating on the common excitation bus.
It is interesting to note, however, that of
the steam turbines, 7500 k\--a. and over,
sold by one manufacturer during the last five
years, about 45 per cent were equipped with
direct-connected exciters; and of the gen-
erators, 1000 k\'-a., and over for watenvheel
drive, made by the same manufacturer, 75
per cent had direct-connected exciters.
Some of the hydro-electric generators of
low speed and large size without direct-con-
nected exciters were equipped with individual
exciters driven by motors.
COMPARISON OF VARIOUS PLANS OF
EXCITATION
Common Excitation Plants
Common excitation plants in which the
exciters are operated in parallel on a com-
mon bus have the advantage, as compared
with individual exciters, that the bus voltage
is kept constant so that a storage battery may
be kei)t floating on the bus at all times ready
to take up the excitation load in case of ex-
citer trouble; also that the constant voltage
exciter bus offers a source for the supply of
lighting, attxiliaries, and sometimes the control
of electrically operated switches. If automatic
voltage regulators are used directly on the
exciters this constant voltage is no longer
maintained and tliis advantage disappears
unless a regulator is used on a booster between
EXCITKRE AND SYSTEMS OF EXCITATION
567
568 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 7
the constant-voltage exciter bus and a varying
voltage bus to which the generator fields are
connected. These common excitation plants
have the disadvantage that any trouble on
the main exciter bus may cause a shut down
on the entire generating station.
Fig. 5. Direct-connected Exciter Mounted at the Generator
Bearing and Collector End of Shaft
Individual Exciters
In the case of individual exciters, where
one exciter is supplied for each machine and
the exciters are not normally operated in par-
allel, trouble on one exciter circuit will affect
only one generator. The exciter circuits are
short and simple and are not liable to trouble.
Methods of Driving Exciters
Whether the common excitation plant or
individual exciters are used, the method of
drive is important.
For individual exciters usually only two
methods of drive are used, namely, exciters
directly connected to the generator shafts
and exciters driven by motors.
In the latter case the motors may be con-
nected to the main bus or preferably they
should be connected to an auxiliary bus sup-
plied by an alternating-current generator
driven by a prime mover. Transformers are
also furnished, so that the motors may be
supplied from the main bus in emergency.
This method of driving individual exciters
is used chiefly for large h\'dro-clt.'ctric plants
where, on accoimt of the low speed of the ver-
tical shaft generators, direct-connected ex-
citers become too expensive.
For individual exciters it may be said that
those which are direct connected are prefer-
able on account of cost, reliability of drive,
and shortness and simplicity of wiring.
Direct-connected exciters large enough to
excite two units are sometimes specified. For
steam tiu-bines, such large exciters may be
undesirable on account of their weight and
size being too great to overhang on the ex-
tended shaft. The exciter drive should not be
allowed to jeopardize the continuity of opera-
tion of the main turbo-generator. For turbines
up to 1800 r.p.m. direct-connected exciters are
reliable machines and have given good ser^nce
records. For turbines of 3000 r.p.m. direct-
connected exciters are often used, but in order
to obtain the best results as to commutation,
and to make such machines as reliable as those
of lower speed, great care must be exercised in
manufacture.
Exciter Drive in Common Excitation Plants
In the case of common excitation plants a
number of arrangements for driving the ex-
citers are in use. The most reliable and ef-
ficient arrangement is the direct-connected
exciter, unless there are reasons, such as too
high speed or too low speed, against using it.
Belted units are widely used in small plants
where the engine speeds are low and the use
of a belt involves ver\' little risk or trouble.
On account of the low engine speed a consider-
able saving of cost and space is made by using
belted instead of direct-connected exciters.
Geared exciters have been proposed for large,
low-head, hydro-electric plants.
The plan generally adopted for a common
excitation plant is to have some of the excit-
ers motor-driven through transformers from
the main alternating-current bus and some of
them driven by separate prime movers.
Another plan which has been used in con-
nection with some large steam plants is to have
the exciters motor-dri\-en from an auxiliary
alternating-current bus supplied by auxiliary
generator units designated as"house turbines."
Transformers connecting the auxiliary bus to
the main bus are supplied for emergency use
or for adjusting the power on the auxiliar>'
bus for heat balance purposes. Tliis aux-
iliary bus is used also for the su])ply of aux-
iliary power for the whole station, such as
circulating water, air and hot-well pumps,
stoker motors, economizer and draft fans,
coal crushers and conveyors, etc. In ver\-
large stations an auxiliary bus and its gener-
ating unit may be supplied in connection with
each main generating unit on the system.
An arrangement commonly used in hydro-
electric plants and used occasionally in steam
plants is to have each exciter connected to a
prime mover and to an alternating-current
EXCITERS AND SYSTEMS OF EXCITATION
569
motor supplied from the main bus, so that
the exciter may be driven by either or both.
This arrangement has been used in steam
plants of moderate size and in hydro-electric
plants for the following reasons:
The reason which applies to both cases,
is to have two separate sources of power for
the exciter drive. In hydro-electric plants
for high head where the exciter waterwheel
nozzles on account of their small size are
likely to become blocked, it has been for many
years the practice to have an induction motor
connected to the bus mounted on the same
shaft as the exciter and the waterwheel, so
that when the waterwheel fails to carry the
load the induction motor will take it up. In
steam plants the chief reason for using this
arrangement is to provide means of adjust-
ing the amount of exhaust steam available
to heat feed water, which is done by adjust-
ing the governor of the exciter unit to take
more or less power, the remainder being sup-
plied from the motor.
This arrangement in steam plants has the
disadvantage that to obtain an efficient tur-
bine the speed must be high, possibly too
high for the proper design of the direct-cur-
rent generator, or of the motor, necessitating
sometimes a geared connection which of
course is disadvantageous for a high-speed
continuous running unit.
The plan of using direct-connected exciters
on the -main generator shaft, exciters not
operating in parallel, the voltage of each gen-
erator controlled by the exciter field, appears
to be the most reliable and simple method of
excitation for large plants wherever the speed
requirements are not prohibitive.
For all large stations using individual ex-
citers, it is desirable to have an emergency
excitation bus with a reserve exciter driven
by a separate steam turbine, waterwheel, or
motor, so that any generator field may be
thrown on this bus in case of trouble with one
of the individual exciters. The question as
to whether a storage battery is necessary will
depend on the number of units in the plant
and on the importance of the service.
VOLTAGE OF EXCITER PLANT
For many years the standard excitation
potential has been 125 volts and this pres-
sure is still standard for small and medium
size plants. In recent years 250 volts has
been coming into use and has now become
standard for large plants for the following
reasons :
The difficulty and expense of building high-
speed commutators for 125 volts, especially
turbine-driven exciters, or waterwheel exciters
which stand double speed. The space occu-
pied by the commutator is reduced at 250
volts. The expense of busbars, machine
leads, circuit breakers, etc., to carry the large
currents necessary at 125 volts, especially in
m o
Fig. 6. Waterwheel-driven Exciter of 600-kw. Capacity
600-r.p.m. and 220-volts
large stations where the field currents are
heavy and in long stations where the dis-
tances are great.
EXCITATION REQUIREMENTS OF
ALTERNATORS
Voltage Range
In steam turbine generators the armature
reaction may be about equal to the no-load
ampere-turns. This means that with 100
amp. field current required to give full voltage
at no load, 200 amperes would be required to
maintain full voltage at full load at the rated
power-factor. Since the alternator fields must
be designed to take not over 125 volts at
rated power-factor full load and maximum
temperature, and because a margin must be
allowed in the design for variation in the
material, etc., the actual machines may meet
their requirements at 90 to 115 volts across
the field, and this means that at no-load full
alternating-current voltage the exciter pres-
sure may go as low as 40 or 50 volts, or about
30 or 40 per cent of the rated voltage. The
exciters and their rheostats and regulators
must be designed for this range of voltage.
For synchronous condensers the range of
exciter volts is down to 10 per cent or less of
full pressure, and for synchronous motors the
range depends on the range of power-factor for
which they are designed.
570 July, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII. Xo. 7
Kilowatts Required
The excitation requirements of alternators
vary according to the design, but for modern
standard lines may be summarized as follows
in per cent of the kilovolt-amperes alternator
rating :
Steam Turbo-generators Per Cent
1000 to oOOO kv-a. 0.5 to 0.3
7500 to 3o,000 k'^•-a. 0.4 to 0.3
W aterwheel-driven Generators
1000 to 5000 ]VK. low speed 1.5 to 0..S
1000 to 5000 kw. high speed 1 to 0.5
7500 to 20,000 kv-a. low speed.. (). 7 to 0.5
7500 to 20.000 kr^--a. high speed .0.5 to 0.4
Motor Generators
1000 to 5000 k\'-a.
1 to 0.5
two to two and one half times the resistance
furnished with ordinary direct-current gen-
erators not used for exciter purposes.
For a common excitation plant with hand
voltage regulation, where the alternating-cur-
rent voltage is controlled by the alternator
field rheostats, the exciter rheostats may be
of ordinary design with resistance points
closely graduated from S5 to 100 per cent of
full exciter voltage and further apart for lower
\'oltage ranges.
For individual exciters, non-automatic volt-
age control, where no generator field rheostats
are used, the exciter rheostats should have
closely graduated resistance steps all the way
down to 30 per cent of the voltage and may
have as many as 100 to 150 steps. When this
method of regulation is used, the alternator
field rheostats may be dispensed with, but it
Fig. 7. Exciter Direct-connected to a 7500-kw., 1800-r.p.m., Curtis Steam Turbine Generator Set
RHEOSTATS FOR EXCITERS
For small exciters hand-operated rheostats
mounted on the back of the switchlx>ard, or
operated by chain drive from a handwheel on
the switchboard, are generally used.
For large plants the alternator field and the
exciter field rheostats are nearly always elec-
trically operated and this method of opera-
tion is recommended for any plant where an
electric control circuit for operating rheostats
is available and where the main control board
is on another floor or distant from the
machines. For all such ])lants convenience
of operation, location and wiring, as well as
cost considerations, will usually give electri-
cally operated rheostats the preference.
For operation with automatic voltage reg-
ulators the exciter field rheostat is generally
made three to four times the ohmic resistance
of the exciter, shunt field winding, or from
is considered better practice to install them
for emergency use in case some other source of
excitation is resorted to; and it will usually be
found that more stable operation at the lower
ranges may be obtained by the use of these
field rheostats to a certain extent to enable
the exciters to work at somewhat higher volt-
age. If the alternator rheostats are used there
is a slight sacrifice of efficiency.
CIRCUIT BREAKERS AND FIELD SWITCHES
As a general i)rinriiiK' no aulimiatio over-
load circuit -breaker or fvise should be installed
in exciter circuits. When the alternator is
short-circuited the altemator field current may
rise to several times normal in the normal
direction and an automatic circuit-breaker
under such conditions might interrupt the
exciter circuit, which must not he allowed.
Short circuits in the .generator field circuits.
EXCITERS AND SYSTEMS OF EXCITATION
571
or on the exciter busbars, are an infrequent
occurrence and should be taken care of by the
operator. It is considered better to risk injur v
to the exciter than to install overload devices
which may operate at the wrong time.
With exciters operating in parallel it is
desirable to have circuit-breakers between
the exciters and the direct-current busses
operated by reverse current in case of trouble
in an exciter or its prime mover.
Alternator field switches should be equip-
ped with discharge resistances. Field switches
and exciter switches should be electrically
operated in all large plants and in other plants
when dictated by convenience of operation,
location, and wiring. It is desirable, of course,
to keep the field switches as near to the alter-
nators and the exciter switches as near to the
exciters as possible, and to locate the exciter
as near to the generator as possible in order
to keep the exciter circuits short, since the
shorter they are the less the chance of trouble
to an liour. This battery may be used as fol-
lows :
(1) Floating on the constant-voltage ex-
citer bus.
(2) In reserve on the emergency bus.
(3) Battery separated into halves, each
floating through a high resistance on the va-
riable voltage exciter bus with automatic
switches to cut out resistance and throw
the two halves of the battery in series when
required for excitation.
Charging a battery may be provided for
by an exciter set designed' for high voltage,
by a special booster or a charging set, or by
separating the battery in two halves and
charging through resistance from the exciter
bus.
In many stations the exciter bus is used to
supply current for the control bus for the work-
ing of motor and solenoid operated circuit-
breaker switches, field rheostats, indicating
Fig. 8. Exciter Direct-connected to a 300-kw., 3600-r.p.m., 2300-volt Curtis Steam Turbine Generator Set
and the better the economy. Hence, hand-
operated field and exciter switches may be
used even in large plants if operated by the
floor men, but it is generally desired to operate
them from the central switchboard.
Field switches should never open on over-
load, but may be made to open automatically
when the main alternating-current circuit
breaker opens by the action of reverse power
or differential relays in the main alternator
leads.
When individual exciters are installed, au-
tom.atic throw-over switches may be used to
throw the alternator field over to a reserve
exciter bus in case of faikire of the individual
exciter.
lamps, etc. It is now considered better prac-
tice to provide a separate control bus, for the
reason that during short circuits on the main
generators transient high pressure of the order
of several hundred volts may exist in the
generator field circuits and alternating cur-
rents of normal or double frequency are super-
imposed on the field currents.
It is difficult to insulate the multiplicity of
switches, lamp sockets, wiring, etc., in the
control circuit for such high voltages, owing
to the limited space requirements for control
boards. For these reasons, in large stations
a separate control bus with a small battery
and motor-generator charging set is generally
installed.
EXCITER BATTERIES
In most large steam stations, and in some
large hydro-electric stations, a storage battery
is provided capable of carrying the excitation
requirements of the station for thirty minutes
SHUNT VERSUS COMPOUND-WOUND
EXCITERS
The relative advantages and disadvantages
of shunt and coinpound-wound exciters have
been frequently discussed; and the selection
572 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 7
has sometimes been dictated by the type of
excitation plant used; sometimes by indi-
vidual preference of engineers.
The matter may be summarized under the
following headings :
"A": Small belted exciters operating in parallel;
no battery
Compound-wound machines are generally
used to keep the exciter bus voltage con-
stant with variation in excitation load due to
either variation in the alternating-current
load or to a change in the niunber of alter-
nators in service. With shunt exciters such
variations would require adjustment of ex-
citer field rheostats as well as generator field
rheostats; and consequently the compound
machines are usually preferred for con-
venience in operation.
If voltage regulators are used either shunt
or compound exciters can be handled equally
well. For this class of exciters standard
belted generators are used, which are manu-
factured and stocked in large numbers com-
pound wound, so that reducing the varieties
stocked is another reason for the choice of
compound winding.
"B": Motor, engine, or waterwheel-driven excit-
ers operating in parallel; no battery
The compound winding is preferred for
same reason as stated under " A. "
"C": Exciters direct connected to the main
generating units; operating in parallel;
no battery
The compound winding is preferred for
above reasons, but stock requirements do not
apply excepting in small units.
"D": Exciters operating in parallel, with stor-
age battery floating on the exciter bus
In this case either compound or shunt
windings may be used.
The compound has the advantage of keep-
ing the bus voltage constant with a change in
the excitation load, and the disadvantage of
possible reversal and motoring of an ex-
citer in case of failure of its source of power.
This contingency may be provided against by
reverse-current relays, so far as motoring is
concerned, but this may not prevent reversal
of the polarity of the exciter due to the sudden
reversal of the current in the series field.
Shunt exciters are safer, when in parallel
with a floating batter}', as regards possible
overspeeding due to motoring in case of fail-
ure of reverse-current relays. They are cer-
tainly less liable to be reversed in polarity
when an exciter slows down due to trouble
with its drive. They require more frequent
adjustment of the exciter field rheostats, but
have the advantage of omitting the equalizer
bus with its extra switches and connections.
Commutating pole shunt-wound exciters
should be adjusted to have a drooping char-
acteristic at all operating voltages, not only
for proper parallel operation, but in order to
reduce the liability of reversal.
Individual Exciters
The shunt winding appears preferable for
individual exciters, whether the alternating-
current voltage regulation is accomplished by
the alternator field rheostat or the exciter
field rheostat. In the former case there is no
reason for compound winding unless for manu-
facturing or stock convenience in the smaller
sizes. In the latter case, when the exciter has
to operate down to a low voltage the shunt
winding has more stability, particularly in ex-
citers of the commutating pole type. The shunt
exciter is also less susceptible to reversal by dis-
charge from the alternator field, or by residual
magnetic effect from the alternator field.
REVERSAL
The reversal of polarity of an exciter, due
to failure of dri^•ing power, has been discussed
in the foregoing. It appears likely that in
case of compound-wound exciters the chances
of reversal due to this cause may be reduced
by exciting the shunt field from the busbars
instead of across the exciter brushes.
The reversal of an exciter has been occasion-
ally observed at the time of shutting down an
alternator. The possibility of reversal at this
time is apparent only on an individual exciter
direct connected or otherwise driven from a
main generating unit. In one case a steam
turbine unit with individual direct-connected
exciter was taken out of service, the field
being left closed to bring the machine to rest
quickly, and the field opened after the machine
had reached a standstill. It appears probable
that residual magnetism in the generator field
structure would persist to a lower point of
speed than that in the magnetic circuit of the
exciter. The flux would be varied in passing
the armature slots, causing weak alternating
currents to flow in the field circuit, which may
account for the reversal of the exciter.
Assuming that the field current docs alter-
nate or reverse, it is apparent that a series field
on the exciter would be effective in reversing
the residual magnetism of the exciter.
In the case of a commutating-pole exciter,
the position of the brushes would detemiine
the influence of the commutating field on
reversal.
EXCITERS AND SYSTEMS OF EXCITATION
573
EXCITERS USED WITH REGULATORS
When a vibrating-contact regulator is used
to control the alternating-current pressure
through the exciter field there appears to be
little choice whether the exciter shall be
shunt or compound wound, so far as the
action of the regulator is concerned. For
shunt-wound machines, the field current
handled by the regulator is greater. For
compound-wound machines, the field current
is less, but a greater range of voltage must be
applied to obtain the same speed of regulation.
The shunt across the series field makes the
latter a damper winding, which impedes sud-
den flux changes, but this is partially neutral-
ized by the action of the series field with
changes in current.
For sensiti\'e regulation the shunt machine
is undoubtedly better, since the entire field
is controlled by the regulator, and on account
of the absence of the damper formed by the
series winding. For most plants the compound
exciter is satisfactory with a regulator, since
its period is usually faster than that of the
alternator field.
COMMUTATING-POLE EXCITERS
Some years ago, during the first period of
experience with commutating-pole exciters,
some troubles were encountered in operating
them in parallel. These were due to incor-
rect design or to incorrect adjtistment of the
commutating field strength or to the position
of the brvishes.
Shunt generators to operate in parallel with
proper division of load must have a drooping
characteristic, and compound machines to
operate in parallel must have a drooping
characteristic without the series winding.
Machines having a rising voltage characteris-
tic without the series winding in operation
will be liable to give trouble in parallel opera-
tion, either as shunt or compound generators,
unless the regulation of their prime movers
is sufficiently poor to overcome the rising
characteristic of the generator.
With commutating pole exciters it was
soon found that if compounded flat in test
at 125 volts full load, 12.5 volts no load,
then when operated at lower voltages (as
frequently happens under control of a reg-
ulator) they had a rising characteristic and
were therefore unstable in parallel operation.
This trouble was sometimes made much
worse by the slight backward brush shift
required for good commutation when the
commutating field was too strong.
To take care of this the expedient adopted
for a time in one manufacturing plant was
to flat compound the exciters in test at 80
volts, thus insuring a drooping character-
istic at higher pressures, the division of load
at lower pressures not being so important.
This expedient involved carrying in stock
generators for ordinary purposes compounded
flat at 125 volts and generators for exciter
purposes flat compounded at 80 volts, or the
delay of testing and adjusting the shunt after
the receipt of an order.
With further knowledge of the character-
istics of commutating-pole machines and
more nearly correct designs, the foregoing
expedient was abandoned, and it is the
present practice to proportion the com-
mutating field so that the machine is not
over compensated; the design is such that
some range of brush shifting is allowable,
and the brushes are given a slight forward
shift so as to obtain a drooping character-
istic at all voltages within the range where
parallel operation is required. With this
arrangement, exciters when compound wound
may be compounded flat at 125 volts and
still operate properly in parallel at lower
voltages.
The stability of a commutating-pole
machine depends greatly on the brush
position, the commutating-pole field strength,
and the voltage at which it operates with
regard to the saturation curve of the unit.
These three factors aff'ect equally the sta-
bility of the shunt and the compound-wound
unit. In addition, the compound-wound
unit is affected by the amount of com-
pounding, the nature of the compounding
curve, and the size of the equalizer connec-
tion, also the amount of resistance in the
equalizer circuit.
Many engineers feel that brush position
alone changes the characteristic of the com-
mutating-pole machine and, whenever a
change is desired, the first resort is always
to change the brush position. In many
cases, the desired effects can be secured in
this way, but nearly always a change in
commutating field strength, together with
a change in brush position, if necessary,
will obtain the results required in a more
satisfactory manner.
It is a well-known fact that shifting the
brushes back from the direction of rotation
on a commutating-pole generator improves
the voltage regulation of the machine, and,
on some machines, it is possible to obtain
practically a flat voltage characteristic curve
on a shunt-wound unit, and, in exceptional
cases, it is possible to obtain a rising voltage
characteristic curve. The latter is obtained
574 July, 19-20
GENER.\L ELECTRIC REVIEW
Vol. XXIII. Xo.
by a combination of o^'e^-compensated com-
mutating field and backward shifting of the
brushes. In fact, the two act together in
that to hold commutation with a backward
shift of the brushes it is necessary to use a
stronger commutating field than would be
required with the brushes
on neutral with proper
com.pensation.
The over -compensated
commutating field tends to
magnetize the m.ain pole,
and, therefore, has a com-
pounding effect. The mag-
netizing effect is due to the
short-circuit current in the
coil undergoing comm.uta-
tion. The reverse is true
for under - com.pensation,
since the short-circuited
current in the coil is re-
versed for this condition.
We have never found an
instance where successful
parallel operation could
not be obtained after the
brush position, as well as
the com.mutating-pole field
strength, where properly
adjusted for shunt-wound
machines, and where these
two conditions were met
and the equalizer connections ])ro])crly made
on com.pound- wound machines.
In this connection also it is desirable to
arrange the switches of a machine so that
the equalizer switch is closed with the line
switches and not before. If closed before
the machine on the bus is operating with an
additional shunt across its series field and
the incoming machine is operating with
series as well as shunt excitation, result-
ing in a lower shunt excitation and, there-
fore, a chance for instability when the series
part of the excitation is changed in amount,
and, in extreme cases, in direction.
VOLTAGE REGULATORS
Large city central stations supplying power
from, the generator busbars at generator
voltage through a multii)licity of feeders
have seldom found it necessar\' to resort
to automatic voltage regulation, because
the sudden changes of load are small in
proportion to the generator capacity.
Exceptions are noted, such as the plants at
Philadelphia, Baltimore, and Pittsburg, where
unusual requirements in intemiittent loads
exist, due to the supply of main railway
electrification or steel mill loads.
In hydro-electric plants, on the other
hand, the power is usually carried through
a few large transmission lines, the interrup-
tion of anv one of which means the loss of
Fig. 9. Vertical Exciter Mounted on an 850-kv a.. 144-r.p.m., 2300-volt
Watcrwheel-driven Generator
a large proportion of the station load, thus
necessitating automatic voltage regulation
of the generators. The vibrating contact
forms of regulators devised by Tirrill are
the only ones in wide use in this country,
and no others will be discussed.
These regulators can be made to take care
of exciters operating individually or in
jjarallel in sizes uj^ to the largest which it
has been found necessar\- to use. The exciter
field current is controlled through relay con-
tacts; the relay coils themselves being actuated
by direct current passing through the main
regulator contact. Up to about four amperes
at 125 volts a single relay handles the exciter
field current: for larger exciters, the field
rheostat is divided into sections short cir-
cuited by a number of relays, the general
tu\l' being that each relay will take care of
two amperes field current ; that is to say,
for a total of ten amperes field current there
will be at least five relays. At 250 volts half
the current is handled. To obtain the best
results the output per exciter should not be
more than 25 kilowatts per relay. When the
field current is more than 20 amperes, it is
EXCITERS AND SYSTEMS OF EXCITATION
575
generally desirable to split the field, so as
to keep the actual field current handled by
the relays below 20 amperes.
Since a 12-relay regulator will handle about
a .'^OO-kw. exciter, and regulators have been
made with as many as 48 relays, which could
handle four such exciters in multiple, it is
evident that the regulator can be made to
take care of very large excitation plants.
For still larger plants, if such should be
contem.plated, other methods of application
of the vibrating contact regulator may be
used, so that we can say there is no "limit
to the size of plant which can be regulated
on this principle.
A single regulator may be used to control
a number of exciters operating in parallel,
or to control a ntmaber of invididual ex-
citers not operating in parallel when the
alternators excited run in parallel. It is
also possible to use individual regulators in
the latter case. The proper division of the
reactive component am^ong the alternators
is then accomplished by a com.pensating
coil on the regulator supplied from a current-
transform.er connected in such a way that
the current is at a right-angle phase relation
to the voltage which the regulator is main-
taining. This last arrangement is favored
for large hydro-electric plants having indi-
vidual exciters, as the individual regulator
may be mounted near each exciter.
Stops may be provided on regulators to
lim.it the field current of an alternator and
this is always done when a regulator is used
with a synchronous condenser to keep the
voltage of the receiving end constant up to
the limit of output, which is determined iDy a
lim.iting field current.
Regulators for large plants are provided
with accessories, such as over-voltage relays
and over-current relays, which cut in an
extra block of resistance in the exciter field,
so that in case of over-speed of the gen-
erators or of the relay contacts sticking,
unduly high voltage will be prevented and in
case of short-circuit the action is to prevent
over-excitation of the fields.
DISCUSSION OF VARIOUS PLANS OF
EXCITATION
In selecting a plan of excitation for any
plant the local conditions will govern to som.e
extent. In either steam or hydro-electric
plants there will be a certain amount of
auxiliary power about the station which
must be supplied. In a hydro-electric plant
the requirements for auxiliary power are not
very exacting as to continous operation un-
less motor pumps are supplied for the step
bearings. Most other motors about such
plants are for intermittent operation and
may be taken care of by an auxiliary bus
supplied by step-down transformers from the
main bus.
In a steam plant the continuous operation
of many of the auxiliaries is of vital im-
portance; and since variable-speed motors
are supplied for many of the auxiliaries in
order to obtain economical operation at
part loads, direct-current motors are fre-
quently used for part of the auxiliaries. At first
sight, the best source of power for such
auxiliaries would appear to be the exciter
bus and, thus, the question of choice of ex-
citation plans becomes involv'ed with the
other auxiliaries in the station.
The earlier practice in this country was to
operate all of the auxiliaries, such as the cir-
culating water pumps, hot-well piunps, feed
pumps, stokers, draft fans, etc., by steam
power, thus insuring an ample supply of
exhaust steam to heat the main feed water.
In some cases this provided too much steam
and some steam had to be wasted; and in
any event the driving of many of the aux-
iliaries by individual steam turbines or engines
is somewhat wasteful, since the small ma-
chines consume steam per horse power out-
put at a rate several times that of the main
turbine. If a certain ntunber of pounds of
exhaust steam is required to heat feed water
it is evidently more economical to use that
steam first in a large and efficient turbine,
so as to get as many horse power as possible
out of it before passing it into the feed-water
heater. Modern practice is now tending to-
ward the operation of as many of the aux-
iliaries as possible electrically, particularly
on account of the convenience, reliability and
freedom of repair of the electric motor itself,
and partly on account of the high efficiency
obtained by this method of operation, whether
the electric power for the auxiliaries is de-
rived from the main busses or from a separate
auxiliary generating source of high efficiency.
The main boiler feed pumps are usually
n.m by steam_, but if all the other auxiliaries
are electrically operated, as appears to be the
modern tendency, there will not be sufficient
exhaust steam from the feed pumps to heat
the feed water. If the electrical auxiliaries
are operated from the main bus the feed
water may be supplied by bleeding the m.ain
turbine at an intennediate stage and drawing
ofl^ sufficient steam to heat the feed water.
576 July, 1920
GENERAL ELECTRIC REVIEW
Vol XXIII, Xo. 7
This is an efficient method of operation,
since the steam is used very efficiently in
producing mechanical power in the main
turbine before it is drawn off. Such an »
arrangement should be operated on the unit
system; that is to say, with a certain bank
of boilers supplying a certain turbine, the
steam drawn from such a turbine should be
used to heat the feed water for its own
bank of boilers.
This method has the disadvantage that
in case of trouble on the main bus many
of the station auxiliaries may be inter-
rupted.
The house turbine arrangement in which a
house turbine with its own boilers supplying
power to an auxiliary bus is installed for each
main generating unit, or one house turbine
for two main generating units in a ^-en,' large
plant, appears to possess m.anv advantages.
When the main units are 15,000'to 30,000-kw.
each, each house turbine may be 1000 to
2000-kw., large enough to obtain efficiency
in the use of steam. The auxiliars- bus may
be connected by transformers to the main
bus with automatic relay arrangements, so
that in case of trouble on the main bus the
auxiliary- bus is cut off and supplied only by
its own power.
This arrangement possesses the advantage
that the supply of power for the auxiliaries,
including the excitation, is independent of
the main supply, and also the advantage that
the heat balance is readily adjustable by ad-
justing the amount of power su]3i)lied by the
auxiliary- generating unit, so that the amount
of steam exhausted by it is just sufficient to
bring the feed water to the proper tempera-
ture, the remainder of the auxiliar\- power
being supplied from the main bus, which is
operating in parallel.
When such an auxiliary house plant is sup-
plied it is generally of alternating current
and the excitation for the main unit may be
supplied from a motor-generator set run from
the auxiliar\- bus. It would seem desirable
that these exciter busses for separate units
of the power house should not be operated
in parallel, but provision may be made for
connecting them in parallel and to obtain
the greatest safety a reserA-e exciter unit and
storage batten,' with emergency bus may be
supplied to which any generator field may
be connected.
For the ver\' largest steam station such a
plan appears desirable, but it is a still better
plan, when an auxiliary house plant is used,
to supply the excitation for the main units
from direct-connected exciters on the main
units. There should still be installed a re-
ser\-e exciter bus with battery- and reser%-e
exciter driven from the auxiliary bus, with
automatic throw-over switches, so that in
case of trouble with any direct-connected ex-
citer the field of that alternator is discon-
nected from the exciter and thro"mi on the
reser\-e excitation bus. With this plan the
excitation is kept separate from other auxili-
aries and the exciter bus is not liable to trouble
originating in a motor, also the exciter and
field connections are kept short and simple as
possible. This plan has the further advantage
that the alternating-current voltage may be
controlled by the exciter fields and the losses
in the main field rheostats eliminated. It is
desirable, however, to supply main field rheo-
stats for emergency use.
In connection with an auxilian.- house plant
generating alternating current for the sup-
ply of most of the auxiliaries, motor-generator
sets have been installed fed from the auxiliary-
bus to produce direct current for the A'ariable-
speed auxiliaries. This complication of an
extra direct-current bus may be done away with
if an alternating-current motor with good
adjustable-speed characteristics were avail-
able. Such a motor is now coming into use.
It is a three-phase commutator motor with
three sets of brushes on the commutator and
the speed is varied by shifting the brushes,
which may be done by distant control by a
small motor geared to the brush shifting yoke.
The motor has a series characteristic and is.
therefore, well adapted for driving fans or
centrifugal pumps, but it is not as good as
an adjustable-speed shunt-wound motor for
applications where it is desired to adjust the
speed through a wide range and to keep the
speed constant for var\ing load.
There has been some experience with such
motors in continuous operation driving mine
fans, which- indicates that they are as good
as direct-current motors with a possible dis-
advantage of more brush wear. Further ex-
perience will undoubtedly justify their use
for the exacting requirements of power-house
ser\ice for such purposes as fans and cen-
trifugal jmnips.
577
Gaseous Conduction Light from
Low-voltage Circuits
By D. McFarlan Moore
Edison Lamp Works of the General Electric Company
The first artificial electric light produced was of the gaseous conduction type, as were its immediate suc-
cessors. Later, the solid conductor or filament lamp came into being and rapidly outdistanced the gaseous
conduction lamp in development and application. The development of the filamentless lamp has not been
neglected, however, as is evidenced by Mr. Moore's description of it in the following article that was presented
last March at a New York meeting of the A.I.E.E. The author explains the problems that have arisen and
describes the various types of lamps produced, concluding with the latest type which is of a size comparable
with the standard Mazda lamp, starts and operates on low voltage without auxiliary equipment, and fur-
nishes light of an amount useful for many purposes. — Editor.
The production of artificial light is one of
the most impoitant activities concerning the
welfare of humanity. It is a very large sub-
ject, since both its practical and theoretical
aspects cover vast fields, yet there are less
than a dozen distinct methods of making
light artificially and some of them are not
developed commercially, although theoreti-
cally they possess great possibilities.
This article is written to consider some of
these methods of producing artificial light,
that have to do with electricity and that
come under Item 3 of the following list :
1. Torch (and candle)
2. Oil
3. Gas
4. Solid electric conductors
5. Gaseous electric conductors.
/\>ye,ecc Corona Arc Discharge
^oiM l/ottago. , -,.^^_ _,.,
/NoGai) ^e77gwa/s VW, etc Coro7i&-Arc Discharge
\6as Rz,nt,^a.ls
/Hg Vapor -HT Starting
v< - -
\Chemica/s
Fig. 1. Diagram Illustrating the Esctension of the
Varieties of Gaseous Conductor Lamps
Electricity can be used to agitate solids,
liquids, or gases into light. The light of the
incandescent lamp is due to electrically-heated
solids; and when electricity is conducted by
a gas under suitable conditions, light also
results. Many varieties of lamps of this
nature, both in design and construction, are
indicated in Fig. 1, the scope of which can be
enlarged almost indefinitely; for example, by
the use of many other gases and vapors.
High-tension lamps require special auxiliary
transforming apparatus to generate the high
potential.
'^^H>
Connection Diagram of Gaseous Conductor Lamp
and Vibrator Employed in 1895
The two major factors in all of these types
are; (1) the electrodes, and (2) the gaseous
conductor.
Both electrodes of alternating-current
lamps can be similar, but in direct-current
lamps the cathode differs from the anode.
Electrode materials differ with the gas used.
It is therefore seen that the construction and
design of each one of the scores of lamps
indicated is a distinct and difficult problem,
the solution of many of which have hardly
been seriously attempted.
As might be surmised, the specific type of
lamp I wish to emphasize is the one in which
I have been most interested recently, but in
order to give it its proper setting, it is neces-
sary to review the past. The first natural
electric light was lightning, or the aurora.
The first artificial electric light was due to
gaseous conduction and was produced with
the revolving glass sphere of Hawksbee in
1750.
578 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo.
A hundred years later, Geissler first oper-
ated his small tubes from an induction coil.
In 1879 Crookes modified them in many
ways, including obtaining high vacua.
About 1891 Nickola Tesla delivered his
famous lectures on "High Voltage and High
Frequency."
onX220 volts resulted in no light whatever.
All I known gases were unsuccessfully tried.
Light from many of the common gases proved
ven.' interesting; for example, the bluish-white
light from COo, the pinkish, hot and almost
non-luminous light of hydrogen, the efficient
orange-yellow light of nitrogen, and the dull
Fig. 3. Various Designs of Early Negative
Glow Lamps
ti
Fig. 5. Seven-foot Vacuum Tube Lamp with
External Electrodes and Vibrator
Due to the rapid and very objectionable
blackening that was deposited' over the inside
of incandescent lamp bulbs in 1893. I first
began thinking and talking about the pos-
sibility of constructing a lamp without a
heated filament — a filamentless lamp.
In connection with the American Institute
of Electrical Engineers, I explained that I
meant a bulb form of lamp, the light source of
which was to be not an incandescent solid
conductor but an enclosed gas or vapor elec-
trically agitated by the low-tension circuits
in common use.
During the twenty-six years that have
intervened, this simple thought has never left
me, though the tortuous road has been verv
X^^
whitish light of oxygen; also many mi.xtures
were tried, together with chlorine, bromine,
etc., and various vapors like those of sulphur
and mercur\-. The prediction was made
that progress would result only after the dis-
cover\- of some of the gases indicated by the
table of the periodic law of the elements. It
was necessary- therefore, in 1S94, to resort to
the high voltage of an induction coil, in
order to obtain some light from the first
gaseous conductor bulb lamp. In 1S9.3, the
vacuum vibrator displaced the induction
coil, and on direct-current circuits the bulb
lamps were filled with negative glow light.
Fig. 2 shows the vibrator and connections.
Fig. :i shows the negative glow lamps, and
Fig. 4. Special Negative Glow Lamps Designed for Advertising Purposes
dark at times, but it is now brighter than it
has been before.
In order that I may not be misunderstood.
I must hasten to say. perhaps sorrowfully,
that it is still far too dim even to think of its
competing in brilliancy with that splendid
array of present day commercial illuminants —
led by the incomparable tungsten lamp.
My first attempts in 1893 to obtain any
light from a lamp without a heated filament
*'A New Method for the Control of Electric EnerKy."
A.I.E.E., 1893. Sept. 20th.
■Recent Developments in Vacuum Tube Lighting," A.I.E.E..
189S. April 22nd.
"Light from Gaseous Conductors Within Glass Tubes —
Moore Light." A.I.E.E., 1907, April 26th.
Fig. 4 depicts the use of negative glow for
advertising purposes, and the means for
increasing its intensity. Detailed information
of this nature will be found in some of my
previous ]iapers.*
After neon had been discovered as hoped
for, and nineteen years later I had made the
first low-voltage gaseous conductor lamp,
there was a certain satisfaction in proving
that my original conception of utilizing the
feeble light of the almost despised negative
glow was correct.
In 1S9(), seven-foot vacuum tubes. Fig. o.
with external electrodes displaced the bulb
GASEOUS CONDUCTION LIGHT FROM LOW-VOLTAGE CIRCUITS
579
lamp. The vacuuin rotator succeeded the
vibrator in 1897 and 1898. Fig. 8 shows the
interior of the historical "Moore Chapel."
The first 220-volt direct-current tubes, started
with a higher potential from both vibrator
and rotators, were then made and used. Fig. G
shows the 5-foot tube which was used in tak-
ing the first instantaneous electric portrait —
Chauncey M. Depew being the first subject.
The anticipated discovery of neon was an-
nounced in 1898, but even samples of it were
impossible to obtain in America. Sir Wm.
Ramsey, Lord Raleigh, Travers and their
brilliant contemporaries announced in rapid
succession the five new monatomic elements,
argon, helium, neon, xenon, and krypton, all
of which will probably ultimately take im-
portant places in the world of commerce and
some of which have already done so.
^ F'>
^
<^
Fig. 6. Five-foot 220.volt Tube Lamp
Vacuum-breaks were displaced in 1899
for a combination of resonance coils and a
low-frequency generator and later high-fre-
quency generator.
In 1902, the "long tubes" (about 100 feet)
appeared, and they were improved in 1903
with internal electrodes.
The beauty of the first long tube was ad-
mired by thousands.
The first rotary high-vacuum oil pump was
developed for the exhaustion of the long tubes
built in situ.
Also a 24-inch CO2 tube lamp provided with
a carbon filament cathode was started with
higher potential on 220-volt direct-current
and the resultant light was highly efficient.
Other, though similar, tubes and lamps
had metallic cathodes buried in lime, etc.,
and it was noted that, when operated on
alternating current, rectification took place.
These interesting types of lamps are shown
in Fig. 7.
It was a great advance in 1904 and 1905
to discontinue the use of a special generator
with each "long tube" installation and to
obtain brilliant illumination from the distri-
bution street circuits by the use of nitrogen
gas. An installation of such lamps is shown
in Fig. 9.
Special electrodes were also constructed
with auxiliary circuits, similar to those later
used in rectifiers, pliotrons, and X-ray tubes.
Fig. 7- Hot Cathode Luminous Discharge Lamp
and Its Connections
The life of these long tubes was extended
to 10,000 hours during the period of 1906 to
1909 by the invention of the electromagnetic
feed valve and over four miles of light-giv-
ing tubing were commercially installed. The
lobby of Madison Square Garden is shown
in Fig. 10. Fig. 11 shows the details of the
magnetic feed valve. No light source known
today equals in efficiency a neon tube l^i
inches in diameter and 200 feet long. The
long-tube system is theoretically correct in
so far as it provides means for generating
light at the exact intensity most suitable for
the eye ; this in contra-distinction to the gen-
eration of concentrated light at an enormous
intensity and temperature that must, before
580 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 7
Fig. 8. Interior of the Moore Chapel Lighted by
Gaseous Conductor Lamps
Fig. 10. Lobby of Madison Square Garden Illuminated
with Low-voltage "Long-tube" Lamps
Fig. 9. Installation of Low-voltnge "Long-tube" Nitrogen Conductor Lamps
GASEOUS CONDUCTION LICIIT FROM LOW-VOLTAGE CIRCUITS
581
it can be used by the eye, be either greatlv
reduced in intensity by means of some kind
of semi-transparent or diffusing screen, or
widely scattered by a reflector. Fundamen-
tally the first cost of a long tube system is less
than that of a complete incandescent lamp
system and its life is longer with a
resulting lower maintenance cost.
It is also simpler.
During 1910 and 1911, the long
tubes in the form of portable artifi-
cial daylight windows made their
appearance. One of these is shown
in Fig. 12.
Between 1913 and 1915, several
types of small tube lamps dependent upon the
new chemical gas feed principle were invented
and marketed for color matching purposes.
Such lamps are shown in Figs. 13 and 14.
The spectrum of this type of color-matching
lamp will never be surpassed as a standard
light by which to judge colors.
Simple neon tubes operable from trans-
formers were designed and made in many va-
rieties. Some were equipped with screw lamp
bases. These outfits consume 13 watts and
are light enough to be screwed into an ordi-
nary incandescent lamp socket. Lamps of
this kind have run without change for over
4000 hours.
Fig 12.
Compact Form of "Long-tube" Outfit
for Portable Use
Fig. 11. Cross-sectional Drawing of the
Electro Magnetic Feed Valve
In the fall of 1 916, there was exhibited the
first portable and thoroughly commercial neon
tube outfit of high intensity and efficiency oper-
ated from a step-up sixty-cycle transformer.
It resembles those shown in Figs. 13 and 14
except that the tube housing is twice as long.
The tube is in the form of a hairpin and has
a total gas column length of 101 inches at J^
inches diameter. The specific efficiency of
this type of lamp is 0.74 watts per spherical
candle-power.
Even this high efficiency can be improved
considerably by using purer neon (that is,
neon gas that does not contain 25 per cent
helium and other impurities) together with
a longer gas column and of greater diameter.
Also the electrode losses can be reduced. But
the photometric measurements of this bril-
liant type of tube lamp showed a total of 180
mean spherical candle-power of 2260 limiens
with 0.162 amperes passing through the gas
582 July, 1920
GENEIL^L ELECTRIC REVIEW
Vol. XXIII, No.
column. Simple straight tubes about 1^
inches in diameter and S feet long could be
arranged as a continuous line of light and
used for the lighting of large interiors or for
streets.
The initial installations would have great
advertising or display worth. The red rays
/*M!P
.,»!...
Fig. 13. Color Matching Tube Lamp Operating
on the Chemical Gas Feed Principle
Still another type of 220-volt direct-current
neon tube was started by using an auxiliary-
current to raise to a high temperature a por-
tion of the cathode. Space will not permit
the listing of many other varieties of gaseous
lamps.
However, attention is to be called to the
type of lamp that has cold electrodes and is
designed to start and operate without using
high potential.
UmmoQ.
rajTOHZTj
m
OiAORAU or CONNECTIONS
rrmr^f I
1/
(xmmr
J ELtCTRCOE
ELECTRODE
H
GAS GENERATOR
RESISTOR
SIMPLIFIED CONNECTIONS
will also be valuable for signaling purposes,
etc.
Various alternating-current tube lamps pro-
vided with two similar electrodes were also
made to operate on 220 volts alternating cur-
rent without a step-up transformer, but they
need a momentary higher voltage to start
the gas coltunn discharge which is most
simply obtained by short-circuiting a series
inductance. The length of the gas column
of this type of lamp is too long (about '.i
inches) to permit 220 volts to pass any cur-
rent, but it will maintain the discharge, which
is positive, for an indefinite length of time
after once started. The necessity for start-
ing apparatus is an objectionable feature
of this particular type of lamp. When the
gap or gas cohmm between the electrodes
of a tube lamp on 220 volts alternating cur-
rent is less than about 1^4 inches, the light
is negative glow.
The direct-current lamp, of the type re-
quiring high potential for starting, involves
even when filled with neon gas a special
cathode of mercury of a KNa amalgam.
ELEVATION
GAS GENERATOR
Fig. 14. Connection Diagram of the Type of
Lamp Shown in Fig. 13
The current of a 220-volt circuit passes
through the neon gas and causes it to give
light. No potential-raising transformer is
used. When the particular problem was
the production of small units of light, its
satisfactor\- solution bv the use of the
GASEOUS CONDUCTION LIGHT FROM LOW-VOLTAGE CIRCUITS
583
transformer was commercially impracticable
but it seemed for many years impossible
to obtain any light without using a trans-
former.
Fig. 15 shows a form of this type of lamp,
for alternating circuits. Scores of modified
designs have been made. It is a novel type
of lamp. I hope that many will see in it,
with me, the possibilities of a lamp of this
kind. In fact, diligent inquiry among scien-
tific men has failed to find anyone who did
not agree that all theory seemed to indicate
the great probability that artificial light of
high efficiency will result from the further
development of lamps of this kind. The hand-
writing on the wall seems to unmistakably
indicate that to further increase the kmiinous
efficiency of light sources in general we shall
need to resort to gaseous radiation, by which
means it may be possible to reduce to about
one tenth the energy now required.
Since the ice on the problem now seems to
be broken, it is my earnest hope that many
of the ablest inventors will become actively
interested and that by the combined knowl-
edge, experience and ingeniousness of all
who have studied and worked on gaseous
conduction phenomena, the many problems
involved will be solved. I believe "that a very
great deal remains to be learned and dis-
covered.
The lamp shown in Fig. 15 resembles an
incandescent lamp in outward form and per-
haps is far niore simple, yet it is not an incan-
descent lamp. Four electrodes made of
aluminum, each 6 in. long, ^.s in. wide, and
iV in. thick, are mounted in a 3-in. straight
sided bulb about a common center. A glass
hub, provided with radial arms of glass,
supports the electrodes, which have holes in
them and through which the arms extend.
The capacity of the solid radiators is objection-
able and yet the effect of a solid radiator is
approached by radiators made of very small
mesh netting.
In designing this lamp, an effort was made
to take advantage of every factor that re-
quired minimum voltage so that it would
operate on 220 volts or less.
The potential is least (volts per centi-
meter) for the negative glow. All of the
light radiated from this new type of lamp is
produced by the negative discharge; not by
the positive column as is the light in all of
the long vacuum tubes when in operation on
either a-c. or d-c. circuits. All text books
and investigators have heretofore considered
that the amount of light given by the negative
glow in any vacuum tube discharge was so
small as to make it entirely negligible as a
light source.
An ordinary long tube discharge consists
of: (a) next to the cathode, the short first
dark space; (b) the short and not bright neg-
Fig. 15. Negative Glow Lamp Capable of Starting and
Operating on Low Voltage Without Auxiliaries
ative glow; (c) the short second dark space;
(d) the long brilliant positive column ex-
tending to the anode. But in the lamps here-
with described, the positive colimm has been
practically eliminated and substantially the
only luminous discharge in the lamp is the
negative glow which appears in the form of
a velvety glow or corona of yellowish light
over the entire surface of the alternating-
current electrodes and also a uniform gase-
ous radiation throughout the interior of the
bulb.
The lamp shown in Fig. 15 is designed for
operation on 220-volt a-c. circuits. From the
line it uses about 0.11 amperes and 21 watts,
but of this amount 3.0 watts at 33 volts is
used by an ohmic resistance about 1 in. long
placed in the skirt of the lamp base, because
due to the impurities principally in the neon
gas and the aluminixm radiators a slight
blackening may form between them in time
and may cause the lamp to short circuit.
The "finished lamp will probably require
no series resistance, but at the present
time the use of such a resistance affords
a convenient method of adjusting the
total watts consumed, the life, and the in-
5S4 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo.
tensity. The specific efficiency of this par-
ticular type of lamp is low. When this lamp
is consuming 17.4 watts, it gives approxi-
mately 1.16 s.c.p., which corresponds to 15
watts per s.c.p. Therefore, the most impor-
tant problem still to be solved is how to de-
Fig. 16.
-w^-
Two Methods of Connecting the Radiators
When Four Are Used in a Lamp
crease the number of heat waves and increase
the number of luminous radiations.
When the line voltage was reduced to 135,
the light was suddenly extinguished.
The neon used had a helium content of
about 25 per cent, but if it had had a nitro-
gen impurity of a fraction of one per cent, the
neon lines would have been greatly reduced.
The pressure of the gas when sealed off was
3.5 mm.
The bulb temperature is about 40 deg. C,
but of course is increased when the watts
are increased.
The color of the light is a beautiful yellow.
Some of the important factors to which
special attention has been given in the design
o*' this new lamp are :
1. The attempt to use a gaseous conductor
of maximum conductivity.
2. Electrodes that are subdivided and of
as large a total area as possible.
3. A gas column (discharge gap) as short
as possible.
4. The planes of the electrodes of oppo-
site polarity placed parallel to each
other.
5. The length of the radiator electrodes
greater than the gas column and per-
pendicular to it.
Since the light is entirely due to negative
glow, cathodic disintegration of the electrodes
is one of the problems in connection with this
type of lamp, but it is practically nil when the
cathode fall equals its minimum value. It is
greater at lower gas pressures and increases
as the square of the current, assimiing a con-
stant electrode area and gas pressure, but
it is not an essential to transmission of current
and seems to be largely due to the occluded
gases, particularly hydrogen. The bulb
blackening is far less with aliuninum radiators
than tungsten, nickel, copper, etc. Carbon
in pure form is difficult to obtain. Iron radi-
ators, as well as various radiators combined
with fluorescent coatings, offer promise.
One of the troubles connected with the
use of the carbon was the difficulty of re-
moving all of the occluded gases. However.
this may be overcome by heating not only
carbon electrodes but all other varieties of
radiators. Radiators of whatever material
should be as pure as possible and be cleansed
in the best manner.
^^
"Fig. 17. Direct-current Lamp
with Concentric Radiators.
the Outer of Which
Emits the Light
Fig. 18. Another Variation of
the Direct-current Radi-
ator Lamp, the Posi-
tive Being Wound
as a Spiral
This corona type of lamp produces a
hmiinosit>' that is not due to arcing or even
pure discharge phenomena, but is due to the
glow of light emanating from electrodes or
radiators that nonnally have a temperature
below red heat. According to the theor\- of
GASEOUS CONDUCTION LIGHT FROM LOW-VOLTAGE CIRCUITS
585
ionization, the temperature within the neg-
ative glow is higher than within the positive
column and the velocity of the negative ion
is greater than that of the positive, and this is
one reason why the potential required to pro-
duce a luminous discharge from a negative pole
is less than from a positive, together with the
fact that in the negative glow the number of
positive and negative ions are about equal.
The exceptional luminous efficiency of neon
makes it unique among light sources. Im-
mediately upon the announcement of its
discovery in 1S9S, I proposed its use for light-
ing purposes. Its great scarcity until recently
has made rapid progress impossible. Within
the last few days announcement has been
made that it can now be boiight in almost
any quantity and of a high degree of purity.
Its luminous spectrum is almost ideally
located to effect the eye in a maximum
manner. It is a splendid example of selective
emission or radiation that eliminates the
long and therefore inefficient waves.
Fig. 19. A Simple Form of
Alternating-current Lamp
Having Four Radi-
ators of Alumi-
num Netting
Fig. 20. A Form of Gaseous
Conductor Lamp Which
in Construction Re-
sembles the Stand-
ard Incandes-
cent Lamp
It does for gaseous conductors just what
the Welsbach mantle or the impregnated arc
lamp electrode does for heated solids.
The maximiun emission is between wave
lengths 590 and 650 which is one of the re-
markable properties of this gas. It produces
over a hundred times as much luminosity
for the same watts as does argon for example.
Its dielectric cohesion is 5.6 which is extremely
low when compared with air at 419. It has
less than one half the resistance of nitrogen.
It is fortunate that the color of the negative
Fig. 21. Lamp with Wire Electrodes Wound Parallel
in the Form of a Square Spiral
glow of neon is different from that of the
positive column.
Neon gas when used as a positive column
of light has a color so reddish that it wovild
be objectionable for many purposes; but when
the same gas is used as a negative glow,
the color is yellowish. It has no blue or
violet or indigo lines and very few infra-red
rays. It is four times better as a light pro-
ducer than the yellow-white light of helium
or the violet of xenon, both of which have
many infra-red rays.
The characteristic crimson of neon has
been displaced by a uniform mass of soft
}-ellow light that somewhat resembles the
color of a high class oil lamp, or that from the
electric incandescent carbon lamp, the in-
trinsic brilliancy of which is theoretically too
great.
586 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo.
The connections of the four radiators are
shown in the upper diagram of Fig. 16, but
the total flux of light is not very much less
when the connections are as in the lower
diagram of Fig. 1 6.
Scores of modifications and varying designs
have suggested themselves. For example.
^
Fig. 22. A Tube Variation of the Radiator Lamp
such a type of lamp that is suitable for alter-
nating circuits will differ from a properly
designed direct -current lamp. But all alter-
nating-current lamps will give some light
on a direct current of the same voltage. That
is, only one of the radiator poles (the nega-
tive) will give any light. Therefore, in the
case of the lamp shown in Fig. 15, only two
of the four radiator plates will be luminous.
The positive poles will remain absolutely
dark. This fact is given recognition in the
design of the direct-current lamp, shown in
Fig. 17. The inner cylinder is of sheet ahmii-
num and the outer cylinder is of aluminum
netting and made the negative pole.
Fig. 18, which is taken from the United
States Patent No. 1,316,967, and which was
applied for November 30, 1917, shows the
positive electrode in the form of a spiral on
the axis of the lamp.
Fig. 19 shows a very simple lamp for alter-
nating circuits that is constructed by insert-
ing into a three-inch straight-sided glass
bulb four right angles made of aluminum
netting of U.052 wire having a mesh of eight
wires per inch. Each right angle is 5 inches
long. Eight glass buttons or spacers keep
all portions of these four angles at a uniform
distance of ys in- from each other, and they
are all held in place by the walls of the bulb.
Just as a final mechanical fonn for the major
designs of the tungsten lani]) was arrived at,
so also will doubtless be the case with lamps
or tubes based on the corona principle.
These lamps should be so designed mechan-
ically that a maximum amount of the light
that is generated has free exit or is reflected
in the best manner.
Fig. 20 has a construction that closely re-
sembles that of a standard tungsten incan-
descent lamp.
Fig. 21 shows wire electrodes wound parallel
to each other on a glass drum.
Fig. 22 shows such a lam]) in the fonn of a
tube.
The dozen most important factors in-
volved in the design of these lamps should
be examined theoretically and a definite
conclusion reached concerning each one, so as
to determine definitely its possibilities.
Some of the important variables are :
The gas pressure; for (a) efficiency, (b) life.
Electrodes; material, form (wire), and area
best suited for a definite voltage, life,
wattage, intensity, and efficiency.
The exhaust program.
The length of the gas column; that is, the
distance between radiators of opposite
polarity.
Volume and shape of the bulb.
When the gas pressure is too high (about
ten millimeters) no light appears; but at six
millimeters it consists of a velvety or lumi-
nescent glow that closely envelops the radi-
ators and extends further and further from
them as the pressure grows less until it fills
the entire bulb with a suffused glow, which,
however, becomes thinner and less luminous
when the pressure is still further reduced.
It seems advantageous to subdivide the
radiator of each negative pole. The lamp
made in accordance with Fig. 23 shows that
far more light is generated between such sub-
divisions than between areas attached to op-
posite poles.
It is apparent that a brighter lamp is
desirable. Photometric measurements of
lamps constructed as in Fig. 21, showed 2.59
spherical candle-power on 220 volts altemat-
Fig. 23. A Form of Lamp with Subdivided Electrodes
ing current. The higher the voltage, the easier
is this problem. Therefore, perhaps it would
be best first to develop a lamp for the com-
mercial 500-volt circuits. When the voltage
is raised abnormally <mi most of these lamps,
they will arc destructively even though the
GASEOUS CONDUCTION LIGHT FROM LOW-VOLTAGE CIRCUITS
oS7
air gap is large. Oftentimes there seems to
be less tendency to this destructive "ball
discharge" arcing when the air gap is small
than when it is large, because then the ohmic
series resistance can be greater. The lamp
shown in Fig. 15 has a discharge gap of 5^ in.
but in other lamps it varies from ^ in. to
one inch.
The lamps that gave the most light on 110
volts are those whose radiators were made of
wire of sm.all diameter and small total area
as shown in Fig. 20.
Photometric data of various types of these
corona lamps have been obtained by the use
7. Multiply Moore lamp readings by ratio
of Mazda "B" reading to Mazda "B"
spherical candle-power.
The tabulation in Table I shows, first,
the performances of four lamps constructed
approximately as in Fig. 15 on alternat-
ing current, and then follows the test
data of several lamps of varying construc-
tions.
Table II shows most of these lamps when
operating on direct-current circuits, under
which circumstances of course but one pole
gives any light.
TABLE I
ALTERNATING-CURRENT LAMPS
LINE
LAMP
LINE
SERIES
Lamp'
Nos.
Volts
Amps.
Volts
Watts
Spherical
Candle-power
Watts per
Spherical
Candle-power
Watts
Watts per
Spherical
Candle-power
Resistance
547
155 (min.)
164
0.03
149
5
0.234
21
4.5
24
500
221
0.105
168
15.5
0.,594
26
21
35.5
264
0.16
184
24.5
1.04
23.5
37.5
36
595
135 (min.)
300
166
0.045
1.53
3.9
0.258
15
4.5
17
•
220
0.11
1S7
17.4
1.15
15
21
17
265
0.18
211
32.9
2.44
13.4
42.6
17
594
127 (min.)
300
167
0.045
153 .
4.4
0.392
11.2
5.0
12
220
0.135
180
21
1.24
17
26.4
21.2
265
0.215
200
37.2
2.4
15
51
21.2
430
220
0.13
155
17
0.715
23.5
26
34
500
605
220
0.11
188
16.7
0,897
18.5
20
22
300
609
220
0.245
195
53
1.825
29
58
32
100
600
220
0.095
172
14.1
0.715
19.8
18.6
26
500
647
220
0.185
164.5
24.5
0.870
28
34.5
39
300
270
220
0.205
179
29
1.047
27
37.5
36
200
669
220
0.135
193
15.9
0.645
24
19.5
30
200
675
220
0.085
126.5
19
0.601
31
16.5
27
1100
673
220
0.22
176
30.3
0.847
35
40
47
200
of an 80-inch sphere. Color corrections
were made by the following procedure:
1. Hold Moore lamp at 220 volts and ad-
just comparison lamp to Moore lamp
color.
2. Set galvanometer on zero and maintain
comparison lamp at above color.
3. Adjust Mazda "B" lamp to comparison
lamp color and note voltage.
4. Ascertain horizontal candle-power of
Mazdz " B " lamp at above voltage.
5. Horizontal candle-power of Mazda "B"
lamp X 0.7S5 — spherical candle-power
of Mazda " B " lamp.
G. Read Mazda 'B" lamp and all direct-
current or alternating-current Moore
lamps against comparision lamp as set.
It is safe only to consider these data, how-
ever, as indicating very broad generalizations
because no two lamps have been made alike,
even as regards their mechanical construc-
tion, and they also differ as regards the purity
of the gas and its pressure, the exhaust pro-
gramme, etc. There were also encountered
difficulties as regards the photometrical and
electrical measurements; for example, when
such a lamp is consuming less than two watts
the amount of the light seems considerable to
the eye in a dark room .
Nevertheless, I believe that complete and
exact specifications should be determined upon
for an ideal lamp of this nature entirely inde-
pendent of itscomparativerelationstootherand
seemingly far superior fomis of artificial light.
588 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo.
Some of the conclusions that may be drawn
are as follows :
1. The efficiency of these lamps is about
the same whether operating on alternat-
ing-current or direct-current circuits.
2. The efficiency is about the same on
alternating-current circuits over a wide
voltage range.
3. The efficiency is about the same on
alternating-current circuits over a wide
range of intensities.
4. The spherical candle-power varies ap-
proximately with the wattage on either
alternating current or direct current.
9. The power-factor of these lamps is
about 85 per cent.
This new form of lamp demonstrates that
useful gaseous conductor light, that to say the
least has advertising value, can now be pro-
duced in a simple manner from ordinar%- com-
mercial circuits. Special uses will be found
for such lamps, for example, as polarity or
potential indicators. Since the internal parts
are all below red heat, gas explosions will not
be caused by bulb breakage. Gaseous light,
due to electrical agitation, has to a limited
extent been emancipated from all necessity
TABLE II
DIRECT-CURRENT LAMPS
LINE
LAMP
LINE
SERIES
Lamp
Nos.
Volts
1
Amps.
Volts
Watts
Spherical
Candle-power
Watts per
Spherical
Candle-power
Watts
Watts per
Spherical
Candle-power
Resistance
547
165
0.017
1.56.5
2.66
0.158
16
2.8
17.8
500
220
0.066
187
12.3
0.444
27
14.5
32.7
265
0.124
203
25.1
0.880
28
32.8
37.3
595
165
0.013
161.1
2.0
0.178
11.7
21.1
12.1
300
220
0.072
198.4
14.2
0.792
17.8
15.8
20
265
0.134
224
30
1.88
16
35.8
18.9
594
165
0.014
160.8
9 O
0.178
12.3
2.3
12.9
300 •
220
0.061
201.7
12.3
0.633
18.7
13.4
20.2
1
265
0.13
261
33.9
1.53
22
.34.4
22.5
605
165
0.01 o
160.2
2.5
0.217
TT.5
2.6
12.2
300
220
0.078
196.6
15.2
0.787
19
17.1
21.8
265
0.146
221
32
1.48
22.3
38.7
26
609
220
0.16
204
32.4
1.48
21.5
35
23.5
100
430
219
0.105
166
16.2
0.796
22.5
21.9
27.4
500
600
220
0.04
198
8
0.387
20.5
8.8
22.5
500
The lamps with a reasonably pure
neon color were not as efficient as those
in which gas impurities made the color
whiter.
The general lamp performance is not
ver\' sensitive to wide variations in the
length of the gas column or gap.
The same lamp equipped with the same
resistance and operating at the same
voltage takes a considerably higher line
wattage on alternating current than on
direct current, which doubtless is prin-
cipally due to the light radiating area
being double.
The candle-power is greater ^\•ith radia-
tors of large area.
for an heretofore ever-present high potential
either for starting or normal operation. The
basic conception of using a gas to supplant
the heated filament in an ordinan*- lamp
seemed wholly impossible, yet this new type
of lamp makes it at least a partial reality.
It constitutes an advance in that it adds to
our knowledge of a very little developed
subject. A new epoch in the histon.- of Gas-
eous Conduction Lighting has been reached.
It is my hope that the great cause of new and
better lighting methods in which my deep inter-
est has been centered for years may be spurred
to rapid advancement in a new direction that
gives promise of reward to an unlimited num-
ber of worthy investigators and inventors.
589
Fundamental Phenomena in Electron Tubes
Having Tungsten Cathodes
Part II
By Irving Langmuir
Research Laboratory, General Electric Company
The preceding section of this article appeared in our June issue. In the present and concluding install-
ment, the author first treats of the fact that electrons impinging on solid surfaces frequently cause the
emission of relatively large numbers of secondary electrons. He then discusses how such phenomena may
play a prominent part in the operation of vacuum tubes and describes striking experiments illustrating these
effects. In the remainder of the article he deals with the manner in which the various fundamental phenomena
co-operate to determine the operating characteristics of electron tubes. — Editor.
Secondary Electron Emission from the Walls of a
Discharge Tube
With relatively high-voltage electron dis-
charges in high vacuum, particularly w^hen the
walls of the vessel are of unusual shapes, so
that the distance between the electrodes is
large and the walls of the vessel interfere with
the free passage of electrons between the
electrodes, it may happen that the bombard-
ment of the walls of a vessel by high velocity
electrons and the consequent emission of
secondary electrons becomes a phenomenon
of vital importance.
When a surface of glass is struck by elec-
trons with relatively high velocities corre-
sponding to 20 to 100 volts or more the energy
of the impact may cause other electrons
already present in the glass to be knocked off
or emitted. Of course, these electrons leave
the glass with less energy than that of the
primary electrons which struck the glass,
and a large part of the energy of the primary
electrons is converted into heat, thus heating
the walls of the vessel subject to this electron
bombardment. Under ordinary conditions,
with a discharge tube in which the walls of
the tube are not so shaped as to interfere
directly with the passage of electrons between
the two electrodes, phenomena of this kind
become important only at extremely high
voltages, of the order of magnitude of 50,000
volts or more, although in the presence of
small amounts of gas the effects of secondary
electron emission arc often a serious disturbing
factor at voltages of the order of magnitude
of 10,000 vdlts.
A glass tube with cylindrical walls, about
two inches in diameter and about eight inches
long, contained a V-shaped tungsten cathode
placed at about the center of the tube, and
a disk-shaped anode of tungsten placed
perpendicular to the axis of the tube at a
distance of about an inch or a little more
from the tip of the V of the tungsten filament.
This tube was exhausted to a particularly
high vacuum, the tungsten anode having
been heated to brilliant incandescence by use
of a high-voltage discharge. It was found
that this tube could be operated in either
one of two ways; (1) If a voltage of a few
hundred volts was applied to the anode and
the temperature of the cathode was gradually
raised, it was found that the tube operated
in a normal manner, the energy of the
discharge being liberated at the anode and
causing a heating of the anode, the walls of
the tube remaining relatively cold, receiving
only heat that was radiated to them from the
anode. Under these conditions the current
between anode and cathode, because of the
space charge and because of the negative
charge which accumulated on the walls of the
vessel, was limited to a comparatively small
value. (2) If a switch in the circuit carrying
current to the anode was opened and then
immediately closed again, it was found that
the characteristics of the discharge were
entirely different. The current flowing to the
anode was now much greater than before, but
the heating of the anode was less than before,
and the energy of the discharge was liberated
in greater part on the walls of the vessel
surrounding the space between two elec-
trodes.
It was found repeatedly that it was possible
to change from one type of discharge to the
other. With the second type of discharge
there was a tendency for the vacutnn to
become poorer, as was indicated by a decrease
in the electron emission from the cathode.
These simple experiments, together with
practical experience obtained in connection
with the manufacture of Coolidge X-ray
tubes and the exhaustion of a large nttmber of
590 July, 192()
GENERAL ELECTRIC REVIEW
Vo XXIII No. 7
thermionic devices, have indicated clearly
that secondary electron emission from the
walls of the glass under influence of bombard-
ment is responsible for phenomena of this
kind.
In the second type of discharge the sudden
application of the voltage to the anode causes
some high velocity electrons to strike the
walls of the vessel. These high velocity
electrons cause the emission of large num-
bers of low velocity secondary electrons from
the surface of the glass, so that the walls of
the glass lose a larger amount of negative
electricity than they received, and thus
tend to become positively charged. The
discharge therefore maintains itself, the
electrons flowing from the cathode to the
surface of the glass, thus causing the emission
of other electrons which then pass to the
anode. This discharge, of course, soon
reaches a stationary condition in which the
number of electrons emitted from the glass is
equal to the number of electrons which strike
the glass. The potential of the glass is then
intennediate between that of the anode and
cathode, usually much more nearly that of
the anode than that of the cathode. The fact
that the glass surface, which has a large
area compared to that of the anode, becomes
decidedly positively charged with respect
to the cathode, greatly increases the current-
carrying capacity of the space between the
two electrodes, in much the same way that a
positively charged grid in a thrcc-clcctrode
device increases the current that can flow
between the cathode and anode.
When devices are made up in which the
glass walls play even a more prominent part
than in the tube that I described, for example,
where the anode and cathode are separated to
considerable distances and the vessel forms a
rather long, narrow tube connecting the
spaces around the two electrodes, the phe-
nomena due to secondary electron emission
have a still more controlling effect on the
characteristics of the discharge. The most
striking difference between the characteristics
of a tube of this kind and one in which second-
ary electron emission plays no part is that
the walls of the tube become heated by the
secondary electron bombardment, instead of
the whole of the energy of the discharge being
liberated in the fomi of heat at the anode.
The bombardment of the walls of the tube
and the resultant heating tend to liberate
gas from the walls, so that if a tube of this
* Annalen der Physik. S3, 67.3 (1910); Leipiigcr Cerichte 63,
34 (I9U); Annalen dor Physik. .53. 24 (19U).
kind is to have constant characteristics,
particularly great care must be used in freeing
the walls of the vessel from gases which
might otherwise be evolved while the dis-
charge is passing.
Lilienfeld* has made an elaborate study of
discharges taking place in high vacuum
through long, narrow tubes connecting two
bulbs, one of which is provided with a hot
cathode, while the other contains an anode.
He measured the potential drop between
sounding electrodes placed in the tube about
three centimeters apart.
The energy of the discharge was consumed
in heating the walls of the tube. This heating
togetherwiththe fluorescence of the tube walls,
and the uniform ])otential gradient along the
tube prove that the discharge depended upon
the emission of secondary electrons resulting
from the electron bombardment of the walls.
Lilienfeld, however, interprets his results as
indicating that empty space is dissociated
into positive and negative electrons by pas-
sage of current through it. By repeating
Lilienfeld's experiments. Dr. A. W. Hull and I
have found that the characteristics are of just
the kind that are to be ex])ectcd as a result of
secondary electron emission. No measurable
discharge takes place in high vacuum with a
tube like Lilienfeld's until voltages above
1000 volts are apjilied between the anode and
cathode. The current then begins suddenly.
At much higher voltages the current increases
about in i)roporlion to the square of the
voltage. This tyj^c of electron discharge is
radically different from that in the pliotron or
kenotron, where the electrons pass dirccth"
from cathode to anode and liberate all their
energy at the anode in the form of heat.
Dr. A. W. Hull* has constructed a device
called the dynatron, in which secondary
electron emission from a metallic anode is
made use of to produce a true negative
resistance.
The device consists essentially of a filament,
a plate, and a jierf orated anode (or grid)
located between the filamelit and plate. The
anode being maintained at a positive potential
of, say 200 volts, attracts the electrons from
the cathode, but most of these pass through
the perforations of the anode and strike the
plate unless this is at too low a potential. If
the plate is at a ])otential of over 2.'> volts,
some secondary electrons of low velocity arc
emitted and tJiese pass to the anode. The
emission of the secondary electrons thvis
decreases the current to the jilate. When
the plate voltage is raised to about lUO volts.
PHENOMENA IN ELECTRON TUBES HAVING TUNGSTEN CATHODES 591
the number of secondary electrons emitted is
about equal to the number of primary
electrons striking the plate, so that the plate
loses about as many electrons as it gains, and
the current falls to zero. With a further
increase of plate voltage the number of
secondary electrons exceeds that of the
primary, so that the effect of the electron
bombardment of the plate is to cause it to
give up more electrons than it receives. Thus
the current to the plate flows in the opposite
direction to the applied potential, but under
certain conditions is proportional to this
applied voltage. Thus the plate circuit has a
Volt-ampere Characteristics
If we determine the volt-ampere character-
istics of any two-electrode electron device or
kenotron, over a wide range of voltages,
including negative anode voltages, we obtain,
in general, a curve consisting essentially of
three parts:
(a) A region in which the current is
determined by the initial velocities of the
electrons; (6) a region where the current is
determined by space charge; and (c) the
saturation region in which the current is
determined by the electron emission from
the cathode. The laws of variation of current
20 —
JO —
-JO B A -0.5
VO/U
characteristic exactly like that of a negative
resistance.
This secondary electron emission is not
normally of importance in the pliotron, but
under exceptional circumstances, as for
instance, when very high grid potentials are
used, the secondary emission from the grid
may produce very remarkable results. Thus
with very high grid voltages the grid current
may reverse in direction, so that electrons
flow from the grid to the anode. If there is
sufficient impedance in the grid circuit the
grid thus becomes still more positively
charged until its potential approaches that
of the anode. In high power tubes this may
lead to extreme electron bombardment of the
grid and undue heat production. vSuch effects
are easily avoided by limiting the positive
potentials of the grid to reasonable values.
• Phys. Rev. r. 141 (1916); Proc. Inst. Radio Eng. ff. 5 (1918).
tSee Richardson. Phil. Mag. 16. 35.3.890 (1908); 17, 813
(1909); IS. 681 (1909).
with voltage and with temperature in these
three regions are totally distinct from one
another.
(a) Current Limited by Initial Velocities
With negative anode voltages the current
varies with the voltage according to a law
derived from Maxwell's Distribution Law.
In this region the current is dependent on
the number and velocities of the electrons
emitted and is therefore extremely sensitive
to filament temperature.
In applying Maxwell's Distribution Law
it is necessan- to know exactly the shapes of
the electrodes, t
In the ideal case of parallel plane electrodes,
the current should increase exponentially with
the voltage of the anode, that is, for each
increment of voltage, the current should
increase in the same ratio. Fig. 1 illustrates
more clearlv the wav that the current varies
592 July. 1920
GENERAL ELECTRIC REVIEW
Vol XXIII. Xo.
with the voltage. The curve A P is an exponen-
tial cun-e. In this cur\^e the anode voltage is
plotted as abscissa and the current passing
to the anode is given as ordinate. It is seen
that this Maxwell's Distribution Law is ob-
ser\-ed only with negative anode voltages.
10-
L -
20 30
VoJtS
Fig. 3
For each increase of 0.2 volt in the anode
voltage the current increases by a little more
than two-fold.
If now. instead of plotting this curve as
current against voltage, we plot the logarithm
of the current against the voltage, we obtain
a straight line as is shown bv the line A P of
Fig. 2. ,
The slope of the line A P in Fig. 2, is not
arbitrary, but can be calculated by Max-
well's Distribution Law. The slope is
inversely proportional to the temperature of
the cathode. It is apparent at once that
there is a very great advantage in plotting
the volt-ampere characteristics over this
range as in Fig. 2, that is, plotting the
logarithm of the current instead of th,e current
itself.
The simplicity of the relationship denoted
by the line A P in Fig. 2 is due to the fact that
between parallel plane electrodes it is only
the velocity component perpendicular to the
surface which determines whether the elec-
trons emitted from the surface of the cathode
are able to pass to the anode. When the
anode or cathode arc not in the form of
parallel planes their relationships are in
general more complicated.
With a small cathode which might practi-
cally be considered as a point, at the center
•Anna'.ender Physik. I,!,. 1011. (1914).
of a spherical anode, it is evident that all the
electrons striking the anode are moving in a
direction perpendicular to the surface of the
anode. Under these conditions it is not the
velocity component in the given direction
among the electrons escaping from the
cathode which is important but rather the
total velocity. This velocity is distributed
according to a somewhat different law from
that which applies to the velocity component,
so that another form of Ma.xwell's Distribu-
tion Law must be applied.
By the dotted line B P in Fig. 2 it is shown
approximately how the characteristics of a
device in the region where the current is
limited by the effect of initial velocities is
modified b>" the use of electrodes which are
not in the form of parallel planes. The exact
shape of this cur\-e depends on the shape of
the electrodes, but in general the curve is
horizontal where it meets the axis O P at the
point P, and tends to be parallel to the line
A P for the larger negative \-oltages.
Schottky* has discussed in detail the
characteristics of a thermionic device having
a straight filament mounted in the axis of a
cylindrical anode. He avoided complications
due to the end effects in the cylinder by
having two auxiliary- cylinders placed at the
ends of the main cylinder; he also avoided
complications due to the potential drop along
the filament by using a rotating commutator.
He found that with currents of the order of
magnitude of one tenth of a milli-ampcrc the
experimental cur\-es depart considerably from
the theoretical curves because of the space
charge effect. With currents as low as 10-'
ampere this effect is, however, not appreciable.
PHENOMENA IK ELECTRON TUBES HAVING TUNGSTEN CATHODES 593
Where the effect is absent the curve is of the
type shown by B P H of Fig. 2.
The part of the curve represented by the
horizontal hne P H corresponds to the satura-
tion current, and is therefore determined by
the electron emission solely. Schottky finds
that in general there is a transition curve
such as that shown by T S H in Fig. 2, between
the part of the curve where the current is
limited by the effect of initial velocities and
the part where it is limited by electron
emission.
With electron tubes which operate with
currents of the order of milliamperes, the
part of the curve in which the current is
limited by the initial velocities is only a very
negligible part of the whole curve.
(b) Current Limited by Space Charge
In the second region the current increases
in proportion to the 3/2 power of the applied
voltage, and is practically independent of the
filament temperature. In this range where
the current is limited by space charge the
relation between the current and the voltage
is like that shown in Fig. 3. If instead of
plotting the volts and the amperes we plot the
logarithm of the current against the logarithm
of the voltage, or if we plot the current and
1800
2/00°
1900' £000"
Temp
Fig 5
the voltage on logarithmic paper, we obtain
a straight line as indicated in Fig. 4. The
slope of this line is not arbitrary but is
always 3/2, that is, for any given increment
of the abscissas the increase in the ordinates is
1.5 times as great. It is readily seen that the
law according to which the current increases
with the voltage, as shown by Fig. 4, is very
dift'crent from that which applies to the cur-
rent limited by initial velocities (Fig. 2). In
the latter case in order to obtain a straight
line, the logarithm of the current was plotted
'/tX/0-'
Fig. 6
directly against the voltage, whereas in the
present case it is plotted against the logarithm
of the voltage. Furthermore, it should be
noted that in Fig. 4, which applies to space
charge, the slope and position of the line are
independent of the cathode temperature,
while the slope of the line in Fig. 2 is inversely
proportional to the temperature, and the
position of the line is also affected by the
temperature since the line must pass through
the point P (Fig. 2) and the ordinate of this
point increases in proportion to the electron
emission.
Current Limited by Electron Emission
In the third region the current is inde-
pendent of the applied voltage but varies with
the temperature according to Richardson's
equation and thus increases extremely rapidly
with the temperature.
The curve which represents the variation
of current with voltage is thus a straight
horizontal line. The current increases with
the temperature very rapidly, as shown in
Fig. 5. The ordinates represent the current
and the temperatures are plotted as abscissas.
This rate of increase with temperature is very
much more marked than that shown in Fig. 3.
594 July, 1920
GENERAL ELECTRIC REVIEAV
Vcl. XXIIL Xo. 7
Furthermore, the curve approaches the axis
0 X asj-mptotically and practically coincides
with it for all temperatures below a certain
value. Above that temperature it departs
very rapidly and cur\-es upward at a high rate.
In Fig. 6 is plotted the same cur\-e as that
shown in Fig. 5, except that the logarithm of
the current is plotted against the reciprocal
of the temperature, so that we obtain a
straight line.*
Let us now consider how the three types of
characteristics just discussed combine to
form the complete volt-ampere characteristic
of an electron tube. With three-electrode
tubes or pliotrons let us assume at first that
the grid is connected to the anode.
A
0.7
P
H
J.H
0.6
<^0.5
/
\o.4
d/
ft
%
/f
^0.3
/«
/t
5
/>
:$£».2
/g
*..
/'C
B/^'
A
J^
<'
.0123
4 5 6
Vo/ts
8 9
Fig. 7
With very low filament temperature, where
the saturation current is of the order of a
microampere or less, the space charge effect is
practically absent. For negative anode
voltages the current is thus determined by the
initial velocities, while at positive anode
voltages the current is saturated. The volt-
ampere characteristics are thus of the type
shown in Figs. 1 and 2. There is normally a
transition cun-e as indicated by T S H in
Fig. 2 between the region in which the current
is (a) limited by initial velocities and that
in which it is (c) limited by electron emission.
When the filament temperature is raised
so that the saturation current is of the order
of milliamperes, the volt-ampere character-
* Strictly speaking, the logarithm of the current divided by
the square root of the absolute temperature should be plotted
in order to get Richardson's equation. However, this square
root term produces an effect that is hardly perceptible in any
ordinary plot.
istics undergo a ver\- fundamental change.
The space charge then becomes the pre-
dominating factor. As an example, let us
consider the characteristics of a kenotron
having a cylindrical anode one inch in
diameter and two inches long with a tungsten
filament of 0.005 inches diameter in its axis.
The full line A B D E F in Fig. 7 gives the
characteristics to be expected for such a tube
when the filament temperature is 19.S0° K.,
giving a saturation current of 0.65 milli-
amperes. The dotted line A P shows the
limitation of current calculated from Max-
well's Distribution Law. This cur\-e corre-
sponds exactly to the cur\-e A P of Fig. L
Thus, if it were not for the space charge
effects the current would vary with the
voltage according to a cur\-e A P J F. As a
matter of fact, if large negative potentials are
applied to the anode so that the current flow-
ing is very small (of the order of 10-' ampere)
the volt-ampere cur\-e does actually follow
accurately the curve A P, but as is seen by
inspection of Fig. 7, the ordinates of the
curve with such low currents are entirely
invisible when plotted on the scale used in
Fig. 7.
If the effects of initial velocities were wholly
negligible no current would begin to flow
until a positive potential is applied to the
anode. The current would then increase
according to Equation 5, that is, the current
would increase in proportion to the 3 2 power
of the voltages as shown by the curve O C H of
Fig. 7, until the current is limited by the
electron emission from the cathode. In other
words, under these ideal conditions, the
current increases according to the cun-e
represented by the dotted line O C J and then
follows J F representing the saturation current.
• Actually, however, because of the initial
velocities there is a certain current flowing to
the anode even when the potential of the
anode is zero. The current under these
conditions is depentlent not only on the
initial velocities, but to an even greater degree
on the space charge produced by the electrons
flowing across the space due to their initial
velocities. The actual volt-ampere character-
istic is therefore of the tvpc given bv the
cur\c A B D E F of Fig. 7. '
The deviations of the curve A B D from the
cun-c O C H arc due to initial velocities. The
curve A B D theoretically follows the cur\-e
A P until a current of a few microamperes is
reached; the cur\-e then departs radically from
the cunc A P and takes the course indicated
by A B and tends to approach the cunc O C
PHENOMENA IN ELECTRON TUBES HAVING TUNGSTEN CATHODES 595
or rather tends to become parallel to it,
differing from it only by a value corresponding
to a few tenths of a volt.
Let us now consider the case where the
filament temperature is raised, so that the
electron emission from the filament is 50
milliamperes. This requires a filament
temperature of about 2360° K., which is
rather lower than the normal operating
temperature of the filament on an ordinary
tungsten lamp.
If we substitute in equation 5, i = 0.010
amperes per centimeter of length (correspond-
ing to 50 milliamperes for a five-centimeter
length) and place r=1.25 centimeters and
solve the equation for T, we find V' = 90 volts.
This indicates that a voltage of 00 volts is
required to overcome the space charge effect
when a current of 50 milliamperes is used,
assuming the initial velocities of the electrons
to be negligible. The volt-ampere char-
acteristic is thus given bv the cur\'c A D F in
Fig. S.
The curve O D J H in this figure represents
the 3 2 power relation calculated from
Equation 5. For voltages above a few volts
the volt-ampere characteristics of the device
considered should follow the theoretical curve
so closely that on the scale used for Fig. 8
the difference between the two curves would
be hardly visible. If the filament temperature
were so "high that a large surplus of electrons
was produced the current would increase
indefinitely along the curve O D J H. Actually,
however, since the filament at 2300° K.
emits only 50 milliamperes, the current cannot
increase above the line represented by P F.
If the voltage on the anode is zero the
effect of initial velocities is to cause some
current to flow to the anode. This current is
limited, however, mainly by space charge
although it is larger than at lower filament
temperatures.
The cur\'e corresponding to the distribution
of the initial velocities of the electrons
according to Maxwell's law is represented in
Fig. 8 by the dotted line A P, which corre-
sponds exactly to A P in Fig. 1. For negative
voltages on the anode of such magnitude
that the current is of the order of micro-
amperes of- less, the actual volt-ampere
characteristic represented by the line A D will
theoretically approach the line A P, but of
course currents as small as this represented
on the scale of Fig. S would give ordinates
much too small to see.
From the foregoing discussion it is clear
that the volt-ampere characteristics of two
electrode devices consist essentially of the
three parts in which the currents are limited
respectively by (a) initial velocities of the
electrons, (6) space charge, and (c) total
electron emission. However, over certain
parts of the characteristic two factors may
operate simultaneously. Thus there are
transition curves like those shown in Fig. 2
by T S H, in Fig. 7 by A B and D E F, and in
Fig. 8 by D E F. This latter form of transition
curve, namely, that between the space charge
and the saturation regions is the one that
concerns us most directly in electron tubes.
The extent or length of this transition curve
depends upon several factors of which the
following are probably the most important :
/O 0 10 20 30 40 50 60 W 80 90 100 110 120 130
Volts
Fig. 8
Non-uniformity of the Field Around the Cathode
In the case of a straight filament in the
axis of a long cylindrical anode, the field
around the cathode is greater at one end
than at the other, because of the effect of the
voltage drop along the filament. Thus as the
voltage of the anode is gradually raised, the
current from some parts of the cathode
becomes saturated before that from other
parts, and this effect tends to extend the range
of the transition curve. A similar effect
occurs in case the strength of field around
the cathode is made non-uniform in any other
way. Thus, if the filament is in the form of a
V or W some parts may be closer to the
anode than other parts. Or again, the grid-
like effect of one part of the cathode or
another part may cause differences in the
field strength.
596 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 7
Lack of Uniformity of Filament Temperature
The cooling effect of the leads causes
difference of temperature in the filament
and the current from some parts becomes
saturated before that from other parts. This
also extends the range of the transition cur\-e.
Heterogeneity of the Surface of the Cathode
This effect, which has been discussed at
length in connection with the electron
emission from filaments, is of particular
importance in affecting the transition cur\-e
between the space charge and the saturation
regions.
The Effect of the Grid in Photrons
The action of the grid is to modify the
effect of the space charge. A positive charge
on the grid partly neutralizes the space
charge of the electrons and thus increases the
current-carrying capacity while a negative
charge has the reverse effect.
In general, when the current is limited by
space charge the electron current / depends
on the anode voltage 1 'a and the grid voltage
Vg according to the equation*
Here A.' and k are constants depending on
the construction of the electrodes.
By taking the logarithm of this equation
and differentiating we can readily find that
the exponent n according to which the
current increases with the anode voltage is
„_ dlogl ^ 3 ^ d I
9
dlog V
^ a
(7)
Thus, if the grid is at the potential either of
the cathode or of the anode, the exponent is
3/2. The same is true if the grid voltage is
increased in proportion to the anode voltage.
I'g I'a remaining constant. If, however, the
grid voltage is kept constant while the anode
voltage is varied. Equation 7 becomes
Thus, when the grid is positive with respect
to the cathode the exponent n is less than 3 2
while for negative grid voltages n becomes
greater than 3 2. Measurements of the
characteristics of tubes show this relation
clearly. A receiving tube (pliotron) with
zero volts on the grid gave for the exponent
n the value L54. For a positive grid potential
of 2 volts the exponent was 1.3 while for a
negative potential of two volts it was 1.9
and for ten volts it was 3.6.
If the grid and the filament are of different
materials there will be in general a contact
difference of potential between them even
when they are connected together. Thus, if
the grid is of nickel and the cathode is of a
material having a much higher electron
emission (such as a Wehnelt cathode or
thoriated tungsten cathode), the grid will
have a negative potential with respect to the
cathode. The exponent should thus be
greater than 3 2 if the grid is connected to the
cathode. In order to bring the grid to the
same potential as the cathode, a positive
electro-motive force should be applied to the
grid sufficient to compensate for the effect
of the contact potential. In the case of a
nickel grid and a Wehnelt cathode this
contact difference should be about 2.2 volts, t
so that there should be a material effect on
the exponent n.
• Langmuir. Proc. Inst. Radio Engs. ». 278. (1915) General
Electric Review IH. (1915).
t The constant b of Richardson's equation for a Wehnelt
cathode (50 per cent /*,,() and .50 per cent SrO) has been gi\-en
by W. Wilson (Phys. Rev. /(). ,9 |1<1171). as approximately
25.000 degrees corresponding to 2.14 volts. According to a
method of calculation given bv L.inRmuir (Trans. Amer. Elec-
trochem. Soc. !0. 16.5-166 |19i61». the probable value of 6 for
nickel is .50.000 degrees, corresponding to 4. .'12 volts. The contact
potential is the difference of these or 2.18 volts.
597
The Safety Car
By W. D. Bearce
Railway and Traction Engineering Department, General Electric Company
Some five years ago a movement was started toward a reduction in car weights and the consequent
reduction in operating cost. In order to provide equipment to compete with the jitney automobiles then
thriving in the West and Southwest the so-called safety car was made unusually hght with ample power for
rapid acceleration and was handled by a single operator. The success of the venture was immediately
apparent and engineers and car designers then started the standardization which has since been attained in
this equipment. Necessarily the preliminary information that was available regarding the safety car was based
on engineering estimates; and while these figures were reasonably accurate actual operating results are much
more convincing. In the following article the author has included figures of this kind from all parts of the
country. — Editor.
During the past few years the electric
railways of the country have been confronted
with rapidly increasing cost of operation
while their gross income has remained practi-
cally unchanged. A vast amount of study and
attention has been given by the engineering
and financial interests to assist the railways in
the continuance of business under the existing
unfavorable conditions.
The most encouraging results achieved by
these studies have been the development and
the many successful installations of the one-
man light-weight safety car. Examples of
what may be accomplished by this rac^ical
departure from the ordinar}^ method of street
railway transportation may be found in
almost every section of the United States.
Briefly stated, the reasons for the success of
this innovation are :
1. Improvement in service.
2. Freedom from accidents.
3. Increase in riding habit.
4. Lower maintenance cost.
5. Reduction in labor cost.
6. Reduction in power consumption.
As a result of these features, the operating
company's net income has shown a marked
improvement in almost every case. This
increase in gross receipts, combined with the
marked reduction in cost of operation, effects
sufficient saving to insure profitable operation
of roads previotisly run at a loss.
Report of A.E.R.A. Committee
The conclusions of the committee on one-
man car operation presented to the American
Electric Railway Association in October,
1919, represent the findings of a competent
body of operating men on this subject :
1. The safety car is one of the most important
improvements in street railway service that has
appeared for many years. Its valuable features in
the order of their importance are :
(a) Greatly improved service to the public, both
as to frequency and safety.
(b) Increased earnings for the company.
(c) Decreased operating expenses.
2. One-man operation alone, while useful in
saving platform expense in the smaller communities,
is not comparable with the improved service that
can be obtained by the light-weight safety car with
its more frequent headway and greater average
speed.
3. The savings obtainable from one-man cars
should be shared with the trainmen in the form of a
higher hourly rate for the operators of such cars than
is paid to the trainmen on two-man cars.
4. When inaugurating one-man car service, it is
good policy to assure the trainmen that no one will
lose his job due to putting in the new cars. They are
installed, as a rule, on a line at a time; and experience
has proved that the company is not burdened with
extra men through this policy.
5. From the nature of the traffic available, the
safety cars can accomplish more in a large city than
in a small one, for the reason that the possibilities of
increasing riding in the small community are
limited. This statement is made to correct the
erroneous impression existing in some minds that
the safety car is useful only for saving expense in
the smaller cities.
6. Where traffic is believed to be too heavy on the
peak to be successfully handled by safety cars, the
larger, heavy cars may be used for tripper service
on the peak, thus making the light cars handle the
long hour runs.
7. Similarly, where snow storms require the use
of the heavier equipment at rare intervals, the
safety cars can still be used to advantage during
other times.
8. The safety car, though light, is just as sub-
stantial and with the same care in maintenance
should last just as long as the former types of car.
It has a steel frame and thoroughly modern, venti-
lated, interpole motors.
9. Regarding the matter of standardization, your
Committee was not unanimous, but the majority
opinion favored adhering to the present standard
design of the safety car in the interest of cheaper
costs through quantity production.
10. Experience has shown that the overwhelming
majority of both riding public and trainmen favor
the one-man safety car; that it can, at one and the
same time, improve the public's service, increase
the trainmen's wages, and raise the company's
profits; that it can be operated for about half the
cost of an ordinary car; and that most of the
companies that have tried it want more. We
predict an increasingly rapid extension of the use
of a device that can make a showing like the above.
598 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 7
^Hr .« I3hP.
T ,'. T
i
'a
THE SAFETY CAR
599
General Features of the Safety Car
The standard safety car which is most
commonly used is approximately 28 feet in
length and seats 32 passengers, when arranged
for double-end operation. By utilizing the
rear end, three additional seats can be
obtained when the car is desired for single-
end operation only. The body is mounted
on a single truck with 2(5-inch wheels and a
wheel base of about eight feet. The con-
struction of the truck is such that the car has
excellent riding qualities and it is possible to
use accelerating speeds comparable to those
of the competing automobile without dis-
comfort to the passengers.
The safety car, completely equipped, weighs
about eight tons. It is of all steel constrvtc-
tion and is built to a standard form and size.
The roof is of the arched type and the sides
are of steel with windows arranged for open-
ing when desired. The platform is on the
same plane as the body floor and folding
doors and steps are equipped with mechanical
opening and closing devices under the control
of the operator.
The electrical equipment of the car consists
of two 2.5-h.p. ventilated type railway motors,
a controller, special light-weight grid re-
sistors, and a motor-driven air compressor
having a capacity of 10 cu. ft. per minute.
The air brakes include various safety features
and labor saving devices. The safety control
equipment is especially adapted to one-man
operation; the brakes, doors, steps, and
Sanders being controlled by a single brake
handle and mutuallv interlocked.
Fig. 5. 25-h.p. Railway Motor Equipped with Ball Bearings
for Light Weight Safety Car
As may be gathered from the foregoing
and from the following detailed description
of the air brakes and safety devices, the
requirements of this type of car have been
studied out with great care; and to quote
again from the report of the American Electric
Railway Association the development of this
equipment has resulted in:
The creation of an entirely new type of car of
low weight, greatly improved safety, and more
rapid acceleration and deceleration. This car of the
light-weight safety type not only saves platform
and accident expense, but permits of an improve-
ment in service, such as well nigh to revolutionize
the street railway business.
Fig. 6. 25-h.p. Railway Motor for Light Weight Car
Equipped with Standard Sleeve Bearings
Improvement in Service
The effect of impro\'ed service by the use
of safety cars is best shown by the actual
results in Table I.
TABLE I
Houston, Texas. .
El Paso, Texas. . .
Tacoma, Wash. . .
Seattle, Wash.. . .
Gary, Ind
Terfe Haute, Ind.
Tampa, Fla
Bridgeport, Conn,
Per Cent
Increased
Service
80
44
45
55
62
77
51
125
Per Cent
Increased
Gross
Receipts
60
43
41
67
46
44
51
100
Power Consumption
Because of the increased cost of power, due
to the high price of coal, labor, and materials,
the reduction in energy consumption secured
by the use of light-weight safety cars is an
important factor in their success. In some
cases the adoption of this equipment has
actually postponed indefinitely the ptirchase
of additional power equipment. The power
consumption is, of course, dependent upon
the weight of the car, the number and dura-
tion of the stops, the speed profile of the
line, etc. It is, therefore, difficult to make
any definite statement as to the actual power
consumed, except for a specific case, but it is
evident that a car weighing eight tons with
two motors shotild operate with an energy
600 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo.
consumption of approximately one third that
of a 24-ton car equipment with four motors.
The average consumption on most city rail-
way systems is approximately three kilowatt-
hours per car mile. According to the A.E.R.A.
report, the actual figures from forty-five
Fig. 7.
standard Platform Type of Controller Used on
Safety Cars
companies show the energy consumption of
safety cars to range from O.S to 1.75 kilowatt-
hours per car mile.
Safety Car Installations
The total number of light-weight safety
cars in operation and on order in the United
States at the present time is approximately
3400, not including rebuilt cars, many of which
have been equipped with safety features and
are operated by one man. In general, the
rebuilt cars have been used only on lines
of light traffic, and their general use is not
recommended.
By taking the results of many investiga-
tions as a basis it is possible to make a study
of the financial results of replacing the
ordinary types of heavy rolling stock, using
present-day costs of operation, and thus
secure a fairly accurate idea of what return
can be counted upon for an investment made
in safety cars. All such studies so far made,
confirmed by actual results in every existing
installation, indicate that the majority of
roads cannot well neglect placing some of
these cars in their ser\dce.
For instance, there are thousands of
standard city cars which weigh about 40,000
lb. and seat an average of 40 passengers.
The safety car weighs 16,000 lb. and seats
32 passengers. Its motive power consumption
is approximately 40 per cent that of the
heavier car. Its maintenance will be about
40 per cent less. In many instances, where
the cars replaced are exceptionally old or
obsolete, the sa\ang in maintenance will be
much greater. The power ratio of about
40 per cent has been repeatedly checked and
verified; and the maintenance records of the
earliest installations indicate that the ratio
shown is accurate after the cars have been in
service from two to three years.
Table II shows the saving in equipment
maintenance and power which can be secured
by the use of safety cars.
TABLE 11
POWER AND MAINTENANCE CHARGES
Cents per Car Mile
40.000-lb.
Car
'A.rt
4.2.
16.000-lb.
' Car
Equipment maintenance.
Power
2
2
Total
7.7
4
A car operating IS hours daily on an S.5
m.p.h. schedule, which is the average for
city ser\-ice in practically all parts of the
countrA% will run approximately 50,000 miles
a year. The hea\-y car costs for power and
maintenance, when making this mileage,
S4312; the safetv car §2240, a saving of
S2072.
Fig. 8
Operating Handle for Safety Car Controller with
Base and Pilot Valves
The platform expense for a two-man car
averages SI. 20 per car hour. An all-day car,
incliuling a five per cent allowance for report-
ing and lay up time, will run approximately
6!U)0 hours per \-ear, costing in wages $S2S0.
It has been customar>- to pay the operator
of a one-man car a higher wage than cither
THE SAFETY CAR
601
member of a two-man car. The average
platfonn expense for a safety car is 66 cents
per hour. At this rate the platform expense
for the safety car would be $4554 annually, or
a saving of $3726 as compared with a two-
man car.
Car for car, therefore, the safety car on
all-day runs can save over $5700 per year
and would pay for itself within 14 months.
Car for car replacement is not recommended,
as the best results are obtained by operating
more cars on shorter headway thus providing
improved service. Experience has proved
that most lines will stand at least 40 per cent
improvement in service. This can best be
accomplished by operating aboiat 30 per cent
more cars and increasing the schedule speed
10 per cent. For instance, instead of operat-
ing ten cars on a ten-minute headway, operate
thirteen cars on a seven-minute headway,
giving 8.5 cars per hour instead of six, a 40
per cent increase. Reduced stops and the
higher accelerating and braking rates of the
safety cars enable such a schedule speed
increase to be easily made.
The costs and effect of such an increase can
be shown as follows, assuming that only the
regular all-day cars are replaced, using exist-
ing equipment for rush-hour trippers:
Ten old cars, running 8.5 m.p.h. make
560,000 car miles annually at a cost for
power, maintenance, and crew wages of
$122,200.
Thirteen safety cars at 9.3 m.p.h. make
795,000 car miles per year; their cost for
power, maintenance, and crew wages will be
$88,300, a saving of $33,900, while providing
40 per cent more service.
The average receipts per car mile on street
railwavs in the United States is 37.7 cents.
Fig. 9. Air Brake Valve for Safety Car Control Equipment
increase, this amounts to $42,200. The com-
bined effect of reduced cost and increased
gross income is a net increase in earnings of
$76,100, or approximately $7600 per car
annually for each heavy car displaced, which
is equivalent to an annual return of 7S per
Fig. 10.
Automatic Air Compressor Governor Controlling
Operation of Motor-driven Compressor
cent on the first cost of fifteen safety cars.
This provides two spare cars. Taking in-
creased fixed charges on the increased capital
Fig. 11. Standard 10-cu ft. Air Compressor with Tee Bolt
Suspension
The total receipts, therefore, for the ten old
cars in this case will be $211, 120. Experience
shows that a 40 per cent increase in service
means approximately a 40 per cent increase
in receipts. Assuming only a 20 per cent
account at 18 per cent to cover interest,
depreciation, taxes, and insurance, there is
still left a profit to the purchaser of better
than 58 per cent annually — enough to wipe
out their cost in approximately two years.
602 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo.
A^'here traffic does not warrant increased
service and the replacement is made car for
car, 11 cars would probabh- be sufficient for
a ten-car line. The net savings would be
approximately $58,000 which is equivalent to
an annual return of approximately SO per cent
of the first cost of 1 1 safety cars.
Probably in wide-spread applications some
lines would fall into one categor\-, some into
the other. An average result would un-
questionably show, after paying all increased
fixed charges including amortization, between
$5000 and $6000 profit for each car displaced, a
sum sufficient to pay the interest at 6 per cent
on $80,000 to SIOO.'OOO worth of securities.
The data in Table III illustrate the
economies and increased earning possibilities
The foregoing values are based on average
costs for labor and on the replacement of the
heavier types of city cars. Average wage
scales in many properties are materially lower
and power consumption less. Many, moreover,
will show lower average receipts per car mile.
Under such conditions the savings of the
safety car become less, but are still remark-
able as is evidenced by the figures in Table
IV, representing about the lowest costs
anywhere in the country today; they are the
a^-erages of representative roads operating
in the smaller cities of the middle west and
south. The average weight of the cars is
30,000 lb.; their average platform expense is
11.3 cents per car mile and their average
receipts 30 cents per car mile.
1 _j
Fig. 12. Floor Plan of Standard Safety Car
for each car displaced. The figures in column
A are based on running equal mileage with
no increase in cars; those in column B are
based on running 40 per cent more mileage
with 30 per cent more cars.
TABLE III
Saings Made With
.A
Equal
Mileage
$ 840
1232
3726
B
40 Per Cenl
Increase
Maintenance of equi]
annual saving
Power
Crew wages
jment,
per cent
safety
$ 370
7(50
2280
Total savings
Increased receipts at 20
$5798
$3410
4222
Increase in net earnings
Annual return on cost of
car. approximately . . .
$5798
80%
$7632
78%
Even under these circumstances, the new
cars would pay for themselves in less than
two years; or if from these increased earnings
be deducted interest, depreciation, taxes, and
insurance, there remains a clean profit of
from $3500 to S5000 for each car displaced.
Motor Equipment
The electrical equipment developed for the
safety car by the General Electric Company
includes two 25-h.i). railway motors, a light-
weight platform type controller adapted for
use with standard safety features, special
light-weight grid resistor, modified straight
air brake equipment, also suitable for use
with safety devices, and a ten-cubic-foot air
compressor for supplying the air brake and
accessory requirements.
Two types of motors have been most
generally adopted for use on safety cars.
These were designed for
this service and are,
for their capacity, the
Hghtest weight railway
motors manufactured.
One has ball bearings
on the armature shaft
and weighs approxi-
mately SS5 lb. The
other has sleeve bearings
of liberal design and
weighs approximately
1000 lb. The continu-
ous capacity of these
machines is so great
that they operate at
unusually low tempera-
tures, and their per-
formance during the
past five years has been
extremely satisfactory.
This controller was de-
signed for use on light-
weight cars, is compact,
occupies the minimum
of platfomi space and
weighs only 135 lb.
Air Brake and Safety
Features
The air brakes with
safety features and
labor saving devices are
of special importance
when the responsibility
for the operation of a
car is placed in the hands
of one man instead of
the usual crew of two.
In the design of this
equipment, every effort
has been made to guard
against accidents that
might be caused by the
disability or the inat-
tention of the operator.
This equipment is a
modification of the
well-known straight air
brake with emergency
features and safety de-
vices which provide for
bringing the car to a
standstill automatically
should the operator
by reason of sudden
physical or mental
disability be unable to
THE SAFETY CAR
603
604 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No.
perfonn his duties properly. Normalh^ the
brakes, doors, steps, and sanders are con-
trolled by the operator by means of a single
brake valve, making it unnecessary^ for him
to remove his hand from the brake valve
handle to open the doors after the car has
been brought to a stop, to close them when
he is ready to proceed, or to manipulate the
automatic sander. The brake valve is so
constructed that a downward pressure on the
handle in any of the several positions will
cause sand to be applied to the rail.
The safety controller handle, which is an
important part of this equipment, is so
interlocked pneumatically with the brakes,
doors, steps, sanders, and a circuit breaker
tripping device as to cause the brakes to be
applied automatically with full force if the
operator removes his hand from it without
having first made a brake application. In
addition, the circuit breaker is opened, sand
is applied to the rail and the doors are
balanced so that they may be opened manu-
ally, if desired.
To relieve the operator of the necessity of
keeping his hand on the controller handle at
all times while the car is in motion, a relief
valve known as the combined foot and cutoff
valve is provided. This valve is installed in
the safety control pipe and is located on the
platform in such a position that the operator
can reach it with his right foot. By holding
this valve closed, the "dead-man" feature
is transferred from the controller handle to
the foot valve. The latter is automatically
held closed when a brake application of
sufficient force to insure bringing the car to a
stop has been made.
It is impossible for the brakes to " leak off "
through carelessness on the part of the opera-
tor in leaving the car with the brake valve
handle in the "lap" position by reason of the
fact that the combined foot and cutolT valve
will automatically open if the brake cylinder
pressure falls below a safe minimum. The
opening of the foot valve under these condi-
tions will result in emergency operation under
which the brakes are applied with full force
and maintained against leakage.
An emergency valve, which is located inside
the car, automatically controls the brakes,
door engines, sanders, and circuit breaker
cylinders under emergency conditions. This
valve is actviated by a sudden reduction in
pressure in either the safety control pipe or
emergency pipe, hence it will operate: (1)
if the operator removes his hand from the
controller handle (or his foot from the foot
valve) when the brakes are not applied, (2)
if the operator moves the brake valve handle
to the emergency position, or, (3) if the pipe
on either end of the car is accidentally broken
or ruptured.
In all positions of the brake valve, except
the door-opening position, the door-closing
pipe is connected to the emergency line, hence
when emergency operation takes place from
any cause, pressure is automatically removed
from the closing side of the door engines
which permits of the doors being opened
manually.
TABLE IV
OPERATING COSTS
Cents per Car Mile
30.000-lb. 1 Safety
Car 1 Car
Maintenance of equipment
Power
2.5
3.4
1.5
2
Total
5.U
11.3
3.5
Platform expense
6.23
All-day Service
Savings Made With
B lual 40 Per Cent
Mileage Increase
.\nnual savings on maintenance $560 $210
.Annual savings on power . 7Xt) 320
.■\nnual savings on platform
expense $2840 $1870
.\nnual savings on total 4180 2400
Increased receipts at 20 per cent 3360
Increased net earnings $4180 $5760
Annual return on cost of safety
car, approximately 58% 63%
In the normal position of the emergency
valve, the sander reser\-oir is connected to
the main reservc>ir, thus keeping the former
fully charged. When the emergency valve
operates, the sander reservoir is connected
to the sanders and sand is blown onto the
rail until the pressure in the sander reservoir
is exhausted. This arrangement limits the
time of automatic sanding in emergency and
thus avoids an undue waste of sand.
Motor-driven Air Compressor
As comi)rcssed air is used for operating the
brakes and all of the safety devices, it is
imperative that the air comjiressor be of such
design and construction as to insure con-
tinuity of ser\-ice. The center-gear type air
compressor is in succes.sful operation on
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
605
approximately 1000 safety cars in all parts
of the country and has fully demonstrated
its many superior qualities in this class of
service where schedule speeds are high and
the demand for air is greater than heretofore.
These machines have duplex cylinders
fitted with single acting trunk type pistons,
and are driven through herring-bone gearing
by series-wound motors having four salient
poles.
Air Compressor Governor
The functions of the air compressor
governor are to start and stop the air com-
pressor automatically so as to maintain the
air pressure in the main reservoir within
predetermined maximum andminimimi limits.
Air compressor governors were developed
after a careful study of the rigid requirements
of electric railway service, and are in success-
ful use on thousands of cars throughout the
country.
This type of governor is essentially a single-
pole switch of the contactor type operated by
means of a rubber diaphragm, a piston, and
a set of levers. The interrupting switch is
provided with an arc chute of highly refrac-
tory material, an effective blowout, and
easily renewable contacts. The principal
bearings are provided with hardened knife
edges to reduce friction to a minimum, and to
insure a quick snap action.
The Production and Measurement of High Vacua
PART II
METHODS FOR THE PRODUCTION OF LOW PRESSURES
By Dr. vSaul Dushman
Research Laboratory, General Electric Company
This installment and the one which will appear in our succeeding issue describe some of the different
types of pumps that have been developed for the production of high vacua. The August installment will
deal mainly with the Langmuir condensation pump and will contain an appendix which gives detailed infor-
mation regarding the actual set-up and operation of an exhaust system. — Editor.
Classification of Methods for the Production of
Low Pressures
The methods for the production of low
pressures may be classified convenientlj^ under
the following headings:
/. Mechanical Pumps
1. Piston pumps
2. Toepler and Sprengel mercury
pumps
3. Rotary mercury ptm:ips
4. Rotary oil ptimps
5. Gaede "Molectilar" pump
II. Mercury Vapor Pumps
1. Gaede "diffusion" pump
2. Langmuir condensation pump
III. Physical-Chemical Methods
1 . Charcoal or other absorbing agent
at low temperature
2. Clean-up of residual gases by
chemical reactions
3. Clean-up of gases by ionisation
methods
^ In writing this section, the author has made extensive use of
the article by Dr. Langmuir on "The Condensation Pump," in
General Electric Review. Dec, 1916, p. 1060.
General Theoretical Considerations Regarding Vac-
uum Pumps'
In comparing vacuum pumps it is nec-
essary to consider the following factors,
which are the main characteristics of a pump :
1. Exhaust Pressure. This is the pressure
against which the pump may be operated.
In general, the higher the degree of vacutim
desired on the "fine" or intake] side of the
pump, the smaller the exhaust pressure should
be. The low exhaust pressure is then
obtained by means of another (so-called
"rough") pump in series with the high vacuum
pump. Two or more rough pumps may be
used in series in order to obtain a sufficiently
low exhaust pressure for the fine pump.
2. Degree of Vacuum Attainable. "This is
the lower limit of pressure which may be at-
tained in a closed vessel connected to the
pump. With most types of pump the degree
of vacuum attainable depends to a large
extent on the exhaust pressure used. This is
usually due to leakage through the pump."
In the cases of the mercury vapor pumps, to
be described in the next installment, there is
theoretically no lower limit to the pressure
606 July. 1920
GENER.\L ELECTRIC REVIEW
Vol. XXIII. Xo.
which may be attained, while in the case of the
Gaede molecular pump the limiting pressure
bears a constant ratio to the exhaust pressure.
3. Speed of the Pump. The law for the rate
of decrease in pressure in a closed vessel
connected to a pvimp is quite similar to that
Fig. 3. Geryk Vacuum Pump. Power Drive
of chemical and physical reactions of the
first order. It may be stated as follows:
If po denote the lower limit of pressure at-
tainable with the pump, then the rate of de-
crease in pressure at any instant is propor-
tional to p-po. where p denotes the pressure
at that instant. That is,
-t = k(p-po)
(IS)
where ^ is a constant. Further considera-
tion shows that with a given pump the rate
of exhaust must var\- inversely as the vol-
imie (V) of the vessel to be exhausted. Thus
we can write
dp S , ,
(ISa)
where S is a constant for the given pumping
system, that is, pump and connecting tubing.
Integrating the last equation we obtain the
relation,
t \pi-poJ
where / is the interval of time required to
reduce the pressure in the volume \", from
pi to p2.
Gaede has defined 5 as the speed of the
pump, and it is ordinarily measured in cubic
cm. per second. It is readily seen that from
the above equation 5 may also be defined
as follows:
With a pump of speed 5. it is possible to
reduce, in each second, the pressure in a
volume S cm.' by 63.2 per cent of the maxi-
muin possible decrease in pressure.
It is necessary- to distinguish between the
speed as defined in this manner and the
actual speed of exhaustion, which we may
denote bv E. The latter is defined thus :
dp E
- -P
or
dt V
c- ^'i P^
C = — 1 M - -
t p«
(20)
It is only when ^0 = 0, that 5 and E are
identical, and remain constant during the
whole period of exhaust. In all other cases
the speed of exhaust gradually decreases
from the value 5 which it has at the beginning,
and as the pressure in the vessel approaches
the limiting pressure, po. E decreases rapidly
until it becomes zero when the pressure has
decreased to po-
The actual speed of exhaust depends not
only upon the design of the pump but also
upon the diameter and length of the con-
necting tubing between pump and vessel to
be exhausted. The pump and tubing
together really constitute a system which is
the equivalent of a pump of lower speed.
Mention has been made in a previous section
of the results of Knudsen's investigations on
the resistance to flow in tubes. According to
these results, the quantity of gas, Q. flowing
through a narrow tube is given by the relation
Q=tr^- (12) and (15)
where U -xp, is the "resistance," and p« — p\
is the difference in pressure at the ends.
Let us now assume that the volume of the
tube is negligible compared to the volume of
the vessel to be exhausted, and that the
limiting pressure for the pump, po = o. Let
p2 denote the pressure in the vessel and />i
the pressure at the jiump intake (end of the
tube). Also let 5i denote the speed of the
pimip itself, and 5; the speed of pump and
connecting tubing. Then since the quantity
of gas taken out each second by the ])ump
is the same as that flowing through the tul>e.
we have the following relations:
n V pi
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
607
Eliminating pi and p-> from these equations,
we obtain the equation,
i=J-+TlV;^ (21)
which shows the effect of the added resistance
of the tube on the speed of the pumping system.
It will be observed that -^ has the same di-
mensions as U'-y/^, that is, the speed of a
pump may also be looked upon as the re-
ciprocal of a resistance to flow of gases through
it, and by analogy with electrical usage we
mav define -^ as the "impedance" of the
pump itself and -^ as the impedance of
pump and tubing. Similarly we may re-
gard Si, Si and 1 /W\/'^,, as "admittances. "
It follows logically from these considera-
tions that "in operating vacuum pumps of high
speed it is essential to use tulaing of large
diameter (and short length) between the
pump and the vessel to be exhausted if
full advantage is to be taken of the speed of
the pump." As an illustration of the effect
of narrow tubes in diminishing the effective
speed of a pump, let us con,sider the case of
a tube 10 cm. long and 1 cm. diameter con-
nected with a pump of speed 5i = 1400 cm.'
per second (which is the value for a molecular
pump under ordinary operating conditions).
The "resistance" of such a tube has been
calculated in a previous section.^ For air at
room temperature 1 /W^/yi = 1070. Applying
equation (21) it follows that 52 = 606, that is,
the speed of the pumping system is about 43
per cent of that of the pump alone.
With a ptimp which has a speed of 4000
cm.^ per second (such speeds are easily
attainable with mercury vapor ])umps) the
same piece of tubing would diminish the
actual speed of exhaust to 844 cm.^ per
second. In order to make effective use of
the speed of this pump, it would be necessary
to use very much larger tubing. Thus, let
us assume that the connecting tube has a
diameter of 3 cm. and a length of 30 cm. (To
use tubing larger than this is usually im-
])racticable, while the length given is about
as short as would be practical.)
I'Fv'7^ = 10-'X1.04
1 /Si =10-^X2.5
Hence, S2 = 2S25 cm.' per second.
These results indicate how seriously the
speed of a mercury vapor pump may be
'■See Qi, Table V, Part I.
limited by the resistance of the tubing un-
less this is of very great size. It also follows
from the above considerations that in the case
of a low speed pump such as the Gaede dif-
fusion pump (5 = SO) or a rotary oil pump
(5=100), the resistance of the tubing, as
Fig. 4. Details of Construction of Geryk Vacuum Pump
long as it is not too large, is not nearly as
important a factor as in the case of high speed
pumps.
MECHANICAL PUMPS
The early forms of exhaust pumps were of
the piston type. As they have been largely
superseded in modern practice, especially for
high vacuum work, no detailed mention of
them need be made in this connection. More-
608 July, 192()
GENERAL ELECTRIC REVIEW
Vol. XX III, No. 7
over, they are described in most elementary
text-books on physics.
Geryk Vacuum Pump
This is a modern form of the piston type
of pump (see Fig. 3), made by the Pulsometer
Fig- 5. Gaede Piston Pump
Engineering Co. The illustration (Fig. 4) and
the following description arc given by E. II.
Barton ;^
"Referring to this figure, A is the suction
pipe, B the air port into the cylinder above
the piston, C is the piston whose bucket
leather is kept up to the cylinder wall by
oil pressing in the annular space D, E is
the piston valve, F an air pi])e to relie\-e
the piston on the first few strokes, G, H and
I collars and cover forming a good joint and
delivery valve combined.
"When the piston is at the bottom of its
stroke as shown, there is a perfectly free
opening from A to B. As the piston rises the
port B is cut off and the cylinder full of air
3 An Introduction to the Mechanics of Fluids, p. 197 (Long-
mans. Green & Co., 1915).
See also Encycl. Britannica. 11th Edition. Vol. 22. p. 646.
* W. Gaede. Phys. Zeits. 14. 1238. 1913.
See also E. H. Barton, loc. cit. pp. 108-9. from whose books the
following description is quoted.
irresistibly carried up to outlet valve G.
No air can get back past the piston as it is
covered with oil. When the piston approaches
the top of its stroke, it lifts the valve G off its
face and gives a free outlet for the air. The
oil on the piston then mingles with that shown
above G, but the right quantity returns with
the piston on the closing of G. L is the plug
for filling up with oil, which is very non-
volatile, moistureless and non-solvent of air
and fills all clearance spaces and seals the
valves. "
With a single-cylinder pump of this type
it is claimed that a pressure of about a
quarter of a millimeter of mercury can readily
be obtained.
Gaede's Piston Pump*
This form, shown in Fig. 5, consists really
of three piston pumps in series. The vessel
to be exhausted is connected at R. As the
piston rod D moves upwards, it carries with
it the three pistons A, B, and C. The air
is thus forced from .V (which communicates
with the tube M) through the valves 0 in the
stationary' partitions c, b, and a into the
chamber K from which it is ejected into the
air by the vent q. The top chamber /\' also
contains a small amount of oil which forms
an emulsion with the water and other vapors
condensed above the piston A. "This emulsion
is forced, together with the air, through the
valve o in the cover a, through the tubePabove
the valve, and thence into the chamber A'.
This chamber is filled with a fibrous mass by
which the oil and water emulsion is separated
into its components. In consequence of its
greater density, the water collects on the bot-
tom M of the chamber, and may be pumped
off as often as necessary by means of a glass
syringe and rubber tube connected to the
tube A' extending upwards out of the pump.
The oil overflows through the tube 5 into
the space between a and M, whence it re-
enters the pump barrel to combine with
fresh quantities of water vapor."
According to Gaede's jjublished account
it is possible with this pump to obtain a pres-
sure as low as O.OOOOo mm. mercury; i. c. 0.067
bar., when exhausting into atmospheric pres-
sure.
Sprengel Pump
The use of a water-jet as a suction pump is
quite familiar. With this pump, the minimum
pressure obtainable is that corresponding
to the vapor pressure of water at the tem-
perature which it has in the supiily line; i. e.
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
609
from 5 to 10 mm. mercury. As the vapor
pressure of mercury at ordinary tempera-
tures is only about 1 to 2 bars it is possible
by means of a stream of mercury to obtain
fairly low pressures, and by interposing a
refrigerating chaniber between the vessel
to be exhausted and the nozzle which com-
municates with the mercury stream it is
possible to obtain still lower pressures. The
Sprengcl mcrcur}- pump operates on this
principle, and some of the simpler forms are
described in most elementary text-books.'
G. W. A. Kahlbaum" has described a form
of Sprengel pump which he states to be ca-
pable of exhausting a 400 cubic cm. bulb in
30 minutes to 0.004 bar. In a subsequent
paper' he gives the following data with
regard to the speed of exhaust of a 500 cubic
cm. bulb :
3 minutes to 0.5 mm. mercury
15 minutes to 0.000165 mm.
30 minutes to 0.000069 mm. = .092 bar
With special care he states that he was
able to get a pressure as low as .0024 bar.
This pressure is evidently that of residual
gas, and does not include the pressure of
the mercury vapor itself, which, as stated
above, would be between 1 and 2 bars.
Geissler-Toepler Pump'
The principle of this pump is fundamen-
tally the same as that used by Torricelli
in his famous experiment. In this type
(Fig. 6) mercury forces the piston and also
opens and closes certain ports, so that no
valves are needed except one rough glass
valve (g) to prevent the mercury from en-
tering the vessel, E, which is being exhausted.
The essential parts of the piunp are made of
glass and the air from E is exhausted by
alternately raising and lowering the mer-
cury reservoir R which is connected to the
tube of barometric length below B. At
each upward "stroke," the gas in B is
closed from E and forced through the tube
F, into the atm.osphere at M. Then on the
downward stroke, the pressure in E is lowered
by expansion of the gas into B. E. Bessel-
Hagen^ has described a m.odified form of
* See Encycl. Brit. loc. cit., aho Winkelmann. Handbuch der
Physik. I. 2. pp. 1314-1-332, contains a very detailed description
of the different forms of Sprengel and Toepler mercury pumps.
« Wied. Ann. oS. 199 (1894).
: See also L. Zehnder. Am. d. Phip. 10. 623 (1903), for a de-
scription of an improved form of Kahlbaum's pump.
" See Encycl. Britannica and" Winkelmann, loc. cit.. also Barton
loc. c;t., from whose books Fig. 6 is taken.
' Wied-Am. 11. 425, 18S1.
'" Other forms of Toepler pump are described by a A. Stock
Ber. deutsch. chem. Gis. 3S. 2182, 1905. and E. Grimsehl. Phys.
Zeits. S, 762. 1907.
Toepler pump with which he claims to have
obtained pressures of residual gas as low as
0.016 bar.i"
Both the Sprengel and Toepler pumps
have rendered \-ery useful service in high
vacutun investigations, and there is no doubt
-^B—TOe^d~£^ /^UMf^ •
Fig. 6. Toepler Pump
that with care it is possible to obtain pres-
sures as low as .02 to .01 bar by their use.
The great disadvantages of these pumps
are, however, two-fold. First, they re-
quire constant personal attention during the
exhaust and second, the speed of exhaust
is extremely slow, as it depends upon the
rate at which the mercury, can be raised and
lowered alternately. It is of interest to note
in this connection the results obtained bv
610 July. 1920
GENER.\L ELECTRIC REVIEW
Vol. XXIII. Xo. 7
Scheel and Heuse'^ in their investigation
of the degree of vacuum attainable with
diflferent types of pump. Theii' used a 6
liter bulb and measured the speed of exhaust
by means of a v^ery sensitive McLeod gage.
(See subsequent section for description of
Fig. 7. Gaede Rotary Mercury Pump
this gage.) In the experiments with a Toepler
pump, each stroke actually required two
minutes, and two more minutes were allowed
between each stroke for equalization of pres-
sure. Table VII gives the pressures at the
end of different intervals of time.
The last cokmin gives the "speed of ex-
haust" as calculated by Gaede 's formula.
Compared with the speed of even 1 00 cm.' sec.
obtained by a Gaede rotar\' mercur\' or
an ordinary oil pump (described below)
the speeds given in Table I, are mani-
festly ven,- low. Considering, furthermore,
that in the case where gas is continually
evolved from the walls, the minimum at-
tainable pressure is given by the ratio S /q
where q denotes the rate of gas evolution,
it is seen that in actual practice it would be
very difficult to obtain pressures below .01
bar by means of a Toepler pump.
Similar results were obtained by Scheel
and Heuse in investigating the rate of ex-
haust of a 6 liter bulb by means of a Sprengel
pump (Zehnder's form).'-
" Zeits/. Instrumentenkundc. S9, 47, 1909.
"Ann. d. Phys. 10, 623 (1903).
" Phys. Zeits. 10, 316, 1909. '
>' Verh. d. deutsch. Physik. Ges. 7, 287, (1905). Phv
758-760, (190,5).
TABLE VII
t
press
„ 2.3 V , p,
^-60t '"^p.
(minutes)
(mm. Hg.)
0
0.0645
2
0.0399 ■
24
0.0254 ,
48
0.0107
60
0.00703
108
0.00141
120
0.000.«
180
0.00024
192
0.00015
240
0.000053
252
0.000038
264
0.000032
300
0.000025
\ J
0.40
0.38
0.35
0.35
0.39
0.28
0.06
Gaede Rotary Mercury Pump
An automatic form of Toepler pump has
been described by U. von Reden," with which
he claims to have exhausted a 500 cm.' bulb
in 13 minutes to a pressure of .00001 mm.
From his data, the speed of exhaust is
found to be about 20 cm.' sec.
In lilO.j, W. Gaede designed a rotar>- mer-
cur\- pump which has been used to a ver>-
large extent in the commercial exhaust of
incandescent lamps and Roentgen tubes until
quite recently. The pump as described in the
first publication'* and illustrated in Fig. 7,
consists of an iron casing (with glass front)
partially filled with mcrcur\-, in which a
porcelain drum is made to rotate. A rough
pump producing a vacuum of 10 to 20 mm.
g
r
g
t
f.
'^l
t
s
R
J2
•J
i I
'■ t
"La
g
t
g
- . 1
■
. Zeits. '
Fig. 8. Gacd; Rotary Mercury Pump.
Vertical Section
is used as fore-piuiip. Fig. S shows a vertical
section of the pum]), and Fig. il a front \new.
The iron case is shown at j;. and (i is a heavy
glass plate through whicii pass the tubes R and
r which connect to the vessel to be exhausted
and the fore-pump respectively. The porcelain
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
611
drum / is built up of two (or more) sections as
shown in Fig. 9 and rotates on the axis a. As
the dnmi rotates in the direction of the arrow
the compartment W is at first increased
in volume and thus sucks in the gas at the
opening /, from the vessel to be exhausted.
During the second part of the revolution,
the opening / becomes covered with mercury,
as shown at p, and the gas is then forced
out under pressure from the compartment Wi
into the space between the walls Zi and i]
and into the rough pump connection at P.
Fig. 10 shows an improved form of the
pump in which the opening to the rough piunp
r is brought in through the iron casing. The
vessel to be exhausted is connected at E,
and a side tube t is provided with Po O5 to
take up water-vapor. The tube MOF acts
as manometer and also makes it possible
to exhaust with the rough pump alone at the
beginning. As the vacuum improves, the
mercury in o rises and seals off the connection
to the rough pump through a Si. The system
is then ready for exhausting to lower pressure
by means of the mercury ]3ump.
Gaede gives as an illustration of the opera-
tion of the pump the following data. With a
pump rotating at 20 r.p. m., and a volume of
6,250 cm.^, the pressures during exhaust
were as follows :
Fig. 9.
Gaede Rotary Mercury Pump,
Diagrammatic View
.„ 2.3 V , Pi
t (min.)
P (mm.)
^-60t '°«p.
0
5
0.03
94.5
10
0.0018
46.5
15
0.00023
34.0
20
0.0001
13.8
25
0.00007
(3.9
30
0.00007
0.0
The speed of this pump is therefore ap-
proximately 100 cm.^/sec. at the maximum,
while the degree of vacuum attainable is
about .00007 mm. or 0.1 bar."
Rotary Oil Pump"'
Figs. 11 and 12 show the construction of a
pump of this type designed by Gaede pri-
'= Later improvements in this pump have been described in
Verh. d. deustch. Physik. Ges. 9, 6.39 (1907). and Phys. Zeits .S.
852, (1907).
'« G. Meyer. Verh. d. deutsch Phys. Ges. 10, 7.53 (1907).
Fig. 10. Improved Form Gaede Mercury Pump
marily for the purpose of acting as a fore-
pump to the rotary mercury pump described
above. It is also shown in Fig. 15 at the
right hand side. The pump consists of a
steel cylinder A which rotates eccentrically
inside a steel casing. The projections at
S are held tightly against the inner wall by
means of springs, so that as the cylinder
rotates the air is sucked in at C and forced
out through the valve D into the oil chamber
O and from there into the -atmosphere at J .
The oil serves as automatic lubricant and
also helps to prevent air from leaking
back into the fine pump side, by forming a
film between the rotating and stationary
members.
Fig. 13 illustrates a standard form of
rotary oil pump used in incandescent lamp
factories and which can also be used as a fore-
pump to higher \-acuum piimjis such as
Gaede's Molecular or Langmuir's Condensa-
tion pump. With a rough side pressure of
about 1 cm. mercury, such a pump is capable
of exhausting to a pressure of approximately
612 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No.
1 bar, and with two pumps in series the fine
side pressure may be lowered to 0.1 bar.''
With a pump of this type operating at
about 400 r.p.m., the speed of exhaust is
100-150 cm.' per second.
Fig. 11. Gaede Rotary Oil Pump, Side View
Gaede Molecular Pump
The Gaede Molecular pump undoubtedly
marks a distinct advance in the design of
pumps for the production of high vacuum.
The diiTerence between his pump and the
types previously constructed has been well
described by Gaede himself in the paper which
he published in 1913:'*
"All high vacuum pumps known up to the
present consist of an exhaust arrangement
which, according to the original idea of Otto
von Guericke, separates a definite volume of
gas from the vessel to be exhausted, and then
gives it up to a fore-vacuum or the atmos-
phere. It is absolutely essential in these
pumps to separate the rough side from the
higher vacuum side as much as possible.
This is accompHshed in the mechanical pumps
by tight-fitting pistons and valves, and in the
case of mercur}- and oil pumps by means of
the liquids themselves. On the other hand, in
" K. T. Fischer, Verb, deutsch. Physik. Ges.. 7, 383 (1905), has
described a form of rotary oi! pump for use in commercial ex-
haust operations. With two pumps in series exhausting into
atmospheric pressure he states that a pressure of about 2 bars
may be obtained.
'» W. Gaede, The Molecular Air Pump. Ann. d. Phys.. 41. 337-
380 (1913). This paper contains a complete discussion of the
theory and construction of the pump. Briefer descriptions may
also be found in the following:
W. Gaede, Physikal, Zcits. IS. 864-870 (1912). and Verh. d.
deuts. Phys. Ges., 14. 775-787 (1912).
K. Goes. Physikal, Zcits.. IS. 1105, and li. 170-2 (1913).
Description of some experiments with the pump and precautions
infusing it.
Electrician (London). 70, 48-50 (1912).
K. Jellinek, Lehrbuch d. Physikal, Chemic. I. 1. pp. 330-333
(1914).
M. L. Dunoyer, Les Idees Modemes sur la Constitution de
la Matiere, pp. 215-271 (1913).
the case of the molecular piomp there is no
separation, whether piston or fluid, between the
high-vacuimi and f ore- vacuum. " The gas is
dragged along from the vessel to be exhausted
into the fore-vacuum by means of a cylinder
rotating with high velocity inside a hermeti-
cally sealed casing. The pump thus repre-
sents a logical development and application
of the laws of flow of gases at ver%- low pres-
sures as investigated by Knudsen, Smol-
uchowski, and Gaede himself.
The fundamental principle of the pump
may be illustrated by means of Fig. 14. The
cylinder A rotates on an axis a (in the
direction of the arrow) inside the air-tight
shell B and drags the gas from the opening n
towards the opening nt, so that a pressure-
difference is built up in the manometer M, as
shown by the mercur\--le\-els at o and p.
Between m and n there is a slot in the case B
as shown in the diagram, while at ever\' other
point A and B are ver}- close together. Now,
at ordinar\' pressures the viscosity is inde-
pendent of the pressure. Lender these con-
ditions, as Gaede shows, the difference in
pressure at o and p depends only on the speed
of rotation u, of the cylinder, the co-efficient
of viscosity of the gas, n the length of the
slot L and It the depth measured radially,
according to the following relation:
pi-pi = 6Lunlh-
At low pressures, however, the number of
collisions between gas molecules becomes
Fig U Gaede Rotary Oil Pump, Front View
relatively small as compared with the number
of collisions between the gas molecules and
the walls. Under these conditions the mole-
cules therefore tend to take up the same
direction of motion as the surface against
which thev strike, if the latter is in motion.
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
613
This conclusion is based upon the investiga-
tions of Knudsen on the laws of molecular
flow, which have been discussed in a previous
section. The relation deduced above is there-
fore found to be no longer true and instead
of the pressure-diflference remaining constant
at constant speed of rotation, the pressure-
ratio is now constant and independent of the
pressure in the fore-vacuimi. Gaede shows
that at very low pressures,
e
where A.' is a constant whose value depends
upon the nature of the gas and the dimensions
of the slot in the casing B of the pump, so
that at constant speed of rotation, n. the ratio
between the pressures on the two sides of the
ptimp is constant.
The construction of the actual pump
based on the above principles is illustrated
in Figures 15 and 16, while Fig. 17 shows the
piunp connected in series with a Gaede rotary
oil pump.
The rotating cylinder A (Figs. 15 and 16)
has 12 parallel slots aroimd the circiunference,
into which project the extensions C from
the outer casing. If A rotates clockwise,
Fig. 14. Diagram Explaining Operation of Molecular Pump
and the depth of the slots vary from 0.15 cm.
in the outer section to 0.6 cm. in the inner
ones. With the cylinder rotating clockwise
as indicated, the vessel to be exhausted is
Fig. 13. Standard Form of
Rotary Oil Pump
Fig. 15. Gaede Molecular Pump,
Front View
Fig. 16.
Gaede Molecular Pump,
Side View
the pressure at m is greater than that at n,
and in order to increase this pressure dif-
ferent sections are connected in series. The
distance between the outer edge of the cylin-
der A and the inside of the shell B is about
0.01 cm. The over-all radius of A is 5 cm.,
connected at 5, while the opening T is con-
nected to an ordinary mercury or oil pump
capable of exhausting to a pressure of less
than 0.05 mm. Hg. As the speed of rotation
of the cylinder is very high (about 8000 r.p.m.)
oil cups are provided at F, and the shaft A"' is
614 July, 1020
GEXEIL\L ELECTRIC REVIEW
Vol. XXIII, No. 7
so designed that the oil in the spiral slot is
driven outwards by the centrifugal action.
The slots in the rotor are so arranged that
the lowest pressure is in the center, and the
pressure increases uniformly outwards until
TABLE VIII
1500 ■
ec
UOO
-
^ ^
1500
-
/^
^s.
1200
-
/
\^
1100
-
A
/
\
1000
k
/
\
900
-
/
V
000
- /
/
Y^
700
-/
\ ^,
eoo
/
500
-
100
;oa
-
I \
200
-
B
1 \
100
-
1
' 1
~ r— .^
Speed of Rotation
Rough-pump Press.
Press, on Fine Side
R.P.M.
mm. Hg.
mm. Hg.
12000
0.05
0.0000003
12000
1
0.000005
12000
10
0.00003
12000
20
0.0003
6000
0.05
0.00002
2500
0.05
0.0003
8200
0.1
Not measurable
8200
1
0.00002
8200
10
0.0005
6200
0.1
0.00001
6200
1.0
0.00005
4000
1.1
0.00003
4000
1
0.0003
10 10 mn
Fig. 13. Effect of Rough-pump Pressure on
Speed of Gaede Molecular Pump
the ends, where it is equal to that produced
by the rough pvimp.
The effect of var\-ing the speed of rotation
or the rough pump pressure on the degree of
vacuum produced by the molecular pump, is
shown in Table VIII.
The pressures on the fine side were meas-
ured with an extremely sensitive type of
McLeod gage except in the case of the first
result given in the table which was estimated.
The writer's own experiments'' with the Gaede
" S. Dushman, Phys. Rev. .5. 224 (1915).
molecular pump at SOUO r.p.m. have shown
that with a rough pump pressure of 20 mm.
the fine side pressure was O.00U4 mm., so that
the ratio of the pressures was 50,000 — a
result which is in accord with figures given
by Gaede above.
The speed of the pimip as defined by the
relation,
5 = 1^/.^
i pi
has been found by Gaede to var>- with the
magnitude of the rough-pump pressure. The
cun-e A in Fig. IS shows that the m.aximum
speed is about 1400 cm.' per second with a
fore-vacuum of 0.01 mm. For comparison
Gaede also shows the curve 6 for his rotar>-
mercun,- pump, which has a speed of about
\'.H) cm.' i)er sec. at the maximum.
{To bf Continued)
Fig. 17. Assembly of Gaede Molecular and Gaede Rotary Oil Pumps
615
Two Years' Service of Battleship New Mexico
The battleship New Mexico, pride of the
United States Navy, and first of Uncle
Sam's fighting fleet equipped with the electric
drive, has recently completed her second year
of active service.
Commander £5. M. Robinson, fleet en-
gineer of the Pacific fleet, of which the New
Mexico is flagship, reviews this two years of
electric propulsion in the following report :
"The New Mexico has been operating for
nearly a year in company with two sister
ships, the Idaho and Mississippi, which
have hulls identical with that of the New
Mexico. During this time it has been pos-
sible to get an accurate comparison of the
relative economy of the three ships and also
the relative maneuvering qualities. In the
latter respect, the New Mexico is decidedly
superior, and the remarkable part of it is
that nearly all of the maneuvering in re-
stricted waters has been done with our turbo-
generator. When this installation was first
proposed, its opponents maintained that,
while a ship like the Jupiter could be satis-
factorily operated with the screws on both
sides of the ship running at exactly the same
speed, it would not be possible to get satisfac-
tory operation with that arrangement on a
ship which had to operate in formation.
But exactly the reverse has proved to be true ;
it has been found that more satisfactory
operation is obtained when using one gen-
erator than when using two, and it is cus-
tomary, when in dangerous waters where
it is desired to take all possible precau-
tions, to use one generator for driving the
ship and to keep the other turning over idle.
If the ship is getting under way from an
anchorage and has to turn, as soon as the
anchor is away the signal is given for standard
speed ahead on one side and the same speed
astern on the other; with this arrangement
the ship will turn absolutely on her wheel
without gaining ground either ahead or
astern.
"The advocates of electric propulsion have
always claimed that it was very superior to
all other forms of propulsion at the cruising
speeds, but even the most enthusiastic of
these have been surprised by the remarkable
showing made. This is doubtless due to
the fact that no one made sufficient allow-
ance for the saving due to sluitting down one
generator and all the auxiliaries that go with
one of the condensing plants. At a speed of
10 knots the New Mexico uses about 1().7 per
cent less oil than her sister ships, or, putting
it another way, her sister ships use about 20
per cent more than the New Mexico; at 13
knots the figures are 29.9 per cent, or 42.7
per cent; at IG knots the figures are 32.3 per
cent, or 47.8 per cent; at 19 knots the figures
are 28.6 per cent, or 40.1 per cent; at full
power the figures are 24.4 per cent, or 32.2
per cent. At 19 knots, also at full power, the
New Mexico uses about .975 lbs. of oil per
shaft horse power, and at 15 knots she only
uses 1.1 lb. of oil per shaft horse power per
hour. This is a remarkable uniform economy.
"In regard to the reliability of the machin-
ery, the New Mexico has had nothing but the
most minor troubles with her electric plant
and there have been no navy yard repairs
whatever."
616 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 7
Electric Power in the Oil Fields as a
Central Station Load
By W. G. Taylor
Power and Mining Engineering Department, General Electric Compasy
In our Mav 1919 issue, Mr. Tavlor thoroughly explained to the oil producer the advantages of motor
drive In the present 'article, which was delivered as a paper at the N. E. L. A. Convention, Pasadena Calif
Mav '^O-'" the author discusses the subject from the standpomt of the central station. He gives detailed
informatron as to the nature of oil field load, the motor equipment for various local conditions, the cost
of installation and the power consumption. In comparing electric drive with gas-engine drive, tables are
included showing the time saved bv the former, the increased production, and the lower operating expense.
Full consideration is given to the electrification of gathering and line pumos. vacuum pumps, compressors
and circulating pumps for casing-head gasolene plants, dehydrators, machine sho.«, and lighting.— Editor.
Oil companies have in general reached a
very receptive mood toward electric drive
for oil field operations. This is particularly
the case in the California, Mid-continent, and
Texas fields, comprising between 70 and ib
per cent of the productive wells of the United
States and producing over SO per cent of the
crude oil in this country and more than half
in the world. Not only are several thousand
wells now being pumped in these fields by
electric power, but motor drive has also been
successfully used for several years in Cali-
fornia for drilling, and has been recently
introduced in the Mid-continent field for this
work with notably good results.
The other principal applications of electric
power in the oil fields are to :
Water pumps.
Gathering and line pumps,
Vacuum pumps,
8PH -n r.-» 5r« 4r>i ir.
Fig. 1. Twenty-four-hour Load Curve of Substation of Midland Counties Public
spa
hpa If* %r» *«« •>•» ■"«
Fig. 2. Twenty four hour Load Curve of Substation'of San Joaquin Light and
ELECTRIC POWER I\ THE OIL FIELDS AS A CENTRAL STATION LOAD 617
Compressors fo r casing-head gasolene plants,
Circulating pumps for casing-head gasolene
plants,
Dehydrators,
Machine shops,
Lighting.
Refining and other operations not directly
concerned with oil production are not covered
in this article.
NATURE OF OIL FIELD LOAD
The almost ideal nature of an oil field load is
well presented in Figs. 1 and 2 by actual curve-
drawing wattmeter records of substations
ser\-ing typical oil fields in California. The
load is practically constant twentv four hours
a day every day in the year, without anv ma-
terial seasonal variation. A slight rise in the
load curve at night is due to electric lighting.
At any individual installation of oil well
motors, there are of course relatively large
load fluctuations, as illustrated by Figs. 3, 4,
and 5, but the diversity factor of a large
ntmiber of installations prevents the peaks
from being felt at either the generating
station or substations.
Oil well pumping comprises the largest
percentage of an oil field load, and accordingly
is the chief factor in determining its character.
The quite varied operations necessary for the
maintenance of a producing well require a
very versatile motor, and these requirements
are fully met only by the two-speed oil-well
motor with both speeds variable, which has
become the standard machine for this work.
Although its chief duty is to pump the well,
the motor is frequently called upon for other
work, particularly when the well is cleaned
out and the rods and tubing must be pulled.
This hoisting work demands power, for short
periods, several times greater than that
required for pumping. Although the motor
develops this high power at its high speed
and is designed for pumping conditions at
its low speed, it is not practicable for it to
have as high a power-factor under these
circum.stances as can be expected of an
ordinary industrial motor, and therefore the
power-factor of the system is correspondingly
affected. An oil field load for this reason
generally has a power-factor, without cor-
rection, of about 60 to 65 per cent.
• • •
Service Corporation Serving a Portion of the Coalinga Oil Field in Califomi
*w »S>v ^V»^ ^v^ y^52^
W^*'^°^^^^^»^ly>>si^^^;;^
tvm ^Hi "^' '
••••••••
9*« S«» 7»l' 51, 5,1, ji.
Power Corporation Serving a Portion of the Midway Oil Field at Taft. Califomii
f>^»l'N "« Swi
618 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo.
The cost of correcting this power-factor in
the low-voltage distribution system would
be prohibitive on account of the scattered
nature of the load and the small capacity of
the units involved. The most economical
correction is therefore obtained by the instal-
lation of synchronous or condenser equip-
on the other hand, high voltage at the motor
is impracticable on account of the necessity
of handling line current in the controller for
reversing the motor.
Si.x large power companies furnish nearly
all of the electric power at present used in the
American oil fields, these being the Southern
-s
iV:
zT
T^
iiH
P3
4=
Mt
Fig. 3. Wattmeter Record of Diilling with Standard Tools and Bailing in a 10-in. Hole at a Depth of 2025 Ft.
r^
'fv
•^-
— ~"~-
f^
H \ —
%
(
g
i' 1 1
T
^ f
1 r
if ■ !
■5"
c "
I \
iF'
|F
S I
1
i'-
1 1
r
' s > 1 ' -J
^
Ma
1
i\
~\
__
-^
1 1
-11
\-
W-
— H~
\\^~
=fc.|lj
%H- Vx
Fig. 4. Wattmeter Record of Miscellaneous Work in Connection with Diilling a 10-in. Kole
with Standard Tools at a Depth of 1840 Ft
Fig. 5. Wattmeter Record of Swabbing in 4 ' .-in. Casing at a Depth of 2165 Ft.
ment in the priman,- circuit, after the load
has developed to such an extent that cor-
rection becomes advisable.
The standard distribution voltage is 440
volts. The cost of both the motor equipments
and the distribution copjjcr would be increased
by the use of any lower voltage, and there
would l^e no comjjcnsating advantages; while
California Edison Co. and the San Joaquin
Liglit and Power Corporation in California.
the Kansas Gas and Electric Co. in Kansas,
The Oklahoma Gas and Electric Co. and the
Oklahoma Power Co. in Oklahoma, and the
subsidiaries of the American Power and Light
Co. in Texas. The large increases in station
capacity and power line extensions, which arc
ELECTRIC POWER IN THE OIL FIELDS AS A CENTRAL STATION LOAD G19
now being undertaken by these companies to
reach and carry additional oil field load,
furnish good evidence of its desirability.
MOTOR EQUIPMENTS FOR
OIL WELL PUMPING
Piunping the well requires
continuous operation of the
motor usually for weeks at a
time without change or shut
down, with a low power de-
mand and at a comparatively
low speed of the rig. Speed
control is necessary to adjust
the number of strokes per
minute to the changing con-
ditions at each well. Pulling
rods and tubing, which is
necessary at intervals to clean
out the well or to replace
broken or worn parts, as well
as all of the other roustabout
work, is done at high speed
to save time, and demands the
characteristics of a high-torque hoist motor.
The two-speed slip-ring induction motor with
both speeds variable, which has proved to
be especially adaptable to all these opera-
tions, has all the flexibility of engine drive
and is an improvement over it in numerous
respects. Figs. 6 and 7 show typical instal-
lations.
f
Fig. 6.
Fig. 7.
Typical Installation in California of a Two speed 30/15-horse power Oil Well
Pumping Motor Equipment Before Completion of Housing
A number of wells of moderate depth are
equipped with 25/10-h.p. motors, but the
machine which meets the greatest variety
of conditions is rated 30/15 h.p. In Cali-
fornia, one of these is pumping a 4S00-foot
well. The higher rating is
developed at a synchronous
speed of 1200 r.p.m. for pull-
ing, bailing, and similar work ;
and the lower rating at 600
r.p.m. for pumping duty. By
means of a pole-changing
switch mounted on the frame
of the motor, the speed is
readily changed as desired for
the work to be done.
A drum controller and spe-
cially designed secondary re-
sistor give the required speed
variation at either high or
low speed. The controller is
installed near the motor and
is operated by a rope wheel
from the "headache post" at
the derrick.
The motor is fully protected
by an oil circuit breaker hav-
ing under-voltage release and
overload trip. The overload
trip coils are double-wound to
provide protection on both
pumping and pulling duty,
A Kansas Installation of a Two-speed 30/15-horse power Gil Well Pumping Motor thcSC COlls being electrically
Equipment, This Being Typical of Those in the Mid-continent Field interconnected with the polc-
G20 July, 19-2()
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 7
changing switch on the motor frame in
such a manner that the coils are always
connected to trip the circuit breaker at an
overload corresponding to the motor rating
in use. This makes it impossible for the
motor to be left without adequate protection,
particularly when running unattended on Ionv
speed for pumping.
An automatic device can be placed in the
base of the controller, if desired, to prevent the
controller from being moved past an intermedi-
ate point until after the current has dropped
below a predetermined value. An unskilled
operator can thus be prevented from abusing
the equipment, and can at the same time be
automatically taught to handle the controller
in the proper manner to get the best results.
Many of the installations in the Alid-
continent field are equipped with an ammeter
mounted on the cover of the oil circuit breaker.
This is of assistance when counterbalancing
the well and ser\'es as an indication jf the
condition of the well when pum_ping. The
ammeter is not in circuit when the motor
is connected for high speed. On most leases
a watthour meter is installed at each well,
and this enables the operation to be watched
more closely from month to month.
Required Motor Capacity for Pumping
There is no apparent way to calculate the
power required to pump a well which will
give figures at all consistent with actual
results, because of the difficulty of determining
the effect of varying well conditions. For
instance, a large amount cf sand in the oil
will increase the power necessary to pump
it; while on the other hand, gas may be
encountered which will help lift the oil. No
numerical value can be placed on these
conditions, so the motor capacity is deter-
mined largely by comparison with results
obtained at other wells; and a ])roper selection
depends largely upon the data at hand and the
judgment of the engineer or salesman.
Other conditions being equal, the power
required for pumping will vary directly as,
(a) The length of stroke.
(b) The number of strokes per minute.
(c) The square of the diameter of the
tubing.
The following figures may serve as a guide
in estimating the motor capacity for pumping
individual wells. Owing to changeable con-
ditions always encoimtered, it is best to have
some reserve capacity in the motor on the
pumping connection.
In California, the following data have been
obtained on 213 wells, but cannot be con-
sidered representative of conditions in deep
territory:
Depth of wells 900 to 3110 ft. (av.
1430)
Length of stroke 29 to 32 in. (2nd hole
in crank)
Strokes per minute 20 to 30 (av. 24)
Diameter of tubing 3 in.
Daily production per well ... 10 to 230 bbl. (av.
122 bbl.)
Power required per well. ... 1 to 5 h.p. (av. nearly
4.0 h.p.)
E xcept ional wells have
required more, some
as high as 17 h.p.
In Louisiana, some heavy pumping wells
have been encountered. One well in the
Jennings field required the following:
Depth of well 2000 ft. (approx.)
Depth of pumping 1100 ft.
Length of stroke 30.5 in. (3rd hole in
crank)
Strokes per minute 40
Diameter of tubing 2.5 in.
Daily production 600 to 800 bbl., 90
per cent water
Power required 9.5 h.p.
Another well in Louisiana in the field at
Hosston required the following:
Depth of well 1050 ft. (approx.)
Depth of pumping 1000 ft.
Diameter of tubing 3 in.
Daily production 500 bbl. (av.), over
90 per cent water.
Length of Counterbalance Average
Stroke Speed Used Horse Power
37.5 25 No 11
37.5 38 No 17.5
30.5 40 No 15
30.5 40 Yes 13
Compared with the California data, these
Louisiana wells have a longer stroke, higher
speed, larger percentage of water, less gas, and
therefore require higher horse power. The
Hosston well, compared with the Jennings
well, has a lower speed but less gas and larger
tubing, therefore requires a somewhat higher
horse power.
From 10 other wells in the Jennings field in
Louisiana, the following data were obtained:
Depth of wells 2050 ft. (approx.)
Depth of pumping 1100 ft.
Length of stroke 31 in. (except in a
few cases)
Strokes per minute 1<S to 44 (av. 27.5)
Diameter of tubing 2.5 in., 3 in. and 3.75
in.
Daily production per well . . . GOO bbl. (av.), 90 per
cent water
Power required per well. . . . 5.2 to 8.0 h.p. (av.
C.4 h.p.)
ELECTRIC POWER IX THE OIL FIELDS AS A CENTRAL STATION LOAD G21
In the Louisiana field near Hosston, data Capacity of Motor for Pulling
on 11 other wells were obtained as follows: As a guide for determining the maximum
depth of well at which a motor of a gi\'en rat-
E^^lh o[ putping-. •.;:::::: lsfti\ooo ft. ing can safely be installed for pulling work the
Length of stroke 30 to 36 in. (3rd hole followmg formula IS of much service. It is
in crank) based on the maximum torque of the motor,
Strokes per minute 28 to 38 (av. 34) j^^^ jjj^g ^ggn found sufficiently Conservative
Ein""prod°ucHo'n"^er-weir.:; Lo' bbl. (av.), 90 that the motor heating will normally not be
' per cent water excessive under the usual operating con-
Power required per well ... . 7..") to 15 h.p. (av. ditions.
lOh.p.)
., • , u . n RXEXLXK
T T -I 11 • .Hi o /-I Maxim_imi depth of well = -;r;r5
In Texas, /4 wells m the Goose Creek wXa
field furnished the following figures: .^ ^^,^.^^^ /? = ratio of motor speed to corre-
Depth of wells 2800 ft. (av.) 3400 sponding bull-wheel speed.
Length of stroke 29 to 32 in'. (2nd hole E = mechanical efficiency of the rig (usually
in crank) varies from (J. .5 to 0. < )•
Dia°meter^of"tub'?ng . . . . . . . . 2^5 in. L = number of load lines used in the tackle
Daily production per well . . 100 bbl. (av.) max. for pulling the tubing.
water 10 per cent . , r i ■ • n r^
Power required per well (mo- W = weight of tubing m lb. per ft.
toi' i"P"'') .5.1 to 5.5 kw. c/ = diameter of bull-wheel shaft in inches.
These wells produce a considerable amount 7.; = ^ constant, depending upon the motor
of sand. used.
T T- ^u -If • The constant A' is determined as follows:
7« A awsas, the power required for pumping
five wells in the El Dorado field was as f ol- „ _ l'260Xh.p.XT
lows: r.p.m.
Depth of wells 2500 to 2940 ft. (av. in which, h.p. = hoTse power rating of motor
, r ^ , no"''^*^ ^'■■' on high speed.
Length of stroke 28 m. "= ^
Strokes per minute 19 to 22 (av. 21) r = max. torque of motor in per cent of
Diameter of tubmg. •.■■.• 3 in. full-load torque.
Daily production per well. . . 300 to 600 bbl. (av.
460 bbl.j r./'.»J.= full-load high speed of motor.
Power required per well. . . . 6.9 to 9.7 h.p. (av. .
7.9 h.p.) The extreme condition which may be
^, -.^ , . . • 1 < ^ encountered is pulling rods and tubing
These Kansas records were obtained about together with the tubing full of oil. This mav
three years ago on comparatively new wells. ^^^ ^^j.^^ -^^^^ account bv detemiining the
Records were checked a year ago on another ^^^^j ^^-^^ j-^^^ ^f ^-^-^ 1^^^ ^^^ ^sjng
group of about bO wells, and the average ^j^j^ ^^ ^^^ .^. -^ ^^^ formula.
per well was o h.p., the maximum being <S h.p.
Similar operating conditions prevailed but the p^^^^ Consumption for Pumping Operations
production had materially declined. ^j^^ kilowatt-hour consumption for deep
T ^, , , „, , , . „ . wells is no more in m.anv cases than that for
/« Oklahoma, at Shamrock, the following shallow wells. The most influential factors
pumping data were obtained from seven are the length of stroke, the speed of pumping,
^'^"^- the diam„eter of the pump barrel, the gravity
Depth of wells 2660 to 2920 ft. (av. of the oil, and the fluid level in the wells.
1 2810 ft.) This is indicated in Table I.
Length of stroke 29^in.^ (2nd hole in 'pj^g ^^^^ ^^ r^^y^^^ j ^^ ^^^ represent all or
Strokes per minute 1870^28 (av. 23) even average conditions in California, and it
Diameter of tubing 2 in. would be a difficult matter to determine
Daily production per well. . . 60 to 100 bbl. (av. them. It is of interest to note, however, that
p . , ■ 11 ^"^ ^^^'^ there are a large number of wells in both
"("^otormpuT) ...''.".. ''.'^ . 4.1 to 6.4 kw. (av. the Midway-Sunset and the Coalinga fields
5.2 kw. with a depth from 10(10 to 2500 feet which
622 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 7
TABLE I
COMPARATIVE POWER CONSUMPTION OF SELECTED GROUPS OF ELECTRICALLY
OPERATED OIL WELLS IN CALIFORNIA
Number of wells
Location (California)
Average depth (feet)
Pumping Speed (strokes per min.)
Hole in Crank — used for pumping
Gravity of oil (Deg. Baum6)
Size of tubing (inches)
Water
Gas
Number of months
Kw-hr. per well per month
Case 1
Case 2
Cases
Case4
Case5
50
5
6
4
6
Kern
Casmalia
Casmalia
Cat Canyon
Santa Maria
1000 to 1100
1800
2000
2900
2800 to 3000
27
18
20
22
20
1st, 2nd & .
1st
1st
1st
1st
3rd
12.5 to 14.9
10 to 11
12
15
19 to 21
■iH
234 and 3
2ii
Much
Much
\'er>' Little
Some
Yes
4
3
3
3
3
5252
4403
3927
4229
3941
use between 2000 and 3300 kw-hr. per well
per month. An average obtained for 366 of
these wells was 2600 kw-hr.
In Kansas there is not such a wide variety of
conditions. Aside from a number of shallow
wells pumped by "powers," the following are
about the average operating conditions:
Depth of wells 2400 to 2950 ft.
Pumping speed 15 to 20 strokes per
min.
Hole in crank used for
pumping 2nd
Gravity of oil 38 deg. B.
Size of tubing 3 in.
Some water
No gas to lift the oil.
In March, 191S, the records of 82 wells
showed a monthly power consumption per
well of 3030 kw-hr. For the following month,
the average for 96 wells was 3340 kw-hr.
In October, 1918, 26 wells, not included in
those just mentioned, required an average of
2430 kw-hr., but seven of these pumped only
part time. The average for the 24-hour wells
was 3110 kw-hr. A later record of 79 wells,
some of which pumped only part time, gave
an average of 2220 kw-hr. per well per month.
In the Burkburnett field in Texas, con-
ditions are about as follows:
Depth of wells 1650 to 1800 ft.
Pumping speed 15 strokes per min.
Length of stroke 20 in.
Gravity of oil 38 deg. B.
Size of tubing 2 in.
No water or sand.
In the Townsite pool in this field, wells are
pumping on the beam for the ftdl 24 hours a
day and require approximately 3000 kw-hr.
per well per month.
In the Northwest pool, with practically the
same field conditions, most of the wells are
still being swabbed from 3 to 10 hours a day,
with a power consumption an^-where from
2.500 to 9500 kw-hr. a month.
The Goose Creek field in Texas is a deep
pool from which the following figures have
been obtained :
Depth of wells 2800 (av.) 3400 ft.
(max.)
Pumping speed 28 strokes per min.
Hole in crank 2nd
Gravity of oil 20 deg. B.
Size of tubing 2 ' 2 in.
Considerable water and gas.
Monthly kw-hr. per well . . . 3500 to 4000
Motor for Driving "Powers"
The well-known method of pumping wells
in a group from a central "power," with
shackle rods extending from the "power" to
the jack at each well, is ver>' well adapted to
motor drive and a large number of such
applications have been made. No special
electrical features are necessary-, and constant-
speed duty is usually all that is required.
Such wells are pulled by portable hoisting
outfits, to which small hoist motors have
been applied in several instances.
The number of wells in a group and the
length of time each is jiumped daily var>- so
widely that no good general data can be
given. A motor load of about 2..") h.p. per
well and an average power consimiption from
30 to 45 kw-hr. per well per day is a rough
estimate of about what may be exi>ected.
ELECTRIC POWER IN THE OIL FIELDS AS A CENTRAL STATION LOAD 623
ADVANTAGES OF ELECTRIC POWER FOR
OIL WELL PUMPING OPERATIONS
For oil well pumping, motor drive has a
number of special advantages which have
already been discussed in detail,* but which
can well be summarized here:
Increased Production
Ftiel Saving
The oil fuel consumption for steam-engine
pumping of individual wells is from 3 to L5
barrels per well per day. This is saved by
electrification and thus in fact amounts to an
increase in net production. More gas is avail-
able for the market where motors replace gas
engines.
Decrease of Shut-downs
Elimination or reduction of many avoidable
shutdowns in oil well pumping operations can
be accomplished by using motors, and pro-
duction can thus be increased in many cases
as much as 15 per cent. Evidence of this is
given in Table II.
TABLE II
COMPARISON OF PUMPING TIME LOST
FROM SHUT-DOWNS WITH GAS ENGINE
AND ELECTRIC DRIVE UNDER SIMI-
LAR NORMAL OPERATING CONDI-
TIONS IN KANSAS, PUMPING
ON THE BEAM
GAS-ENGINE
ELECTRICDRIVE
DRIVE
EL DORADO
AUGUSTA FIELD
FIELD
Nov.,
Feb.,
Oct..
Nov.,
1917
1918
1918
1918
No. of wells
208
216
26
27
Percentage of available
pumping time lost, all
causes
2.3.3
28.2
10.7
9.8
Percentage of available
pumping time lost.
engine or electric
troubles only
4.8
8.15
1.98
0.63
Experience has demonstrated that the
value of motor drive in accomplishing these
results lies in the following points :
(a) Electric troubles do not cause over
two per cent loss in time due to
shut-downs.
*' 'The Operation of Oil Wells by Electric Power and the Result-
ing Gain to the Oil Producer, " by W. G. Taylor, General
Electric Review, vol. XXII. May, 1919, p. 384.
(b) There are no gas, water, or freezing
troubles with electric drive.
(c) The time lost from rod breakage is
usually cut in half when motors are
installed .
(d) Valve and cup troubles are reduced
several per cent with motor drive.
(el There are occasionally some other
troubles which electric operation
remedies to a considerable extent.
Time Saving
By reducing or eliminating many delays,
electric drive makes more pumping time
available and thus increases the production.
In this respect the following are included
among the advantages of a motor over a
gas engine or steam engine.
(a) No delay from steam lines full of
water after an idle half hour.
(b) No time required to get up steam after
long idle periods.
(c) A motor cannot stick on dead center.
(d) A motor, unlike a gas engine, will
always start without difficulty.
(e) A motor does not materially slow
down on the heavier " pulling " work
and hence pulls the first "stand"
of tubing as fast as the last one.
(f) The more accurate control obtained
with motors results in quicker work
■ in handling rods and tubing.
(g) After drilling is com]3leted, less than
an hour is ordinarily necessary to
change to electric pumping when
the proper arrangements are made.
Production lost at the flush period
during the long time required to set
a pumping engine is thus nearly all
saved.
Uniform Pumping Speed
Production is much reduced by variations
in engine speed. The more uniform speed of a
motor maintains full output of every well.
Actual examples are given in Tables III and
IV.
G24 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo.
TABLE III
COMPARATIVE PRODUCTION WITH STEAM
ENGINE AND ELECTRIC DRIVE UNDER
IDENTICAL OPERATING CONDITIONS
ON THE SAME WELL, PUMPING BY
ENGINE AT NIGHT AND BY MOTOR
IN DAYTIME. BURMA OIL COM-
PANY, SINGU FIELD, UPPER
BURMA, INDIA
Oil pumped by motor.
Oil pumped by engine .
Aug.. 1916
Bbl.
Per
Cent
1311
1777
42.5
57.5
Sept.. 1916
Bbl.
1.310
1587
Per
' Cent
45.4
54.6
Total time motor oper-
ation 271
Total time engine oper-
ation 473
Barrels per hour.motor 4.84
Barrels per hour, en-
gine 3.75
Increase in production
due to motor drive . . .
Per
Cent
36.5
63.5
28.5
Hrs.
270
450
4.82
3.52
Per
Cent
37.5
62.5
36.0
TABLE IV
INCREASE OF PRODUCTION OBTAINED
WITH ELECTRIC DRIVE BY AN OIL
COMPANY IN THE SPINDLETOP
FIELD, TEXAS. PUMPING
FROM A "POWER"
Eight wells on steam,
January and Febru-
ary, 1918
Same eight wells, elec-
tric power, March and
April, 1918
Increase (11.6 per cent)
Total Bbl.
9346
10,791
Bbl.
Per Day
158.4
176.9
18.5
Bbl.
Per Well
Per Day
19.8
22.1
2.3
Lower Operating Expenses
0]:)eratinjj fij^iires for a number of oil
comimnies have been published to show the
comparison between engine and motor drive,
and they all indicate a remarkably large
saving with the latter. F'our dilTerent com-
panies, for example, show savings respectively
of 22 per cent (12 wells), 24 per cent (12
wells), 40 per cent (107 wells) and ()3 per cent
(number of wells not stated). Comparisons
for five other companies show average
savings, obtained when electric drive was
substituted for steam engines, var\'ing from.
$450 to S277.> per well per year, the average
for all these being approxim.ately S1275. As
the complete cost of a standard two-speed oil
w^ell pumping motor installed is from SI 000 to
•'§2()00 per well, depending upon the kind of
installation, the comparisons indicate that
in nearly all cases this can be fully paid for
from the savings in less than two years, even
though the greatly diversified conditions
encountered in the oil fields cause a wide
variation in the costs of operation.
The items taken into consideration in these
comparisons are only those affected by the
change; viz., fuel, power, labor, water, and
maintenance. A few comments on each of
these will be of interest.
Fuel and Power
As previously m.entioned, the oil field
consumption for steam-engine pumping opera-
tions is from 3 to 15 barrels per well per day.
The electric power usually required is from
00 to 150 kw-hr. per well per day, though in
exceptional cases it m.ay reach about 200 kw-
hr. maximum. From this it is clear that at
prevailing power rates electric power is much
the cheaper. It may also be cheaper than gas
fuel where the latter has any market value.
Labor
Motors require much less labor expense
than engines. One i)umper can usually look-
after 15 or 20 motors, but cannot properly
handle more than S to 12 gas engines or 10 to
15 steam engines under the same conditions.
One electrician can take the place of several
gas-engine and boiler repair men. and firemen
are needed only in proportion to the number
of boilers retained on the lease.
Water
Water is scarce and expensive in many oil
fields, and such as is obtainable is usually bad
for boilers. The use of motors eliminates it
at a saving often in excess of the cost of
electric power.
M at tiic nance
The a^-erage annual repair expense on oil
well motor equipment does not reach one
per cent of the first cost, even over periods of
operation up to 12 years or more. Gas engine
equipments not over four or five years old
have an average annual maintenance expense
of more than II i>er cent. A I'nv figure for
ELECTRIC POWER IN THE OIL FIELDS AS A CENTRAL STATION LOAD 625
steam engines and boilers is five per cent.
Another important matter is the investment
necessary for a suitable stock of repair parts.
For motors, this is not over 25 per cent of that
required for gas engines, due to the lower rate
of depreciation and the fewer wearing parts.
Other Advantages of Motors
(1 ) A motor cannot run away when the rods
part.
(2) Explosions are eliminated and the fire
risk is reduced, thus lowering insur-
ance rates.
(3) Accidents are fewer.
(4) More reliable speed control is obtained.
(5) Better motion of cleaning-out tools is
produced by motors than by gas-
engines.
(6) Motors have a simpler method of
control than engines.
(7) Electric power consumption can be
accurately measured.
(S) Electric drive is cleaner and quieter than
engine drive.
MOTOR EQUIPMENTS FOR OIL
WELL DRILLING
A different type of electrical equipment is
used for drilling than for pumping. Drilling
requires a motor of larger capacity than is
necessary on a producing well, and the method
of control is somewhat different. It is there-
fore the practice to use separate equipments
e.xclusively for drilling, and, as each well is
completed, to put in a pumping motor as a
permanent installation, moving the drilling
motor and control apparatus to the next new
rig-
For standard cable-tool drilling, an ordinary
slip-ring induction motor gives the best results.
An auxiliary' controller provided in addition to
the main controller giv-es the very fine adjust-
m.ent of speed necessary to make the move-
ment of the walking-beam accord with the
natural period of vibration of the drilling
line. The two controllers are operated
independently by rope-wheels from the
headache post. Fig. 8.
In cable drilling the beam must overspeed
and allow a "free drop" of the tools on the
down-stroke to get the most effective blow.
To accomplish, this the m_otor must slow down
on the up-stroke and speed up on the down-
stroke. This characteristic is very satis-
*"The Application of Electricity to the Production of Crude
Oil." by W. G. Lane. The Oil Age, vol. XVI, Jan. 1920, p. 10.
factorily obtained by so proportioning the
pulleys that some secondary resistance is -in
circuit when the motor is operating at the
proper drilling speed.
For rotary drilling the same type of motor
is used for the draw-works and turntable, but
very fine speed control is not necessary-. The
slush pumps are also driven by standard
slip-ring motors.
Drilling itself is a fairly steady load on the
motor, but the other work, particularly the
handling of casing, is heavy and very inter-
mittent in character. This may be seen in
Figs. 3, 4, and 5. For wells much over 2000
ft. in depth, a 75-h.p. motor has been found
by experience to be of suitable size. It has
been used for drilling to a depth of 4412 ft. in
California and apparently has ample capacity
for still deeper drilling. For 2000 ft. or shal-
lower wells, 50 h.p. may be sufficient, depend-
ing upon local conditions.
Power Consumption for Drilling Operations
It has been found that after a well has
reached a depth of 300 or 400 feet, the amount
of energy required per hundred feet increases
with the depth of the well. As the well grows
deeper the drilling tools used are smaller in
diameter and lighter in weight, and a larger
amount of water is usually carried in the hole,
so that the power required to swing the tools
grows less. On the other hand, the length of
time required for bailing increases in pro-
portion to the depth, and the "dashpot"
effect in pulling out the bailer increases due
to the larger amount of water in the well.
It is also usually necessary to work the casing
more frequently as the depth increases, in
order to keep it from "freezing." Both of
these conditions cause a considerable increase
in energy consumption. Furthermore, prog-
ress becomes slower as the well deepens. The
power consumption as a whole, therefore,
increases more rapidly than in direct pro-
portion to the depth.
Mr. W. G. Lane recently published* very
reliable power consumption figures which are
based on actual meter readings taken not
only over a considerable period of time, but
also on a number of different rigs in various
fields. These show the average energy con-
sumption per 24-hour day to be as follows:
lOOO-ft. territory 150 to 170 kw-hr.
1500-ft. territory '. . . . 180 to 21.5 kw-hr.
20aO-ft. territory 200 to 235 kw-hr.
2.500-ft. territory 230 to 270 kw-hr.
3000-ft. territory 250 to 285 kw-hr.
Over 3000-ft. territory 265 to 350 kw-hr.
626 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 7
3 "s
o o
s -
0 -i
- J
1°
U a
td Q
2 ■:
Q E
o =
H^
I "
Q ^
X o
„ -a
5 =
0 O
o
<
a .• _>--■■ .'
C 6^
u
II
i £
= I
03 g
° J
li
5 S
m
ELECTRIC POWER IN THE OIL FIELDS AS A CENTRAL STATION LOAD G27
Advantages of Electric Drilling
To the power company it is of advantage to
introduce electric drilling and thus get power
lines into the fields early in the game. Electric
pumping then follows as a logical step, as oil
men all consider it more feasible to adapt motor
drive to pumping than to drilling operations.
As to the success and advantages of drilling
by electricity, a large oil company in the Mid-
continent field recently ])ublished* convincing
comments and data. It was stated that
"results obtained in the drilling of Stokes No.
27 and a subsequent well show conclusively
that a combination of motor and control
apparatus has been perfected to a degree that
causes even experienced drillers to say electric
equipment is superior to steam." The driller
himself reported: "Having worked on Stokes
No. 27 from start to finish, my candid opinion
is that electric power for drilling is great.
From a standpoint of economy and reliaiaility
it has no equal. In spudding, drilling, bailing
water, pulling tools or landing casing, the
motor gave us not the slightest difficulty."
The cost of installing and operating this
drilling equipment compared with what it
ivould have been with steam-engine drive is
given in Table V.
WATER PUMPS
Practically all oil companies using motors
pump their own water with electrically
driven pumps, and it has often been the case
that these were the first motors put in. In
♦"Drilling by Electricity" in The Empire (published by Empire
Gas and Fuel Co., Bartlesville, Okla.), Oct. 30, 19)9.
the aggregate this amounts to a considerable
power load, but for any particular property
a small unit furnishes all the water necessary.
The large companies usually have an extensive
water supply system.
Fig. 12. Pumping 450 Barrels of Oil a Day From a Depth
of 1900 ft. with a Two-speed Oil Well Motor in the
California Midway Field
TABLE V
COST OF INSTALLATION AND OPERATION OF MOTOR EQUIPMENT FOR DRILLING
2440-FOOT WELL, COMPARED WITH STEAM-ENGINE DRIVE
Boiler and Engine
Motor
Loss
$335.53
Saving
Initial cost . .
.$1,862.00
432.50
290.00
480.00
2,160.00
$1,625.00
♦768.03
32.50
60.00
574.93
$ 237.00
Cost of installation (including belts, etc.)
Estimated depreciation per well
Cost of water . .
257..50
420.00
Estimated cost of fuel oil at $36 per day .
Cost of electric power
Saving in cost of power
Saving in installing pumping motor in
same house, on same foundation ....
Saving in oil production during change
to pumping
1,585.07
186.16
1,. 305.00
Totals
$335.53
$3,990.73
Net estimated saving of electric drilling
$3,655.20
*The installation charge of the motor drilling equipment was high, due to the fact the equipment was new and changes had to be
made which involved labor charges that will not be necessary in future outfits. It also includes the cost of building the motor house.
G2S July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo.
GATHERING AND LINE PUMPS
Pumps requiring motors up to lUO horse
power, and often built for high pressures, are
used for gathering crude oil from the wells and
sumps and transferring it to the pipe lines or
tank f anns. These furnish a considerable power
load, though it is not large in com.parison with
oil well piunping. The small units generally use
squirrel-cage motors, but slip-ring m.achines
are necessan,- to obtain variable speed on the
larger pumps on account of the variable
pressure encountered in cold and hot weather.
The application of electric drive to main
pipe line pumping makes a ven,- desirable
load for the central station, but current is
seldom available at even,- pumping station
along the line. The general tendency up to
the present has been to install uniform pump-
ing equipment at all stations on a pipe line,
and accordingly there are only a few instances
where motor-driven pimtps have been in-
stalled. The larger stations require several
hundred horse power and the load is uniform
and practically continuous. Power com-
panies will therefore be warranted in mak-
ing strong efforts to obtain this desirable
load.
VACUUM PUMPS, COMPRESSORS AND
CIRCULATING PUMPS FOR CASING-
HEAD GASOLENE PLANTS
The electrification of casing-head gasolene
plants is a comparatively recent develop-
ment, but in som.e fields, particularly in
Oklahoma, it has prom.ise of exceeding oil
well pum_ping as a central station load. In
the DrtuT-right field alone there is more than
32,000 horse power in gas and oil engines in
these plants, and there are good reasons for
anticipating that m.any of these will be
changed to electric drive within a brief period.
Most casing-head gasolene plants are now
driven by gas-engines and use the residue
gas for fuel. In m.any cases this gas would
be blown oR and wasted if not so used, and
it would accordingly be expected that m.otor
drive would receive scant consideration.
However, engine troubles are found to be the
cause of a big loss in gasolene production, this
more than orfsetting the cost of the electric
power which would be consumed. The
engines are not only shut down se\-eral days
a month for repairs, but their speed variation
results in reduced output of the compressors.
It is ver\- necessar>' that a definite speed be
m.aintained at all times in order to handle the
maximum amount of wet gas.
One example of the better results obtained
with motor drive is a plant at Los Angeles.
Three 200-horse power gas engines driving
the compressors were replaced, after less
than a year's operation, by three 200-horse
power slip-ring induction motors. An average
shut-down of about four days per engine
per m.onth was thus eliminated and this
increased the production an amount just
i
Panoramic View of the Gook Creek Oil Field in Tezaa, Wherr
ELECTRIC POWER IN THE OIL FIELDS AS A CENTRAL STATION LOAD (L'l)
about sufficient to pay the monthly power
bill. In addition, the steadier speed at which
the motors drove the compressors resulted
in greater production per day. This con-
vinced the operating company of the supe-
riority of electricity and they accordingly
electrified their entire lease.
Operators consider motor drive for vacuum
pumps even more necessary than for com-
pressors. This is due to the fact that if the
vacuum pumps are not operated continuously
at a fixed speed, a loss in vacuum results
which enables the neighboring leases to secure
the gas and also a certain amount of oil.
Every engine trouble which causes a shut-
down means a probable loss in production to
the plant for 10 to IS hours, this being the
time required to again build up a vacuum
on the lines so that it balances that of the
neighboring companies.
ELECTRIC DEHYDRATORS
Electric dehydrators were developed to
provide a more economical method than that
usually employed for breaking up emulsions
of water and oil which cannot be separated
by settling. The oil containing the emulsion
is passed through an electric field, between
two electrodes having a difference of potential
of about 1 1,000 volts. The discharge between
the electrodes breaks down the emulsion and
the water then settles out. This raises the
Bauine gravity of the oil to practically its
original figure and thus increases its market
value. The saving thereby m_ade is usually
sufficient to pay for' the treater within a
period of operation of nine months or less.
The average power consiunption is in the
neighborhood of 2000 kw-hr. per month,
though it m.ay vary with the conditions from
1200 to 5000 kw-hr. The amount of power
required per barrel of cleaned oil ranges from
22 to 05 watthours. The power-factor is
about 98 per cent leading, due to the con-
denser effect of the highly charged elec-
trodes. The dehydrators may consist of
from two to eight treater units, but this
does not affect the power consumption to
any great degree.
MACHINE SHOPS AND LIGHTING
Little need be said about these applications
of electric power in the oil fields, as they pre-
sent no unusual features. A machine shop is
a necessity in oil field operations and motor-
driven tools have well-known advantages.
The fire risk is of course an important
consideration in connection with oil pro-
duction and for this reason electric lighting,
particularly in gassy territory, is especially
favored. When it is necessary to keep all
wiring as far away from a drilling rig as
possible and to avoid even the danger of
ignition from the breaking of a bulb, flood
lighting furnishes practically the only safe
method of illumination.
Wells are Now Pumped by Two-speed Oil Wei! Motors
630 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 7
Theory of Speed and Power-factor Control of Large
Induction Motors by Neutralized Polyphase
Alternating-current Commutator Machines
By John I. Hull
Designing Engineer, General Electric Comp.\ny, Schenect.\dy, X. Y.
This article discusses the theory of induction-motor control under the headings of: single-range (below
synchronism only) speed and power factor control by means of a constant-speed series commutator motor,
by means of a constant-speed shunt commutator motor, by means of a constant-speed compound excited
commutator motor; double-range (all speeds above or below synchronism) speed and power factor control by
means of a constant-speed shunt commutator motor; and double-range (either above or below synchronism)
operation remote from synchronism. This article was presented as a paper at the A.I.E.E. convention,
White Sulphur Springs, W. Va. — Editor.
A stationary polyphase wound-rotor induc-
tion motor is merely a static transformer
arranged so that the primary coils are all on
one part of the magnetic circuit, and the
secondary coils on another part of the
magnetic circuit, the two parts thus being
arranged so as to permit relative motion.
The reluctance of the magnetic circuit is
kept as low as possible by imbedding both
primarv' and secondary winding in slots,
thereby permitting the "teeth" between the
slots of the primar\- iron to come as close to
the "teeth" of the secondary iron as safe
mechanical clearance permits. The necessity
of some clearance or "air gap" makes the
reluctance, hence, the magnetizing current
and kv-a. larger than for the static trans-
former of similar capacity, voltage and fre-
quency, while the separation of the primar\'
winding from the secondary- winding and the
imbedding of both in slots make the leakage
reactances larger than for the corresponding
static transformer. It is thus evident that
the induction motor may logically be con-
sidered from the point of view of a trans-
former so arranged as to permit the forces
set up in the secondary conductors to cause
rotation, at the cost of an increase in the
magnetizing current and the leakage react-
ance.
The flux common to both and set up by the
resultant of the primary and secondary
magnetomotive forces is the link of mutual
influence between the primary and secondar\'.
This influence manifests itself as electro-
motive forces set up in proportion to the
effective number of turns, the flux, and the
frequency in the circuit in question. (Of
course, this is rigorously true only with the
usual asstunptions of sine wave distribution,
etc.)
The torque is proportional to the sum-
mation of the products of the secondary turns,
and the components of current in quadrature
with the common flux of each, and to the
common flux.
The usual theor\' of the induction motor
takes into account only the phenomena within
the motor itself, performance with adjustable
external secondan,- resistance being analyzed
by considering it to be merely an addition
to the normal resistance of the secondary'.
In Fig. 1, we reproduce a circle diagram,
a vector diagram for analyzing induction
motor performance, which lends itself to the
introduction of electromotive forces etc., of
concatenated machines.
AH — 1 1 (proportional to priman,- current)
represents the flux linking priman,- and second-
ary- which would be produced by the primarj'
current alone. (Saturation neglected.)
Fig. 1 . The Circle Diagram of an Induction Motor
H I = C\Ii represents primary- leakage flux.
H G = I E = h (proportional to the second-
ary- current) represents the flux linking sec-
ondan*- and iirimar>- which would be jiroduced
by the secondan current alone. (Saturation
neglected.)
POWER-FACTOR CONTROL OF LARGE INDUCTION MOTORS 631
G D = C'i.Ii represents secondary leakage
flux.
AE is thus the resultant of all the primary
flux and that secondary flux, linking the
primar\-, so that neglecting primary resist-
ance drop £ is a fixed point for constant line
voltage and frequency, as .4 E is the flux which
generates the counter e.m.f. to balance the
applied voltage e^.
H G intersects AE at B, and as
ABIAE =
h
Ii(l+Ci)
we see that .4 B is constant, making B also a
fixed point . .4 L> is resultant of /i and Jo 4- C2/2
and therefore generates all secondary electro-
motive forces except resistance drop which is,
therefore in phase opposition to the voltage
62, set up in the secondary in quadrature to
flux .4 D. This makes .4 F parallel to H D and
to I E and further makes A D H a right angle,
which taken in connection with the fact that,
as shown above, A and B are fixed points,
demonstrates that the curve traced by point
D is the arc of a circle.
A line parallel to ^4 I from point D intersects
prolongation of .4 £ at C and prolongation of
7 £ at A'.
CD =CK+KD
CK _AI
EK IE
„„ CJiXliCl+G)
C K = J
= JlC2(l-|-Cl)
K D =H I — Cili (since they are parallels
intercepted by parallels)
CD =/,[G+r2(H-G)]
BD =BG+GD
BG_ ^lE
E'G I A
EGXIE ^ ^ ,, h
BG =-
I A
■ = CiIiX
= /2X
ii,(i+co
1+Ci
GD =CJi
BD=L.(C2+J^;)
h
-,^^X[Ci4-G(i4-Q]
£.4 is the flux whose counter e.m.f. balances
all the applied line voltage as noted above.
Let it be designated Im- B A is the mutual flux
at running light and, if denoted by 7o, we have
J-9—'
l+Ci
CB --
CE
EA ''
CE
EB
EA
■ CE+EB
EK
''EI
= EAX
I A
EK
= /„
EI
EB = I.
Cxh
1C2
,x
/i(l + Ci)
CB =/,
= 1,
Ci
X
1+Ci
'(^^+r+k)
14- G
-X[Ci4-C2(l+C0]
7T-r = -=-7 = C onstant = C2 X /2 //a
= /o[C:4-C2(14-CO]
Further, C is a fixed point since:
CE^EK
EA~ EI
CA=EA+CE = Im+In.C, = I„ (I + C2)
_In. (14-C2)[Ci4-C2(14-Ci)]
Ci+G (l+Ci)
Summing up we have :
CD = Ii[Ci+C2 (1+C,)]
I.
DB =
CB =
,^CXICM
(1+Ci)]
X[Ci+C2(l+Ci)]
1+Ci
= Jo[Ci+C2(H-C,)]
^, /m(l+C2)X[Ci-t-C2(14-Ci)]
Ci4-C2(l-|-Ci)
£A = J«
BA = Io
With proper scale, Im could be made to
represent the magnetizing current for a total
flux Im, (which is the quantity commonly
calculated, as the primary reactance and
resistance drop are usually omitted) Jo could
be made to represent the true running light
current, priman,' reactance drop considered,
/i the primar}^ current and /2 the secondary
current. If we now change the scale of the
diagram by the factor Ci + d (1 +Ci), we may
say that magnetizing current im divided
by 1-l-Ci equals CB, equals true running
light current to ; primary current n equals C D;
secondary current divided bv 14-Ci equals
DB.
G32 Jiily, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 7
At standstill, with zero secondan- resist-
ance A D, the resultant secondar^^ flux must,
of course, be zero, as its generated voltage
is zero, which means that D coincides with
.4 and CD = CA, so that we have the ideal
short-circuit or standstill current, with zero
secondary resistance, equal to C A.
Fig. 2. Simplified Diagram
As Ci li is defined as prirnar\- leakage flux,
the primary reactance drop with current im
is Ci ei, since im produces a total flux whose
e.m.f. is equal to e\. The primary reactance
drop is further equal to },„ A'l, if A'l be the
primary reactance, thus:
Ci ei = im A'l
Ci = — _— and similarly
G =
im A'l
Thus, to draw the diagram of the motor,
we need to know the primarA- and secondary
reactances A'l and A'-> and the nominal magnet-
izing current im- We need then only so much
of Fig. 1 as is shown in Fig. 2.
Having chosen a scale, lay off:
C B equal to !o
/-I 1 i tm (1 +(-2) J
C .4 equal to ~ . ^ 1^ , n\ ^"^^
Ci-t-C2 m-Cij
draw a circle with B A as diameter. C M is the
in phase or watt component of input current
for any considered lead. M D parallel to C A
then locates /) and the remainder of the
diagram. The primary current is then i\ =
CD and the secondary current /'; is then
(1 -j-Ci) D B, or if we use as the unit for «« the
unit denoting the other currents divided by
1 +Ci, we can let D B=ii.
To find the secondary' voltage ?". = .4 F, we
can first determine its value reduced to full
freqiiency. The voltage generated by flux
.4 E of Fig. 1 is ei, and that generated by .4 B
is, therefore.
ei
So in Fig. 2, knowing ei.
l+Ci"
we can say that .4 B units of length correspond
to volts, and can regard .4 B, etc., as
1+C 1
measures of voltage. So .4 D is -j-b times
volts at standstill frequency. If the
l+Ci
secondary resistance is known, the actual
value of secondarv induced voltage e^ is, of
course, ii r^, so that per cent slip is s =
AD-
" SvTichronous watts " torque is B D times
.4 D, output is (1 -5)BI>X-4£>,efficiency (1 -s)
^^ , AD ,,^ , , MC
B D -I AI C, power factor 7^-p:.
ei L JJ
If the ratio of secondary- turns to primary
turns is other than 1 to 1, the diagram is.
of course, of necessity drawn for all factors
reduced to either primar\- or secondary terms,
secondary terms being usually used for work
of the present sort. Thus, the primar>- voltage
to be expressed in terms of secondary- must, of
course, be multiplied by ratio of secondary- to
Fig. 3. Induction Motor Diagram
With regulating voltage er introduced into secondary' in phase
opposition with total induced c.m.f.
primary effective terms, primar\- reactance by
the square of this ratio, etc.
In the demonstration of Figs. 1 and 2, it was
pointed out that jjoint D traces the arc of a
circle whose diameter is /? .4, because .4 /) B is
a right angle, due to the phase opposition
of e« and /'• ^2 wlien the only e.m.f. in the
POWER-FACTOR CONTROL OF LARGE INDUCTION MOTORS
033
secondary circuit, other than that induced by
the total secondary' flux, is resistance drop. If
as in Fig. 3, another e.m.f. than the resistance
drop as Cr be introudced, then D will still
trace the circle with the diameter A B when
and only when the introduced e.m.f. is in
phase with or in phase opposition to e^. In
this case, for given values of ii, i->, etc., e^ must
be equal and opposite to the algebraic sum of
12 rn and the introduced voltage; hence, since
the inducing flux of e^ is determined by the
currents, — its inducing frequency and the
slip and speed must follow variations in the
algebraic sum of 12 fs and the introduced volt-
age. It is, therefore, evident that varying the
introduced voltage, while maintaining it in
phase with t2 gives a means of varying the
speed of the motor without effecting its power
factor torque, etc.
If now, as in Fig. 4, the introduced e.m.f., er
be of difl^erent phase from that of £"•>, point D
departs from the circumference of the circle
whose diameter is i? .4, as shown at D' because
12 is no longer in phase opposition to €«, hence,
.4 D' B is no longer a right angle. It is seen
that in addition to regulating the speed, the
power factor of the motor may also be regu-
lated by proper selection of phase as well as
magnitude of the introduced e.m.f.
It is clear, of course, that the frequency of
the introduced or regulating e.m.f. must at
Fig. 4. Induction Motor Diagram
With regulating voltage er introduced into secondary out of
phase with total induced e.m.f.
all times be exactly that of (?2, in order to
maintain the phase relation shown.
Thus, if we can introduce at exact second-
ary frequency a regulating voltage of con-
trollable phase with respect to €2 and con-
trollable magnitude, we shall be able to
regulate either speed, power factor or both.
Single-range (Below Synchronism Only) Speed and
Power-factor Control by Means of a Constant-
speed, Series Commutator Motor
In Fig. 5, we show schematically at D, a
three-phase series neutralized commutator
machine whose terminals are connected to
the secondary slip rings of main motor .4.
Mechanical
Load
Fig. 5. Neutralized Series-excited Three-phase A-c. Commuta-
tor Machine and Connections for Automatic Single-
range Regulation of Induction Motor Equipped
with Flywheel to Reduce Peaks on Line
The speed of D is held practically constant
by generator E.
Neutralizing winding Ci, Co, C3 balances the
armature reaction (magnetomotive forces)
of armature Aw, and so of necessity neutralizes
the e.m.fs. set up in Aw by the transformer
action of the fluxes induced by series exciting
windings SFi, SF2, SF3. (Ci, C2, C3 being in
series with Aw carry the same currents as
Aw, hence, for a balanced condition of
magnetomotive forces must have an equiv-
alent and opposite niunber of turns, so the
e.m.fs. also cancel.) Thus the e.m.fs. ap-
pearing at the terminals of D are the leakage
reactance drop, resistance drop and the rota-
tion e.m.f. induced by the rotation of the
armature Aw. The rotation e.m.f. is, of
course, proportional to the flux and the speed
of rotation, the flux, neglecting saturation
being proportional to the main currents which
flow through series exciting windings SFi,
SF2, SF3. This arrangement can then be
seen to be such that the speed of A will be
reduced with the increase of load, provided
the rotation voltage, as Cr in Figs. 3 and 4, be
634 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 7
given a suitable component in phase with the
resistance drop, thereby having the same
effect on the main motor speed, as increasing
the resistance.
Up to the point of the magnetic satura-
tion, two laws may be seen to inhere in the
machine D.
Fig. 6. Circle Diagram of Induction Motor with Constant-
speed Series A-C. Commutator Regulating Machine
deg. — angle G F .4 , we see that j3 = angle B D A
is also constant.
Thus, as A, B, and C are fixed points, point
D traces a circle whose center must be at the
intersection of the perpendicular bisectors
of .4 B and B D.
We have remarked that Figs. 1, 2, 3, 4 and 6
are rigorous when and only when the iron of
the machine is unsaturated, that is, when
the flux may be regarded as proportional to
the ampere turns. This condition is closely
enough approximated in the main induction
motor so that saturation may be neglected
without much loss of accuracy. For the
series machine of Fig. 6, however, to be of
economical proportions, considerable satura-
tion will be attained within the working range ;
hence, it becomes desirable to investigate its
effects. In Fig. 7, A, B, C has been deter-
mined as was done for Fig. 6, and B D A is
the corresponding circle determined bv angle
\ C
B D A, which is determined bv ratio W^ and
F G
design angle a, saturation neglected.
Now with current B D. we can calculate a
new value for X2+c+cs which will hold only for
this one current, since in expression
la'-a. to excite regulating motor
1. The flux, hence the rotation voltage at
constant speed, is proportional to the ciurent.
2. The phase angle between the current
and the rotation voltage (hence, the angle
between resistance drop and rotation voltage)
is constant (it can only be changed by chang-
ing the construction of the machine).
These are the basis of the circle diagram
of Fig. 6.
Points A, B, and C are determined exactly
as in Fig. 1, except that for A'2 we now sub-
stitute X2+C+CS where X2+c+cs = X«+Xc+Xc^
and Xc = leakage reactance of regulating motor
at priman.' frequency
,. _kv-a. required to excite regulating motor
The k\--a. is at primary' frequency and
unsaturated iron of regulating motor is
assumed. Obviously the performance of an
induction motor is not changed for our pur-
poses, having a part or all of the rotor leakage
reactance external to the machine. Angle a
= FGA, between resistance drop FG and
rotation e.m.f. G A is constant by law No. 2,
and G A is by law No. 1 proportional to B D
and hence to F G. For these reasons angle
G F A is constant and since angle B D A = !)()
Fig. 7. Diagram of Induction Motor and Constant-speed Series
A-C. Commutator Regulating Machine
Taking account of saturation of irun cf reflating mAchine
we can determine kv-a. (at full frequency, of
course) from the known or assumed magnet-
izing cur\-e of the regulating machine. With
this new (and decreased) value of A'2-Hf-(-f,
we calculate .4' C instead of .4 C. If the new
value of X2+c+t, were constant, our new circle
POWER-FACTOR CONTROL OF LARGE INDUCTION MOTORS
635
would be B D' A', but as it varies, D' may be
the only load point upon it. Triangle A', F',
G' is, of course, equal to triangle A F G, and
angle B D' A' is equal to angle B D A.
A second effect of the saturation is that the
ratio of rotation e.m.f. to field current (field
current being the same as the main current
for a series machine) is reduced, so consider-
ing this, A' G" F" is the e.m.f. triangle with
angle A' G" F" still equal to angle A G' F'
and A G F.
Since .4' G" is less than A' G' and G" F" =
G' F' with constant a we see that angle
G" F" A' is less than angle G' F' A', hence,
angle B D" A' greater than angle B D' A' and
A G"
the circle, if X2+c+cs and ^„ ^„ were constant,
would be B D" A'.
We thus see that the two effects of satura-
tion of the regulating machine partially offset
one another, as the reduction of X2+c-^c^
makes the imaginary circle larger and the
power factor more leading, while the reduction
A' G' A' G"
of ratio 7=7-79 to 7=^™, makes the imaginary
G t G r
circle smaller and the power factor more
lagging.
The point D" cannot be located by rule
and compass unless we calculate triangles
A' D" B and A' F" G" which can be done as
follows :
A' G", G" F" and angle A' G" F" are known.
A' F" = 'SJA^l7'- + (7F"''-2XA' G"XG" F"
XcosA'C'F"
, _ A'G"XsinA'G"F"
bin A t G A' D"
Angle F" A' G"
= 180 deg.— (angle A' F" G"+A' G" F")
determining triangle A' F" G".
In triangle 5 D". 4', B A' and BD" are known
and angle B D" A' = 90 deg.— angle A' F" G" .
BD"XsmBD"A'
SinBA'D" = -
B A'
Angle D" BA'
= 180 deg.— (angle B A' i?"-f angle B D" A')
^,^„_BD"XsinD"BA' A' B sin D" B A'
sinBA'D" °'" sinBD"A'
Knowing, thus, A D" and B D", we can
find point D" with compass.
We can now construct the curve traced
by D" by assuming values of current B D",
calculating for each value A' B, A' G" and
A' D" as described.
For the designs ordinarily encountered,
this yields a curve so closely approximating
for the working load the original circle B D A
in which saturation is neglected, that it is not
necessary to go beyond the construction
oi B D A to get a good idea of the charac-
. A' F"
teristics except slip which is ., ^„. If the
scale used for A' F" is not that of A' D",
A' F"
then, of course, slip is . , „„ multiplied by the
proper ratio of scales.
The combination in Fig. 5 is suitable to
service in which there are rapid and wide
fluctuations in load which it is desired to
absorb as much as possible by the flywheel
B. This arrangement is superior to the use
of a resistance across the slip rings because
instead of being wasted as in the resistance,
the slip energy can all be returned to the
power system except for the machine losses
of D and E. When applied to a motor with
secondary resistance the flywheel reduces the
peak loads by delivering torque as it is re-
tarded. The return of most of the slip energy
to the line by the regulating set decreases the
peak loads still more. A further advantage
for the regulating set is the means which it
affords of materially improving the power
factor of the main motor.
Single-range (Below Synchronism Only), Speed and
Power Factor Control by Means of a Constant-
speed, Shunt Commutator Motor
The series regulating set is, of course, the
simplest form, but it is not adjustable without
tapping the field winding or external apparatus
and as it imparts to the main motor the
characteristic of a material reduction of
speed with the assumption of load, it is not
suited to the majority of industrial uses in
which variable speed from large induction
motors is required. In the greater number
of cases, it is desired to adjust the speed to a
value suited to the momentary requirement
of the process, and have the speed remain
at approximately the adjusted value irrespec-
tive of load variation.
The total induced secondary e.m.f. of an
induction motor including the secondary
reactance drop is proportional to the "rotor
field" (see A D oi Fig. 1) and the slip. So,
as is well appreciated, within the working
range the slip is about proportional to the,
torque as the torque is about proportional
to the rotor current, the current being pro-
portional to the total induced rotor voltage.
If at a given load we obtain speed reduction
636 July, 19'2()
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo.
by an increase in resistance or by the use
of a "series" regulating set, in which cases
an increase in secondary induced e.m.f.,
hence, slip is required to overcome the addi-
tional resistance drop, or the rotational e.m.f.
of regulating set, plus resistance drop, then,
Power Suppl
Fig. 8. Neutralized Three-phase Shunt Constant-speed A-C.
Commutator Machine and Connections for Adjust-
able Speed Control of Induction Motor
Below Synchronism
returning the energy deriv-ed from D to the
line. B is an auto transform.er excited
from the slip ring circuit and provided with
suitable taps to apply pre-determined per-
centages of the slip ring e.m_.f. to the shunt
exciting windings Fi, f 2, F3. Assume that the
resistance drop in the F\, F2. F3 circuit is
negligible and that "B" applies to Fi, F«, F3
the selected percentage of the total secondary
induced e.m_.f . This with the further assump-
tion that the reactance drop of the regulating
motor is included in the slip ring e.m .f . (which
is supplied to B) and that the resistance
drop of the main motor rotor is not included
in the slip ring e.m.f., will give what may be
termed for our purposes, "pure shunt excita-
tion." The effects of these assiunptions will
be pointed out later. The counter e.m.f. of
Fi Fv, Fi thus consists of the e.m.f. set up
by the flux excited by it, and is, therefore,
proportional to the flux and the frequency
(frequency itself being proportional to slip 5).
i. e., ef = kX<t>Xs
The e.m.f. applied to the field is proportional
to the total induced e.m.f., as remarked
above, and the total induced e.m.f. is equal
to its own standstill value for the main motor
current (hence torque) in question times the
slip, so we see that, total induced e.m.f. at
standstill X5 = ifeX«Xs.
<i> is thus independent of frequency and
proi)ortional only to total standstill induced
as soon as the load disappears, the main motor
speeds up to synchronism, since the secondary
resistance drop and the rotational e.m.f. of
the regulating motor vanish with the cur-
rent. If the rotational e.m.f. of the regulating
motor could be made independent of the load
and of the slip, then with the departure of load
it would remain constant, so that the speed
would only rise enough to make the total
induced secondary e.m.f. equal to the rota-
tional voltage of the regulating motor, lea\-ing
no resultant to circulate load current. With
load fluctuations, the speed would then
fluctuate by only such small amounts as to
cause at all times th ^ small difference between
total induced secondary e.m.f. and the
rotational e.m.f. of the regulating motor to
overcome the small resistance drop of the
windings; usually only a few per cent of
synchronous speed.
In Fig. S is shown an arrangement to
approximate these conditions. .4 is the
main motor, D a neutralized three-phase
shunt commutator motor, whose speed is
held practically constant by generator E.
Fig. 9. Circle Diagram of Induction Motor and Pure Shunt
Excited Constant-speed A-C. Commutator Machine
e.m.f. Referring to Fig. 9. A. B. ('is con-
structed exactly as Fig. 1, including the
leakage reactance of the regulating motor with
that of the secondary of the main motor, so
that instead of A'j, we shall have A's+£=A*»+
POWER-FACTOR CONTROL OF LARGE INDUCTION MOTORS
637
A'r and instead of Ci, we shall have C2+f.
The circle B, D, A would thus apply with
regulating set D and E of Fig. S, stationary.
With the regulating set running, we get
secondary current B D', total induced e.m.f.
,4 F of main motor secondary proportional
to slip and the .4 D' and rotation e.m.f. of
regidating motor .4 G proportional to ,4 D'
and at a constant angle y from .4 F, angle y
being determined by the connections of the
transformer B and exciting winding Fi, F^, F3
of Fig. 8.
Resolve resistance drop F G into the
component F H in phase opposition to ^4 F
and H Gin quadrature to .4 F, the correspond-
ing components of secondary current B D'
being B D andD D'.
D D' is proportional to H G, H G = A G si a
7, so // C is proportional to A G (7 constant)
which is proportional to A D', hence, to D D'
is proportional to .4 D'.
A D = A D'-D D', hence, .4 D is pro-
portional to .4 D' .
B D and B' D' are both perpendicular to
A D', hence, ^^—^7 =^-rTT = constant, and B'
A B A JJ
is fixed point. So curve traced by D' is a
circle.
.4 F
Slip "5" is equal to 7-^7 ■
A U
At running light (zero torque) B D and
H F become zero. (For proof of this see
Fig. 1. The torque of the motor is pro-
portional to the mutual flux ,4 G and the
component of secondary current in quad-
rature with it. This is the same thing as the
total secondary current and the component
of the mutual flux in quadrature to the
current, which component is equal to A D,
the "rotor field." The torque is, therefore,
zero when B D is zero, and in Fig. 9. B D is
the torque producing component of B D' .)
A H
Thus, running light slip 5o = x^-pn-, and the
additional slip sx, due to the load is thus,
HF
A D"
It is thus seen that at running light the
main motor runs at slip Sa, determined by the
.4 G
angk 7, and the ratio ^ — pr-,, which conditions
are adjusted by the connections at B, Fig. 8.
The load slip Si, is the same for all values of
Tj r^
So. provided angle 7 be so chosen that ^ — fz,
A D
remains constant, and is the same as would
obtain for a normal motor whose circle is
B' D' A and whose short-circuited secondary
has the resistance corresponding to the
current B' D' and the drop H F. Thus it is
evident that the main motor, regulated as in
Figs. S and 9 would retain practically the
same load slio-torque, power factor-torque
Power Sup|)l.v
Fig. 10. Neutralized Compound-excited Constant-speed Three-
phase A-C. Commutator Machine and Connection for Adjust-
able-speed Single-range Speed Control of Induction
Motor Giving Automatic Drop in Speed With
Increase of Load
and input-torque characteristics as with short-
circuited slip rings, but would have no-load
1 r
speeds equal to the synchronous speed X — z—^.
It will be noted from Fig. 9 that the primary
power factor can readily be improved and that
at the same time the pull-out torque of the
main motor can be increased.
Single-range (Eelow Synchronism Only) Speed and
Power Factor Control by Means of a Constant
Speed, Compound Excited Commutator Motor
Occasionally, in processes where the peak
loads are high and of brief duration and of
sufficient magnitude in proportion to the
capacity of the supply system to be objection-
able, it becomes desirable to have a larger
drop in speed due to load than would be
obtained with a shunt commutator motor,
so that a fly-wheel can be effectively added
to smooth out the peak loads, and at the same
time retain the adjustability of the speed.
Fig. 10 illustrates a method of compounding
the regulating motor.
638 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 7
In the shunt excitation, neglect the same
factors as in the case of the shunt regulating
motor, (the secondary- resistance drop of the
main motor and the absence of the reactance
drop of the regulating motor). In order
to get the compounding action it would
c,
B'\
v
[b \\
\
\/
h
'^Sx^^J
\i
7^
"^
^\^ ^-^z
\
'^^-^ -""^
Fig. 11. Circle Diagram of Induction Motor with Com-
pound-excited Constant-speed AC. Commutator
Regulating Machine
not do merely to put in some series turns as
in a d-c. machine, since there is applied to the
shunt field a fixed percentage of the total
induced secondar\' e.m.f., which means a flux
proportional to the rotor field of the main
motor as already pointed out. Thus, the
ampere turns of the series winding would
merely be balanced by a change in the
shunt current, F,, Fj, F, ser\-ing as the
priman,' of a transformer. It, therefore,
becomes necessary- to change the field volt-
age in response to load, in order to change
the flux and hence the rotation e.m.f. This is
done by the series transformer H, which has
an air gap in the magnetic circuit, so that its
flux is proportional to the resultant of the
primarj^ and secondar\^ ampere turns. Of
course, a proportionate effect upon the flux
of the regulating motor, will accompany any
changes in the flux in the series transformer
(slip transformer) since the alteration in
flux produces a proportionate alteration in
•Point H is on .4 F because at zero load B B becomes B B'.
tbe total secondary current and H C and C / the only e.m.fs.
to close the gap between / .1 and .-t F. As B E is proportional
to ^ D" and hence A I and at constant angle thereto, H must
remain on A F.
Time
yoltage applied to regulating motor field, and
this in turn a proportionate change in flux
thereof. This then means that the rotation
e.m.f. will contain a component proportional
to the resultant ampere turns of the series
transformer, which component, of course, may
itself be resolyed into components corre-
sponding to the components of the resultant
ampere turns of the series transformer.
The performance of the motor, controlled
as shown in Fig. 10, is illustrated in Fig. 11, in
which .-1 , B and C are again found as for Fig. 1 ,
including the reactance of ma-
chine D with that of the sec-
ondary- of .4. B is the no-load
point, such that B B' is the
no-load secondarx- current, and
D' is the load point under con-
sideration. Resolye secondar\'
current B D' into component
B E parallel to A D' and E D'
part of B' D'. Component B E
produces no torque. / .4 is the
rotational e.m.f. of regulating
motor at constant angle from
total induced secondary- e.m.f.
.4 F of main motor and bearing
constant ratio to A D' being
due to the application of total
induced e.m.f. .4 F to the shunt
field circuit of D, Fig. 10.
H G* is resistance drop oi B E
and I G is the rotation e.m.f. produced by D
of Fig. 10 by the yoltage introduced into the
field circuit by the existence of current B E
in primarA- of H. The transformer flux com-
l^onent due to the secondan.- current of H
(magnetizing current of D) may clearly be
regarded merely as field leakage flux, and has
no significance except for the yalue of the still
/ .4
constant ratio , ^,. The angle a (I G H) is
A U
determined by design and is, of course, con-
GI
stant, as is the ratio
HG-
F K is total resistance drop, and F L and
L K are components, due to E D' and B E.
K G = L H is compounding effect, due to
E D\ just as C / is due to B E, so angle H L F
= angle I G H = a and is constant. Further
H L
L F
is constant, so angle L F H = fi is con-
stant and as F L is parallel to B' D' and .4 F
is perpendicular to .4 D', angle B' D' .4=5
= i)0 deg. — ^ = constant.
Therefore, D' traces the arc of a circle.
POWER-FACTOR CONTROL OF LARGE INDUCTION MOTORS
639
.4 F
The slip 5 is equal to xTy ^'^"^ '^^ running
light (zero torque) E D', F L, H L and K G
A I
become zero . „, and angle y are constant.
A L>
Angles of triangle I H G being constant and
G H A being 90 deg., we see that I H A is also
a constant angle, hence angle A I H is con-
A H
stant. So -j—fy- is constant and as this is the
expression for Sq, the running light slip where
H F is zero, we see that the slip, due to load
H F .
5i = ,, consisting of L F, due to the resist-
A D
ance and H L, due to the compounding action
of the slip transformer. We thus see that the
running light slip So is adjustable by means
of B in Fig. 10, while the load slip 5i has been
Fig. 11a. Effect on Fig. 11 of Correctly Locating Resistance
of Main Motor Secondary and Reactance of
Regulating Machine
MF HF
increased from j-jy to -j—fc, by the slip trans-
former. Further, it is apparent that by
controlling the angle a we can make the
power factor get more leading or more lagging
as load comes on, and thus, also, increase
or decrease pull-out torque of the motor.
In defining the conditions assumed for
Fig. 10, we mentioned that the leakage
reactance drop of the regulating motor was
supposed to be included in the voltage
applied to the exciting winding. The actual
effect of excluding it from this circuit can now
be shown in Fig. 11a, a modification of part
of Fig. 11. The reactance in the regulating
motor is, of course, not applied to its field,
and hence the actual rotation e.m.f. for shunt
excitation should not include /' /, the rotation
e.m.f., due to the application of reactance drop
of 5 £ to the shunt field. Note that I' A is
the rotational e.m.f. of pure shunt excitation,
and that as triangle B E B' is similar to tri-
angle A D' B', . ^, =-^ — ^. hence /' / is pro-
portional to I A and to A D' so that A H
is still proportional to A D'. G G' and H H'
are the rotational e.m.f. due to application
of reactance drop of E D' to shunt field, and
hence proportional to E D', so they may be
excluded from A' G' and L H', and for them
may be used instead A' G and L H. As
angle G' G K and H' H L are fixed and as
H H' and G G' are proportional to H L and
G K, we see at once that H L is proportional
to L F, angles H LF ( = angle H LH' -\-a) and
H F L are constant, hence, 8=B' D' A =90
deg. — /3 is constant and D' still traces a circle.
Thus when we consider the actual effect
of the leakage reactance of the regulating
motor we see that it is merely to alter the
amount of compounding. Hence, to consider
this in the case of Fig. 9, would mean to
change it to a diagram like Fig. 11, with a
small amount of compounding, the "pure
shunt excitation" being only a hypothetical
condition.
The magnetizing current of the regulating
motor has so far been neglected. Neglecting
regulating motor saturation, this is propor-
tional to and in phase with A D' of Fig. 11. As
it flows through the armature and compensat-
ing windings of the regulating motor only, its
reactance drop can be added to the com-
pounding just as was done in Fig. 11a at
/ /' and its resistance drop, proportional
to B E can be added to the resistance drop
of B E. Thus, we still would get our circle
diagram. However, it does not usually pay
to consider so small an element except as an
interesting theoretical consideration.
The effect of the inclusion of the main motor
resistance drop in the voltage applied to the
regulating motor field of Figs. 8 and 9 may be
040 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No.
treated similarly, providing we confine our-
selves to operation so far from synchronism
and at such a range of loads, that Si is a fairly
small part of So, in which case the component
of rotation e.m.f. caused by the resistance
drop is approximately proportional to the
current components.
[Fig. 12. Neutralized Three-phase Shunt A-C. Commutator Machine and
Connections for Adjustable-speed Control of Induction Motor
for Operation Both Above and Below Synchronism
Further the accuracy of the diagram devel-
oped so far hinges on the assiunption that the
values of 5 (distance from synchronous speed)
are so great as to cause the variations in the
relative values of the resistance and react-
ance drops of the shunt field circuit to be
relatively small, which will mean a small
variation in H G oi Fig. !), since variations in
the phase relation of field current and "total
induced e.m.f." A D' means variations in
angle y. As the resistance drop is larger
and larger compared to the reactance drop
the smaller the slip and frequency, it there-
fore appears that Figs. 9 and 1 1 are accurate
only at fairly large values of slip, and small
ratios of resistance drop to reactance drop in
the field circuit, becoming inaccurate as
synchronism is approached. Consideration
of these effects has lead the writer to the use
of a constant voltage frequency changer
and adjustable resistance for overcoming and
regulating the resistance drop of the field
circuit, and of an auto-transformer with taps
(and alternative devices) for overcoming
the reactance drop, leading in turn to a
feasible way of regulating the main motor
through and abo\-e its s>Tichronous speed
as well as below.
Double-range (All Speeds Above or Below Syn-
chronism) Speed and Power Factor Control by
Means of a Constant-speed Shunt Commutator
Motor
Se\-eral advantages of regulating the main
motor speed above as well as below its
synchronous value appear at once. The
capacit}- of the regulating set for a given
maximum speed variation and maximum
speed is reduced 50 per cent, provided the
synchronous speed of the main motor is
half way between the extrem.es. For if, Smax,
Smin and Ss, represent the maximum, mini-
mum and synchronous speeds of the main
motor and H Pmax be the horse power capacity
at speed S„ax, we ha\-e for single range, —
Si = Smax and capacity of set is : —
H P,„ = H P„a, X ^"""^ ~ ^"^
^max
Now for double range, as above, we have:—
H Ps,t = H PsX
H Ps y5wBI-
But as
H Ps
H P„
Sn
H Ps., = 1
■max
O U p
Ss
we have,
showing that the capacity of the double-range
set is one half that of the single range. Thus,
not only will the first cost be materially less,
but the machine losses will also be greatly
decreased.
A second important advantage is that the
synchronous speed of the main motor is in the
middle of the speed range, so that often times
many processes may be carried out running
as plain induction motor with the set shut
down, with consequent saving of wear and
tear on it.
The apparatus, shown in Fig. 1 2. is the same
as Fig. S, except that instead of going to a
star point the ends of shunt field coils Fi. F.,
Fz are carried through the adjustable resist-
ance M. to the frequency changer H mounted
upon the shaft and wound for the same num-
ber of poles as main motor .4. This machine
has a single primar\- winding connected to a
POWER-FACTOR CONTROL OF LARGE INDUCTION MOTORS
641
commutator exactly as in the armature of a
d-c. machine, and has collector rinj^s tapped
in at points 120 electrical degrees ai)art (for
three-phase power) . The secondary is a smooth
laminated ring without windings which may
or mav not rotate with the primary. Ob^•iously
a "revolving field" is set up in this machine,
which at standstill, rotates at synchronous
speed of .4 and H. With 120 electrical degrees
brush spacing on the commutator, we get
three-phase full frequency voltage of the
same value as we apply to the collectors
neglecting machine drop, and the phase
relating between the commutator and col-
lector currents depends upon the position
of the brushes on the commutator. Assume
A to rotate synchronously in opposite direc-
tion to the rotation of flux of H. which carries
said flux backward mechanically at the same
rate that it is turning electrically, leaving it
stationary in space, and permitting H to pro-
duce direct current at commutator like a
synchronous con\'erter.
Thus, it is seen that H is automatically
a source of constant voltage at exact slip
ring frequency.
If we regulate A at no-load (for simplicity)
we see that the rotation e.m.f. of D, hence
both its flux and field current are proportional
Fig. 13. Circle Diagram of Induction Motor Running Below
Synchronism with Regulating Machine Receiving
Constant Shunt Excitation Without No-load
Power Factor Improvement
to slip 5. Hence the reactance drop compon-
ent of the impedance drop of the field circuit,
being proportional frequency as well as flux
is proportional to 5^ while the resistance drop
is merely proportional to the field current and
to 5. By connecting to taps of B whose
distance from the star point is proportional
to 5, we get a voltage proportional to s^,
since the total e.m.f. of B is itself proportional
to s. By changing taps on resistance M so
that the entire resistance of the circuit is
proportional to 1 /s, we just permit constant
Fig. 14. Circle Diagram of Induction Motor Running Below
Synchronism with Regulating Machine Receiving
Constant Shunt Excitation. Such as to Give
No-load Power Factor Improvement
voltage frequency changer H to supply the
resistance drop balancing e.m.f. while auto-
transformer B furnishes reactance drop balanc-
ing e.m.f. In practice, one set of switches can
be arranged to vary both M and B simul-
taneously.
With M operating at a considerable distance
from synchronous speed, the field resistance
drop can be exactly balanced for a given load
by H, so that as B supplies the reactance
drop, the conditions previously assumed are
attained. From Fig. 9 it will be noted that
the phase of A F alters with load, while that of
the voltage from H in Fig. 12 remains fixed.
This only introduces a comparatively small
discrepancy for working loads, the main
effect being a slight alteration of the load slip.
Let us now, on the other hand, consider the
case of running near synchronism, where the
reactance drop of the field, varying as 5^, and
very nearly balanced by B has become
practically ineffective. Fig. 13 is the simplest
circle diagram for these conditions, the con-
stant excitation of D, Fig. 12, from the fre-
quency changer H being so chosen that, in
642 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 7
Fig. 13, rotational e.m.f. H A is perpendicular
to line ABC found as usual. A F is the total
induced e.m.f. of main motor secondary, F H
the resistance drop in main secondary circuit
for secondary current B D' .
So
H .4 sin a AB X sin a
Fig. 15. Circle Diagram of Induction Motor Running Below
Synchronism with Regulating Machine Compoxmd-
excited, the Shunt Excitation Being Constant and
Having No-load Power Factor Improvement,
and the Series Excitation Yielding E.M.F.
90 Deg. Ahead of the Current
Resolve F H into components F C in phase
opposition to .4 F and G H perpendicular
to .4 F and B D' into corresponding com-
ponents B D and D D' , (Point D, thus traces
the circle of the main motor with regulating
motor reactance included, but with ^4 H left
out of the circuit)
GH = H A sin a
G H HA sin a
FH= .
sin 6 sm 5
B D = A 5 sin a
B D' =
Further, B D' =
BD AB
sm a
cos 5
F H
cos 6
r2+fXsin6
cos 5
And
tan =
HA.
.X-
1
A B' n+c
As H A, A B and r^+c are constant,
Angle 5 must be constant.
Angle B D' A is 90 deg. + 5, hence D' traces
a circle.
In Fig. 14, we have given the excitation, and
hence, the rotation voltage A H s. shift /J from
its position in Fig. 13, so as to improve the
power factor.
Resolve the resistance drop F H' , of the
secondars' current B' D' {A, B' and C being
the usual fixed points) into H' H perpendic-
ular to H A perpendicular to .4 B. and F H,
corresponding current components being B
B' and B D'. Since H' A and angle /3 are
constant, H and B are fixed points. Now
resoh-e F H into F G along .4 F and G H per-
pendicular to .4 F, also B D' into correspond-
ing components B D and D D' . D' may now
be shown to trace arc of circle B D' A as in
Fig. 13. We note the power factor and pull-
out torque are better than for Fig. 13.
As we are considering operation at rather
small values of slip where the shunt excitation
is all from H and M of Fig. 12 and is not
effected by B, we can compound by the use
of plain series windings as with d-c. machines
without interference by transformer action.
In Fig. 15, we have shown the effect of such
compounding, brought about by making
changes in the neutralizing winding G, C., Cj.
It can be demonstrated that a pol\-phase
armature turning above its own s^-nchronous
speed in the field, set up by its own reaction
generates a rotation voltage leading the
current by 90 deg. Hence, by weakening
C\ Ci Ci, we can get this sort of compounding
in any desired degree.
.4 B' C denote the usual fixed points of the
main motor with reactance of regulating
motor included. H' A is the fixed excitation
from frequenc>- changer and rheostat and
B' B is running light secondani- current, due
to H' I. B' D being load current considered.
..4 F is the total secondar\- induced e.m.f.
The secondar^' current B' D consists of a
constant component B' B and variable com-
ponent B D. I H' is the constant resistance
drop and H /.the constant leading rotation
e.m.f. due to component B' B. making H
a fixed point. G H is resistance drop and
F (r leading rotation e.m.f. of variable
component B D. Now the resistance drop and
the variable rotation e.m.f. are each pro-
POWER-FACTOR CONTROL OF LARGE INDUCTION MOTORS
643
portional to the current (the iron of the
regulating motor is always at low densities
for these near synchronism conditions) hence,
angle H F G =£• is constant. Since H is a.
fixed point, angle H F K = 6 can be shown
Fig. 16. Diagram of Induction Motor Running Loaded at
Synchronous Speed
The same conditions are represented as infinitesimally below
and as infinitesimally above synchronism.
to be constant as in Fig. 13. Angle A" F G
= 0 = (;+5 = constant. Angle B D A \5 also
equal to d, since .4 D is perpendicular to .4 F
and B D is perpendicular to F G. So D traces
are of a circle, passing through .4 and B.
It will be seen then that the power factor
and pull-out torque can be improved as well
as the speed regulated by this method, when
regulating near synchronism, as well as
remote therefrom. When we regulate the
speed, we so adjust the taps of B, Fig. 12, as to
get the desired percentage of slip 'voltage
from Fi F2 F3 to overcome the reactance drop
and then so adjust resistance M that the
field current corresponding to the desired
conditions will have a resistance drop equal
to the voltage supplied by H. As the field
current is about constant over the working
range of loads we can thus get an even better
approxim.ation to Figs. 9 and 11 than without
H and M. As we regulate the speed, we thus
transfer gradually from the condition of Figs.
9 and 11 to those of Figs. 13, 14 and 15.
We have drawn Fig. 16 to examine the
phenomenon of regulating the speed of the
main motor while loaded from an infinitesimal
amount below synchronism to an infinitesimal
amount above synchronism. As the slip is
negligible, the total induced e.m.f. is also
negligible and the rotational e.m.f. of the
regulating set A H just supplies the resistance
drop H A , the main motor being assumed to
be a trifle below synchronism. Let us now
assume it to be an infinitesimal amount above
synchronism.
All vectors are referred to the secondary
whose phase rotation has been reversed.
although the physical conditions in the
motor remain unchanged. If we select the
phase of .4 C as the phase of reference for both
phase rotations, then the components of all
vectors in phase with it will not be altered by
reversal of the phase rotation, but the quad-
rature components of all vectors will be
reversed, as a vector which would not reach
its maxim_um until 90 deg. after A C will, in
reversed phase rotation, reach its 90 deg.
ahead of .4 C. This law yields us H' A D' B C
to represent the same phenomena in terms of
reversed phase rotation as are shown by
H, A, D, B, C with original phase rotation in
the secondary.
In Fig. 17, D is a load point with motor
nearer synchronism than its natural slip,
as the rotation e.m.f. of the regulating motor
H A has been reversed so as to have a large
component in phase with the total induced
e.m.f., .4 F. The bulk of the resistance drop
F H is, therefore, supplied by ff .4, so that
.4 F and consequently the slip are reduced.
In these conditions the motor would pass
through and above synchronisin as the load
dropped off.
Let us now increase H A until the main
motor runs above synchronism (with reversed
Fig. 17. Circle Diagram of Induction Motor and Constant-
excitation Regulating Machine with One Value of Exci-
tation Such That Speed for the Load Point Shown
is Nearer Synchronism Than the Natural Slip
Value and Another Value of Excitation
Such That Speed for the Load Point
Shown is Above Synchronism
secondary phase rotation). As H A was in
quadrature to A C, the line of the phase of
reference, its new A^alue will be shown with
reversed direction at H' A. Load point D'
is above line ^4 C for motor torque for the
same reason. The total induced e.in.f. would
644 July, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 7
• also be represented with its quadrature com-
ponent above A C instead of below, but its
direction is actualh^ reversed, as shown at
^4 F' , since at any instant any given con-
ductor now cuts the flux in the opposite
direction.
Fig. 18. Circle Diagrams of Induction Motor Running Below
Synchronism and Above Synchronism with the Same
Characteristics. Controlled in Each Case by a Com-
pound Constant-speed Regulating Machine
Note. — For the develi)|)ment of the circle diayram. particular
mention should be made uf the works of Behrend. Blondcl .ind
Arnold-LaCour. Meyer- Dclms has also written concerning
what the writer has termed Single-Range Regulation.
We note that with no initial quadratiu-e or
power factor component in H A, and H' A,
the motor characteristic when running above
synchronism is better than below in respect
to power factor and maximum torque, while
for the generator characteristic the converse
is true.
Dcuble-range (Either Above or Below Synchronism)
Operation Remote from Synchronism
In Fig. IS, we represent operation both
above and below synchronism, with the
sam-e speed — torque and speed — power factor
conditions. The configuration indicated by
the plain letters is for using a compound com-
mutator motor similar to that of Fig. 1 1
except that angle a has been decreased so
that the compounding is m.ostly in the way of
power factor improvem.ent and adds ver\^
little to the slip. The primed letters indicate
the relations for operating above synchronism
and angle .4 D' E' equals can be shown to be
constant just as in the case of angle .4 D B'.
Keeping the phase of .4 C as the phase of
reference, we note as before that the repre-
sentation in secondan- terms with reversed
phase rotation requires the reversal of the
components in quadrature with .4 C of all
vectors, and as .4 F' is further actually
reversed in passing above synchronism, we
see that its quadrature components are still
in phase with that of .4 F. But as the regulat-
ing rrachine must furnish power to the main
motor secondary in order to satisfy the con-
senation of energy, the total current B D'
must ha\-e a component in phase with the
rotation e.m.f. /' .4, which we see is the case,
thus requiring that /' .4 be larger than .4 F'
fulfilling the condition that the regulating
macliine fimction as a generator.
Fig. 1<). Speed-regulating Set for a 1600 h. p. Motor
TWO DOLLARS PER YEAR
TWENTY CENTS PER COPY
GENERAL ELECTFIC
REVIEW
VOL. XXIII, No. 8
Published by
General Electric Company's Publication Bureau,
Schenectady. N. Y.
AUGUST, 1920
One 35,000-kw. and Three 20.000-kw. Curtis Turbine Generator Units in the River Station of the
Buffalo General Electric Company. The stable parallel operation of high-power
stations is discussed by Dr. Steinmetz in this issue
For
Fractional H. P. Motors
THERE is but one true measure ot value
— the capacity tor service. It is a recog-
nition ot this tact by responsible manufacturers
of high speed electrical machines which has
made "NORfflfl" precision bearings the accepted
standard in the equipment thev build — ma-
chines subject to hard, dav-after-dav driving
service under severe conditions which exact
the utmost trom the bearings. And "NORmfl"
serviceability is helping these machines main-
tain their nation-wide reputation tor rcliabilitv.
See that your Motors
are "NQRmfl" Equipped
Ball, Rollei-, Thrust and CombinafiQn Bearings
General Electric Review
A MONTHLY MAGAZINE FOR ENGINEERS
.. ., o „,^„ „,. ,^uv- T, Tir^i^-T.-T^',- Associate Editors. B. M. EOFF and E. C. SAXDERS
Manager. M. P. RICE Editor JOHN R. HEW ETT t *-i. / aj .• • n -kit t^^m^t?
■'In Charge of Advertising, B. M. EOFF
Subscription Rates: United States and Mexico, $2.00 per year; Canada, $2.25 per year; Foreign. $2.50 per year; payable in
advance. Library and Student Rotes: United States and Mexico, $1.50 per year; Canada, $1.75 per year; Foreign, $2.00 per year;
payable in advance.
Remit by post-office or express money orders, bank checks, or drafts, made payable to the General Electric Review, Sche-
nectady, N. Y.
Entered as second-class matter, March 26. 1912. at the post office at Schenectady, N. Y., under the Act of March, 1879.
Vol. XXIII, No. S *,.c.„^.TS?;;,fco«/-a„.v August, 1920
CONTENTS Page
Frontispiece: The White Mazda Lamp in Display Illumination 646
Editorial : Control and vStability of Stations in Parallel 647
Commercial Statistics and Their Value to the Executive 648
By G. P. Baldwin
Flywheel Effect for Synchronous Motors Connected to Reciprocating Compressors 6.53
By R. E. DoHERTY
Melbourne Suburban Electrification, Australia 6(J2
By W. D. Bearce
A New Short-circuit Calculating Table <J69
By W. \V. Lewis
The Production and Measurement of High Vacua
Part III: Alethods for the Production of Low Pressures (i72
By Dr. Saul Dushman
Five Thousand Horse Power Electrically Operated Pumping Plant 684
By E. Bachman and W. J. Delehanty
Power Control and Stability of Electric Generating Stations — Part I 688
By Dr. C. P. Steinmetz
The Penetration of Iron by Hydrogen 702
By T. S. Fuller
The White Mazda Lamp 711
By Earl A. Anderson
The Reward for Efficiency 715
By E. O. Edgerton
In Memoriam: George Allan Woollev 719
GENERAL ELECTRIC
REVIEW
CONTROL AND STABILITY OF STATIONS IN PARALLEL
We are today well entered upon the ful-
fillment of the prediction that our steadily
increasing national demand for power will be
satisfied by electric power generation, trans-
mission, and distribution. In this develop-
ment, from the small electric light station of a
few horse power capacity — still within the
memory of our generation — to our present
generating systems ha\-ing over half a rr illion
horse power machine capacity, problem after
problem had to be solved; old problems,
which worried the central station man of
a generation ago, vanished, but new problems
and difficulties arose in their stead, and som.e-
times these were practically the antithesis
of their predecessors. For instance, in the
early^ days of lighting, foremost attention
was given to attaining good inherent regula-
tion, that is, constancy of voltage under great
and sudden variations of load. This problem
vanished when our station capacities rose
into the hundred thousands of kilowatts,
and in its place arose the reverse problem, the
limiting of the amount of power that can
accidentally be concentrated at any point in
the system, and the reduction of its destruc-
tiveness. The need for such protection is
apparent when it is considered that a system
having a capacity of several htmdred thou-
sand kilowatts, but no power-limiting equip-
ment, may concentrate several million kilo-
volt-amperes at a fault — a power as large as
that of Niagara. With the growth of the
central station, experience with the increasing
destructiveness of short circuits forcibly im-
pressed upon the engineer the need for the
solution of this problem.
It was solved ten years ago by the intro-
duction of power-limiting reactors, in gen-
erators, busbars, tie lines, and feeders.
Power-limiting reactors ha"i.'e come into gen-
eral use in high-power s"ystem_s, and have made
it possible 'i'a increase indefinitely the size
of the system without any increase of danger
from power concentration.
These ten years experience with the use of
power-limiting reactors have proved their
value in limiting the effect of short circuits
and other troubles; as a result cf this experi-
ence, we are able now to determine the ]Droper
and most economical \ alue of povver-li^riting
reactors for the various ser > ice conditions
with far greater certainty than when their
use was first introduced. At the same time,
experience has shown a number of efi^ects of
the use of power-limiting reactors, which
were not contemplated ten years ago, some
advantageous, som.e disadvantageous, par-
ticularly in their effect on synchronous
operation. Since all our modem high-power
generating equipm^ent is of the synchronous
type, as is also much of our power lead, it is
essential that stable operation of these
machines be maintained under all conditions,
even such abnormal ones as short circuits.
The effect which power-limiting reactors exert
on synchronous operation, the limitations
with regard to power, spe^d, stability, etc.,
thus are worthy of extensive study. The im-
portance of this was forcibly impressed upon
engineers by troubles connected with syn-
chronous operation, experienced by high-
power systems and referred to in Dr. Stein-
metz's article in this issue. It can be seen
that reactors, in limiting the destructive
short-circuit power which rray flow across
them from station bus to station bus, also
may limit the synchronizing power between
the stations, where this has to pass over the
reactors, and thus under certain conditions
may reduce the stability. On the other hand,
power-limiting reactors, by limiting the power
flow at times of accident, tend to localize
the voltage drop, to maintain higher voltage
on the station busbars, and to give a more
rapid voltage recovery after short circiiit,
any of which will increase the synchronizing
power and thus tend toward an increase of
stability.
Thus these mattters are of great im.por-
tance and are being studied by the most
prominent engineers, with the view of adjust-
ing the distribution characteristics of large
stations operating in parallel, for the purpose
of securing with safety a more economical
and reliable supply of power. In this issue
of our magazine and in the one following,
we are presenting a comprehensive analysis
and discussion of the subject by Dr. Stein-
metz.
648 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 8
Commercial Statistics and Their Value to the
Executive
By G. P. B.\LD\viN- •
District Manager Philadelphia Office, General Electric Company
This article is developed along the following line of thought: The executive of today has giown up with
his job and therefore has a first-hand and intimate knowledge of its workings; the executive of tomorrow
cannot be prepared according to this program, and therefore for guidance in his management will have to
depend mainly upon reports prepared for him by his subordinates; the figures of these reports comprise
Commercial Statistics, from which special compilations will reveal the status of any matter under considera-
tion; statistics may be presented in figures but usually are amenable to a more ready conception bv a graphic
type of display, such as a map, chart, curve, etc. — Editor.
It has been said that a good executive is
a man who decides quickly and is sometimes
right. If asked the reason why he gives a
certain decision, that same executive would
reply that it is based on his best judgment.
Judgment ranges in degree from prejudice
to profound reason depending upon the
proportions possessed by the man who
makes the decision.
As a rule our present executives are
extremeh' fortunate in that they have
grown up with their jobs. They have grown
bigger as their jobs grew bigger. They have
had experience in various departments, have
been close obser\^ers of the operations of
others, and have seen the trend towards a
greater degree of specialization, a greater
division of labor, and a greater consolidation
of executive direction.
It is questionable whether the executive
of the future can be made by this process for
business has assumed proportions that practi-
calh- make it impossible. The executive of
the future will be compelled to resort to
statistics, which will be prepared in shape
and form for his instantaneous tise; and his
ability to analyze them will be one of his
qualifications as a good cxectitive.
All concerns whose magnitude is great and
whose product is diversified are susceptible
to the application of a great amount of
statistical data, and these statistics can be
built up and summarized into a few that arc
vital. For example, of the statistics that
guide the directors of various de]5artmcnts.
the few that represent the gist are extracted
to compile the statistical report of the single
manager of those departments and so on up
the scale until the collective activity of the
entire organization is co-ordinated and rep-
resented by a single sheet of abbreviated
statistics.
CLASSIFICATION OF STATISTICS
Product, Orders and Expenses, and Customers
Electrical sales department statistics for
the use of the executive readily segregate
themselves into three classes: those that
have to do with the product, with the orders
and expenses, and with the customers.
We may further state that the product is
susceptible to further segregation; into large
apparatus, small and fractional horse power
motors, supplies, renewal parts, and lamps.
Expenses segregate themselves into depart-
ments: sales department, order department,
credit dej)artmcnt, collection dcpartm'^nt.
etc.; and customers into classes: industrial
customers, lighting customers, railway cus-
tomers, and merchandise customers. It is
possible to get through the use of these
segregations a large number of combinations.
For instance, a segregation of renewal parts
into customer's classifications will show where
that type of business is coming from. The
activities of the sales department when
segregated into components will show what
its branches produce along their special lines.
It's possible to combine these features into
any number of combinations which arc of
infinite use. It would bo interesting to know
for instance what the clerical expense would
be in connection with an apparatus order, a
supply order, and a lamp order when sold
to an industrial customer, a lighting customer,
or a merchandise customer. Orders, volume,
and expenses can be combined in any
amount of detail.
Units and ratios are derived from the
foregoing classifications. We should bear in
mind however that it would not be altogether
fair to compare cost-of-sales ratios among
districts because the mechanics of selling
is aft'ected bv manv factors. The three
COMMERCIAL STATISTICS AND THEIR VALUE TO THE EXECUTIVE 049
U m
f £
P O
O -^
.9 J
CO
650 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. S
Umps
\
y
""■■■■■
;
Public t
Mercha
d-sci^—
S«MX.es
/
/
\
,,-
\J.
r^/
"N
^-
:
\
Repair R rts
■
.......
/--^i
^-^ f^
Fig. 5.
Curves Plotted to Represent the Monthly Trend in the Sales of Three Commolities
Segregated with Respect to the Classification of the Purchasers
principal ones are: the electrical population,
the application of electricity to the pre-
dominant industries, and the geographic area.
Electrical Population
The electrical population (being the popu-
lation sen,'ed by central stations) varies from
94 to 21 per cent in different states. The
total population of the State of New York
according to the last census was about 00
per cent greater than that of the States of
Pennsylvania, Virginia, West Virginia, Mar\'-
land, Delaware, and North Carolina. The
electrical population was about the same.
There are approximately 390 central stations
in the State of New York and OTO in the other
territory just mentioned. The geographic
area of these states is syi times as great as
that of New York; therefore, to get a given
volume of sales from them (assuming that the
organizations are equally effective), there are
more customers to receive attention, more
territory to be co\"ered. and more relations
to be established. It will require more people
in the sales organization and consequently
more expenses.
Electrified Industries
Data as to the source of the orders furnish
an extremely interesting sheet and this will
show that the largest volume comes from the
predominating industries. It will also show
some enlightening comparisons between
supply and apparatus sales. It will demon-
strate that supply sales do not vary in direct
proportion to apparatus business; in fact, as
a territory becomes more thickly settled and
more industrial business becomes established,
the proportion of supply business from the
industrials as against the total will be
constantly accelerating. One may therefore
draw the conclusion that in a sparsely
settled territory the supply business will
come from utilities rather than from
DEPT
Oept
Apr,
»v
-i-' ,
DtfT.
=tpr-
-^^
*f
H.J,
t'
B**T.
■r«
Mn
r,o
H.r
*<>'
•«W
Fig. 6. A Graphical Record of the Departmental Expensrs of a Sales OrganixatiDn
COMMERCIAL STATISTICS AND THEIR VALUE TO THE EXECUTIVE (351
industrials; and as the population grows
and industrials become established, business
with them increases. This is a fact that
probably everyone who has looked into the
subject knows and realizes, but very few
realize its proportions, and it is possible to
realize these proportions only by the examina-
tion of statistics.
It is interesting to note some of the ideas
of supply salesmen in this connection. If
one were to ask how much of their time they
spent on industrials, on utilities, and on
merchandise customers, the answers would
vary, but an average of the replies given
by twelve experienced men was that they
spent about .50 per cent on utilities, 25 per
cent on industrials, and 25 per cent on
merchandise customers. However, an exact
check on their calls showed that they spent
necessary to keep the manufacturer's name
constantly and favorably before the trade.
The number of orders received by one
district as against another would show a
yearly comparison that can be listed against
the number of industries and the number of
customers, and can be developed to show a
unit that will indicate the progress made.
Incidentally, if the number of orders is
compared with the number of quotations it
would bring to mind that the number of
quotations made in large order business will
compare rather favorably with the number
of orders received; but with small order
business the number of orders received will
greatly exceed the number of quotations, as
much in some cases as 300 or 400 per cent;
while with a specialized commodity like
lamps there is an enormous volume of sales
SaJ.i
Feb. Mar. Apr. May
Feb. Mar Apr. May
L-Pens .& ^^_^,
Jan, Feb Mar Apr.
May
Fig. 7. Statistics of the Number of Orders Taken, the Amount of Sales Made, and the Expense
Incurred by a Sales Office as Plotted for the Information of Its Manager
75 per cent of their time on utilities, 10 per
cent on industrials, and 15 per cent on
merchandise customers. Here, then, was an
enormous supply business from industrial
customers being received with comparatively
very little direct supply solicitation.
Whether solicitation is necessary, and to
what extent, is of course another question.
To answer this question, statistics according
to classification of product would become
involved. However, they would probably
show that direct solicitation was necessary
in specialties and special application, and
that specialties and special ap]3lication for
industrials were comparatively few as com-
pared with those used by the utilities and
merchandise customers. They would prob-
ably also show what is necessary for the
further development of industrial business,
particularly in small orders, and that it is
built up with practically no special quota-
tions. In other words, quotations largely
disappear in small order business and that
element of cost is eliminated.
Geographic Area
If we would take a given territory and
make three special maps, one of which we
will call a central station map, another an
electrified community map, and another an
industrial map, the combination would indi-
cate rather clearly the target of the sales
department. Typical maps of this character
are shown in Figs. 2, .3, and 4. In the territory
illustrated there are 250 central stations,
abotit 400 electrical communities, and over
1000 industrials. Such facts as these are not
shown by the ordinary geographical map.
Fig. 1, and therefore it would not occur to
anvone that in that territorv there are located
652 August, 192)
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 8
industries of such magnitude and vokimc as
the special maps indicate. It may even be
further stated that should one pass through
this territory on a train it wouldn't occur to
him that six of the largest individual plants
of their type in the world are located in this
section; at the same time the territory is
comparatively isolated. It is not in what we
would call a metropolitan district, and the
electrical population is about 21 per cent of
the total.
Application of Charts and Curves to Present Statis-
tics
An immense amount of thought and
time has been devoted during the last few
years to charts and curves; to graphic
methods of illustrating facts. The eye
is more receptive to a cur\'e than to a
mass of figures. More information can be
presented on a given sheet, fluctuations are
immediately obser\'ed, and tendencies are
indicated at a glance. The cun"es in Fig. 5
show the source and changes in local lamp,
supply, and renewal part sales. From such
curves as these the eye can read the situation
at a glance, and on comparison with earlier
curves can readily observe the trend from
year to year. Fig. 6 indicates the con-
tributing expense of a sales organization;
the various clerical departmental expenses.
It will show that they are increasing in some
lines and not in others. Fig. 7 presents cur\'es
of orders, sales, and expenses for a local office,
showing that the orders are remaining about
the same, but have a tendency to decrease
in the last few months, so is the volume of
sales, but the expenses are increasing.
Cun^es of this character are of inestimable
value to an executive. He obtains a knowledge
of the situation at a glance. If he has data of
all his districts on a like basis he can prepare
for his own use a chart which will show his
total expenses, and when conspicuous varia-
tions are noted he can by means of these
segregated cur\-es trace the changes to the
exact point where they occurred. If orders
are decreasing and volume is increasing it
should mean comparatively less clerical
expense, and the executive immediately has
a reference" on which to check his opinion.
There is no question but that among his
other qualifications the executive of the
future must be able to assimilate and digest
statistics and direct accordingly: and, obvi-
ously, management must take its cue from
the auditor, the accountant, the statistician,
or from some one who is specifically charged
with this duty.
The executive cannot proceed without
the auditor, nor the auditor without the
executive for that matter; but perhaps it is
well to remember that none of us can pro-
ceed without the customer. Therefore, in
the application and use of statistics, the
human element must not be forgotten.
If wc arc going to know the value of our
relations individually, collectively, and com-
paratively and what they produce we cannot
neglect the value of statistics.
6.53
Flywheel Effect for Synchronous Motors Connected
to Reciprocating Compressors
B\- R. E. UOHERTY
AlTERNATINC-CI RRENT ENGINEERING DEPARTMENT, GENERAL ELECTRIC COMPANY
When driving reciprocating machines by synchronous motors it is important to avoid a condition that
will produce hunting, or a periodic oscillation of the revolving element ahead of and behind the normal posi-
tion. In cases where the natural oscillating frequency of the motor is near the frequency of the predominat-
ing impulses of the reciprocating machine, hunting may occur of such amplitude to throw the motor out of
step. The remedy lies in the selection of the correct flywheel. This whole subject is very clearly explained
by Mr. Doherty with the aid of mechanical analogies, and a general method for determining the proper fly-
wheel effect for any given set of conditions is outlined. The performance of some installations is very puzzling
and troublesome, and this article will prove of great assistance in attacking problems of this sort. — Editor.
The purpose of this paper is to show why
flj-wheels are necessar\^ on synchronous mo-
tors which drive reciprocating compressors,
and to outline the method of determining
the proper flywheel weight.
For convenience in analyzing the problem
I shall divide the discussion into two parts.
torque consumed, the speed m^ust be constant.
This m.eans that at any given instant the
rotating mem.ber of the m.otor has a definite
position in space. To illustrate this point,
consider an analogy. In Fig. 1, two trains
are running side by side at constant speed
and abreast of each other. At anv instant
Air Compressor Driven by a Synchronous Motor. The Flywheel for Effecting Stable Operation
is Shown to the Right of the Synchronous Motor
dealing first with the m.echanical aspects and
later with the electrical. I do not wish to
impart the idea that they are entirely sep-
arate matters, for they are not, as I shall
presently explain, but it is convenient to
treat them- separately.
When the torque developed by a syrichro-
nous m_otor is just equal at all tim-Cs to the
of time these two trains will occupy a definite
position along the track and will retrain
abreast of each other. Under such conditions
the driving force of the trains is exactly
equal to the opposing forces of friction,
windage, etc. The position thus held m.ay
in the analogv' be tenned the "stable posi-
tion" and corresponds to the space position
654 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 8
Fig. 1.
of the motor referred to. It is important to
hold in your mind the idea of this stable
position as the necessar\- reference to which
varying motion may be referred.
Referring again to Fig. 1, assume that
trains "A" and "B" are running abreast
of each other and at constant speed,
and that at an instant designated by
O the driving force on .4 is instantly
increased by an amount indicated in
the figure, and that this excess force,
existing for a time t, is suddenly re-
moved. Let us inquire how this
affects the velocity and relative dis-
placement of trains .4 and B.
In Fig. la, the excess force, which
is consumed entirely in the accelera-
tion of the train, can also, if drawn
to proper scale, represent the acceler-
ation of the train. Therefore the area
under the cun.-e, representing the
product of acceleration and time,
represents the accumulated excess
velocity. Plotting in Fig. lb ordi-
nates which represent the accumu-
lated area in a will give a cur^-e of
the excess velocity at any instant
during the time t. Likewise, since
the area in Fig. lb is the product of
velocity and time, ordinates representing
the accumulated area at any instant, when
plotted in c, give the curve of excess dis-
placement of train A ahead of train B.
That is, at the end of time t train B will
have reached >«, whereas train .4 will have
gone further, that is, to point n. The
excess displacem.ent mn is represented by
the maximum ordinate at the end of cur\-e c.
The idea here is that it is possible to find the
change in both velocity and displacem.ent
from the cur\-e of unbalanced force impressed
upon a body simply by integrating the area
of the loops in the unbalanced force diagram.
The first integration gives velocity, the second
gives displacement.
Carr^'ing the illustration further, assume
that instead of the momentar\' excess force
described, a periodically var>-ing force is ap-
plied. That is, at equal inter\-als of time
first an excess force and then an insuffcient
force is applied to train .4 such that the
resulting average velocity of train .4 is the
same as the constant velocity of train B.
In other words, train .4 will alternately and
at equal successive inter\-als have an excess
and insufficient driving force impressed upon
it. This is represented graphically in Fitr.
Id. Since the area of the first excess loop at
any instant represents added velocity, it
follows that at the end of the first interval
the velocity must be a maximum, and that
likewise at the end of the second inter\-al
the velocity must be a minimum.. The re-
sulting cun'e of velocitv will therefore be
.-„w^- „««,«* TviiinniDiiinni \ Totoforeo rgprKsanispnxfuct.occ«fanai/onXtim*
A -Max. excess nJocity
OrdinoUs excess
displace fnent^
Totof area represents excess enptocemanl.
e-Aero^ te/oc/tir"
yeJocitij of B —^
f-^lyero^e or stable
rmtiofi'postt ton ofB^ ■'
Illustroting theeffecl ofo variable applitd fore* and o constant forctcansumibyload
Diagram Showing How Periodic Relative Displacement of Two Trains
is Derived from Curve of Unbalanced Driving Force
as shown in e. By the same process the dis-
placement cur\-e showing the position of
train .4 first ahead and then behind the
stable position is shown in /.
One part of the problem of determining
the proper fl\-\vheel effect for synchronous
motors direct -connected to reciprocating com-
pressors is illustrated in this way, viz.; that
it is required to find what mass train .4 must
have in order to limit to a definite amount
the plus or minus displacement as shown in
/ when a given periodic variation in the
driving force is impressed upon the train .4.
The idea of angular deviation of the s>-n-
chronous motor can now be approached from
a more direct point of view. Consider two
duplicate synchronous machines, one driving
an air comjircssor, the other driving a blower.
The latter will have constant speed; the for-
mer, on account of the pulsating torque of
the compressor, will have a pulsating speed.
Suppose that reference arrows are placed on
corresponding poles of these two machines.
In this case, just as in the case of two trains,
the variable torque on the motor which
drives the compressor will cause it to be
dis]ilaccd first ahead and then behind the
position of the refea-nce arrow on the rthcr
motor.
FLYWHEEL EFFECT FOR SYNCHRONOUS MOTORS
655
For electrical reasons, explained later, it
is necessary to set limits to this displacement
or deviation. The larger the number of poles
on the motor the smaller correspondingly
must be the lim.it. That is, in the case of a
60-pole mooter, the allowable angular dis-
placement is one half of that for a 30-pole
motor, because the allowable value is a func-
tion of the electrical, not mechanical angle.
Therefore, the problem is to determine how
much flywheel is required to keep the angular
displacement within given required limits.
piston base. Curve / shows the forces in c
developed on the piston base. These forces
on the piston are converted to tangential
forces on the crank pin by multiplying the
different ordinates of curve / by correspond-
ing ordinates in curve e. The resulting curve,
shown in g, is the turning effort curve or tan-
gential force on the crank pin. Most compres-
sors are, of course, double acting. This simple
case was taken merely for an illustration.
Fig. 3a shows the tangential force of an
actual double acting machine. By integrating
^
' ♦>
m-+-
J
a
/
ft
\
x
X
?
4
Anqulor
\
/
■'Ton
gent lal Effort I\
'/icceJeration
V
X
\Zer6 Line
;
s
0
30\
60
30
izo\ no
leo
210
210
270
300
330
360
>
1
1
1
1
-
Average Tangential Effort i5 Developed bt^ J_
Motor and Consumed bij Load. Hence it i
Involves No Anqular Acceleration. f~
, , Unbalanced rurning Effort
-
FlLi^jheel Effect , [
%
1
1
r >
1
/
\
1
\
1
\/plnri/,jA.
'
/
^
'
t
\
b
\
>
/
\
\
/
/
0
\
30\
60
\3Q
120
150
, ■ ■ ■
210 \ 240
llOf
50a\ 330
5
\
1
t
1
'
J
\
\
/
\
■L'
/
J
S /
!
(
s
!
r
Maximum
Excess Dis
}
\
\
C
plac
"" -'
\
\
\
/
V
i^
/
\
/
0
70 60
\30 1201
ISO
ISO
210
Md\
ho\
300
330360
\
y
4nqu/arDi5f
/
\
]
j
•^
>
\
/
y
!
/
^ii-
Fig. 2. Method of Determining Tangential Effort on
Crank Pin from the Indicator Card
and Inertia Forces
I- 3. Velocity and Displacement Curves as
Derived by Successive Integrations of
Tangential Effort Curve
This is easily worked out fro:n the turning
effort diagram of the compressor.
The process of obtaining the turning effort
curve for a reciprocating compressor is illus-
trated in Fig. 2. For simplicity, the case of
a single crank single acting com„pressor is
chosen, a represents the inertia forces, that
is, the forces required to stop and start the
reciprocating m.asses. Positive forces are
taken as those tending to m.aintain motion,
negative as those tending to retard motion.
b shows the forces taken from the indicator
card, that is, the forces acting on the piston.
c shows the sum of a and b, all plotted on the
the areas under the plus and mJnus loops of
this curve in the m_anner as described in
Fig. 1, the velocity curve is obtained as shown
in Fig. 3b. Integrating the loops of this curve
in like manner gives the ordinates of the dis-
placement curve, as shown in Fig. 3c. Thus
it is possible, with a given flywheel eff'ect in
the motor, and a given tangential curve as
shown above, to determine accurately what
angular displacement would occur under the
assumed conditions; or conversely, what
flywheel effect would be required to limit
the angular displacement to a definite
value.
656 August, 1920
GEXERA.L ELECTRIC REVIEW
Vol. XXIII, Xo. 8
Inasmuch as the integration of the loops
of the crank effort and velocity cur\'es by
planimeter is a ven,- laborious process it is
rarely resorted to. Yet any method that
does not in>-olve a proper integration of the
entire curve is at best an approximation.
We have found the analytical m.ethod to oe
very useful. This gi\-es the same results as
the graphical integration. It consists in
expressing the unbalanced component of the
crank effort diagram, which to proper scale
is the acceleration diagram, into a Fourier's
series of sines and cosines, and integrating
this equation twice for displacement. Thus
the unbalanced crank effort or torque can
be expressed as,
T= (oi sin ift + Oo sin 2 u.i + as sin
Swt + . . . + bi cos Ti't + ^2 cos
2wt +b3 cosji Ti't)
Where ii' = 2 irnt
n = revolutions of crank per second
i = time
The angular acceleration is
T
Where /= moment of inertia of rotor.
♦This scheme is described by H. C. Lehn. Gener.kl Electric
Review. March, 1915.
Thus the velocity equation is
r= I a dt=^^ ( — ai cos ui— -^
cos 2u't— — cos 3 li'i — ... +61
o
sin u.i + -■ sin 2 ui + :^ sinSii-t-rf
And the angular displacement in m.echan-
ical radians is
6= I V di= — — - I —Qi sin u-t — -^
»' U'-l \ 4
5IH 2wt— TT sin .3 «.•<+ . . . — ^i
y
cos •Wt—-r COS 2u!t — - COSSwt— }
The coefKcients ai 02,03.... 61, bo. tz
are obtained from an analysis of the crank
effort curve by the use of a schedule shown
in Fig. 4.
By proper substitutions the maximum
angular displacement in electrical degrees
becomes,
10' K frZ„
S» WR'-
A' = lb. on crank pin per inch scale of
crank effort cur\'e.
AhlALYSIS OFCRA/SK EFFO/erCU/PV^C
y
0, 'a.s,nt.jt
e. cos u r • <3, 3
rOfS Z 57-f?OKC CYCLC
n£l^r.o^cos^^^r•a^s^nJi/^r■'ff^co33
of ci-anic ettor* cv^m m .-T<A»f s
^r'<x, sirfa
wf-ft,C05«e»r
a, 1 b.
a^
^?^
a^
t?^
a^
t>^
wC 1 V
■-*-._* ,s^,^ 'PC—*-, ncwr
T%^c^ f^e^
1 *a»»A vtio^^
r'..-«-J fA^.PMr
i<»«-i
ol
- 0 * : -
•\ioo& * ; ■
* 0
./^ * -
-^ 1
'
.
/COO
.'T'''
* JOOO
f^
*
MB i
.
,9*5
'.SOO ;
•iH
•
707
,
7*77
■ "^
•' *»; 1
SO
■SOO 1
.
«**
'.0t^
'H
.
1000
,
0
-
• «•*
- SOO
'
^s
.
70J
707
-iOOO
.1™
,
TOT
.
.3W7
• <?
•]tooo
€>0
:*«j
• JOO
• se^
-\soo
,
0
_
/OOO
- M«|
- JW 11
7^
• Msi
• zsa
* SOO
'VB**
.
.707
_
707
' #•«.
• aeo
90
./cod
• 0
* 0
-\ooc
-
tooa
_
0
- 0 '
*'JOOO,
/OJ
.
M9
'Zsa
- SOO.
-
107
•
T07
:m^
.!»>'
IZO
.
0M
-.SOO ^
- tf «*|
-\soo
.
0
«
fOOO
• a»m
- JDOJ
I3S
li
r^
- 707
-JOOO\
0
•
.W7
«
707
• 0
-iooet,
t^O
r'w
■ «A«
' «««,
.
SOO
iooa
,
0
-1*^
- soa\
y&S
- zsa
- 9*3
- SOO
,
0A»
.
wr
.
.Ttrr
-?••*
-i joej
tao
. 0 , '
■ taoo
- 0
»JOOO
0
- '4000 1
.,«J
/»Ji
- ^s9
-»^
. SOO
- -#••!
.
w/i
- 7WT,
• #«*
ZIO
■ ,joa
-.«*«
• «««•
• SO€k
-.«>•■ '
.
0 '
''#•«'
-\SOO\
zzs
■ 70T
■ T07'
-/«»;
• 0
- 707
*
-Krrl i
-1 » 1
-l.<«>|
zao
- St*-
-.JOO
• ««•
- SO€
.
«
<
too^ 1
■H
-|.jJ
zss
- s*i
- zsa
• SOO
-^«M
•
jwr
•
«» 1
-W
.1*3
Z70
-t(>oa
- 0 ;
• 0
-yooo
.
/OOtf
• "-!
- >
•i"H
za^
- j*rf
'.Z3a
' SOO j
.
TOT
-'.Tor
•'•H
• IJDOl
300
"r***
• SOO
- 0»«
-\soo
,
0
-|«>oo
-:w 1
3'S
.[Z
■ TOT
'JOOO 1
-1 0
.
lai
-l-TiyT'
•; « 1 1
-»«»
330
-xa
•.a«*
-.*"■, 1
'Isoo
.
tooo
-lo 1
-l#*«j
- 309,
3^^
-,.?«'
. r>4Mj
- SOO
>!«*«
-1 yor
• -OT '
.#•*
.i3
ry-
3V/T. 1
j>^^
Svrr-
5w»t
Jwm 1
J.« 1
»**»
S-*— 1
V^
''Ji- '
-c
=.-.f.-
^..^^~.
"••-w"-
«.-*^^-
...-V*
0,-^,— -
e.,.,^Y«..
<...-««-
»---v— •
S ' fftif per ntm /('•lbo^c'~3i^'t^ei^perirtcH9€omafcraritt^0rtcvrY^ ^J^*t> ^poirstrfOe'r^fxrfar i^'' Jbfa/ lb fi' r- Civ^yt p*nro^4VO ^■^ ^
Fig. 4.
Schedule for Determination of Coefficients of Fourier's Series Expressing
the Tangential Effort as a Function of Ti.-ne
FLYWHEEL EFFECT FOR SYNCHRONOUS MOTORS
()57
/' = frequency of supply \-()lta^'e in cycles
per sec.
r = crank pin radius in feet.
Zm = maximum value of bracket quantity
in equation of displacement, that
is, rraximmn value of the series.
S = r.p.m.
n'i?- = total flywheel effect in lb. ft.-
It is not my purpose, as intimated at the
outset, to explain in detail the derivation and
application of formulas, because that is a
long story in itself, but rather to outline
what the problem, is and show the nature of
the solution.
The foregoing considerations cover what
I have chosen to call the mechanical aspects
of the problem, and describe one part of what
seems to be the most practical method of
getting at the complete solution. There is an
additional part which will now be discussed
as the electrical aspects of the problem.
As already stated, there are really not
two separate phenom.ena involved. There
is only one, as shall presently be shown; but
as a matter of convenience both in presenting
the problem and in the practical solution of
any giv-en case, it is desirable to consider that
there are separate parts. In the first, the
effect of the electro-magnetic force (usually
called the synchronizing force or torque)
is neglected. In the second this force is
taken into account.
Consider the character of this force. Have
you ever watched an electric motor whirling
in s]3ace, receiving energy into the stationary
member and deli\'ering that energy to the
rotating m.em.ber, with no m.echanical con-
nection between the two members — have
you ever watched such a motor and wondered
just what was going on in the small air space
between the members, and also what was
the character of the m^edium through which
that energy transfer takes place? It is very
much as if one end of invisible rubber bands or
small spiral springs were attached to the poles
on the rotating member, and the other ends
of these elastics were carried circumfer-
entially around the face of the stationary
member at constant speed. At no load there
would be no stretch in the elastic elements,
and the roto^ would maintain a definite stable
position in^Votation, Suppose a constant
load is put on the shaft, the ends of the elastic
elements would nevertheless continue to be
carried around the stator face at an unaltered
speed. Since the other ends are attached to
* Direct proportion'ility ceases at large angles, but holds for
the range considered here.
the poles, the rotor would be pulled along
also at the same speed, but on account of the
stretch in the elastic due to the added load,
the rotor would assume a new and lagging
position in space, such that the forces in the
stretched elastics balance the load. Ob\ iously,
>3/r /Pese r voir Cor nith rahtivett/ small moii_
Air
t-jmprossor
Spring \
TOMJOJ
Motivapower ^elotivcly
vary large iTtasi
CorJI
CorB
rjriS with spring, rapresar^ts a si^nchronous motor dnv/ng on
oir compressor. Ihe moss of the cor corresponds to the flt/wnea/
affect of the motor, the spring represents the magnetic flue
CarB. represents the large ponar sgstam which supplies
foworto the sgnchronous motor.
Fig. 5. Analogy Representing an Air Compressor Driven by a
Synchronous Motor Which is Connected to a
Large Power System
if the rotor is displaced from the no load
position, whether by load or by some mom.en-
tary impulse, there is a force exerted upon the
rotor that is proportional and opposite to the dis-
placement. This is the characteristic of syn-
chronizing force.* The magnetic lines of force
em.anating from, the surface of the poles and
entering the stationary m.em.ber are elastic like .
rubber bands or springs, and beha , e in a simi-
lar manner, and they constitute the medium
through which the energy is transferred from
the stationary to the rotating member.
I wish to emphasize another mechanical
analogy as a fixture to which subsequent ideas
may be attached. Fig. 5 is intended to show
a heavy railway car which runs at constant
speed regardless of what else happens. It
draws, through a spring connection, a smaller
and lighter car which has a drive wheel con-
nected to an air compressor. That is, the
source of energy for the work of compressing
air is the large car. The variable force ex-
erted on the drive-wheel by the ccm.pressor
is obviously reflected in the spring tension.
As an analogy, the large m.assi . e car running
at constant speed represents the power supply
system to which the motor is connected. The
sm.all car with spring represents the syn-
chronous motor; the mass of the car corre-
sponds to the flywheel effect; and the spring
to the magnetic lines of force. The entire
sm.all car equipment represents a synchro-
nous m.otor driving an air compressor. The
speed of the drive w-heel corresponds to the
speed of the motor. The analogy is prac-
tically complete. As load is added the spring
stretches; the car assumes a new, lagging
position with respect to the large car. The
variable torque exerted on the drive wheel
658 August, 1920
GENER.AL ELECTRIC REVIEW
Vol. XXIII, Xo. S
by the compressor produces periodic dis-
placement altemateh' ahead and behind the
average, or stable position. If it were pos-
sible to add more and more load the strain or
stretch in the spring would increase, until
ultimately the elastic lim.it would be reached,
the spring would give way, and the car would
stop. Just so w ith the m_otor. If the load is
increased sufficiently, whether momentarily
or gradually, the elastic limit of the m_agnetic
lines of force is reached and the m.otor breaks
out of step.
The speed of the large car corresponds to
the voltage, the spring tension to the electric
current and the product of the two obviously
corresponds to the power. Thus the variable
load, as represented by the var\-ing spring
tension, is manifested at the switchboard by
swinging meter needles. It is beginning to be
clear, therefore, why angular displacem,ent
must be limited. It is a question of setting
a limit to the pulsation of power which we
are willing to accept. This will be discussed
later on.
Another extremely im.portant point to
be drawn from this analogy is that if a sudden
change in load occurs it will be attended by
an oscillation, and the oscillation nill be at a
definite frequency. This follows from the
characteristic of synchronizing forces already
explained. Suppose, for convenience of
illustration, that both cars are stationary';
that the connecting rod of the compressor
is temporarily disconnected from the drive
wheel; and that the small car is pulled back
from the large car, stretching the spring,
and is then released. Ob\ious]y the result
will be an oscillation of the small car with
respect to the large car, and, as in the case of
a pendulum, there will be a definite number
of oscillations per minute. The same thing
would happen even if the cars were traveling
forward. Imagine, then, the two cars to
be nmning along the track at, say fifteen
miles an hour, and the sm.all car to be oscil-
lating or "hunting" with respect to the large
one; you will then have a physical conception
of the hunting of a synchronous motor.
Moreover, if the drive-wheel is now connected
to the compressor, bringing its variable torque
into action, you will have the disturbing in-
fluence which, under certain tmhappy condi-
tions can cause excessive hunting or surging.
For instance, a sudden change in load starts
an oscillation which as already stated occurs
at a definite frequency. Sujjpose that each
time the small car swings back and forth in
( oscillation relative to the large car, it re-
ceives, in perfect time harmony, impulses
from the compressor which are in a direction
to amplify the swing. Little by little the
oscillation will build up to large amplitude,
that is, will become violent. These conditions
exist if the natural oscillating frequency
(oscillations per minute) is exactly equal to
the revolutions per minute of the drive-
wheel. If, however, the impulses occur either
faster or slowor than the oscillations, so that
now and then the impulse acts against, instead
of with, the oscillation, the result will obvi-
ously be less violent oscillation. And as the
frequency of impulse, that is, the speed of the
dri .e-wheel is made to differ m.ore and more
from the oscillating frequency, the amplitude
of the oscillation becomes less and less.
Although by the m.ethod described in Fig. 1
(neglecting the effect of the sprinp) the rela-
tive displacem.ent of the small car \\ ith respect
to the large one may have been calculated as.
let us say, two inches, we nevertheless find
that at resonance between the impulses and
natural oscillations the actual displacement
becomes many times larger; also, that as the
difference between the impulses and oscil-
lations becomes greater the displacement
becomes less, ultimately approaching the
value calculated as in Fig. 1 (assuming that
the unbalanced forces acted on a free mass,
that is, neglecting the effect of the spring).
All of this applies to a synchronous motor
driving a reciprocating comjjressor.
The characteristic that synchronizing force,
like spring tension, is proportional and oppo-
site to the displacement of the rotor from the
no load position, leads us to the ven,- con-
venient fact that the motion of the rotor in
os(:illatit)n must be harmonic. Hence the
well known fonnula for the period of simple
harmonic motion can be applied, and the
natural oscillating frequency for any com-
bination of synchronizing force and flywheel
effect can be calculated. This formula will
be given later.
Such oscillations of the small car as arc
described above can be limited to some extent
by the use of a damping device, such as a dash
pot. For instance, in Fig. ."). if an adequate
dashpot were jjlaced between the two cars so
that any relative movement bet\y,een the cars
would involve a change in the daslipot piston,
the free swings or oscillations (that is. those
occurring at the natural frequency) would be
dam])ed out. leaving only the increased dis-
placement occurring at each im])luse. Bui
even this displacement, although less than
that without dashpot. is many times that
FLYWHEEL EFFECT FOR SYNCHRONOUS .MOTORS
659
which would occur if the natural oscillating
frequency were considerably different from
the impulse frequency. Such a dashpot
represents the amortisseur winding in the pole
face of a synchronous motor in its action of
damping out oscillations.
Fig. 6 illustrates this condition. It shows
the angular displacement for different values
of flywheel effect, first, in the dotted line,
neglecting the effect of synchronizing force,
and second, in the full line cur^e, including
the effect of synchronizing force, and assum-
ing the free oscillations to be damped out.
Since, as intimated above, the oscillating
frequency is a function of synchronizing
nizing force. The beginning of this curve is
interesting. Suppose the small car in Fig. 5
had zero mass, then whatever instantane-
ous variations occur in the torque on the
drive-wheel would be impressed in full and
in phase upon the spring; but there is nothing
impossible about, this except the zero mass.
I mean that the periodic displacement from
the av-erage position would obviously not be
infinite, but a value determined by the con-
stant of the spring. Thus, if the momentary
excess torque is, say 50 per cent of normal,
then with zero mass the spring at that instant
would si^nply be stretched 50 per cent more
than at normal torque. Thus, in Fig. 0,
U5 M
I
Cscilt
/^ngu/ar Dispfocement
^eg/eci/riQ Synchroniz in^ Force
r
r^
Linjtt of /ingu/ar
DisplocemenC for
Success fly/ Operation
2000 \ \ 4000 1 MOO I XOO
^eiosocisoaa 3io xo
Includes Effect
of SunchronuinQ Force
lOOpO 12000 leOOO \l6O00 'isOOOIbft'
Z50 200 Periodsperi^n.
6. Curve Showing Approximate Effect of Synchronising Force Upon
Angular Displacement for Different Values of Flywheel Effect.
Drawn for an Assumed Compressor Unit of 250 r.p.m.
Synchronizing Force Corresponds to 10-kw. per
Electrical Degree Displacement
force and flywheel eft'ect, it is possible to plot
the natural oscillating frequency, as well as
flywheel eft'ect, as abscissae. The curve as-
sumes a 250-r.p.m. compressor connected to
a synchronous motor whose synchronizing
force corresponds to 10 kw. per electrical
degree displacement. You will obser\'e two
peaks of displacement, much in excess of the
dotted line values.
These correspond to a natural oscillating
frequency equal respectively to the revolu-
tions of the crank, that is, 2.50 per minute,
and to twice the revolutions, or 500 per
minute. Beyond 15,000 W'B? and between
the peaks, the displacement is about the same
as the calculated value, neglecting synchro-
starting at zero flywheel efl'ect, there is a
definite periodic displacement determined by
the momentarv lead and the synchronizing
force. Adding W'R- reduces the displacement
until, by approaching a critical natural oscil-
lating frequency, the displacement is again
increased, passing through a peak at reso-
nance— and so on. Obsen-e that when a
difference of 20 to 25 per cent exists between
the im.pulse frequency and the natural oscil-
lating frequency, say at 16,000 W'B? or at
6000 W'R-, the displacement approaches the
values as calculated, neglecting synchroniz-
ing force.
Thus in going over a whole range of values
of ]^7?^ we find that the problem is really one
660 August, 1020
CxENERAL ELECTRIC REVIEW
Vol. XXIII, No. 8
connected story, not two; yet at the same
time the story itself suggests a division for
practical calculation. It is this:
(1) Calculate from the crank effort curve
the value of \VR- required to limit
angular displacement to a given
value, neglecting the effect of
synchronizing force.
(2) Then see if this WR' causes the natural
oscillating frequency to fall in a
critical range, that is, within 20
per cent of the revolutions, thus
causing greatly increased displace-
ment. If so, the WR must be in-
creased until the required differ-
ence is obtained.
For reasons discussed later, the present
limit for angular displacement which we
consider necessar>' for satisfactory operation
is plus or minus 3.5 electrical degrees.* For
instance, in Fig. 6 the calculated WR re-
quired to limit the displacement to plus or
minus 3.5 electrical degrees is, by dotted
curve 10,500. But this value would cause a
natural frequency, as noted on abscissa, equal
to about 250, that is, equal to the revolutions
per minute. Hence the U'R- must be in-
creased to about 15,000.
The formula for calculating the oscillating
frequency is simple, but the determination
of one of the factors in it is a very difficult
matter. It has been shown that the char-
acteristic of synchronizing force is, for our
consideration, the same as that of a spring.
That is, this force on the rotor at any angular
displacement is proportional and opposite to
the displacement. This is the definition of
harmonic motion. Hence the well known
formula for the period of harmonic motion
can be used. Thus, the period is
T = '2t^I - seconds
where.
/ = moment of inertia
(T = ratio of torque to angular
displacement.
With proper substitution, the formula for
natural frequency of oscillation becomes,
^_ 3.5200 jp;;~f
'' -RPMyjvrR'- periods per mm.
* To reduce this to mechanical degrees divide by one half the
number of poles.
where,
Po = factor depending upon syn-
chronizing force.
/"= frequency of supply voltage
in cycles per second.
ir7<= = fl3-wheel effect on lb. ft.*
For motors with uniform airgap the factor
Po is easily determined from calculation or
test. But in salient, or definite pole machines,
which constitute practically all that are now
built, this factor is extremely difficult to cal-
culate and can be determined experimentally
only by very elaborate tests. An investiga-
tion extending over some three years has
given a method of calculating Po, and nec-
essan.- tests have been made to confirm the
calculation. Hence, although the calculation
is rather involved, it is nevertheless worth-
while, if we would keep out of trouble.
I have outlined the factors which are in-
volved in the determination of the proper
flywheel effect ; or in other words, have stated
the problem, and also have indicated roughly
the method which has been found to be the
most practical one for solving any particular
case. I shall now touch upon some of the
more general aspects of the problem.
The problems for those of us to answer
who are interested in the electrical side
of this problem (which includes the cus-
tomer) is this: What value of power pulsa-
tion are we willing to accept as reasonable"'
This is the equivalent of asking: What
periodic angular displacement are we willing
to allow? To answer these questions we must
make inquiry as to what harm such power
pulsations produce. The first thing that
naturally comes to our mind is the disagreeable
feeling which seizes us when we see an am-
meter or wattmeter needle surging across
the scale, a feeling that tells us at each new
pulsation that something serious is certainly
going to happen at the next swing. But it
never happens, at least rarely happens. I
have ne^er seen a motor "kick" out of syn-
chronism as a result of surging or hunting.
although, of course, such a thing is i^ossible.
The jjoint is, it is an extremely rare occurrence.
We therefore face this interesting point: If
we have to look at the ammeter to tell
whether there is trouble, can we iustly call
that trouble? Moreover, the complete ab-
surdity of the point is brought out by the
fact that ammeters can be made, and are
made and used, which will not follow the
pulsations. The current may vary 100 per
cent from the average, vet the ammeter mav
FLYWHEEL EFFECT FOR SYNCHRONOUS MOTORS
661
show only 5 per cent. Hence, in such cases
meter swinging in itself surely cannot be
taken as the cause of any contention. Con-
versely, and this is very important, we must not
assume that just because the ammeter is quiet
there are no power pulsations. We must
therefore drop this point as of no value in
determining a limit for angular displacement.
Ho\ve\-er, there are good reasons for setting
a limit. The principal one is the vibration
of certain parts of the motor caused by the
pulsation, and all that the vibration leaves
in its wake. I think it is true that almost
any machine, of whatever manufacture,
designed for normal service would ultimately
fail if subjected to incessant, serious vibration.
The failure may be either mechanical or
electrical. Parts which are mechanically
strong enough and adequately held for normal
service, may, when subjected to such vibra-
tion, gradually work loose. On the other
hand, the pulsating torque, exerted on the
annature coils, may cause them to work
loose in the slot. Breathing of the projecting
portion of these coils will cause chafing of
the insulation, and ultim.ate failure. Even
if the projecting portions were bound firmly
to supporting rings (which construction on
the slow speed motors in question is not
required for any other reasons) the continued
reversal of stress in the insulation, just as
in metal, would gradually produce fatigue,
in this case both m.echanical and dielectric
fatigue, with the result that the life of the
coil would be seriously shortened.
Serious pulsation also lowers the efficiency
and causes unnecessary heating of the arma-
ture coils. A pulsating current, delivering
the same average power as a steady current,
will cause more loss in the windings than the
latter. Moreover, the oscillations cause loss
in the arr.ortisseur or darr.ping winding in
the pole face. Another point is that the
pulsating current taken from the line iray
cause corresponding fiucttiations in the line
voltage, if the impedance of the supply lines
is appreciable. This, however, is not a
serious matter if the motor is connected to a
large power supply system through low im-
pedance feeders.
What, then, shall we call the limit? In
the early days 2.5 electrical degrees, plus or
Fig. 7. 32-pole Rotor for Synchronous Motor Showing
Construction and Amortisseur Winding
n'inus, was established. Step at a time, and
guided by experience, we have raised that
limit to 3.5 degrees. We know that with the
average run of machines one degree deviation
will cause from 3 to 6 per cent of full load
current; hence the above limit means a
pulsation of about plus — m_inus 10 to 20 per
cent, a total variation of from. 20 to 40 per
cent. Present experience indicates that this
limit should not be exceeded. It is, however,
entirely a matter of experience, which the
future mav modify-
662 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 8
Melbourne Suburban Electrification, Australia*
By W. D. Bearce
R.\iLW.\y .\ND Traction Engineering Department, General Electric Company
Before selecting the equipment to be used in electrifying the extensive suburban steam lines out of Mel-
bourne, Australia, a careful analysis was made with the result that high-voltage direct-current equipment was
decided upon in view of the fact that this system of electrification would not entail so great an initial outlay
and annual expense as would the single-phase system. The analysis is of especial weight because of the magni-
tude of this suburban enterprise — one involving the electrification of 336 single-track miles (153 route miles),
the equipping of 400 motor cars and 400 trailer cars, and the installation of 15 substations having a total
ultimate capacity of 81,000 kw. Mr. Bearce has, in the following article, summarized the principal features
of the electrification. — Editcr.
The electrification of the Victorian Railway
lines radiating from the City of Melbourne,
Australia, is the most extensive suburban
steam road conversion in the world. The
project was initiated in 1913 by the State of
Victoria, which operates all of the steam roads
in this State. The decision to electrify was
followed by exhaustive study of the available
systems by Merz and AIcLellan. Consulting
Engineers, and a decision to use 1500-volt
direct current as the most economical both in
first cost and cost of operation. Estimates of
the relative cost of 1.500-volt direct-current
and 11,000-volt single-phase equipment were
published in December, 1912, by the Consult-
ing Engineers after a thorough analysis of
proposals submitted by twenty different
manufacturers. These figures included the
cost of power plant, transmission, substations,
overhead distribution, rolling stock and
alterations to existing equipment.
The following figures, taken from this
report, show the relative cost of installation
and annual cost of operation (including power,
maintenance, and interest charges).
Initial Cost
Annual Expense.
Direct Current ; Single Phase
$11,636,964
1,235,286
$14,857,137
1,578,241
The recommendations of the General
Electric Companv as to the general features
of the electrification were accepted in practi-
cally every respect, and the motor and control
equipment of the 400 motor cars is of standard
General Electric design.
General Features
The Melbourne Suburban System consists
of approximately 336 single-track miles in-
cluding parallel tracks and sidings. A larger
part of the system is equipped with double
♦Abstracted from a series of articles in The Commonwealth
Engineer (Melbourne), by E. P. Grove, in charRe of the installa-
tion for Mertz & McLellan. Consulting Engineers.
track and there are some four and six-track
lines, distributed as shown in Table I. This
table also indicates a total of 153 miles of route.
TABLE I
MILES OF ROUTE AND SINGLE TRACK
MELBOURNE SUBURBAN RAILWAYS.
Miles
Single Track
Ba.sis
Miles
Route
6 track 18.90
4 track 21.44
3 track 6.69
2 track 220.00
1 track 33.00
Sidings 36.00
3.15
5.36
2.23
110.00
33.00
Total , 336.03 153.74
The standard construction consists of ICO-
Ib. T section rails double spiked to untreated
ties 0 ft. by 10 in. by 5 in. and spaced 2 ft. 10 in.
centers with 1 ft. S in. spacing at joints. The
road bed is rock ballasted to a depth of 15 in.
and the tracls arc located 11 ft. S in. apart.
The track gauge is 5 ft. 3 in. and the cur\ature
is limited to a maximum of approximately
10 deg. which occurs on the Flinders Street
viaduct. The line between Sandringham and
Broadmcado^^s. which includes the initial
electrically operated line, may be taken as
typical of the system. This branch contains
a maximum grade of 2 ])er cent, and for a
distance of approximately 9 miles the average
grade is O.S.") per cent. The maximum speed
allowable on this section is 52 m.p.h. with
slow downs on some of the curves. Practically
all of the curves have a 150-ft. easement
approach.
The contracts for electrification which were
made in 1913 contem])latcd the electric
operation of this entire suburban district.
Owing to the precipitation of the European
War actual constmction was seriously handi-
capped, and the official opening took place
May 2S, 1919. The first electrical operation
MELBOURNE SUBURBAN ELECTRIFICATION. AUSTRALIA
663
inckidcd the section of line between Sand-
ringham and Essendon. This line extends
approximately due north from the Flinders
Street terminal to Essendon and south to
Sandringham. With the exception of the
portions around the terminal station, the line
is practically all double track.
The population of the city of Melbourne,
including suburbs, according to the 1912
census, was approximately 70(1, ()(l(). There
is a large outlving residential district which is
supplied with frequent and high-speed train
service, handling a very heavy suburban
traffic. Electrification was adopted in pref-
erence to the construction of additional lines
and parallel tracks to handle the rapidly
increasing business.
Instead of locomotives and trailing pas-
senger coaches, the entire passenger traffic
will be handled by multiple-unit motor-car
trains. At present no provision is being made
for handling freight traffic; and steam loco-
t
ViCT'Ji.AN Railways.*
M- OF suat;Ra-.N system.
-e-
Fi^. 1. Map of Electrified Lines
664 August, 1920
GENER.\L ELECTRIC REVIEW
Vol. XXIII. Xo. S
motives will be employed for this purpose
until the passenger service has been fully
taken care of.
Rolling Stock
The suburban trains on the Melbourne
system are made up of units each consisting
Fig. 2. Multiple Unit Passenger Train on 1500-volt Electric Zone
of a motor car and a trail car which can be
operated from either end. The normal six-
car train will thus be made up of three-units
(three motor cars and three trailers). The
initial orders for rolling stoc'v included the
equipment for 400 motor cars and 400 trail
cars adapted to operate either from the motor
car or trailer. About 4.5 per cent of these cars
are of the compartment type with swing doors,
the remainder being of a combination type
with sliding doors and cross seats, and a
corridor running the full length of the pas-
senger section. A larger part of these cars
are partitioned into sections to form smoking
and non-smoking compartments. The seating
accommodation and weights of the cars are as
given in Table II.
TABLE II
The number of cars at present equipped for
electric traction is as shown in Table III.
Electrical Equipment
The electrical equipment for the 400 com-
plete motor coaches and 400 trailer coaches
was furnished by the General Electric Com-
pany and consisted briefly of the
following :
Sixteen hundred GE-237. 140-
h.p. 7r>0 1500-volt ventilated
motors with gears and pinions.
Four hundred current-collect-
ing equipments including sliding
pantograph with devices for
raising and lowering and acces-
sories.
Four hundred Sprague GE
type M control equipments in-
cluding circuit breakers, con-
uctors, reversers and master
Mntrol equipment.
Four hundred auxiliary equip-
ments including dynamotor, air
compressor, auxiliary devices for
control and lighting circuits.
Each motor car is equipped
with four GE-237 motors, the
motors on each car being con-
nected two in series for operation
on the 1.5()0-volt trolley. The
field coil windings are provided with taps so
that 20 per cent of the field can be cut out
Fig 3. GE.237 Ventilatei Railway Motor
Seats
2U0U-lb.
Tons
Equipped
Sliding door motor car
Swing door motor car. . ...
84
80/70
84
80/70
94
90
53.57
52.41
Sliding door driving trailer
Swing door driving trailer
Sliding door non-driving trailer.
Swing door non-driving trailer . .
30.88
29.43
30.69
28.48
TABLE III
Swing-door
Type
Sliding-
door Type
193
29
161
Toul
Mot
Dri\
ors ... .
ing trai
-driving
crs
164
27
359
56
Non
trailers
126
287
317
385 I
702
MELBOURNE SUBURBAN ELECTRIFICATION, AUSTRALIA
665
automatically for high-speed running. The
gear ratio is 7-1/23, giving maximum speed
of approximately 52 miles per hour on level
tangent track.
The collector is of the pantograph type
having two sliding pan shoes. These shoes are
ing current for each motor coach, can be
collected without sparking. The working
range is from 14 ft. (i in. to 21 ft. 6 in.
The pantograph is raised by admitting
compressed air to the working cylinder
mounted on the base of the collector. The
Fig. 4. 1500-volt Contactor Box for Type M Control
spring supported and free to move independ-
ently. Each shoe has two strips of contact
copper about 2 in. wide and i^ in. thick, so
arranged as to form a pan between them
which is filled with graphite grease. The
contact strips are replaceable, and the design
is such that the greatest amount of surface
is provided in the center where the maximum
i^^^^SR^
Fig. 5.
One of the 1500-kw.. 1500-volt Synchronous Converters
Installed in Jolimont Substation
running occurs. The over-all width of the
contact surface is 45 in. The upward pressure
of the pantograph on the contact wire is
approximately 25 lb. and with a reasonably
clean wire a current of considerably more
than 500 amp., the maximum normal operat-
valves controlling this cylinder are of the
electro-pneumatic type, mounted near the
master controller, and they can be remotely
controlled from the leading cab so that panto-
graphs can be raised or lowered on all cars
simultaneously.
Control current at ap]3roximatcly 750 volts
is supplied by the dynamotor for operation of
the various control circuits and for lighting
accessories. The main contactors are actuated
from the master controller containing a
single-contact cylinder with stationary fingers.
The action of the contactors is automatic, the
master controller having only four points
forward and two reverse. These points are
1 nown as switching, series, parallel lap, and
parallel. The switching and lap positions
include resistances in circuit. Each car is
provided with an automatic line circuit
breaker, which is also used to make and break
current to the motors, being the last switch
to close the motor circuit and the first to open.
When tripped by overload, this switch is
reset by an electro-magnet controlled from the
motorman's cab.
The reverser is of the drum type, also
electrically operated and is controlled by two
electro-magnets, one for each position. This
switch is connected to reverse the fields of the
motors. Train acceleration is entirely auto-
matic, being controlled by current-limiting
relays which insure the completion of each
step before the next step is taken.
The air compressor is of the standard
General Electric center-gear type operated
666 August, 1920
GEXER\L ELECTRIC REVIEW
Vol. XXIII. Xo. 8
MELBOURNE SUBURBAN ELECTRIFICATION, AUSTRALIA
667
directly from the L500-volt trolley. It has a
capacity of 25 cu. ft. of free air per minute,
and is controlled automatically by an air
compressor governor operated by an air
cylinder. The normal pressure on the
rcser\-oirs is 100 lb. per sq. in. The air brakes
are of the compressed air type commonly
used in the United States for multiple-unit
train service. Two pipe lines are installed,
one connecting the reservoirs and the other
known as the train pipe line. The air com-
pressor governors are arranged to operate
simultaneously so that the work is unifonnly
divided between the several compressors.
Power Station
The central power station for the system
is located at Newport on the River Yarra.
This location is also adjacent to an arm of
Port Phillip Bay, from which a plentiful
supply of cold water can be obtained for con-
densing purposes.
The electric power is obtained from steam-
turbine generator sets which deliver three-
phasa 25-cycle alternating current at 3300
volts. Provision is made for six 10,0()0-kw.
units operating at 210-lb. steam pressure with
a normal vacuum of '2S^4 inches.
The 3301)-volt three-phase current is step-
ped up through a bank of three single-phase
transformers, for each generating unit, to
20,000 volts. The 20,000-volt feeders are 13
in number, all three-phase lead-covered
armored cable. These feeders are laid under-
ground in trenches in the congested parts of
the district, while overhead transmission lines
TABLE IV
1500-VOLT DIRECT-CURRENT SUBSTATION
EQUIPMENT
Sub tation
No. Units
tJltimate
Size Unit;
Kilowatts
Total
Kilowatts
Tolimont
6
4
3
2
3
3
3
2
2
3
2
3
2
3
3
3000
1500
3000
750
1500
750
3000
750
750
3000
750
3000
750
1500
750
18000
Middle Brighton
Newmarket
Glenroy
Newport
Albion .
6000
9000
1500
4500
2250
North Fitzrov
9000
Reservoir.
1500
Macleod
East Camberwell
1.500
9000
1.500
Caulfield
9000
Springvale
1500
4500
Seaford.
2250
To'-al.
44
81,000
are used on some of the outlaying portions
of the distribution network. Coal is brought
in from Spotswood over a branch line from
the main tracks. There are two boiler houses,
each equipped with 12 boilers of the Babcock
and Wilcox marine type. Coal bunkers are
4
Fig. 10. 1500-volt Direct -current Line Breaker
provided with a total capacity of 3000 tons.
Coal from these bunkers is fed by gravity
chutes direct to the chain grate stokers.
Substation Equipment
The present plan for electrification includes
the construction of 1.') substations for deliver-
ing 1.500-volt direct current to the various
sections of the line. These stations are
located as shown in Table IV.
In addition there are one or two (JOO-volt
substations connected to the transmission line
for supplying local tramway systems. Owing
to the delays caused by war conditions, only
five of these stations have so far been placed
in service. In the Jolimont station four
l.jOO-kw. General Electric units are operating
temporarily, pending the receipt of 3000-kw.
units for this station. These machines are
of the standard commutating-pole type,
starting from high-voltage taps on the
transformers, and wound for 1500 volts
direct current per commutator. Two 1500-
kw. units of English manufacture have also
been installed in the Middle Brighton sub-
station and two 3000-kw. units in the New-
market station. There are also three 1500-
kw. units installed in the Newport sub-
station and similar equipment at North
Fitzroy. The total substation equipment
068 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. S
provided under the plans now drawn includes
44 units with a total capacity of 81,000 kw.
Direct-current Distribution
The 1500-volt current is transmitted to the
motor cars through overhead contact conduct-
ors of the catenary type. Over the main
suburban tracks the normal construction
consists of a 0.25 sq. in. hard-drawn grooved
copper contact wire supported from a stranded
hard-drawn copper cable of either 0.25 sq. in.
or 0.375 sq. in. cross section. This gives an
equivalent sectional area of 0.5 or more
square inches, corresponding to from 650,000
head construction is supported on steel
bridges with anchor structures approximately
3000 ft. apart on which the tensioning device
is located. The intermediate supports are
located at inten-als of 300 feet on tangent
track with somewhat closer spacing on cun-es.
Special construction is necessary- where the
structures are required to carr\- signal
equipment and also on the four and six-track
sections.
The construction on sidings is similar to the
main line, except that a contact conductor of
0.125 sq. in. is used, and a stranded steel
messenger.
Fig. 11. Slider Pantograph Collector
to 800,000 circular mils section. There are no
paralleling feeders other than the catenary
supporting wire. Owing to the poor con-
ducting qualitv of the fexible supporting
droppers which employ a section of link
chain to obtain flexibility, the catenary sup-
port is connected to the contact wire at
inter\-als of about 600 ft. The contact wire
is maintained in constant tension at 25(0 lb.
per sq. in. by cast-iron weights arranged at
each end of sections about 3000 ft. in length.
This scheme is intended to provide for changes
due to temperature variations. The ovcr-
Ftg. 12. Double Track Section in Electric Zone
All of the electrical equipment for the
rolling stock including motors, control and
compressors and auxiliary switching is of
General Electric design. The General
Electric Company has also furnished some of
the substation equipment. The remainder of
the installation has been supplied from firms
in Great Britain and from local Australian
manufacturers. The engineering and design
for the complete system has been carried out
under the direction of Mcrz and McLcllan.
Consulting Engineers of New Castle, England,
under the direction of E. P. Grove.
669
A New Short-circuit Calculating Table
By W. W. Lewis
Power and Mining Engineering Department, General Electric Company
A description of an entirely new development for calculating short-circuit currents in large power networks
was published in the General'Electric Review for October, 1916. This calculating table was later improved
and enlarged, and a description of the table and methods of employing it for the calculation of short-circuit
problems was published in the General Electric Review for February, 1919. The value of this calculator
was at once recognized bv large power companies and a number of them have been built and sold to companies
operating large transmiss'ion systems, although the device was not originally developed as a commercial article.
Further improvement in this calculating tabic is described and illustrated in this article. — Editor.
The short-circuit calculating table and its
use have been described in previous issues of
this magazine.* Since the Original table
was built in 1916, many improvements have
been made to increase its accuracy and
simplify its operation.
A new table has recently been installed
in the Power and Mining Engineering Depart-
ment of the General Electric Company
which,. for the present at least, is the last
one giving a ten per cent setting and the
left-hand one a one per cent setting, so that
it is possible by this means to set accurately
on the nearest per cent reactance. A total
of 104 rheostats are provided, 20 of these
being connected as generators and 84 as
lines. Six of the line rheostats may be con-
verted into generator rheostats by small
switches on the horizontal part of the table.
The network to be studied is set up in min-
1 :r?K:«i^i"r^
B.i.vfi.i ii:
Fig. 1.
Short-circuit Calculating Table with Set-up of
System Shown in Diagram. Fig. 3
Fig. 2. Back View of Short-circuit Calculating Table
Shown in Fig. 1
word in this sort of device. It is well
illustrated in Figs. 1 and 2, which show
^especti^•ely front and rear views. The
rheostats consist of tubular resistance ele-
ments with taps brought out and connected
to buttons on the back of the board. On the
front of the board is an etched dial plate
calibrated in per cent reactance. Each
rheostat has two handles, the right-hand
*Gener.\l Electric Review. Oct.. 1916. p. 901; Feb., 1919.
p. 143.
iature with the assistance of telephone cords,
plugs and jacks.
When the telephone cords are not in use
they are held out of the way by weights,
only the plugs protruding from the front of
the board. Each rheostat is wired to an
ammeter bus and by pressing a button the
current passing through any rheostat may
be read. Thtis the total short-circuit current,
the current giv-en by each generator, and the
670 August, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, Xo. S
67^/
vMT.
iso%
/VW,
■^
—L
■^
22 l7-r^ 6 37.
A/\J\A
/r
wlv'^-
/I
B .^^-L
IWVi
VWV
VWV
Avv^
I TJr. "
r-*-f-Q— *
■XF
o%
§_/^ 119. S 7.
2 «
A/Wv
«):
2<i7.
r. -^|0'*«''-
#/^?i
"tMI-
-o
J7.5%
</.zr.
^'Oli-
^'X
2264%
Base for Reactance tSOOO Hv-a.
X/ndicates Points of Short
Circuit TaHen One at a Time.
Fig. 3. Diagram of Transmission System as Set up on Calculating Tabic and Illustrated in Fig 1
A NEW SHORT-CIRCUIT CALCULATING TABLE
671
Fig. 4. Enlarged Calculating Table Made by Adding
to the Original Table Built in 1916
current passing through each portion of the
circtiit, may be read in a very short time after
the system is set up. A three-position switch
gives three ranges on the ammeter and a
switch is provided for reversing the polarity.
A key is also provided for converting the
ammeter into a voltmeter for reading the
potential of the supply circuit.
The table is designed for operation on a
100- to 125-volt, two-wire, direct-current cir-
cuit. A small lamp over the ammeter serves
to light the instrument and also acts as a
voltage indicator. Switchboard lamps at the
top of the table furnish general ilhmiination.
Fig. 3 shows the diagram of the trans-
mission system which is set up on the cal-
culating table illustrated in Fig. 1 .
Two sizes of tables have been designed,
the one previously described with 10-t rheo-
stats and a smaller one with 50 rheostats.
They can also be made with fixed instead of
adjustable resistors. These are suitable for
some systems whose lines are more or less
stable. The following is a list of the com-
panies and institutions now having cal-
culating tables:
Alabama Power Co.
General Electric Co.
Georgia Institute of Technology
Hydro-Electric Power Commission of
Ontario
New England Power Co.
Public Service Co. of New
Jersev
Turners Falls Pwr. & Elec. Co.
Westinghouse Elec. & Mfg. Co.
The New Jersey, New England
and Ontario Companies have tables
with fixed resistors, the others with
adjustable resistors. All the tables
are modelled more or less on the
original one described in the Octo-
ber, 1!)1(), General Electric Re-
view.
;^ s a m.atter of interest there are
reproduced in Figs. 4 and 5, the
original table of 1916 as enlarged
and rebuilt in 1917, and an im-
proved table built for a power
company in 1918.
It may be readily appreciated
that the labor of solving short-
circuit problems is greatly reduced
by the latest table, and that its
conveniences have materially added
to the speed and accuracy of the
work connected with such prob-
lems.
r
Fig. 5. Improved Calculating Table Built for
a Power Company, 1918
072 August. 1C20
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 8
The Production and Measurement of High Vacua
PART III
METHODS FOR THE PRODUCTION OF LOW PRESSURES— ( Cont'd)
By Dr. Saul Dushmax
Research Laboratory, General Electric Company
The first installment of this series discussed the bearing which the fundamental principles of the kinetic
theory of gases has upon the production and measurement of high vacua. The second installment discuss;d
the fundamental theory of vacuum pumps and described the construction and operation of the mechanical
types of pump. The present section of the article deals with Gaede's diffusion pump, Langmuir's condensa-
tion pump, and others of th? mercury vapor type. An appendix furnishes remarks relative to the care and
operation of exhaust equipment. — Editor.
MERCURY VAPOR PUMPS-"
The fact that a reduction in pressure can be
obtained by a blast of steam or air has been
known and applied in the industn.- for a long
time. In steam aspirators or ejectors such
as are used for producing the low pressure
required in the condenser of a steam turbine,
"the high velocity of the jet of steam causes,
according to hydrodynamical principles, a
lowering of pressure, so that the air to be
exhausted is sucked directly into the jet."
An analysis of the action of the aspirator
shows, according to Langmuir, that in its
action two separate processes are invol-\-ed.
" 1. The process by which the air is drawn
into the jet.
"2. The action of the jet in carr\-ing the
admixed air along into the condensing
chamber.
"The aspirators cease operating at low
pressures because of the failure of the first
of these processes. If air at low pressure
could be made to enter the jet. and if gas
escaping from the jet could be prevented
from passing back into the vessel to be ex-
hausted, then it shovild be possible to con-
struct a jet piunp which would operate even
at the lowest pressures."
This problem has been solved in two
different ways by Gaede and Langmuir.
In the pumps devised by each of these, a
blast of mercur\- carries along the gas to be
exhausted into the condenser (process 2).
In order to introduce gas into the blast
of mercun,', Gaede has used diffusion
through a narrow opening. On the other hand,
Langmuir has made use of the fact that the
"The introductory remarks arc based largely upon the dis-
cussion of this subject by I. Langmuir in his paper on. "The
Condensation Pump. .\n Improved Form of High Vacuum
Pump." General Electric Review. 1916. p. 1060. also Journ.
Franklin Inst., ISi. 719 <1916). Phys. Rev. H. 48 (1916). An
excellent discussion of the mercury vapor pumps described in this
section has also been written by A. Gehrts. Naturwissenschaften.
7. 983 (1919).
" Ann. Phys.. je. .357-392 (1915).
mercury atoms on colliding with the gas
molecules must impart to the latter a portion
of the momentum which they possess in
virtue of their high average kinetic energ\-,
while the mercun." atoms themselves are
removed rapidly from the stream of mixed
gases by condensation on the cooled walls.
Gaede's Diffusion Pump-'
The action of Gaede's "diffusion" pump
can best be illustrated by referring to Fig. 19.
A blast of steam is blown through the tube
AB, in which is fixed a porous diaphragm C.
The ve.ssel to be exhausted is attached at E.
\^
Fig. IQ. Diagram Illustrating Principle of DiFTutioi Pjmp
Water vapor diffuses through the capil-
laries in the diaphragm into the trap D
where it is condensed by some refrigerating
agent, while air diffuses through the dia-
phragm in the opposite direction into the
tube A B, from where it is drawn awav
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
673
rapidlv by the blast of steam. The result
is that the pressure in E decreases and finally
reaches a very low \'alue.
A study of the phenomena of diffusion of gas
through mercun,' \-apor in narrow tubes led
Gaede to the conclusion that at sufficiently low
pressures, where the mean free path L of the
air molecules in mercury vapor is comparable
with the diameter d of the tube, the volume
of air, r, diffusing through a tube of length I,
per unit time, is given by relation of the form :
V = kTrd'/l
where /? is a constant for any given gas. In
other words, the speed of exhaust is independ-
ent of the actual pressure of the gas in the
vessel to be exhausted, and the relative
decrease in the pressure per unit time there-
fore remains constant as the pressure in the
svstem is decreased.
The actual construction of the diffusion
pump is shown in Fig. 20. The porous
diaphragm is replaced by a steel cylinder C
with a narrow slit S whose width can be
altered by means of the set screws H. The
cylinder is set in the mercury trough G,
which forms a seal between the low and high
pressure parts. The mercury at ^4 is heated
and the stream of vapor passes over the slit
in the steel cylinder in the directions indi-
cated by the arrows. The air or other gas
from the system to be exhausted (connected
at F) diffuses into this mercury stream at S,
and then passes out through E into the
forepump which is connected at I'. Any
mercury vapor passing out through 5 is
condensed on the glass in the immediate
neighborhood, by means of the water cooling
jacket A'l A^. The opening V connects
with the fore-pump or other source of rough
vacuum and is used for exhausting the system
until the pressure gets low enough for the
operation of the diffusion pump to become
effective. As soon as this stage is reached
the mercury in the trap automatically closes
this opening and the exhaust there continues
by means of the diffusion pump.
According to Gaede's theory the maximum
speed of the pump is attained when the width
of the slit 5 is of the same order of magnitude
as the mean free path of the gas molecules
in the slit and when the vapor pressure of
the mercury is only slightly in excess of
the pressure in the fore- vacuum (at V).
Consequently the temperature of the mercury
vapor has to be maintained at a fairly con-
stant value. For this purpose a thermometer,
T, is placed inside the tube B.
The effect of varying the temperature of
the vapor (and consequently its pressure) on
the speed of exhaust is shown by the data
(given by Gaede) in Table IX and the plot
of these in Fig. 21. These results were ob-
tained with a slit width of (1.1112 cm. The
Fig. 20. Gaede Diffusion Pump
maximum speed of SO cm.^ per sec. was at-
tained at a temperature of the mercury
vapor of 99° C. At this temperature the
pressure of the mercury vapor is 0.27 mm.,
while the mean free path for air in mercury
at this pressure according to Gaede's calcu-
lation is about 0.023 cm.
Table X (t=10G°) and Table XI (t = 110°)
show the effect of varying the width of
the slit. The noteworthy fact is that the
speed of the pump remains constant as the
674 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. S
pressure in the exhausted system is decreased,
a result which Gaede previously deduced
from theoretical consideration, as mentioned
above.
The great advantage of the diffusion pump
over all the previous types of pump consists
speed and the necessity of carefully regulating
the temperature of the mercur}- vapor.
too
—
so
-
0
Hj bo
u
u
i so
§
- A
\
r
-
^\
ZQ
-
^^^--,,.^
10
-
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 .J_
o z OA oe> OB to I z
Pressure of Mercur t^ in mm.
Fig. 21. Effect of Mercury Vapor Pressure on Speed of Diffu-
sion Pump
TABLE IX
T
P (mm.)
S
T
P
0.72
S
90° C.
0.165
13.4
118.5
51
94
0.20
60
127.5
1.10
38
97
0.24
70
134
1.51
23
99
0.27
80
139
1.84
15
113
0.55
62
143.5
2.2
11
therefore in the fact that there is theoretically
no limit to the degree of vacuum which can
be attained by its operation. In the case of
the Gaede rotary pump and all mechanical
pumps the speed of exhaust decreases with
decrease in pressure. In the case of the Gaede
molecular pump the minimum pressure at-
tainable depends upon the pressure in the
fore-vacuum as the ratio of the i)ressures
is constant for the pmn]). There is thus
with all these pumps a fixed lower limit
to the lowest pressure attainable in the
exhaust system. While there is no such
limitation with the diffusion pump, it does
have the double disadvantage of low exhavist
" I. Langmuir. loc. cit.
TABLE X
TABLE XI
WIDTH OF SLIT =
0.025 CM.
WIDTH OF SLIT =
=0.004 CM.
t
S
P
5
0.025 mm.
77
0.07 mm.
52
0.009
72
0.028
48
0.0025
67
0.006
40
0.0008
72
0.0015
38
0.0002
73
0.0004
41
0.00006
70
0.00007
40
Langmuir's Condensation Pump
Both these disadvantages are removed in
the type of mercur\' vapor pump designed
by langmuir, while the advantages of the
diffusion piunp are retained. In constructing
and operating a pump of this type it occurred
to langmuir that "the limitation of speed
could be removed if some other way could be
found to bring the gas to be exhausted into
the stream of mercurv' vapor. "- As stated in
the introductory section, Langmuir comes to
the conclusion that the ejector pump must
become inoperative at low pressure, since at
these pressures, "according to the kinetic
theon,- of gases, the molecules in a jet of gas,
passing out into a high vacuum must spread
laterally, so that there would be no tendency
for a gas at low pressures to be drau-n into
such a blast. "
Furthermore, under these conditions, the
mercuPi- atoms condense on the walls of the
inner tube near the inlet and owing to the
latent heat of evaporation raise the temper-
ature of the walls so that condensation
ceases, and the mcrcur\- atoms are merely re-
flected from the walls in all directions.
Consequently there is just as much tendency
for the mercury to diffuse back towards the
exhaust system as away from it, and the
air molecules are thus jjrcvcnted fron enter-
ing into the mcrcun,- blast at the nozzle.
These considerations and the results of his
previous investigations on the mechanism
of conden.sation of gas molecules on solid
surfaces led Langmuir to the conclusion that
the mercur\- atoms could readily be prevented
from diffusing back in the direction from which
the gas molecules arc diffusing by simply
cooling the walls of the tube near the mercury
vapor outlet. Under these conditions the
mcrcun,- atoms ought to be rapidly condensed
as they strike the walls. At the same time the
gas molecules diffusing in from the system to
be exhausted would collide with the high
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
675
speed mercury atoms at the jet and thus
acquire a velocity component from the latter
which would remove them rapidly from the
spare around the jet opening. The whole
action of the pump constructed on the basis
of this reasoning thus rests on the fundamental
principle that the mercury vapor is rapidly
condensed as it leaves the jet and the tem-
perature is maintained so low that the mer-
cury does not re-evaporate to any measure-
able extent. Langmuir has therefore sug-
gested that pumps based on this principle
should be designated as "Condensation"
Pumps.
The best type of glass condensation pump
as constructed bv Langmuir is shown in
Fig. 22.
' ' In order that the pump may function
properly it is essential that the end of the
nozzle L shall be located below the level at
which the water stands in the condenser J.
In other words, the overflow tube K must
be placed at a somewhat higher level than
the lower end of the nozzle as is indicated in
the figure. The other dimensions of the
piunp are of relative unimportance. The
distance between L and D must be sufficiently
great so that no perceptible quantity of gas
can diffuse back against the blast of mercury
vapor, and so that a large enough condensing
area is furnished.
"The pump may be made in any suitable
size. Some have been constructed in which
the tube B and the nozzle L were one and a
quarter inches in diameter while in the other
pumps this tube was only one quarter of an
inch in diameter and the length of the whole
pump was only about four inches. The larger
the pump the greater is the speed of ex-
haustion that may be obtained.
" In the operation of the pump the mercury
boiler A is heated by either gas or electric
heating so that the mercury evaporates at a
moderate rate. A thermometer placed in
contact with the tube B, under the heat in-
sulation, usually reads between 100 and 120
degrees C. when the pump is operating satis-
factorily. Under these conditions the mer-
cury in the boiler A evaporates quietly from
its surface. No bubbles are formed so there
is never any tendency to bumping.
"Unlike Gaede's diffusion pump, there is
nothing critical about the adjustment of the
temperature. With an electrically heated
pump in which the nozzle L was % in. in
diameter, the pump began to operate satis-
factorily when the heating unit deli\-ered
220 watts. The speed of exhaustion remains
practically unchanged when the heating
current is increased even to a point where
about 550 watts is applied.
"The back pressure against which the
pump will operate depends, however, upon
the amount and velocity of the mercury vapor
Fig. 22. Langmuir Condensation Pump, Glass Form
escaping from the nozzle. Thus in the case
above cited, with 220 watts, the pump
would not operate with a back pressure ex-
ceeding about 50 bars, whereas with 550 watts
back pressures as high as SOO bars did not
aft'ect the operation of the pump."
Condensation Pumps Built of Metal
For most practical purposes a glass pump
has many disadvantages. Langmuir has
therefore applied the same principles to the
construction of a metal pump.
One such type of pump which has proved
relatively simple in construction and efficient
in operation is shown diagrammatically in
Fig. 23. "A metal cylinder A is provided with
two openings, B and C, of which B is connected
to the backing pump and C is connected to
the vessel to be exhausted. Inside of the
cylinder is a funnel-shaped tube F which rests
on the bottom of the cylinder A. Suspended
676 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 8
from the top of the cylinder is a cup E in-
verted over the upper end of F. A water
jacket, J, surrounds the walls of the cylinder
A from the level of B to a point somewhat
above the lower edge of the cup E.
Fig. 23.
Diagram of Construction of Condensation Pump
Metal Form
"Mercury- is placed in the cylinder as in-
dicated at D. By applying heat to the bottom
of the cylinder the mercur\- is caused to
evaporate. The vapor passes up through F
and is deflected by E and is thus directed
downward and outward against the water-
cooled walls of A. The gas entering at C
passes down between A and E and at P meets
the mercur\- vapor blast and is thus forced
down along the walls of A and out of the
tube B. The mercur\- which condenses on
the walls of A falls down along the lower part
of the funnel F and returns again to D through
small openings provided where the funnel
rests upon the bottom of the cylinder. A
more detailed drawing of the pimip as actually
constructed is shown in Fig. 24."
A pump in which the funnel F is 3 cm. in
diameter and the c>-linder A is 7 cm. in dia-
meter gives a speed of exhaustion for air
of about ;?000-4()()U cm.' per second. It
operates best against an exhaust pressure of
10 bars or less and requires about 3(10 watts
energA^ consumption in the heater circuit.
Degree of Vacuum Obtainable
"The condensation pump resembles Gaede's
diffusion pump in that there is no definite
lower limit (other than zero) below which the
pressure cannot be reduced. This is readily
seen from its method of operation. A lower
limit could only be caused by diffusion of
-' This is apparent when we consider that no appreciable number
of atoms pass up into the space £.
gas from the exhausc side (X in Fig. 3) back
against the blast of mercur\- vapor passing
down from L. The mean free path of the
atoms in this blast is of the order of magni-
tude of a millimeter or less and the blast is
moAnng downward with a \-elocity at least as
great as the average molecular velocity (100
meters per second for mercun,-).-'
" The chance of a molecule of gas moving a
distance about 4.6 times the mean free path
without collision ^s only one in a hundred.
To move twice this distance the chance is
only 1 in 100-. etc. If the mean free path were
one millimeter the chance of a molecule
moving a distance of 4.6 cm. against the blast
without collision would be 1 in 10^. In other
words, an entirely negligible chance."
Actual obser\'ations with the ionisation
gauge (described in a subsequent section) in this
laboratory, ha^"e shown that it is possible with
the Langmuir condensation pump to obtain
Fig. 24. Langmuir Condensation Pump. Metal Form
pressures which are of the order of 10~*
bar or less. The limiting factor which ordi-
narily makes it possible to obtain pressures
as low as this, is the continuous liberation of
gas from the glass walls or metal pans, so
that it becomes extremely difficult to obtain
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
(377
vacua in excess of the above order of magni-
tude. The necessary precautions in using
the pump are discussed at greater length in
Appendix I.
Other Forms of Mercury Vapor Pumps
Other forms of mercury vapor pumps have
been described by H. B. WilHams-'', Chas. T.
Knipp", and L. T. Jones and N. O. Russell. ='^
The construction used by the latter is shown
in Fig. 25. The advantage of this form is
that it " pennits using the pump as a mercury
still at the same time that it is being used for
exhaustion purposes. Two barometer col-
umns introduce the mercury into the arc, the
arc being started by blowing in one neck of
the Woulff bottle. As shown at B the
mercury vapor is driven through the nozzle
N, and condenses in the chamber surrounded
by the water-jacket, J. The condensed clean
mercury is then drawn off at O." With a
current of 10-15 amps., a speed of exhaust of
400 cm.'' per sec. was obtained.
A simple construction for a condensation
pump has also been described by W. C.
Baker." In this form as well as Knipp's the
main object of the design is to simplify the
glass blowing.
J. E. Shrader and R. G. Sherwood^** have
used a modified form of Langmuir condensa-
tion pump made of pyrex. Full details with
all necessary dimensions are given in the
original paper. The speed of the pump was
measured for different amounts of energy
input into the mercury heater and was found
to be a maximum at about 400-500 watts
-^ j|-
Fig. 25. Condensation Pump, Arc Type
"Phys. Rev. 7. 583 (1916).
'» Phys. Rev. 9. 311 (1917). and /i, 492 (1918).
'« Phys. Rev. 10. 301 (1916).
" Phys. Rev. W, 642 (1916).
» Phys. Rev. IS. 70 (1918).
"Phys. Rev. ilj. 557 (1917).
■" J. Am. Chem. Sec. 39. 2183 (1917).
" J. Washington Acad. Sciences 7. 477 (1917).
BJ
input. With the speed of exhaustion pur-
posely cut down by a special constriction,
the maximum speed observed was around 225
cm.^ per sec. and pressures as low as 2X 10"^
mm. Hg were obtained after care had been
taken to heat up all the
glass parts to a tempera-
ture of 500° C. for a long
time.
An interesting form of
mercury vapor pump is
that devised by W. W.
Crawford-^ and shown in
Figs. 26 and 27. "The
mercury vapor gener-
ated in the boiler B at a
pressure of 10 mm. of
mercury or more, es-
capes through the nar-
row throat T (Fig. 26),
ahead of the point of
entrainment. The vapor
expands in the diverging
nozzle N, and the issuing
jet passes through the
tube E, which it fills, and
condenses in D, mostly,
where it is found at the
upper end. A slight
amount of vapor escapes
into the chamber A and
condenses there. The
condensed vapor drains
back through the tubes
a and b, to the boiler."
The vessel to be exhausted is connected at c,
while D connects with the rough pump. The
speed of the pump in series with 10 cm. of
tubing, 19 cm. in diameter was observed to
be around 1300 cm.' per sec. at a boiler pres-
sure of 10 mm. of Hq.
A two-stage mercury vapor pump to work
against a primary vacuum of 2 cm. given by
a water aspirator, has been described by
C. A. Kraus.'" It consists essentially of two
Langmuir condensation pumps in series.
The pump is very rapid and is capable of
exhausting 1500 cm.' to less than 10"'' mm.
in 10 min.
H. F. Stimson has also constructed a two-
stage pump along the same principles'^ , which
is illustrated in Fig. 28. "The operation of
the pump is as follows : Cooling water enter-
ing at tube A flows up through the water jacket
B above the lower end of nozzle F, up through
the water jacket C above nozzle G, and out
tube D. Mercury vapor from the boiler enter-
Fig. 26. Crawford's Form
of Condensation Pump,
Vertical Type
678 August. 1920
GEXER.\L ELECTRIC REVIEW
Vol. XXIII. Xo. S
ing through tubes E flows through the nozzles
F and G, is liquified in the condensation
chambers H and I. falls into the tubes K. and
returns to the boiler through tube L. Gas
from the vessel to be exhausted enters at M,
flows past nozzle F, is compressed by the jet
Fig. 27. Crawford's Form of Condensation'Pump.
Horizontal Type
of mercur\^ vapor in the condensation chamber
H, and flows up through X to the intermedi-
ate pump. From here it flows past the
nozzle G and is compressed through O into
the chamber I to a pressure measured by the
attached manometer, then out by tube P
to the water aspirator. "
The speed of the pimip as defined by
Gaede's equation was obser%-ed to be about
250 cm.' per sec.^-
General Remarks Regarding Exhaust Procedure
As has been pointed out in a previous sec-
tion, the vacuum actually attained by the use
of any pump is dependent, first, on the type of
pimip used, and second, on the rate at which
gases are given off from the walls of the
vessel to be exhausted and metal parts inside
it. In the case of the Gaede molecular pump,
as mentioned above, the degree of vacuum at-
tainable (Pi) is dependent upon the exhaust
pressure (Po) produced by the rough pumj).
As the value of the ratio Pj Pi is about 50,000,
it is evident that even with a rough pump
pressure of one bar, the pressure attainable
with this pump is less than 10~* bar. In the
case of the mercur\- \'apor pumps there is theo-
retically no lower limit pressure, and the only
limitation is therefore that due to the second
cause mentioned above.
_ » M. Volmer, Bcr. o2. (b). 804 (1919). has also constructed a
similar form of two-stage condensation pump, which is de-
scribed briefly in :.n abstract in J. Chem. Soc. lis. ii, 225 (1919).
The original article containing an illustration of the construc-
tion is not available to the writer.
The gases occluded on the walls of the
glass vessels consist for the most part of
water vapor and carbon dioxide gas along
with slight amounts of carbon mono.xide and
other gases which are not condensible at the
temperature of liquid air. Metal parts usu-
ally contain carbon monoxide and hydrogen
gases. In order to eliminate these gases it is
essential to heat the glass walls and iretal
parts to as high a temperature as practicable.
The longer the duration of the heati.ng and
Fig. 28. Stimson's Form of Condensation Pump
the higher the temperature, the lower the
pressure of residual ga.ses.
A usual procedure is to heat the glass
vessels in an oven, during exhaust, for an
hour or longer. For lead glass, the temper-
ature in the oven should not exceed 360° C,
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
679
in the case of lime glass, the temperature may
be raised to 400° C., and for vessels made of
pyrex, the oven temperature may be increased
to 500° C. The oven may be either heated by
gas flames, or more conveniently by electri-
cally heated grids, as the latter method permits
of a more uniform distribution of temperature
inside the oven and also is more convenient
for regulation.
Where a very high degree of vacuum is
desirable it is possible to heat the glass to
temperatures higher than those mentioned
above, by reducing the pressure of air in the
oven itself, so that the glass walls will not
collapse because of external pressure. For
this purpose Dr. Langmuir devised the form
of oven shown in Fig. 29.*
It consists of a metal chamber 7, which is
open at the bottom but rests upon a base
plate 8, with which it makes an air-tight
joint. The chamber is provided with a heat-
ing coil wound on the inside and separated
from, the walls by heat-insulating lining. The
leads for this heating coil are shown at 11.
Uprights 9 are provided for the purpose of
allowing the oven to be raised or lowered.
As the chamber 7 is to be exhausted it is nec-
essary to make the joint between it and the
base plate 8 air-tight. This may be accom-
plished by means of a rubber gasket, and in
order to prevent injury to this by the heat,
the chamber and the base plate are cooled
by water, which flows as indicated by the
arrows through the tubes 23, 24, 25 and 27.
Openings are provided in the base plate for
the connection between the vessel to be
exhausted and the pump, and also for ex-
hausting the chamber itself. A rough pump
is, of course, all that is necessary in the latter
case.
With this type of oven it is possible to
heat the glass about 100° C. higher than in
an ordinary ov^en, so that the residual water
vapor and other condensible gases are re-
moved more completely.
In the case of metal parts the elimination
of gases is a more difficult matter. Where
these parts are so constructed that current
can be passed through them (wires or fila-
ments) they ought to be heated to as high a
temperature as the metal will stand without
injury. In the case of hot cathode devices
the anodes can be heated to incandescent
temperatures by electronic bombardment. f
* I. Langmuir, U. S. Pat., 994,010, May 30. 1911.
t For illustrations of this the reader may refer to the following
publications:
1. Langmuir. Phvs. Rev. g. 450, 1913.
S. Dushman. Phys. Rev. 4. 121, 1914.
Heating the metal parts in a vacuum furnace
before putting them in the glass vessel also
assists materially in the subsequent exhaust
on the pump. Special care should be taken
to remove all traces of grease and oil from
machine-made parts, by washing in acetone
To Pump'
Fig- 29. Vacuum Furnace for High Temperature Exhaust
and alcohol and drying thoroughl\- before
assembling in the glass vessel.
In order to eliminate mercury vapor and
condensible gases emitted from the grease
used on the ground glass joint between the
pump and the vessel to be exhausted, it is
necessary to interpose some form of refriger-
ating chamber in which these vapors are
condensed, as shown at G in Fig. 22. (See also
Appendix I.)
The most efficient refrigerating agent is,
of course, liquid air, which is now available
in a large number of laboratories. Failing
this, a suspension of solid carbon dioxide in
acetone or ether may be used. In the latter
case it is well to insert a tube containing PsOs
between the oil pump and the fine pump to
take care of water-vapor. Observations in
this laboratory have shown that in using a
condensation pump it is possible to obtain
practically as low pressures, with solid CO2
and P2O5 as with liquid air, but the interval
of time required to attain this low pressure is
ordinarily much longer with the former.
680 August, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, No. 8
The temperature produced by liquid air
evaporating freelv into the atmosphere is
about -190° C. (93° K.), but varies from a
lower value for fresh liquid air to a higher
value as the nitrogen boils away and leaves the
oxygen behind. With solid carbon dioxide a
temperature of — 78° C. is obtained. Table XII
shows the vapor pressures of a number of
Table XIII gives the vapor pressures of car-
bon dioxide, ice and mercun,-. In all cases the
data for the lower temperatures have been
extrapolated in the same manner as those
given in the previous table.
As is evident from these data, ice and mer-
cury have no appreciable vapor pressure at
— 190° C, while carbon dioxide has a vapor
Fig. 30. ArranKcment of Exhaust System
different gases at these low temperatures.
In the case of methane, ethane, ethylene and
carbon dioxide the pressures given have been
extrapolated from the values in the standard
tables for higher teinperatures, by plotting the
value of log P against 1 T. where T is the abso-
lute temperature. The values plotted in this
manner are found to lie on a straight line, thus
making the extrapolation an easy matter.
pressure between 0.001 and 0.000 1 bar at this
temperature. Under these conditions any of
the latter gas if condensed in the liquid air
trap would produce a constant pressure of
residual gas of at least 10"' bar. Under the
same conditions ethylene and ethane would
produce pressures that might exceed 10 bars.
However, these gases are not met with in
ordinary exhaust operations. The other gases
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
681
are non-condensible at the temperature of
liquid air, and are therefore removed by the
puinp.
It is also evident that at the temperature
of evaporating solid CO-j ( — 78° C.) the vapor
pressure of mercury is negligibly small, Ijut
that of ice is quite high, hence the necessity
for using P2O5 in order to take care of this.
It is also advisable in all cases of exhaust
operations not to use the refrigerating agent
during the initial stages of heating, so that
the bulk of condensible gases will be re-
moved by the pump. Otherwise the vapors
may be condensed in the trap and maintain
a constant pressure of residual gases in the
system for a very long time.
TABLE XII
Temperature
Degrees Centtg.
Absolute
Temperature
Pressure in
mm. Hg.
Oxygen (Oa)
- 182.9
-211.2
90.2
61.9
760
7.75
Nitrogen (No)
-195.8
-210.5
77.3
62.6
760
86
Carbon Monoxide (CO)
-190
-200
83
73
863
249
Methane (CH,)
-185.8
-201.5
87.3
71.6
79.8
50.2
Argon (A)
-194.2
- 186.2
78.9
86.9
300
760
Ethylene (C2H4)
— 175.7
97
0.76
-188
85
.076
-197
76
.0076
-205
1
68
.00076
Ethane (C^Hs)
-1.50
123
7.6
-180
93
.076
-190
83
.0076
-198
75
.00076
Temperatures below —190° C. may be ob-
tained by working with liquid air under
reduced pressure. As is evident from Table
XII, it is possible in this manner to obtain a
temperature as low as — 'iOO" C. At this tem-
perature the vapor pressure of solid CO2 is
less than 10"^ bar, and it is therefore possible
under these conditions to obtain extremely
low pressures of residual gases.
Either liquid air or solid CO2 prevents the
diffusion of vapors emitted by grease around
the joint, as these vapors are all condensible
at these temperatures.
Determinations in this laboratory of the
vapor pressures of \-acuiun pump oils have
shown that these range around one bar at
room temperature, decrease to 0. 1 bar at 0° C,
and are negligibly small at — 78° C. Stopcock
grease and similar compounds possess even
lower vapor pressures.
Further details regarding the actual oper-
ation and care of typical exhaust systems are
given in Appendix I.
TABLE XIII
Temperature
Degrees Centig.
Absolute
Temperature
Pressure
in Bars
Carbon Diox
de
(CO2)
-148
-168
-182
-193
125
105
91
80
1
101)
1
0.01
0.0001
Ice (H.O)
-20
253
1045
-.30
243
380
-40
233
127
-.50
223
39
-60
213
9.6
— 75
198
1.0
-89
184
0.1
-100
173
.01
-110
163
.001
Mercury (Hg)
-I-.30
303
3.7
-f-20
293
1.6
+ 10
283
0.65
0
273
.25
-10
263
.087
-20
253
.029
-40
233
.0023
-78
195
4.3 X
io-«
-180
93
2.3 X
io-«
G82 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 8
APPENDIX I
Fig. 30 shows diagrammatically an arrange-
ment of the Langmuir condensation pump
and accessor}' connections which is conven-
ient for high vacuum exhaust operation, while
Fig. 31 is a photograph of a set-up such as is
Fig. 31.
Photograph of Set-up for Exhausting Coolidgc
Xray Tubes
used for exhausting Coolidge X-ray tubes of
the standard types. The connections to the
McLeod gauge and oil-pumps are to be made
with as large tubing as practicable; ordinarily
about half-inch tubing may be used. For the
sake of the illustration an ionization type of
gauge is shown connected to the condensation
pump ; and the oven is indicated diagrammati-
cally. As previously mentioned, the tubing
in this case ought to be as wide as possible.
Care of the Condensation Ptnnp: Accord-
ing to the instructions, ()2() grams of abso-
lutely clean mercury should be poured into
the pump. Lender normal conditions, with
300 watts input into the heater and a
* A good stopcock grease may be made by heating approx-
imately equal parts of pure rubber and vaseline. The rubber
should be cut into very fine pieces and the heating continued
until the mixture has about the consistency of heavy molasses.
flow of 1000 cm.' per minute through the
cooling coils, only the lower portion of the
pimips should run warm. If for any reason
the flow of water ceases and the grease around
the joint is melted, the pump should be re-
moved, the mercury emptied out, and the
pump cleaned with gasolene as directed in the
instructions. Ver\- little mercury should
condense in the glass grinding above the
pump. If. however, the grinding rapidly
becomes covered inside with mercur\- and
feels warm, it is an indication that the pump
is not exhausting and a cleaning is required.
Seme observations carried out in this
laboraton,' on the variation in exhaust pres-
sure with energA- input into the heater are of
interest in this connection. The following
table gives the watts used and the correspond-
ing minimum pressure obtained as measured
by an ionization type of gauge. The oil
pvimp pressure was 6.0146 mm. of Hg.
Watts in Heater
Minimum Pressure in Bars
130
0.27
150
0.07
170
0.04
180
0.023
200
0.013
220
0.007
240
0.002.5
2S0
0.002
300
0.002
_ -
Care of Rubber Tubing: All rubber tubing
for use in vacuum systems should be cleaned
well inside before use. This is best accom-
plished by washing with a warm 10 per cent
solution of caustic soda, then with water and
alcohol, and finally dr>'ing thoroughly by
blowing air through it, or exhausting with a
rough pump. The rubber tubing between
the condensation piunj) and glass tube leading
to the mercur>' trap should be as short as
possible. Heavy vaseline or grinding grease*
should be used to make the junctions air-
tight, and after the tubing is in place a little
castor oil should be dripped over the joints,
and over the rubber tube itself.
Mercury Trap: This is essential in order
to prevent mercury from getting into the oil
pumps, where it would gradually disintegrate
the bearings. A very good scheme which has
been used successfully in this laboratory is
to make this trap quite large (about 1 liter
volume) and fill it with broken pieces of glass
tubing, thus increasing the surface for the
condensation of any mercury vapor that
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
683
I
diffuses out through the rough side opening
of the condensation pump.
Detection of Leaks: A small Tesia coil, one
end of which is grounded, is a useful accessory
device in all exhaust operations. The high
tension terminal of the coil is encased in rubber
tubing and provided with a wooden handle so
that it can be touched to any part of the glass
system. A pin-hole leak will show up by the
direct passage of a spark to this point, while
at all points there will be a uniform glow if the
pressure is over a few bars and there will be
no glow at all when the pressure is one bar or
less.
Seal Off Procedure: Constrictions in glass
vessels at points where they are to be sealed
off after exhaust should not be boo thick
walled, otherwise a large body of glass will
have to be heated during the seal-ofif, causing
the liberation of a great deal of gas. Further-
more, the constriction should be torched till
it is almost melting and the ptimp allowed to
exhaust the gas thus liberated for about two
minutes, after which the sealing off should be
performed as rapidly as possible without heat-
ing the glass any more than absolutely neces-
sary.
The importance of these precautions may
be iudged from the following observations:
An ionization gauge, having a volume of about
100 cm.'\ was exhausted till the residual gas
pressure was less than 0.001 bar. On sealing
this off without observing the precautions
mentioned, the pressure in the sealed off tube
was 0.25 bar, whereas by torching the con-
striction first and sealing off afterwards, it
was possible to obtain a residual gas pres-
sure in the sealed off gauge of less than 0.01
bar.
(To be continued)
684 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 8
Five Thousand Horse Power Electrically Operated
Pumping Plant
By E. Bachman and W. J. Delehanty
San Francisco Office, General Electric Company
The centrifugal pumping plant described in this article is of particular interest in that it is probably the
largest of its type in the world and is electrically operated. The authors explain the purpose of the instal-
lation, describe the construction of the building (which is especially heavy), the pumping and electrical equip-
ment, the power supply and its protection and control, and the lighting system em.ployed. — Editor.
ments. standing 36 ft. 6 in. from the main floor
level to the roof. It is supported on reinforced
piles spaced approximately on 5-ft. centers
longitudinally and 4 ft. centers crosswise of the
building. Built upon this is the floor which is
4 ft. thick over the entire area. The walls are
4 ft. 6 in. at the floor line, tapering to .30 in. at
the high-water ele\-ation, and from that point
tapering off to lo in. at the roof-line. In
addition to the extremely heavy walls, there
are six buttresses on the north and south walls
and two on the east and west walls. The lower
portion of the building is entirely without
openings, the windows and doors being above
the high-water level.
Running at the high-water level through
the entire length of the building is a mezza-
nine platform, built of reinforced concrete,
access to the main-floor level being obtained
by steel stairs. At the east end of the build-
ing, at the same level, is a switch gallen-.
the switchboard being placed on the main
floor immediately below. An operating
platform runs along the south wall just
underneath the mezzanine floor, constructed
entirely of structural iron and checkered floor
plating so that it may be easily removed to
give access to machiner\-. The roof of the
btiilding is a flat reinforced concrete slab sup-
ported on 2()-in. I-beams, the roof being
finished with an imjicrvious coat of asphalt
and gravel. The building is well lighted,
having three large steel frame windows
approximately ]'2 by 14 ft. on both the north
and south walls and two windows in the east
and west walls, one of these windows being
removable to admit the machiner>- into the
building.
For the handling of machinen.' there is
installed a l.>-ton crane, having a motor-
driven hoist, with the bridge and trolley
hand operated. The motor is a 110-volt
3-phase (iO-cycle variable-speed induction
motor, callable of raising 15 tons at the rate
of 20 ft. per minute.
A reinforced concrete platform is built
alongside the east wall of the building to
During the development of the Sutter
Basin Reclamation Project, a large pumping
station was erected at Knight's Landing,
California, to drain the vast area of land which
is protected by levees. The district lies be-
tween the Sacramento and Feather Rivers,
and is commonly known as "Sutter Basin."
The area comprises approximately 68,000
acres.
The plant is probably the largest centrif-
ugal ptmiping yjlant in the United States, if
not in the world, has a total capacity of
676,000,000 gallons of water per day, and
requires 5000 horse power for its operation.
The plant operates approximately 30 to 60
days during the normal year, during which
period it may be run partly or as a whole,
depending upon the amount of water to be
handled. The larger part of this will be
rain water with a certain amount of seepage
through the levees.
On account of some of the adjacent lands
not being protected by levees, the Reclama-
tion Board took every precaution to limit the
liability of failure of the pumping equipment.
The plant is located at the extreme south-
ern end of the district, on the Sacramento
slough, about l}/2 niiles above its confltience
with the Sacramento River, approximately
22 miles from Sacramento city, via the Sacra-
mento River.
Fig. 1 shows the general lavoul of the jjlant
as finally constructed. It will be noted that a
timber piling bulkhead is built at the outer
and inner toe of the levee at the nearest
point to the river jiroper, thus forming a
suction forebay. The jnimping plant is
located with its long side ])arallel to the canal
at this point, the i)tnn])s taking their suction
from the canal and discharging through tlie
levee into the Sutter Basin by-pass.
The btiilding is of extremely heavy con-
struction, being designed so that the total
weight of the building and machinery will
overcome the tendency of the structure to
float at high water. The building is Oil ft. 6 in.
long and 2.') ft. 6 in. wide, inside measure-
FIVE THOUSAND HP. ELECTRICALLY OPERATED PUMPING PLANT 685
.1111 ei ^%i ^^fri^ - '^
O £
01 <
2 J
U) CO
1°
i 6
I §
> u
ii C
c
•5 J
Of
•■2«
■S 2
o w
c
^ "o.
■B B
u 3
3 0,
Z
B
*" u
X i!
in 5
3 «J
a. ,"
■5 >>
U o.
J= CO
^ u
> s
- K
0. Id
<
686 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 8
support step-down transformers as shown in
Fig. 2. The floor of this structure is just above
the high-water line and is supported by
concrete columns, in turn resting on rein-
forced concrete piles.
The pumping plant consists of six 50-in.
pumps, direct connected to induction motors,
each pump having a capacity of 175 second
feet of water when pumping against the maxi-
mum head of approximately 29 ft., the lower
limit of head being approximately 10 ft. The
static head varies from approximately 2 to
25 ft., with approximately 4-ft. discharge
friction. The pumps are of the double-suction
type, similar to the pumps installed in other
large reclamation projects in this section.
pumps, having suctions connected to th2 top
portion of the casings of all the centrifugal
pumps. When priming the pumps, the gate
valve in the discharge line is closed and, after
priming, the pump is started. After reaching
full speed the gate valve is opened.
The suction pipes are approximately 60 ft.
long, running directly out through the south
wall of the building, a distance of 32 ft., and
then dropping down with easy bends into the
suction forebay, the diameter at the pumps
being 50 in. and tapering to a diameter of 96
in. at the lower end. These suction pipes are
of ? g-in. plate steel construction, the lower
end of the pipes resting on a grillage work of
structural I-beams provided with proper
^-^5A■l/-/4.\I/
iiooo/aaoo
Volt.'
I- KV. Auxiliary
I. COO/ I IV j--r V ' ^ ' ^ — ■> v_.^ ^^.^ v_^
. ^'P^ . I ^ - ■ Six eOOKeZ<eR.P.M.Sauirrel Cage Motors
Line -T 1 1 T T
Bus llOVolts SPhase eOCyclas
,5 tation A uxiliary and L igh t s
Fig. 5. One-line Diagram of the Power Connections in the Pumping Station
The large size of these pumps can be seen
by reference to Fig. 3, which is an interior
view of the pumping house.
The pumps are direct connected to General
Electric squirrel-cage induction motors, oper-
ating at 248 r.p.m. through flexible leather
link couplings. The motors are rated at 800
h.p., at 40 deg. temperature rise, with 25
per cent overload guaranteed for two hours.
The pumps are so designed that the horse
power over the entire range of j^umping from
the lowest to the highest head varies only
from 800 h.p. under minimum head to 825
h.p. at the maximum operating head.
For priming purposes there is installed two
12 by 12-in. duplex motor-driven ^•3cuum
grizzlies to prevent foreign matter entering
the pumps.
In the discharge pipes just outside of the
building are placed motor-operated gate
valves. These valves are controlled by
General Electric direct -current motors. They
are capable of opening and closing safely in
two minutes imder the conditions of usual
operation, and each motor with its valve is
equipped with limit switches and an indicator
showing the amount of gate opening. This
indicator is visible from the switchboard,
where the control of all valves is centralized.
The Power Company's 60.000- volt lines run
to the step-down transformers immediately
outside of the building. From the trans-
FIVE THOUSAND H.P. ELECTRICALLY OPERATED PUMPIXG PLANT 6S7
fomiers, fi\-e leads run through the building
walls to busses forming one 2200-volt and one
1100-volt circuit, from thence through dis-
connecting switches to the starting and run-
ning switches and thence to the motors. The
starting and running switches for the motors
of the main circuit are ;JUO-ampere, 2.300-volt,
triple-pole, single-throw, hand-operated, re-
mote-control, oil-break switches, with over-
load and undervoltage protection. The
switches are installed in groups of two,
mechanically interlocked. The starting
switches are connected to the 1 1 00-volt bus
and all switches are operated from the main
switchboard on the lower floor of the puni])
house.
Incoming power is taken from the Pacific
Gas & Electric Company's 63,000 Y-volt,
3-phase, 60-cycle line through a General Elec-
tric automatic, outdoor type, oil circuit breaker.
This circuit is then run to three Y-delta con-
nected, loOO-kv-a. 6:_), 000-1100 /2200-volt self-
cooled, outdoor type transformers, and thence
to the station switchboard.
An outdoor type oxide film lightning
arrester, as shown in Fig. 4, has recently been
installed to further insure the station against
shut-down. This arrester has been in service
a sufficient length of time to demonstrate the
advi-sability of placing an arrester at this
point.
For operation of small motors, lights, and
other station auxiliaries there are installed
two 2o-kv-a., 2200 to 110-volt distribution
transformers. From these transformers a
10-kv-a. motor-generator set is used to charge
a 56-cell, 112-volt chloride accimiulator type
storage battery which is used to operate the
valve motors, emergency lighting service, and
supply tripping current for the outdoor oil
circuit breaker. The storage battery is used
to operate the valve motors in order to
provide safety in case the main power supply
should fail.
The switchboard consists of eight panels
of blue Vermont marble mounted on pipe
frame work. All vacuum gauges, pressure
gauges, voltmeters, ammeters, power-factor
indicators, etc., are finished alike, and are
all mounted on the switchboard proper.
With this arrangement it is only necessary
for the operator to remain at the sA^itch-
board, making an occasional inspection of the
bearings at the pumps while operating.
The lights on the mezzanine floor, on the
valve-motor platform, and the Sivitch gallery
are equipped with Holophane steel reflectors
and 150-watt Mazda lamps. The bracket
fixtures along the north wall are equipped
with 26-in. Mazda economy diffusers, sus-
pended from ornamental iron brackets by
links; each fixture is equipped with five
100-watt Mazda lamjjs. The main entrance
fixtures near the door are equipped with 100-
watt Mazda lamps. The lighting circuits are.
all controlled from the lighting panel, which is
installed near the main sA^itchboard. The
direct-current circuits are equipped with a
three-point switch at the main entrance with
the corresponding switch on the lighting panel.
This permits the lighting of the station from
either end, thereby making it unnecessary for
the operator to grope in the dark in case of
power failure.
The building and concrete discharge pipes
were constructed by the Sound Construction
& Engineering Company, and the pumping
plant contract was let to Charles C. Moore &
Co., who sub-let the electrical machinery' to
the General Electric Company.
688 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. &
Power Control and Stability of Electric
Generating Stations
Part I
By Charles P. Steinmetz
Chief Consulting Engineer, General Electric Company
About a decade ago our larger central stations had grown to such size as made necessary the adoption
of reactors to limit the enormous amount of power which might accidentally be concentrated at a fault in the
system's lines and thus prevent its destructiveness. This revision in the scheme of operation has given rise
to a new set of conditions with respect to maintaining the stability of connected synchronous machinery. For
the purpose of furnishing a complete mathematical analysis of the subject, Dr. Steinmetz presented at the
last annual convention of the A.I.E.E. a paper of which the following is a reprint of the first half. The re-
mainder will appear in our September issue. — Editor.
POWER LIMITATION
With the increasing use of electric power,
the size of electric generating systems has
steadily increased from the small electric
lighting stations of the early days to the huge
metropolitan systems having several hundred
thousand kilowatts capacity in steam turbine
alternators.
The problem of close inherent regulation
of the generators has ceased, since no possible
sudden change of load — less than short circuit
— is sufficient appreciably to aflfect the voltage
of these big systems. The reverse problem
however has become serious, that of limiting
the power which can accidentally be concen-
trated at any point of the system, and of limit-
ing its destructiveness.
With the increasing size and extent of
systems, they were divided into a nimiber of
generating stations, more economically to
cover the territory, as under present con-
ditions there seemed to be no material gain
in going much over hundred thousand kilo-
watts in one station. Thus two or more main
generating stations are generally used, together
with a number of smaller secondar\- generat-
ing stations to stabilize the power at the end
of long feeders, in outlying centers of dis-
tribution, etc.
Economy and reliability of operation
demand parallel operation of the entire system,
and synchronous oiieration of all the generat-
ing stations thus is the universal custom.
In the former 2.")()-volt direct-current gene-
rating systems, from which most of the large
metropolitan systems have developed, sub-
division into a number of generating stations
limited the power which could be developed
at any point and thereby its destructiveness,
by the resistance of the feeders and mains.
In the present three-phase systems, inter-
connected by and distributing through under-
ground cables at fWiOd to 22,1100 volts, the
impedance of these cables is entirely insuffi-
cient to limit the power concentration possible
at any point of the system, and special means
of limiting the possible power concentration in
these systems thus became necessary. This
problem became aggravated by the inherent
characteristics of the high-speed steam-tur-
bine alternators which have completely super-
seded the former low-speed engine-driven
machines.
In the belt-driven (iO-cycle alternators of
former days, the output was from lo to 30 kw.
per machine pole, in the 2.'>-cycle low-speed
engine-driven multipolar alternators such as
were installed in the first Metropolitan Rail-
way Station of New York City, etc., the out-
put was about 100 to 12,') kw. per machine
pole, while in the modem high-speed steam
turbine alternator values of l.i.OOO to 20,000
kw. per machine pole have become necessary-.
This means enormously larger magnetic fluxes
and correspondingly larger armature re-
actions per pole. But with increasing output
per pole, the effective or equivalent reactance
of armature reaction (which is not instanta-
nous, but requires several seconds to develop)
increases at a faster rate than the true or self-
inductive reactance of the armature (which
latter is instantaneous, and thus is the only
reactance which limits the momentani' short-
circuit current of the machine). Thus, while
in the early high-frequency alternators the
ratio of effective reactance of armature
reaction to true self-inductive armature
reactance was less than 0.,^ to 1 , it has risen in
the large low-frequency turbo-alternators to
values (if 20 to 1, and more. That is, while in
the early high-frequency turlio-altemators the
momentary short-circuit current was very little
larger than the permanent short-circuit cur-
rent, in large low-frequency turbo-alteniators
the momentary short-circuit current maybe 20
or more times the permanent short-circuit cur-
POWER CONTROL AND STABILITY OF ELECTRIC GENERATING STATIONS 689
rent. Thus in a high-power system of several
hundred thousand kilowatts connected steain-
turVjine generator capacity, without power-
limiting devices, the momentary short-circuit
current may represent several million kilo-
\-olt-amperes, with corresponding electrical,
themial, and magnetic stresses. It is not the
question of whether a circuit breaker can be
designed to open such power safely , but it is the
fact that such a circuit breaker would in size
and cost be econom.ically impracticable, when
considering that the hundreds of feeder cables
and interconnecting cables of such systems
would require several hundreds of such circuit
breakers
The practice of giving the circuit breakers
a considerable time limit, so that they open
onlv after the momentary short-circuit cur-
rent has greatly decreased, greatly relieves the
stress on them, but at the expense of the
svstem which is exposed to the full momentary
short-circuit stresses, usually resulting in a
shutdown. The use of group circuit breakers in
series to the circuit breakers in the individual
feeders (and usually of larger interrupting
capacity than the latter) reduces the number
of high-power circuit breakers required and
increases the reliability, and thus is extensively
used, but by itself does not solve the problem
as the still large number of group circuit
breakers places an economic limit on their
interrupting capacity, and the required tim_e
limit of their operation leaves the system
exposed to the full destructive eflfect of the
momentary short circuit.
Thus power-limiting devices, in some fonn
or another, have become necessary and are
universally used in all modern high-power
systems.
Such devices comprise :
(1) Power-limiting Generator Reactors. Be-
sides designing the generator for the high-
est possible internal self- inductive reactance,
which can be given to it without serious
sacrifice of its other characteristics, reactors
are inserted into the leads between the gener-
ator and the busbars, so as to limit the power
which the generator can feed into the busbars
in case of short circuit at or near tiie busoarc
and to limit the power which the ousoars can
feed back into the generator in case of accideni
to the generator.
Such power-limiting generator reactors
are used wherever the internal self-inductive
reactance cannot be made sufficiently high
(10 to 15 per cent). The latter is frequently
the case with (iO-cycle machines.
Internal reactance of the generator, wher-
ever it can be secured without material
sacrifice of other characteristics, has the
advantage of saving the space and cost of the
external reactance, but it is not quite as good
in protective value since, in case of an accident
in the generator, its internal reactance is more
or less eliminated and thus does not protect
against the busbars feeding back into the
generator.
The amount of generator power-limiting
reactance necessarily is limited to that value
which does not materially increase the total
(or synchronous) reactance of the generator.
Thus, with many generators running in
parallel on the system, even with the power
limitation of the individual generators, the
total power which may be developed in case
of a short circuit on or near the busbars
becomes excessive. The economic limit of
generator power, which may be concentrated
on one busbar, probably is between oU. 000 and
100,000 kw. Beyond this value it becomes
necessary to cut, or divide the busbars, and
since parallel operation is necessar\-. this may
be done by:
(2) Power-limiting Busbar Reactors. These
are reactors that are inserted into the
busbars to limit the power which can fliow
along the busbars from one side to the other
side of the reactor, without interfering with
the flow of such current along the busbars as
may be required for synchronous operation,
etc., under the variations of load.
For economical operation, the busbars are
naturally arranged so as to require the mini-
mum average flow of power along them . That
is, the feeders which carr\- power from the
busbars to the load intermingle with the leads
which bring power from the generators to the
busbars. The power flowing along the bus-
bars thus is the difference between the in-
coming and the outflowing power. Theoreti-
cally, with a ring bus, cut into sections by
power-limiting reactors, the maximum power
690 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. S
which may have to flow over any busbar
reactor is one quarter that of the smallest
alternator connected to the section adjoining
the reactor ; and m.ay rise to twice as much if
the busbar sections are not connected into a
closed ring but into an open chain.
The transfer of power from one busbar
section to another over the dividing
reactance does not mean a drop of volt-
age, but with the same voltage on two
busbar sections, the transfer of power occurs
bv a phase displacement between the volt-
ages of the two sections. That is, if the
load on one busbar section increases beyond
ths output of the generators connected to it,
or decreases below it, power begins to flow
orer the busbar reactors connecting it with
the adjoining sections. The
voltages of the adjoining
sections howe.er are kept
constant by the control of
the alternator field excita-
tion at the san^e \alue c.
and the reactance voltage
ix of the current i passi.ig
over the busbar reactance
X thus forms an equilateral
triangle with the two vclt-
ages e of the adjoining bus-
bar section. (See Fig. 1.)
That is.i.vis approximately
in quadrature with the
section voltages e: and as
ix, as reactance voltage, is
in quadrature \\ ith the cur-
rent i, the current i is (a')- "" '""
proximately) in phase w ith
the generator voltages e,
that is, it is an energy current. The phase
angle '2a) betiveen the t 'o voltages c of the
t .vo adjoining busbar sect: ns the.T is gi ren by :
ix
sin w = -^
2e
As the synchronizing power between the
adjoining generator sections is a maximum
for 20) = 90 deg., and decreases beyond this,
no danger of breaking out of synchronism
exists, as long as Ju is materially less than
90 deg. Thus with a ijhase angle between the
generator section voltages e. of ia) = 3() deg.,
that is, fairly srrall phase displacement.
As theoretically / may be limited to one
quarter of the full-load current, io. of the
smallest generator on the section,
'^ = 2.0S
e
that is. the maximiimi theoretically pennis-
sible busbar reactance, at a m.aximum. of 30
deg. phase displacement between the busbar
sections, would be 200 per cent referred to the
sm.allest generator on the section, as far as
concerns energy transfer from section to
section with negligible phase displacement,
1.5 deg.
As the power-limiting generator reactances
are 10 to 15 per cent, or an average of 12.5
per cent, it is seen that much larger reactances
Fig. 2
1 J
-
^k-
1 Ml
^.-1 ^ 1 ■
1
^
^
' 1 ^"^^
i*|-
E
M
7M0
—
1
— i
tx
2e
= sin 15° = 0.26,
or
'-^' = 0.52.
e
Fig. 3
may safely be used in power-limiting busbar
reactors than are permissible in power-limiting
generator reactors.
It is advisable to use as large busbar
reactances as possible, to limit to the maxi-
mum extent the shock of a short circuit at or
near a busbar section, that is, to affect the
remainder of the system as little as possible.
Where a number of stations are connected
together, operating on the same system, that
is, tied together by interconnecting cables into
one bus, preferably a ring bus. it is advisable
as far as possible to locate the power-limiting
busbar reactors in the connections between
the stations, that is, tie the stations together
over power-limiting reactors. In this case it
is advisable to install one half of each of the
busbar reactors at each end of the inter-
connecting cable, since the probability of short
circuits in the interconnecting cables is far
POWER CONTROL AND STABILITY OF ELECTRIC GENERATING STATIONS 691
greater than the probabiHty of short circuits
at the busbars, and the division of the reactor
into one half at each end of the cable limits
the efTect of a short circuit in this cable on the
generating stations connected together by it.
(3) Feeder Reactors. Even with generator
power-limiting reactors and busbar dividing
reactors, the effect of a short circuit at or near
the busbars is very severe, at least on that
section of the system operated from this bus-
bar, and will probably shut down this section.
However, short circuits on the busbars are
very much less frequent than short circuits
in cables. The installation of proper feeder
power-limiting reactors, by preventing short
circuits on feeders from, directly affecting the
1
fm4
L«
.* ..
_.
^tad
fi^V
-—^
^
^m^ ^IMH
^^^
^^^
S—U
1 ,
'**•
-*-w^
*^^
'^
—
—
m 1
1
-* — '
9000
—
E
—
1
'CCO
S
=:
rm
sit
J^M.
J
"M
4J
»/•«■
4/
'M
jiafM
j/^-s-
»00D
-e
^
-«j
t
^
^^
"^
^
"^
^
k
5«
^^
tMC
3
=1
"■**"
tFM
St
fiA
t
S
fiM.
4
K^M
*/
•M
JZ
Fig. 5
busbars, even when these short circuits occur
very near the busbars, thus eliminates m.ost
of the severe short circuit shocks from the
generator sections, and is therefore economi-
cally very desirable. While the reactance of
the feeder reactor may be only a small per-
centage of the feeder rating, it usually is very
much larger than the combined reactance
of the generators feeding into the sections, and
a short circuit beyond even a small (in per-
centage) feeder reactor will thus result in a
very much smaller short-circuit current than
would occur without the feeder reactor, and
thereby the shock will be greatly reduced.
Furthermore, without the feeder reactor, a
short circuit in a cable near the busbars means
practically zero voltage at the busbars, that
is, the dropping out of synchronous apparatus.
With a short circuit beyond a feeder reactor,
however, considerable voltage is retained at
the busbars on the affected generating station,
so that synchronous apparatus is not affected,
that is, the short circuit passes without
material effect on the system, especially if
the circuit breakers are set with short time
limit which is permissible due to the greatly
reduced current which they have to open.
By the proper use of power-limiting
reactors in generator leads, busbars, and
feeders it has become possible to operate the
modern huge power systems with a high
degree of safety and to give the possibility
or unlimited extension of the system., that is,
a power system of several million kilowatts of
connected generator capac-
ity will be just as safe as
regards the limitation of
the possible destructive-
ness of short circuits and
other accidents as a system
of less than hundred thou-
sand kilowatts generator
capacity.
When thus sectionaliz-
ing the system in installing
reactors between the gen-
erators, stations, or station
sections, these reactors are
very low in absolute value
of reactance (of the magni-
tude of an ohm), and thus
permit ample current to
flow over them for all re-
quirements of the shifting
load without giving appre-
ciable voltage dro]3or phase
displacement between the
sections. But, relative to
the station capacity, these reactances must
be fairly high to fulfill their function in power
limitation. Thus a reactor of 1 .75 ohms react-
ance, connecting a 9000-volt station section of
72,000-kw. generator capacity, passes a maxi-
mum of 45,000 kw. of energy, at the limits of
synchronizing power, that is, materially less
than the rated generator capacity.
The question then arises, what effect this
necessary sectionalizing of the system by
reactors has on the synchronizing power
of the system and thus on the stability of
operation, the more so as in case of accidents
or disturbances a local and temporary drop of
voltage may occur and a corresponding de-
crease of synchronizing power.
As illustration is given in Figs. 2 to 5, the
voltage record during a trouble on September
692 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. S
IS, 1919. in the Commonwealth Edison
Company in Chicago. Fig. 6 gives the
diagram, of the station connections. The
system consisted of four sections, .4, B. C. and
D, interconnected in chain connection, from
A to C and from C to B by power-lim.iting
station .4 (Fig. .5). Interesting also is the
wattmeter record of the power exchange
between stations over the tie cables between
B and D. (Fig. 7) : while usually considerable,
practically no power or current exchange
occurred, during the trouble. An excessive
current however flowed o^•er the power-limit-
ing reactor between B and C. This reactor
was opened after seven m.inutes, thus cutting
off stations .4 and C from stations B and D.
As the result of this operation, the voltage
Fig. 6
reactors of 1.75 ohms per phase; from B to D
by six underground cables of 0.31 ohms joint
resistance and 0.074 ohms reactance per
phase. The busbar voltage was 9000, and the
load almost entirely 23-cycle synchronous
converters. The connected generator capac-
ity during the trouble was 237,000 kw.,
nearly full load. A dead short circuit close to
the busbars of section B dropped out the con-
verters on sections B and D, and some con-
veters on sections .4 and C; the circuit
breakers in the substation opened promptly
and cut off the substations, and the short
circviit was opened in a very few seconds, so
Fig. 7
recovered in .4 and B. but still stayed at zero
in B and D. without any apparent reason,
until seven minutes later (or after about a
quarter of an hour of zero voltage) just as
suddenly full voltage reappeared again in
both stations B and D. without any apparent
reason.
What happened in this case, as the investi-
gation showed, was that under short circuit
the stations B and D momentarily dropped to
zero voltage and lost their synchronizing
power. The steam turbines speeded up, cut
otT steam by closing their emergency valves,
but were put back on the steam governors:
90
to
1
i
^
1 ;
irfa
-•^ >
h"-* — -
I^B..^
^^
]\
1
ki.^i
lL /
Ml
M
1 I
1
1
1
, 1
Fig. 8
that the system was clear again in three to
four seconds and the voltage should have
come back. But it did not come back, it
stayed at zero in both stations B and D (Figs.
2 and 3*), and showed a permanent great
drop in station C (Fig. 4), and a lesser drop in
» While the charts do not read below 6niX) volts, the station
voltmeter showed that there was no appreciable voltaKC during
the entire period.
seoo
XAM
Fig. 9
their speeds however were already too far
apart to pull each other into step promptly,
and while the unafTcctcd stations .4 and C
stayed in step with each other, the stations B
and D not only broke out of synchronism with
each othcrand wiih.4andC",but the individual
machines in B and D broke out of synchro-
nism with each oilier. The stations B and D
POWER CONTROL AND STABILITY OF LLECTRIC GENERATING STATIONS 693
and the indi\-idual machines in these stations
then kept driftiog past each other indefinitely,
nnable to pull into step until some of the
machines happened to drift into phase with
each other, caught in synchronism and thereby
established some voltage, and then quickly
pulled all the other machines into step, and the
voltage then came back suddenly.
Figs. S to II show the voltage records of the
same four stations during a trouble on Mav
Ifl, 1919, and Fig. 12 the wattmeter record
of the tie cables betw-een B and D. The
station arrangement was the same, the con-
nected generator capacitv 2.30, ()()() kw., about
% load.
In this case, a generator short circuited in .4
(Fig. 11), pulling the voltage down to practi-
cally zero, but was cut olT by the circuit
breakers and the system cleared in less than
two seconds, so that the voltmeter record of
station .4 shows only a momentarv' drop to
iODO-
Fig. 10
zero voltage. Nevertheless, a voltage dis-
turbance resulted in all four stations, lasting
for over a quarter of an hour, that is, the volt-
age greatly dropped, and wildly fluctuated;
most at the source of the trouble, station .4 ;
least at the remote end, in station D; and the
voltage rem.ained low and fluctuating for no
apparent reason, for IS minutes, and then
suddenly recovered and steadied down, with-
out any apparent reason also. An excessive
current passed during the disturbance be-
tween stations D and B as shown by the watt-
iffneter record going off the scale, and an exces-
sive current between stations B and C as
shown by the heating of the reactor. In this
case, the stations did not break out of step
with each other, but stayed in synchronism.
In appearance these records look ver\- much
like hunting, or surging of the stations against
each other, and thus are rather disquieting to
the station operation. It is questionable, how-
ever, whether it is a real hunting.
The matter of the synchronizing power of
these big stations and in general of all phenom-
ena of synchronous operation, as affected
by the impedance between the machines,
thus is of fundamental importance for the
safe operation of our modern large powder
svstems.
1
A
M^^
Ussafl
^^ ^ /L. 1 ■{ f
^^•^^^•^"1
1
j 1
^1 1
lA-r
)
■■■
3000
taoo
e/lM.
Fig. 11
PARALLEL OPERATION OF SYNCHRONOUS
MACHINES
A: Steady Strain
Let two alternators or groups of alternators,
such as stations or station sections, of the
same terminal voltage, be connected with each
other through a reactance, or more general
through an impedance, and in synchronism
with each other.
We may assume the alternators to be of
equal voltage, since a voltage difference
merely superimposes on the synchronizing or
sioo
Fig. 12
energy current flowing between the alter-
nators a reactive m.agnetizing current, with-
out materially changing the energ>" relations,
and the equations thus are of the same
general characteristics, merely a little more
complicated.
694 August, 1920
GENER.\L ELECTRIC REVIEW
Vol. XXIII, Xo. 8
If the loads on the two alternators equal
the power output of the respective machines,
no power flows over the impedance between
them. If, however, the load on the one
alternator is greater than on the other by the
same amount less than its outi^ut, power must
flow over the impedance. The load on the
alternators varies with the changing con-
ditions in the system; the relative output of
the alternators or groups of alternators how-
ever is fixed by the speed governors of their
prime movers and can be varied only in steps,
by shutting down a machine or starting an
additional machine. Thus the output of each
generating section cannot always equal the
load on it. and an exchange of power must
occur between the generating sections, that is,
power must flow over the impedance be-
tween the generating sections.
Let:
P = the power flowing from the under-
loaded to the overloaded alternator, over a
circuit of impedance z and let:
2aj = the phase displacement between the
two alternators, caused by the flow of power.
The e.m.fs. of the two alternators then may
be represented by:
ei = eo cos (</) — oj)
€2 = Co cos (<l>+03)
where eo = niaximum value of e.m.f.
and 0 = 27r//.
The resultant e.m.f. acting in the circuit
between the two alternators, then is :
e = ^1 — «'2
= eol cos (<t> — w)—cos (0+co) > (2)
= 2«'o sin CO sin 4>
that is, in quadrature with the average vc It-
age of the two alternators.
The interchange current between the two
alternators then is:
. e
(1)
2^0 . . ,^ X
= — stn w sin (<t>—a)
(3)
where :
r = resistance
.V = reactance
of the circuit between the two alternators,
including their internal resistances and re-
actances.
The phase angle a is given by:
X
tan a = -
r
The effective value of the current i is then
gi\-en by:
T V 2 fo ■
1 = sm w
.V
or, if
£ = effective value of generator e.m.f.
e(i = E\/2, and
2£
/ = ^— sin 03
(4)
The power consumed in the resistance r of
the circuit is:
P' = r-r
■iE-r
sin- 03
4E-
stn- CO cos a
Co)
The power of the first alternator then is:
pi = eii
2fo-
= - — 51 H w sin {4> — a) ccs (0 — co) (6)
The power of the second alternator:
P,=
sin u sin (4> — a) cos (0-|-aj) (7)
The sum of the power of the two alternators
then is:
= - — sin u sin (tt> — a) [cos [<p — w) —
~- I'di" (0 + 0})]
4r„
-sin- 03 sin <j> sin (0 — a)
J. Co
- — sin- 03 [cos <t>—cos (2 <(> — a)
and its average value thus is:
cv. p =—^sin- 03 cos a
4E-
stn- 03 cos a
P'
thai is, the sum of the powers of the tw<^
alternators is the power consuired in the
resistance of the circuit between them, as is
obvious.
The difference of the powers of the two
alternators is:
2 /J = pi - />.
= ——sin 03 sin (<l>—a) [cos {<t> — 0))-^
cos {<f>+03)]
4fo- .
= sm 03 cos 03 cos <t> sin (0 — a)
POWER CONTROL AND STABILITY OF ELECTRIC GENERATING STATIONS 695
'2 p= — r^'" - ^ [■^""' oc — siii {.2 0 — a)) (S)
and its average value thus is:
av. 2 p= —sin 2 co sin a
2E- . .,
= sin 2 CO sin a
= 2P (9)
that is, the power transfer between the two
alternators (or generating stations or sections
of generating stations) is:
F= — sin
z
(10)
and the leading alternator, e^, delivers power
to the lagging alternator, e\.
The power P thus is zero for w = 0, increases,
reaches a maximum of
P'
-sin a
(11)
for CO = 4.5 deg., or 90 deg. phase displacem.ent
between the alternators, and then decreases
again to zero at co = 90 deg., or phase opposi-
tion of the alternators.
Beyond aj = 90 deg., the synchronizing
power Pm becomes negative, with the same
values, that is, the alternators s^-nchronize at
the next pole.
The synchronizing power P is zero for
a = Q, that is, if the circuit between the alter-
nators contains no reactance but only resist-
ance, and is a m_aximum when the resistance
is negligible compared with the reactance,
that is a = 90 deg. :
P"„, = -sin
X
w
(12
Substituting in (10)
SI n a --
grves :
P= „ sin
(l.'«
that is, with a giA'cn impedance s, and thus
given synchronizing current between the
alternators, the synchronizing power P is
directly proportional to the reactance x of the
circuit between the alternators.
The maximuni synchronizing power be-
tween the alternators thus occurs at phase
angle w = 4o deg., that is, 90 deg. phase dis-
placement, and negligible resistance, and is:
Pm=^ ■ (14)
at current (effective) :
-1 m
(15)
and resultant e.m.f. :
£m = Ev'2 (Ifi)
In this case, the phase angle 2co between
the alternators or station sections is constant
during operation, but varies with change of
load between the station sections, and can be
kept very small by properly apportioning the
number of generators in operation in each
section to the respective load on this section.
B: Oscillation
Consider again two alternators or groups of
alternators, such as stations or station sections
which are running in synchronism with each
other, that is, have the same frequency, /, but
are connected together while out of phase
with each other by angle 'ico, or thrown oi;t cf
phase by some sudden change of lead,
momentary short circuit, etc. / s v ell
known, the alternators then oscillate against
each other with (practically) constant fre-
quency of oscillation pf and gradually
decreasing amplitude of oscillation, and
finally steady down in phase with each other,
or at the constant phase angle co° determined
by the condition of steady power transfer
between the alternators.
Since under normal conditions of operation
the stead}' phase angle co" must be small, we
may assume that the oscillation occurs
s\Tnmeterically around the position of the
alternators in phase with each other, that is,
the one alternator has the phase </> — to when
the other has the phase 0-fco.
The same equations then pertain as in the
foregoing section Steady Strain, that is:
The e.m.fs. of the two alternators are:
ei = eo cos (<^ — co)
e^ = eo cos (0+co) (1)
The e.m.f. acting in the circuit between the
two alternators:
^ = 2 ^0 sin CO sin 4>
with effective value :
E> = 2Esin(ji (17)
The current flowing between the two
alternators :
. 2 Po .
-sin CO (sin <i> — a)
with effective value :
2£
/ = sin CO
(3)
(4)
696 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 8
where z is the impedance of the circuit be-
tween the two alternators or groups of alterna-
tors, including their internal impedance. The
power transferred between the alternators is:
£2 .
p = — sm
0) [sin (2 4>~a)—sin a]
(8)
The first term of equation (8) is of double
frequency, '2f. It thus does not represent
energy transfer between the alternators, but
merely represents the energ}.- storage and
return, twice per cycle, occurring in any
inductiA'e circuit. It thus is of no further
interest, and it is;
Power transfer between the alternators:
£2
Substituting (19) into (10) gives as the
periodically var\-ing power transfer or syn-
chronizing power:
P = — sin a sin 2 (coo sin p<t>) (20)
where Wo is the maximum amplitude of this
oscillation.
The average value of P during the half cycle
of oscillation m.ay be represented by :
Pc=av. P
E- . 1 —cos 2wo .,..
= — sin a
2aJo
P =
sin 2 o) sin a
(10)
and as the duration to of one half cycle of
oscillation (during which the power transfer
remains in the same direction) is given by
half a cycle of pit>, that is:
p4> = 2-Kpfta = IT
EH
sin 2 oj
(12)
In this case, however, the phase
angle co of the e.m.f. i§ not constant,
but pulsates at the approximately
constant frequency of the beat and
decreasing amplitude.
Let:
aJo = Wooe"" (18)
be the maximum value of the phase
angle during each oscillation, (decreas-
ing from its initial maximiam value coon
by the expotential of time t"^*-).
We may then represent the
gradually decreasing amplitude of
the phase angle co by :
aj = Wo sin p<t>
= 0)00 6^"' sin p4>
■* • V-tK/ jfr
^g^lllliw^^^^g^^^
n 1 1 1 M M I M M I M I I M M I M M M M M 1 1 M \MVr
fton s^ncfirffAOiUDrt/Onf
SyncArontua OaciMtimna
Fig 13
It is:
(19)
'o = ;
1
(22)
where :
p/ = frequency of the beat, or the (com-
plete) periodic variation of the phase angle oj.
In reality, the equations (3), (4), (.S), and
(10) of section A are not strictly correct for the
conditions under investigation in section B.
since in the derivation of these equations in
section A, w was assumed as constant. In the
case of section B, oscillation of the alternators,
o) varies periodically, isa function of 0 and thus
additional terms appear in these equations.
Since however the frequency of variation of
w is very low comi)arcd with the frequency of
/, that is, p equals a small ciuanlity, these
additional terms are small; and the foregoing
equations thus are correct with sufficient
approximation, especially in the present case,
where we are essentially interested in the
magnitvidc of the power relations.
2pj
and the energy transfer between the two
machines or groups of machines, during each
half cvcle of oscillation, is thus given bv:
lI- = /oPo
E-
= sin a (\-cos2cio) (23)
4p/a;o2
This is a maximum for Wo = 90 deg. =-;^, and
then is:
\V„=:^^sina (24)
U'm thus is the maximum energy which can
be absorbed by ihc machine or group of
machines, without being thrown out of syn-
chronism. In other words, if a sudden demand
greater than \\'„ is made on the machine, or if
more encrg\- than 11 '„ is given by the steam
supply to the machine or group of machines,
after the load has been lliniwn otT and before
POWER CONTROL AND STABILITY OF ELECTRIC GENERATING STATIONS 697
the steam has been cut off, the machine is
thrown out of synchronism ; otherwise it
remains in synchronism and after an oscil-
lation settles down again in phase.
As seen from the equations, during each
complete cycle of oscillation of frequency pf,
the current twice rises and falls, thus reaching
two maxima, and the power P twice reverses, so
that the energy W flows one way during half
the cycle, and in opposite direction during the
other half cycle of oscillation. The frequency
of the rise and fall of the current thus is 2 pf.
Curves I and II in Fig. 13 show the current
i and the voltage ei of the oscillation for the
1 1 M M M M 1 1 1 1 1 M
■'
'
1 ' '
1 1 1 f
-
~'
r^
^
>
T"
s
/
V
^\
,
.-
■-f'
,
■' 1
rfi-
■1"
r
-
\
^
f*
■
.;-
■p'
Oirt
■^
-J
„
^
V
y
s
'■
■■
--
"
/
/
t
^
.'
*.
^
^
n-
/
f*
•n
'••
■>
•■*
fi
i>
1
i
^^
1
r'
/
OscilJotiO
7(y eo'
_j
^
'i
■ ■\
r*
f
N
,.
-1
r-
^*
/
r
>
■
'■^
*'
/
*
^
"N
k
^
f
'■
-
'
/
*
S
r
,*
Hi
?*
'«
\
777
/
.^
1
*
'
'v
/
\
/
^
'
i"
1
f
Jy
i
'
rJ*
^
/
\
■'
"
\
/
\
,
•'
/
\
/
\
*
/
y
^
/
^
t
H
/
,f
*
\
/
,•
/
,
V
'•
^
/
k
(^
^
-
-
f'
\
N
>
,
'!
N
\
'
/
1
\ """"
'
y
«;
\
/
\,
/
■»
\
'*
J
S
^-^
/
L
J
_
_
L
U
L
u
_
u
U
L
u
_l
u
u
u
u
L
-
_
J
_
_
_
Fig. 14
(exaggerated) value p = 0.\. and for coo = 45
deg. and Wo = 90 deg.
It is interesting to note from equation (20)
that the power transfer P reverses twice per
cycle of oscillation (for p(t> = 0 and 180 deg.).
If a;o = 4o deg. or less, that is, 90 deg. or less
maximum phase displacement during the
oscillation, then the power P has two maxima,
at the maximum phase displacement mid-
ways between the reversal of power, as seen
in Cun.'e I of Fig. 14. If however coo is
greater than 45 deg., that is, more than 90 deg.
phase displacement, then the power transfer
decreases again at m„aximum phase dis-
placement, midways between the reversals
of power, and the power transfer has four
maxima separated by two reversals and two
minima, as seen bv Curve II of Fig. 14, and
finally at Wo = 90 deg. (Curve III in Fig. 14),
the power reaches four maxima and four
zero values during each cycle of oscillation
but reverses only twice. That is, at the
moment when the two alternators are in
phase, the power transfer is zero, the power
reverses, and the current is zero, and in phase
with the voltage. With increasing phase dis-
placement, power and current increase; the
power reaches a maximum at 90 deg. phase
displacement between the machines, where
the current is 45 deg. out of phase with the
voltage. With further increase of phase dis-
placement during the swing of oscillation,
the power decreases again to zero at 180
deg. phase displacement or phase opposition;
but the current continues to increase and
reaches a maximimi at phase opposition, with
the phase angle between voltage and current
steadily increasing, to 90 deg. or zero power,
in phase opposition. Then, without reversal
of the flow of power, the phase angle between
voltage and current again decreases, the cur-
rent decreases, but the power increases again
in the same direction as before, to the second
maximima in the same half cycle, at 90 deg.
phase displacement, and then the power
decreases again to the reversal. This is well
illustrated in Figs. 13 and 14.
C: Slipping
Consider now the case that two alternators,
or groups of alternators such as station sec-
tions, are connected together while different
from each other in frequency by 25, that is,
one alternator has the frequency (1— 5)/, the
other the frequency ( 1 -|-5)/,and the alternators
thus are slipping past each other with the fre-
quency 25/.
We m.ay again assume the alternators as of
equal voltage, since a voltage difference
merely superposes on the synchronizing
energy current a reactive magnetizing cur-
rent, without materially changing the energy
relations.
The e.m.fs. of the two alternators then may
be represented by :
^1 = ^0 cos (1 ~s)(j) \
f2 = eo cos (l-)-s)</) /
The resultant voltage in the circuit between
the two alternators then is:
(25)
e = ei — ei
= eo [cos (I-5) </)— cos (l-|-5)<^]
= '2eo sin s <i> sin 0
= '2 E\'^ sill s </> 5/;; <t>
(2G)
69S August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. S
and its effective value :
E" = 2 E sin s 4> (27)
where £ = effective value of generator e.m.f.
Assume now that 5 is a small quantity
(just as we assumed in section B, that p
is a small quantity), that is, that the two
alternators have nearly the same frequency.
The change of sin s<i> then is slow compared
with that of sin 0, and for all phenom.ena
of frequency /, sin 5 </> may be assum.ed
as constant, and the reactance of the circuit
may be assumed as the sam_e, x = 2irfL, for
both component e.m-.fs., ei and €2. that is, for
both frequencies (1— 5)/and (1+5)/.
The interchange current between the alter-
nators then is:
2p = pi~p->
t = sm s 4> sin (0 — a)
z
_2EV2
sin s 4> sill {(p — a)
hence, the effective value:
2 E
I = ~ — sin s (j>
where :
(28)
(29)
tan a = —
r
With regard to the e.m..f. of one of the
alternators, for instance, Ci, this current
always lags. Its lag is 90 deg. when the cur-
rent is a maximum. With the decrease of
current, the lag decreases from 90 deg. in the
one, and increases in the next beat, and
approaches in-phase respectively in-opposi-
tion, when the current is a minimum. The
power-factor thus varies from zero at rraxi-
m.um. current, to unity at zero current, and its
average thus is low. Fig. \:i shows as Curve
III the relation of ei to i for the exaggerated
value 5 = 0.09.
The jjower of the one alternator then is
given by:
P:=eii
^■2e^
z
4 £2
= sm s <t) si)t (<t>~ a) cos ( 1 —5)0 (.'50)
sin s 4> sin (4> — a) cos (/ — 5)<^
that of the other alternator:
pi = e-ii
4 £2
= sin s <t> sm (0 — a) cos (\+s)4> (.'51)
z
and the power transfer between the two
alternators then is gi\-en by :
8 £2
2£=
sin s 4> sin {<i> — a) cos s </> cos <^
= - - - sin 2 5 0 [sin {2 4) — a) —sin a] (.32)
The first term, with sin (2 0 — a), again is a
double frequency term, representing the
periodic storage and return of the energy-
during the half cycle of voltage, thus does not
represent any power transfer and the power
transfer between the alternators is thus
giv'en by:
P = — sin 2 s 4> sin a
(33)
L'sually it is approximately: a = 90 deg.,
that is, the reactance is large com.pared with
the resistance, and equation (33) then
becom.es :
£-
P = — sin 2 50
(34)
During each cycle of the frequency sf, of
ths slip from synchronism, or average fre-
ciuency, the am.plitude of the current i thus
twice becom_es zero and in phase, and twice
reaches a maximum, when the alternators are
in opposition, and the power P four tim.es
reaches a maximiun and four times becom.es
zero and reverses, twice when the current
comes into phase with the e.m.f. but the
current becomes zero, and twice when the
current is a maximiim but in quadrature witli
the e.m.f., and the power thus becom.es zero.
The power transfer between the alternators
thus reverses four times per complete cycle of
sli]), sf, that is, is of the frequency 25/, with
two positive and two negative maxima.
The average value of the power is:
o 2 £-
IT
-P =
w z
-sin a
(3.->)
and as the duration of one quarter cycle of
slip is '0 = 7-7, the energy transfer between the
two m.achines, during a quarter cycle of slip,
thus is:
■isf IT
E^
= - ^ sm a (36)
2 7r sfz
The difference between the slipi)ing of
alternators past each other out of synchro-
nism and the oscillation of alternators against
each other at synchronism is thus that in the
slipping the power fluctuation and the
reversal of the energy is of twice the frequency
POWER CONTROL AND STABILITY OF ELECTRIC GENERATING STATIONS
699
of the current fluctuation, while in the oscil-
lation of the alternators against each other
at synchronism, the power fluctuation or
reversal of energy flow is of the same fre-
quency as the current fluctuation
If two alternators are connected together
while out of synchronism, and slowly slip past
each other, during each half cycle of slip, or
beat, while the two machines e.m.fs. pass
from in-phase, to in-opposition, to in-phase
again, a periodic energ\' transfer takes place.
During one quarter cycle of slip (that is, while
one alternator e.m.f. slips behind, the other
pulls ahead of the mean frequency by one
quarter cycle, and the two alternators,
e.m.fs., thus slip against each other by one
half cycle) the alternators are partly in phase
with each other, and the slower machine
receives energA^ from the faster machine.
The two machines are thereby brought nearer
to each other in speed, pulled towards syn-
chronism. During the next quarter cycle of
slip, however, the two alternators are partly
in opposition, and the faster machine receives
energ>' from the slower one. The faster
machine then speeds up, the slower machine
slows down, and the two machines pull apart
again, by the same amount by which they
pulled together in the preceding quarter cycle
of slip (if their e.m.f. is constant). Thus the
two machines can pull into step only if the
energy transferred during one quarter cycle
of slip, ir, is larger than the energy required
to speed up the momentum that is, kinetic
energy M of the machine to full synchronism.
Due to the energy transfer IT between the
machines, resulting in an alternate speeding
up and slowing down, the slip s is not constant
but pulsates periodically between the mini-
mum value s— 5i, at the end of the quarter
cycle during which the machines pull together
and the beginning of the quarter cycle during
which the machines pull apart, and a maxi-
mum value s + Si, at the end of the quarter
cycle during which the machines pull apart
and the beginning at the quarter cycle during
which the machines pull together; where Si
is the amplitude of the pulsation of slip. As
the energy required to accelerate the momen-
tum M of the machine by the speed 25] is
■isiM, it follows:
ir = 4siM
Sl =
^\-
w
AM
E^ sin a
' SwsfzM
is the amplitude of the speed fluctuation of
the two alternators during the slipping past
each other out of synchronism with the slip s.
Si = s gives as minimum slip s — Si = 0, that
is, the m.achines pull into synchronism.
The maximum slip 5i from which the two
machines pull into synchronism with each
other is thus gi\-en by substituting 5i = 5 in (37)
Sa
_E ■! stn a
~^\2Tr fzM
(38)
(37)
St, thus is the limit of synchronizing pou-er.
In Fig. 14, four curves of power and of
current (effective value) are shown, the
fonner drawn in full and the latter in dotted
lines, for oscillation; coo = 30, 60, 90 deg., and
slipping.
As seen, the single maximmn power, Cur\-e
J, with increasing swing of the oscillation,
becomes a double maximum with a minimum
between the maxima. Curve //; the minimum
then decreases to zero. Curve /'//, at the
limits of synchronizing power; and the power
curv-e then overturns. Curve IV , that is, the
alternator instead of swinging back into phase
again continues to slip and drops into phase
again by slipping one cycle, etc., and thereby
the power transfer curve doubles its frequency
by one of the two lobes of Curve /// over-
turning, while the current curve rem.ains the
same, of the frequency of the beat or slip.
D: Pulling in Step
When the two machines are out of syn-
chronism with each other by a greater speed
difference, 25, than that from which the
machines can pull each other into synchro-
nism within one quarter cycle of slip, from
the equations of the foregoing section C it
would follow that the m.achines can never pull
each other into synchronism, if the voltage
£o is constant, but must indefinitely continue
to slip past each other coming nearer together
during one quarter cycle of slip and dropping
apart again by the same amount during the
next quarter cycle of slip.
This, however, is under the assumption
that the machine e.m.f. E is constant. In
reality, however, E is not constant but varies
periodically with the same frequency as the
current fluctuates. The current in the cir-
cuit between the machines and thus the
armature reaction in the machine varies in
amplitude and in phase difference against
the machine voltage, and the machine volt-
age varies with the am.plitude and the phase of
the armature reaction.
700 Au£;ust. 1920
GENER.\L ELECTRIC REVIEW
Vol. XXIII. No. S
Consider, as an approximation, the
armature reaction as proportional to the
quadrature component of the current. The
e.m.f. of the machine would then be expressed
by an approximate equation of the form:
E'
=e[\-c
sin s<f> sin 8
(39)
where c is a constant and 5 is the phase angle
between the current and the e.m,.f., and 5<^
represents the amplitude of the current pulsa-
tion by equation (29), thus sin s <j> sin 5 rep-
resents the quadrature component of the
armature current.
It is, however, by (25) and (29) :
5 = {<t> - a) - {I -s)<l>- 90 deg.
= 5</)-a-|-90 deg.
thus :
E' = E ■ 1 -\-c sin s4> cos (s4> — a) )
(40)
Substituting equation (40) into the expres-
sion of the power of the alternator equation
(33), the equations still remain alternating,
that is, there is no resultant synchronizing
power, but equal positive and negati/e values
of power alternate.
However, equation (40) assumes that the
magnetizing effect of the armature reaction is
instantaneous, that is, that the e.m.f. E at
any m.oment is the value corresponding to the
arm.ature reaction existing at this mom.ent.
This, however, is not the case; the armature
reaction is not instantaneous but requires an
appreciable tim.e, several seconds, to develop,
and the magnetizing or demagnetizing effect
of the armature reaction on the voltage there-
fore m.aterially lags behind the armature
reaction.
Let then or = the angle of lag of the voltage
change behind the arm.ature reaction which
causes it. It is then:
E' = E<. \+c sill s4>cos {s(t> — a—a)
(41)
and substituting equation (41) into (33) gives
the power transfer between the machines:
!,£-•<,'• !l+csins<t> \«
z [ cos (s<f> — a—(T] !
or approximately, considering c as a small
quantity :
E- 2cE-
P = — sin 2s(t> sill a+ ~ sin '2s(t> sin a
sin s<t> cos (s<f> — a—(X>
(42)
The first term :
Er
— sin 2s<t> sill a
is the slowly alternating energy transfer
between the machines, discussed in sec-
tion C, which causes their speed to fluctu-
ate, but does not pennanently bring them
nearer to each other, that is, exerts no s>ti-
chronizing power unless and until during
these speed fluctuations they reach complete
synchronism and then catch into step.
The second term:
2cE-
P\= sin 2s(j>sin a. sin s4> cos{s4>— ot—a)
n 2s<psin a[ sin {2s.t>-a-<J J
cE- .
cEP . , . ._ .
= — stn 2s<i> sm a sm ( a+a)-\ — r— 5»m a
2 23
Icos (■ls<l> — a—a)+cos (a-rffil
cE-
cE-
cE--
= —;^stn2s(i>sin asm {a-\-<T) + -^ sm a
cos {■iS(t>—a — <T)+-^ sm a
cos (a+a) (43)
The first two terms also are slowly alternat-
ing, at double and quadruple the frequency
of slip, as they contain terms with 2s(t> and 4s<i>
and thus represent no continuous power trans-
fer; the third term, however.
„ cEr .
Po = -^ Stn a cos (a+a)
(44)
is constant, that is. represents a continuous
synchronizing power.
If a = 90 deg., that is, the resistance is
negligible compared with the reactance:
(45)
If thus two alternators or station sections
are considerably out of synchroni.sm with each
other, they continue slipping past each other
with large fluctuating currents flowing be-
tween them, and the speeds of the machines
fluctuate with the fluctuations of the current.
These currents do not decrease in amplitude,
but remain of practically constant value, but
their period of fluctuation gradually becomes
slower, that is, the fluctuation gradually
becomes slower, while currents slowly pull the
machines nearer into synchronism with each
other, that is, decrease their frequency differ-
ence, until the critical frequency 2so is reached
(where the acceleration during a quarter cycle
of slip, 2si, reaches full synchronism). Then
the machines suddenly drop into synchro-
ni.sm. but oscillate in phase against each other
with an approximately constant frequency
of oscillation, but with a current fluctuation
which steadily (and usually rapidly) decreases
until steady conditions of speed, current, and
voltage are reached.
POWER CONTROL AND STABILITY OF ELECTRIC GENERATING STATIONS 701
The armature reaction of the alternator is
represented by the difference of the syn-
chronous reactance xo and the true reactance
Xi, that is, by an efifective reactance of arma-
ture reaction.
X2=Xo — Xi.
The coefficient c in the synchronizing
power, Po in equation (44), is that fraction of
the reactance of the armature reaction X2
which appears during the short time of the
current fluctuation. Thus c is larger, the
slower the fluctuation, that is, the less 5.
In other words, c increases with decreasing
slip, that is, increasing approach to syn-
chronism.
Inversely, since ff is a maximum and practi-
cally 90 degrees for large values of 5, where
the voltage fluctuation lags practically 90 deg.
behind the fluctuation of the armature
reaction, and decreases with decreasing s,
that is, increasing approach to synchronism,
c siti a, and thus the synchronizing power,
Po in equation (44), should be a maximum at
some moderate slip .j and decrease for larger
as well as smaller slips.
Assume that it takes ^0 seconds for the
field to build up to correspond to the armature
reaction. With the current fluctuating with
the frequency 2sf, and assuming that the
magnetizing effect of the armature reaction is
sinoidal, it would be:
1
4sfto
and:
thus :
ina = \l\ — I - — r- I
Po-
Eo-
S 35/^0
However, secondary effects occur
or less modify the value Po. such as
of secondary currents induced in
structure by that component of the
current which is due to the e.m.
other machine, and which gives an
motor torque tending to pull the
together into synchronism.
E: Equations
3= \/r^ + x''
(40)
and more
the effect
the field
armature
f. of the
induction
machines
total impedance of circuit be-
X
tween alternators. tan a =
r
CO = phase angle between alternator e.m.fs.
/)/= frequency of oscillation.
5/ = frequency of slipping.
Eff:
\/2
.4; Continuous Power B.' Oscillation C and D.' Slipping
Transfer in Synchronism Out of Synchronism
Alternator e.m.fs.:
?1 = fii cos (<() — <j) «(! COi (<> — w) Cis co% {\ — s)ii>
Pz = fii f oi (* -H m) en coi (<() -|- u) es>cos{\-\-s)^
ta en
Resultant e.m.fs.:
e =2esin w sin tt> 2 e sin a sin <)> 2 c sin S(t> sin <t>
Eff:
£" = 2 £" sinu 2 £" sin u 2 £" sin s <}>
Resultant Current:
. 2 f n . . 2 1,1 . 2 CT .
1= sin tit sni itp ~ a) stn i^ sin {if — a) — sm s ip sin {ifi — a)
s s s
Eff:
2E 2 E
I = — sin u - — sin w
Continuous Power Transfer:
E-
Pv = — sin 2w 51 H a
2£ . ^
— si n s <t>
sm a COS { a -f o
Low-frequency Power Fluctuation
P =
Low-frequency Energy Transfer
ir =
Attenuation:
E" . . E-
— sin 2 (Jo sin a — sin2 s <l> sin a
z z
E- . 1 —cos2ic
-- sina
4 pfwo 2 TT sfz
u} — ttiasin pti>
= wooe""' sin p(t>
Pulsation of Slip:
Si =
Critical Slip:
So =
Pulsation of Armature Reaction:
c =
Lag of Armature Reaction:
sin a =
E' sin a
8 TT sjzM
— -»! ^
jzM
1
4i/fa
\ l-c=
It is interestitig to note that the limit case
7r •
of ir, in section B for w = — , and in section^C
for 5=5o. must coincide: irB = irf. This
gives:
/£= . l-fOs2M,, / , / £- • /
/ — sma —-7 ' = 2 ' J- sjn a
I 2 -t />/".. /co„ = -^ 12-^sjz js = Sl,
Hence:
p=Sa
and, substituting for so'-
A = — ^i" "
^ 2\2,r^M
is the frequency of oscillation.
(To be Continued)
702 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. S
The Penetration of Iron by Hydrogen
By T. S. Fuller
Rese.\rch L.^boratory, Gener.al Electric Comp.\ny
.This article describes the results o£ tests made to determine the eflfect of current, treatment of the iron,
temperature, electrolyte, and surface coating on the penetration of iron by nascent hydrogen. A new and
convenient form of apparatus was developed for the experiment. It was found that the rate of penetration
increases with increase in temperature and is greater for iron immersed in one per cent sulphuric acid than
for iron electrolyzed as cathode in a similar solution. Copper is impervious to hydrogen at ordinary tempera-
tures. Coating the iron with tin increases the rate of penetration, while coating it with zinc or copper has
the opposite effect. — Editor.
Some time ago the writer published the
restilts of som.e plating experiments on steel
springs in which he concluded: "The facts
are all in accord with the assumption that the
absorption by the steel of atomic or nascent
hydrogen liberated at the cathode is the cause
of the embrittlem-cnt of steel springs in the
plating-bath." The results of these experi-
ments, together with those of Charpy and
Bonnerot', Coulson-, Merica', and others led
to the study of the factors which govern the
penetration of iron by atomic hydrogen,
namely, current, treatment of the iron, tem-
perature, electrolyte and surface coatings.
It is well known that iron at room tem-
peratures is impermeable to gaseous or molec-
hydrogen expressed in cubic centirreters per
hour, collected in the steel tube:
Temperature
1 cc. of H: per hr
350 deg.
C.
1.1
450 deg.
c.
3.2
.5.50 deg.
c.
8.5
750 deg.
c.
30.0
850 deg.
c.
42.0
Pressures as high as 26 atmospheres were
obser\-ed by these writers in an iron con-
tainer which was made cathode in an acid
solution.
Fig. 1. Diagram showing Construction of Apparatus for Measuring the
Rate of Hydrogen Penetration
ular hydrogen (Ho), but that at higher tem-
perature it becomes more or less permeable.
It is not so well known, perhaps, that iron at
room temperature is permeable to nascent or
atomic hydrogen (H). These problems have
been investigated by Charpy and Bonnerot'.
Their apparatus consisted of a steel tube ()■')
mm. thick, connected to a pump which
maintained a constant pressure of 0.2 mm.
inside the tube. This tube was placed inside
another one of porcelain, which was heated
and through which gaseous hydrogen at
atmospheric pressure flowed, surroimding the
steel tube. The iron was found to be imper-
meable up to a temperature of ;52.j deg. C.
The following table taken from Charpy and
Bonnerot's data shows the amount of gaseous
■ Trans. Am. Elictrochem. Soc. VoL 32 (1917), p. 247-2.V5.
' Compt. Rend. 154. 592-4 (1912), 1.56. 394-6 (1913).
• Trans. Am. Elcctrochem. Soc. Vol. 32 (1917). p. 2.37-24.i.
• Met. and Chem. Eng.. Vol. 16 (1917). p. 496-503.
Apparatus
A very convenient as well as novel form of
apparatus has been developed and used for
the experiments described in this paper.
It consists of a seamless iron tube plugged
at the bottom, and sealed at the top to a glass
U tube having one arm closed and cali-
brated and the other open. After assem-
bling, the apparatus is completely filled with
m.ercur>% an(i when in operation hydrogen
penetrates the iron tube, and rising quickly
displaces the mcrcviry in the closed and cali-
brated arm of the U. Fig. 1 is from a draw-
in;.; showing the construction of one of the
units. .4 is a section of seamless iron tubing
12 in. long, and \^ in. outside diameter,
with a wall ,'g in. thick, i^isa cone-shaped
tube of nickel steel having its small end brazed
to the iron tube A . The large end of B makes
a \acuum tight seal with the bottom of the
THE PENETRATION OF IRON BY HYDROGEN
703
Fig. 2. Apparatus Completely Filled
with Mercury
Fig. 3. Apparatus Showing Partial Displacement
of Mercury by Hydrogen
Fig. 4. Arrangement for Measuring Hydrogen Penetration at Different Temperatures
r04 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. S
glass tube C, which terminates at the top in
the U tube D-E. D is closed at the top and
calibrated, and E is open. The capacity of D
is about two cubic centimeters. F is a plati-
num wire which is sealed through the glass
and dips into the mercury. G is an iron plug
which is brazed into the bottom of the iron
tube A. H iscL section of rubber tubing which
protects the brazed joint between A and B,
as well as B itself, from the action of the
electrolyte, and 7 is a coating of rubber
cement which protects the brazed joint be-
tween ^4 and G. The finished unit is so
arranged that the electrolyte comes in contact
with only iron and glass. When the unit is
complete it is sealed to a vacuum pump and
tested for leaks. None have been used which
could not be pumped out to a pressure of
0.005 m_m. of mercurs' without showing a leak.
After filling with mercury the units are ready
for use.
Fig. 2 is a photograph of one of the units
completely filled with mercury, and Fig. 3 is
a photograph of another unit, the mercury in
the measuring tube of which has been par-
tially displaced by the hydrogen which has
collected. Fig. 4 shows the scheme for
measuring the rate of hydrogen penetration
at dift'erent temperatures. Six units are
immersed in glass tubes, containing the
electrolyte, these tubes being in turn
immersed in an electrically heated water-bath.
Effect of Current
To determine the relation between current
and the rate of hydrogen ])enetration, four
units were made cathode in one per cent
sulphuric acid at room temperature, the units
being connected in the circuit by means of the
platinum wire F. Platinum anodes were used.
Two of the units were electrolvzod with a
current of 0.2 amp. and two with a current of
0..5 amp. The potential across the former
cell was 2.5 volts and across the latter 2.8
volts.
Two of the units were electrolyzed a second
time.
In every case a current of 0.5 amp. allowed
more hydrogen to penetrate the iron than did
a current of 0.2 amp., but not 2 J/2 times as
much. The velocity of hydrogen penetration
is not a straight line function of the current.
The rate is influenced by the current; the
higher the current, at least for such densities
as were used in this experiment, the greater
the velocity of penetration.
No Electrical Connections
One of the units was immersed, without
electrical connections, in a solution of one per
cent sulphuric acid, in other words "pickled."
(Table II.)
At the end of 4S hours, hydrogen somewhat
in excess of 2 cc. had collected. This rate of
penetration is higher than the rate for the
units made cathode with currents of 0.2 or
0.5 amperes.
Effect of Repeated Electrolysis
The same unit was made cathode in a one
per cent sulphuric acid solution under the
same conditions for four successive runs with-
out intervening rest periods. (Table III.)
The velocity of penetration increased with
each successive electrolysis.
Effect of Acid "Pickle"
A unit which had been pickled for 4S hours
in one per cent sulphuric acid was immediately
made cathode in the usual manner. Electro-
Ivte was of one per cent sulphuric acid.
(Table IV.)
TABLE I
Description of Unit
(1) Iron — New
(2) Iron— New
(3) Iron— New
(4) Iron — New
(5) Iron — same unit used in (3)
(6) Iron — same unit used in (4)
Current
Temp.
Volume of
Hydrogen Collected
0.2 amp.
19 hr.
20 deg. C.
0.55 cc
()..5 amp.
19 hr.
20 deg. C.
1.85 cc
0.2 amp.
24 hr.
20 deg. C.
0.15 cc
39 hr.
20 deg. C.
0.90 cc
42 hr.
20 deg. C.
l.(K) oc
0..T amp.
24 hr.
20 deg. C.
0.40 cc
39 hr.
20 deg. C.
1 .35 cc
42 hr.
20 deg. C.
1.50 cc
0.2 amp.
24 hr.
20 deg. C.
0.30 cc
33 hr.
20 deg. C.
1.00 cc
48 hr.
20 deg. C.
2.40 cc
0.5 amp.
24 hr.
20 deg. C.
0.30 cc
33 hr.
20 deg. C.
1.47 cc.
48 hr.
20 deg. C.
2.40 cc.
THE PENETRATION OF IRON BY HYDROGEN
ro.-)
TABLE II
Description of Unit
Time
Temp.
Volume of
Hydrogen Collected
24 hr.
27 hr.
33 hr.
48 hr.
20 deg. C.
20 deg. C.
20 deg. C.
■ 20 deg. C.
0 cc.
0.02 cc.
0.42 cc.
2.00 cc.
TABLE III
Description of Unit
Current
Time
Temp.
Volume of
Hydrogen Collected
Iron — New
0.5 amp.
0.5 amp.
0.5 amp.
24 hr.
39 hr.
42 hr.
20 deg.
20 deg.
20 deg.
C.
C.
C.
0.40 cc.
1.35 CC.
1.5 cc.
Same unit used above
0.5 amp.
0.5 amp.
0.5 amp.
24 hr.
33 hr.
48 hr.
20 deg.
20 deg.
20 deg.
C.
C.
C.
0.30 cc.
1.47 cc.
2.00 cc.
Same unit used above ...
0.5 amp.
0.5 amp.
0.5 amp.
5.75 hr.
6.25 hr.
21.5 hr.
20 deg.
20 deg.
20 deg.
C.
C.
C.
0.38 cc.
0.46 cc.
1.82 cc.
Same unit as above
0.5 amp.
0.5 amp.
20 minutes
7.5 hr.
20 deg.
20 deg.
C.
C.
0.08 cc.
2.10 cc.
TABLE IV
Description of Unit
■"Pickled" 48 hr. in on? per cent
H, SO4
Same unit used above
Current
0.2 amp.
0.2 amp.
0.2 amp.
Time
5.25 hr.
20 minutes
7.5 hr.
Temp.
20 deg. C.
20 deg. C.
20 deg. C.
Volume of
Hydrogen Collected
2.20 CC.
0.20 CC.
2.10 cc.
The acid "pickle" facilitated the passage of the hydrogen enormously. The same volume of gas penetrated
this unit in 5-7 hours as penetrated a new unit in 48 hours under the same conditions.
TABLE V
Description of Unit
Current
Time
Temp.
Volume of
Hydrogen Collected
Iron — Previously electrolyzed five
times, followed by a rest of 72
hours
0.5 amp.
0.5 amp.
0.5 amp.
23 hr.
28.5 hr.
31.5 hr.
48 hr.
20 deg. C.
20 deg. C.
1 20 deg. C.
20 deg. C.
0.25 CC.
0.60 cc.
0.92 cc.
2.20 cc.
TABLE VI
Description of Unit
Current
Time
Temp.
Volume of
Hydrogen Collected
Iron — Previously electrolyzed five
times, followed by heating in air
to 130 deg. C. for 4 hours.
0.2 amp.
0.2 amp.
0.2 amp.
0.2 amp.
24.5 hr.
47.5 hr.
60 hr.
67 hr.
20 deg. C.
20. deg C.
20 deg. C.
20 deg. C.
0.03 cc.
0.50 cc.
1.28 cc.
1.9 cc.
706 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 8
Effect of Rest or Heating
Rest — A unit which had been electrolyzed
five times, and whose penetration value w as
high, was allowed to rest for 72 hours and
then made cathode in one per cent sulphuric
acid. (Table V.)
Heating — Another unit which had been
electrolyzed a like number of times, and
whose penetration velocity was therefore
high, was heated to 130 deg. C. for four hours
in air and made cathode in one per cent
sulphuric acid. (Table VI.)
Both of these units, after their respective
treatrr.ents, behaved like new ones, that is,
they showed penetration velocities equal to
those of new tubes. The velocity of the latter
unit ■« hich was electroh-zed with a current of
0.2 amp. was less than the former whose cur-
rent was 0.5 amp.
If units which have been operating as
cathode continuously for several days, and
which show a high penetration value; or units
which have been pickled for several days in
acid, be heated to a temperature of 130 deg. C.
for four hours, or be allowed to rest for 72 hours,
they will behave like new units; that is, their
hydrogen penetration value will have been
reduced to that of new units.
Effect of Temperature
A rise in temperature increased the penetra-
tion velocity ver\' greatly. A unit was made
cathode in one per cent sulphuric acid at a
temperature of 90 deg. C. (Table VII.)
The volume of hydrogen which collected in
this unit in 33.9 hours was equal to the amount
collected in a similar tube electrolyzed under
similar conditions at room temperature in
48 hours.
Two m.ore units were electrolyzed simul-
taneously under similar conditions at a
tem.perature of 80 deg. C. The two checked
nicely with each other, but the rate of
penetration was somewhat slower than the
rate of the unit which was electrolyzed at
90 deg. C. The slower penetration was
probably due to the difference in temperature.
(Table VIII.)
The time required to collect 2 cc. of
hj-drogen at different temperatures, other
conditions being equal follows:
Temperature
20 deg. C.
80 deg. C.
90 deg. C.
•18 hr.
4.75 hr.
3.5 hr.
Effect of Various Electrolytes
The rate of penetration varies with different
electrolytes. Units were electrolyzed in one
per cent solutions of sulphuric acid, potas-
sium sulphate, potassium hydroxide and tap
water. The electrolysis in potassium sulphate
was done at a temperature of 20 deg. C. and
is therefore compared to the behavior of a
unit in sulphuric acid at that temperature,
while the electrolyses in potassium hydroxide
and tap water were done at 8.5 deg. C. and
consequently are com.pared to the behavior of
sulphuric acid at that temperature. (Tables
IX. X, XI, and XII.)
The rate of penetration for the unit
imjnersed in potassium sulphate is ver>- slow,
being 0.0019 cc. per hour. The rate for the
unit in one per cent sulphuric acid is 0.024 cc.
per hour, or at this temperature twelve and
one half times the rate of the potassium
sulphate electrolysis.
Electrolyte of One Per Cent Sulphuric Acid
Under the heading. "The Effect of Tem-
perature," it was pointed out that a unit
electrolyzed as cathode in one per cent
sulphuric acid with a current of 0.2 amp.
required 4.7.3 hours at a temperature of 80
deg. C. and 3. .5 hours at a temperature of 90
deg. C. to collect 2 cc. of hydrogen. If it
be assumed that the mean time, or 4.12.') hours,
would be required to collect 2 cc. of hydrogen
at the mean temperature, or 8.) deg. C the
relation shown in Table XIII will hold.
The time necessary to collect 2 cc. of
hydrogen from an electrolyte of one per cent
sodium hydroxide or tap water, other con-
ditions being the same, is 1.5 times the time
required to collect 2 cc. from an electrolyte
of one per cent sulphuric acid.
Hot Water and Steam
Hydrogen produced by the reaction be-
tween tap water and iron, or steam and iron
collected in the iron units at temperatures
ranging from .50 deg. C. to the boiling point.
Tubes without electrical connections were
immersed in water at ,50 deg. C IMl deg. C.
and in steam. (Tables XI\', X\', and XVI.)
A straight line results from plotting the
time required to collect equal volumes of
hydrogen from the water-steam system
against the corresponding temperatures,
measured by the centigrade scale. The
velocity of penetration of the hydrogen result-
ing from the reaction between steam and iron
is greater at steam temperature than the
penetration of the hydr<.)gen of units made
THE PENETRATION OF IRON BY HYDROOEN
TABLE VII
707
Description of Unit Current
Time
Temp.
Volume of
Hydrogen Collected
Iron — Had been electrolyzed once 0.2 amp.
followed by a rest of 72 houri. 0.2 amp.
0.2 amp.
0.2 amp.
2 hr.
2 hr. 50 min.
3 hr. 20 min.
3 hr. 30 min.
90 deg. C.
90 deg. C.
90 deg. C.
90 deg. C.
0.20 CC.
0.98 cc.
1.78 cc.
2.10 cc.
TABLE VIII
Description of Unit
Current
Time
Temp.
Volume of
Hydrogen Collected
Previously electrolvzed heated to
130 deg. C. for 4" hours.
Previously electrolyzed heated to
130 deg. C. for 4 hours.
0.2 amp.
0.2 amp.
0.2 amp.
0.2 amp.
0.2 amp.
0.2 amp.
0.2 amp.
2 hr.
3.5 hr.
4.5 hr.
4.75 hr.
2 hr.
3.5 hr.
4.5 hr.
4.75 hr.
80 deg. C.
80 deg. C.
80 deg. C.
80 deg. C.
80 deg. C.
80 deg. C.
80 deg. C.
0.10 cc.
0.80 cc.
1.69 cc.
2.00 cc.
0.10 cc.
0.60 cc.
1.65 cc.
2.00 cc.
TABLE IX
ELECTROLYTE OF ONE PER CENT POTASSIUM SULPHATE
Description of Unit
Current
Time
Temp.
Volume of
Hydrogen Collected
Iron — New
0.2 amp.
168 hr.
216 hr.
264 hr.
312 hr.
20 deg. C.
20 deg. C.
20 deg. C.
20 deg. C.
0.2 cc.
0.25 cc.
0.45 cc.
0.58 cc.
TABLE X
ELECTROLYTE OF ONE PER CENT SULPHURIC ACID
Description of Unit
Current
Time
Temp.
Volume of
Hydrogen Collected
Iron — New
0.2 amp.
24 hr.
39 hr.
42 hr.
20 deg. C.
20 deg. C.
20 deg. C.
0.15 cc.
0.90 cc.
1.00 cc.
TABLE XI
ELECTROLYTE OF ONE PER CENT SODIUM HYDROXIDE
Description of Unit Current
Time
Temp.
Volume of
Hydrogen Collected
Iron — New ( 0.2 amp.
0.2 amp.
0.2 amp.
0.2 amp.
31.5 hr.
40 hr.
49 hr.
60.5 hr.
85 deg. C.
85 deg. C.
85 deg. C.
85 deg. C.
0.11 cc.
0.37 cc.
0.41 cc.
2.00 cc.
TABLE XII
ELECTROLYTE OF TAP WATER
Description of Unit
Current
Time
Temp.
Volume of
Hydrogen collected
Iron — New
0.2 amp.
0.2 amp.
0.2 amp.
0.2 amp.
21.0 hr.
30.5 hr.
43.0 hr.
61.5 hr.
85 deg. C.
85 deg. C.
85 deg. C.
85 deg. C.
0.05 cc.
0.85 CC.
1.51 CC.
2.00 CC.
70S August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 8
cathode in the usual manner in one per cent
sulphuric acid at room temperature, and less
than the hydrogen of units electrolyzed, in
one per cent sulphuric acid at 90 deg. C.
The relative rates follow:
Unit electrolyzed in one per cent sulphuric acid
at 90 deg. C 14
Unit immer:;ed in steam 2.5
Unit electrolyzed in one per cent sulphuric acid
at 20 deg. C 1
PENETRATION OF COPPER AND NICKEL
STEEL UNITS
Copper Unit
A unit such as is shown in Fig. I was made
up, having the tube .4 of copper instead of
iron. After running as cathode under the
usual conditions, namely, one per cent sul-
phuric acid, 0.2 amp. and 20 deg. C, there
was no evidence of gas having collected at the
TABLE XIII
Electrolyte
Time Required to Collect 2 cc. of Gas
Under the Same Conditions at So deg. C.
Relative Rate of Penetration
1 per cent Sodium Hydroxide Tap Water
1 per cent Sulphuric Acid
60.5 hr.
61.5 hr.
4.125 hr.
1.02
1
15
TABLE XIV
WATER AT 50 DEG. C.
Description of Unit
Time
Temp
Voiuiif <_.i
Hydrogen Collected
Iron — New
3hr.
50 deg. C.
0.10 cc.
24 hr.
50 deg. C.
0.19 cc.
84 hr.
50 deg. C.
0.22 cc.
112 hr.
50 deg. C.
0.40 cc.
172 hr.
.iO deg. C.
0.51 cc.
180 hr.
.50 deg. C.
0..55 cc.
188 hr.
50 deg. C.
0.60 cc.
TABLE XV
WATER AT 90 DEG. C.
Description of Unit
Time
Temp.
Volume of
Hydrogen Collected
Iron — New
24 hr.
40 hr.
48 hr.
88 hr.
90 deg. C.
90 deg. C.
90 deg. C.
90 deg. C.
0.30 cc.
0.50 cc.
0.69 cc.
1.50 cc.
TABLE XVI
STEAM AT ATMOSPHERIC PRESSURE
Description of Unit
Time
Iron — New
2hr.
:> hr. 10 niin.
4 hr. 20 min.
5 hr. 25 min.
6 hr.
7 hr. 15 min.
7 hr. 50 min.
9 hr. 10 min.
10 hr. 25 min.
Temp.
100 deg. C
100 deg. C
100 deg.
1(H) deg.
100 deg.
100 deg.
100 deg.
100 deg.
100 deg.
Volume of
Hydrogen Collected
0.15
cc.
0.20
cc.
0.29
cc.
0.38
cc.
0.40
cc.
0.50
cc.
tl.,52
cc.
0.57
cc.
0.60
cc.
THE PENETRATION OF IRON BY HYDROGEN
ro9
end of 3S4 hours. At this point mercury had
amalgamated with the copper to such an
extent that the electrolysis had to be dis-
continued.
Nickel Steel Unit
Another unit was made up having the
tube ^ of 3 per cent nickel steel. This was
electrolvzed under the usual conditions.
(Table XVII.)
The penetration rate was roughly the same
as the rate for iron under similar conditions.
Penetration of Iron Units Coated in Different Ways
Four iron units, having the tube .4 of each
unit treated with a different coating, were
immersed in steam at atmospheric pressure.
The coatings were tin (dipped), zinc (galva-
nized), zinc (sherardized, and copper (dipped).
In this experiment the hydrogen from
steam penetrated the tinned iron unit much
more rapidly than it did a unit of iron. The
passage of the hydrogen was evidently
facilitated by the presence of the tin. The
rate for the two zinc-coated units was less
TABLE XVII
Description of Unit
Current
Time
Temp.
Volume of
Hydrogen Collected
Nickel Steel
0.2 amp.
0.2 amp.
0.2 amp.
0.2 amp.
4hr.
23.5 hr.
29.5 hr.
54.5 hr.
20 deg. C.
20 deg. C.
20 deg. C.
20 deg. C.
0.04 cc.
1.00 cc.
1.20 cc.
1.80 cc.
TABLE XVIII
TINNED IRON
Description of Unit
Time
Temp.
Volume of
Hydrogen Collected
Iron—
— Dipped in molten tin
3.0 hr.
8.5 hr.
16.0 hr.
22.5 hr.
100 deg. C.
100 deg. C.
100 deg. C.
100 deg. C.
0.40 cc.
1.12 cc.
1.62 cc.
2.00 cc.
GALVANIZED IRON
Description of Unit
Time
Temp.
Volume of
Hydrogen Collected
Iron-
- Dipped in molten zinc
3.0 hr.
20.0 hr.
31.5 hr.
45.5 hr.
100 deg. C.
100 deg. C.
100 deg. C.
100 deg. C.
0.20 cc.
0.72 cc.
0.90 cc.
1.10 cc.
SHERARDIZED IRON
Description of Unit
Time
Temp.
Volume of
Hydrogen Collected
Iron-
—Heated in zinc powder
1.5 hr.
13.0 hr.
28.0 hr.
45.0 hr.
100 deg. C.
100 deg. C.
100 deg. C.
100 deg. C.
0.15 cc.
0.40 cc.
0.45 cc.
0.60 cc.
COPPERED IRON
Description of Unit
Time
-r. Volume of
temp. Hydrogen Collected
Iron — Dipped in molten copper
40 hr.
44 hr.
47 hr.
103 1 r.
100 deg. C. 0 cc.
100 deg. C. 0.10 cc.
100 deg. C. 0.14 cc.
100 deg. C. 0.72 cc.
710 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. S
than for iron. The rate for the coppered
unit was very slow, and probably no hydrogen
would have penetrated had the coating been
thick and uniform. It is likely that the
hydrogen went through the iron and not the
copper, in small areas where the latter had
corroded away.
The comparative rates for iron, and for
iron with the various coatings, in steam at
atmospheric pressure follow:
Description
of Unit
Time Required to
Collect 0.60 cc.
of Gas
Relative Rate
of Penetration
Tinned Iron
Iron
Galvanized Iron
Sherardized Iron
Coppered Iron
4.0 hr.
10.5 hr.
15.5 hr.
45.0 hr.
86.0 hr.
21.0
8.2
5.5
1.9
1
Experiment with Barium Chloride and Potassium
Dichromate
Evidence that the hydrogen which collected
in the units was produced outside the tube
and forced through the metal under pressure,
and was not produced by the acid leaking into
the tube, and later reacting with the metal,
was furnished by two experiments.
(1) An iron unit was immersed, without
electrical connections, in one per cent sul-
phuric acid for 22 hours, during which time
2.4 cc. of hydrogen collected. The unit was
then emptied of mercury and 10 cc. of dis-
tilled water poured in and shaken so as to
come well in contact with the inside surface
of the iron tube. This water solution gave a
negative test for sul])hates with BaCl;,
indicating that no sulphuric acid had come
in contact with the inside of the tube.
(2) Another unit was immersed, without
electrical connections, in a solution of one per
cent sulphuric acid to which had been added
one per cent of K2Cr207. No gas collected
in this unit after immersion for 9() hours.
The fact that the oxidizing action of the one
per cent KoCr./)? in one per cent sulphuric
acid solution prevents the passage of hydrogen
through the iron, at least for 9(1 hours; and
that the inside of an iron unit jjickled for 22
hours in one per cent sulphuric acid showed a
negative test for sulphates, furnish additional
evidence that the hydrogen which collects
inside the units is formed on the outside and
forced through the iron walls, and is not
formed on the inside by reaction between the
iron, and some suljihuric acid, which has
by some seemingly imjiossible means leaked
through to the inside of the tube.
Composition of the Gas
A sample of gas, which was collected in an
iron unit electrolyzed as cathode in the
usual manner, showed the following analysis;
Oxygen Xone
Carbon Monoxide. None
Carbon Dioxide ... None
Hydrocarbons None
Hydrogen Sulphide None
Hydrogen | ^j-? Per cent
^ \ 94. o per cent
The remaining 5 per cent was an incom-
bustible gas such as No.
Another sample of gas, which was collected
in an iron unit immersed in steam at atmos-
pheric pressure, consisted largely of hydrogen.
Conclusion
It has been shown that hydrogen penetrates
iron at temperatures between 20 deg. C. and
100 deg. C. under a great variety of con-
ditions, all of which influence the rate. The
velocity of hydrogen penetration is greater
for a unit immersed, without electrical con-
nections, in one per cent sulphuric acid than
for units electrolyzed as cathode in a like
solution, with such current densities as were
tried. The rate for electrolyzed units is
influenced by the current; the higher the cur-
rent, for such densities as were used, the
higher the rate, but the relation is not a
straight line function. The penetration
velocity increases with each successive elec-
trolysis, provided rest periods do not inter-
vene, or w-ith acid "pickling." The effect of
rest or moderate heating upon units which
have been electrolyzed or jMckled is to restore
the original resistance of the iron to the pas-
sage of hydrogen. Temperature has a marked
effect, the rate of penetration increasing with
the temperature. The rate at 90 deg. C. for
an iron unit made cathode in one per cent
sulphuric acid with a current of 0.2 amp. is
14 times its rate under similar conditions at
20 deg. C. The velocities of penetration for
units electrolyzed in one per cent solutions
of potassium suljihate and sodium h\droxide,
and in ta]) water, are about equal and are
1, 12-1 15 the velocity of units electrolyzed
in one per cent sulphuric acid.
Hydrogen produced by the reaction between
tap water at temperatures from 50 deg.
C.-IOO deg. C. and iron, or between steam and
iron, penetrates the metal at a rate depending
dirccth- upon the tcmjieratiiro. Tlie oxide
which is one of the products of the reaction
forms a coating on tlie metal, whicli becomes
thicker and thicker, and tinall\' protects the
THE WHITE MAZDA LAMP
iron from further action. If it were not for
this coating of iron oxide, iron pipes carrying
hot water or steam would continually give off
hydrogen and quickly deteriorate.
The velocity of penetration for 3 per cent
nickel steel is the same as that for iron.
Hydrogen does not penetrate copper at a
temperature of 20 deg. C. The rate for
tinned iron is greater than for iron; and foi
galvanized, sherardized and coppered iron is
less. It seems that the passage of hydrogen
is facilitated by the presence of tin and
retarded by zinc and copper. No evidence of
sulphates could be found inside a unit which
had been pickled in sulphuric acid and in
which 2.4 cc. of hydrogen had collected. No
hydrogen penetrated a unit immersed in a
solution of one per cent sulphuric acid plus one
per cent potassium dichromate in 96 hours.
The gas which was collected in one of the units
was analyzed and found to contain 95 per
cent hydrogen and 5 per cent of an incom-
bustible gas, possibly nitrogen.
The facts will admit of the following
explanation of the manner in which hydrogen
is forced through iron tubes having walls
1^6 in. in thickness. Atomic hydrogen (H)
which has been liberated by the current, in
the case of units which were electrolyzed, or
by the reaction between metal and solution
in the case of units which were not electro-
lyzed, penetrates the iron, where gaseous or
molecular hydrogen (H-.) is later formed.
Iron at room temperatures is impermeable to
the latter. The atomic hydrogen continues
to penetrate the surface of the metal rapidlv
and to form molecular hydrogen. The latter
can escape only very slowly and as a pressure
sufficient to force the gas through the metal
is built up. It is a pressure built up in this
way which also results in the well known
phenomenon of the cracking of hardened steel
when "pickled" in acid.
The writer hopes that the experiments
which have been described in this paper will
help to focus the thought of electro-chemists
on these problems, and that as a result a
more comprehensive understanding of the
mechanism of the passage of hydrogen
through iron at temperatures equal to or
below the boiling-point of water may be
had.
The White Mazda Lamp
By Ea'rl a. Anderson
Engineering Dep.^rtment, National Lamp Works, General Electric Company
Of late years, absence of glare is becoming recognized as one of the essentials of good lighting. A number
of schemes have been developed to eliminate glare by manipulating the light after it has left the lamp; for
example, variations of semi-indirect and indirect lighting. The latest method in connection with moderate
wattage incandescent lamps is the reduction of glare at its source by the use of the lamp described in the
following article. — Editor.
Within the past few years many illuminat-
ing engineers have been turning their atten-
tion more and more strongly toward the
elimination of that bugbear of manj^ lighting
installations — glare. Although much has
been done to minimize this evil in industrial
installations, until recently little considera-
tion has been accorded it in residential light-
ing. The newly developed white Mazda
lamp, on account of the softness of its light,
has proved especially effective in reducing
the glare so often found in lighting units in
the home.
Distinctive Features
The outstanding characteristic of this lamp
is the diffusion of its light. As shown in Fig.
1, the bulb is made of a special white glass,
designed expressly for the purpose of minimiz-
ing glare. The large volimie of light which the
filament emits is dift'used to the point where
the bulb itself appears liuninous. The bright-
ness of the bulb is about 13 candle-power per
square inch over the brightest square inch of
area which is, of course, far below that of the
filament of a Mazda lamp.
It has been pointed out by many illiuninat-
ing engineers that glare (which, however
defined, is ultimately light that hurts the
eye) is to a considerable extent a matter of
brightness contrast. The illustration of auto-
mobile headlights, which are glaring at night
but which are scarcely noticeable during the
day, mav be recalled. Because of the soft-
ness of its light, the white Mazda lam_p can
be used satisfactorilv in locations where other
712 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. ,S
incandescent lamps unless frosted would be
objectionably bright. Frosting the bulb has
always proved an effective means of reducing
the brightness of the incandescent lamp, brt
the practice has not been widely followed
largelv because the frosted bulb collects dust
Fig. I. White Maz.la 50-Watt Lamp
and dirt more quickly than a clear bulb and
is more difficult to clean. The bulb of the
white Mazda lamp is smooth and is as readily
cleaned as a clear-glass bulb.
The white Mazda lamp is made in the
50-watt size, and, notwithstanding the
low brightness of the bulb, has an output
of approxirrately 450 lumens. As shown
in Table I, the efficiency of the lamp is
about 9 lumens per watt. Its m-aximvim
dimensions are about the same as those of the
40 and oO-watt Mazda B lamps; it is 2]^
inches in diameter at the largest point, as
compared with 2^ inches for the 40 and 50-
watt Mazda B lamps. The lamp is designed
for use on standard lighting circuits between
110 and 125 volts, and, in common with all
incandescent lamps, operates most effectively
at the voltage specified on the lamiJ by the
rr.anufacturer.
An especial feature of the new lamp is its
tipless construction. This is produced by an
ingenious method of manufacture in which the
lamp is exhausted through a tube attached to
the stem at the base of the lamp. The
absence of the tip results in an appreciable
reduction in lam.p breakage, and presents a
smooth surface.
The 50-watt white Mazda lamp can be
burned in any position.
US'
-/7
/>C '"^
Y^
/.
iP
^«;
\ v-"*'
rh
9S*
y^^
w
-V-1— ''
' ',
■ — \ — Ill:r~w
r^w
85-^
p^^
>.5^<^
— jl <J5-
wv
3cO?Ov7
i'°^
r-l-J
y\\X/>
III ""
\ \
><%S/j
-ao-VV
yiCy
r\
*^^^
-iiUV-^
yLJ /**■
V
^
3
4
■> ■ — T
/~~-^Z^
'^r--^
\J
5-^ Zi-
^"/5' 5-
5*
*5'
es-
35-
• li-
^
7-'
j*sy\'\J±
o*
/43-
A \
IIS-
3i'-l
-^L
U
^
SJ
A
— — r~r*7 I
1 ■ — '
J. 90*
\ lu
\
%
^
/ ^**
»5*\ ^
\
^
'x^^SL/ "'
^
1
>^
0-
J
0
5-
0-
/»•''
g5'
\
~~~~;5- — S'
a;
White Mazda
Per Celt T.ii.il
Clear L.imp
Zt'tlc
Mazda B
Cle«r Lauip
48
101)
Fig 2. Candle-power Distribution Curves of Enrc White Mazda Lamp a.id Bare Mazda B Lamp
0 lid
M.",
20
0- ill)
2a,-
.TO
91) -1811
2:t5
.■50
l)-l.SO
47i
100
THE WHITE MAZDA LAMP
TABLE I
DATA ON THE WHITE MAZDA 50- WATT LAMP
Wattage
Voltage
Approximate
Total Lumens
Approximate
Lumens Per
Watt
Bulb
Type of Bulb Diameter
Inches
Maximum
Over-all
L.ngth. Inches
Base
50
110
to
125
450
9
Pear 2^
Shape
SVs
Med.
Screw
Light Distribution
The curves given in Fig. 2 show the dis-
tribution of Ught of the bare white Mazda
lamp compared with that of the bare Mazda B
lamp. As would be expected, due to the
diffusive quality of the bulb of the white
Mazda B and with white Mazda lamps. This
is due to the fact that the bare white lamp
gives a slightly larger percentage of downward
light which compensates for the light lost by
cross-reflection between the reflector and the
white glass bulb of the lamp.
IIS''/ y^
'«• r
\^
5
"~~>C/45'
V'\ \"5-
?' ~
t
t
/^
^
w
^
^
-TW-
-^
, — tI
1
J —
J5-V^
r
Ci-T/^
<y^'oCN
\^cr7^
rj~J~e5-
\ vJu-V'
V \-
■\j^O
al^'°vS
v^/~y
"" \\ >
^
W^
/T^°\^
^5^
/ T~T-7s-
esAT \V
Si
Y>-
^"Vy
s/ / /65-
5S^\N/
§
^
^50- —
■ 90 ■
^
\//s5-
«■ \l/^
^Z5-
'"^/s*
'
—
15'^'
Z5- '
~xJ
S'v,
;^_
5*
5*
Hi'/ / /s
H^'J\
^^ /«o' 7
~7^'-«5"
VVv^.
9i-f__/Lj
/^
^
^
1 95-
"
\
n«»
as-H
r
r
i^O^/o v>
-j-j
— <ss-
\ \\Jr
yy/-/<>r^
/~J/ / /
6S-\ \X
S
sL / ' ' ' \
<V
55- \ y
N^
■^\//s5'
^
f Is
T
/
Kyy.
«• \ V,
^^^
^^/5-^
. — 1
/5
25-
^X^ /AS'
.3
V
£1^
^2
3-
, 5'
5'
<
35
0\
White Mazda
Mazda B
Zone
• Lumens
Per Cent Total
Clear Lamp
0- 60
0- 90
90-180
0-180
207
277
146
423
42
56
30
86
Zone
Lumens
Per Cent Total
Clear Lamp
0- 60
0- 90
90-180
0-180
197
264
143
407
41
So
30
85
Fig. 3. Distribution Curves of White Mazda Lamp and Mazia B Lamp, Both with Bowl-shaped Opal Glass Reflector
lamp, the distribution curve shows a greater
( andle-power end-on than in the case of the
clear Mazda B lamp.
In Table II are giren the results of tests
m.ade to determine the effect of the diffusing
bulb upon the output of typical lighting
units. Figs. 3 and 4 are exam.ples of dis-
tribution curves for two com.m.on reflectors.
It will be noted that there is but little dif-
ference in the absorption of any of the units
tested, when equipped respectively with
TABLE II
COMPARATIVE DATA ON LIGHT OUTPUT
Type of Unit Tested
OUTPUT IN PER CENT
OF BARE LAMP OUTPUT
Mazda B
White
Mazda
Glass Bowl, 6-inch Diameter
Glass Bowl, 7-inch Diameter
Enclosing Unit
Enameled-steel Bowl
85.4
84.8
77.5
60.5
87.6
86.1
76.4
61.3
14 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. S
Field of Application
The white Mazda lamp can be used to
advantage in place of the similar sizes of
Mazda B lamps in the same reflector equip-
ments. The effect produced by using white
lamps in semi-indirect fixtures is particularly
pleasing, for distinct shadows of the bowl
edge, of the bowl suspension, and of the leads,
and all striations on the ceiling are eliminated
because of the larger area from which the light
comes. For the same reason, the white bulbs
are also particularly desirable for portable
lamps, where their use will eliminate the
formation of grotesque, and frequently annoy-
ing, shadows upon the walls or upon the pages
of a book. For example, fringe shadows,
which are often very disagreeable, are
eliminated.
With regard to the ser\-ice which m.ay be
expected from white Mazda lamps, it may be
said that experience shows a satisfacton.-
degree of ruggedness for the lighting of
homes, ofF.ces, hotels and public buildings.
Increased usage of these lamps is found in the
m.any pleasing effects obtained by their
installation at motion picture and hotel
entrances. In m.any decoratiA-e types of
fixtures the use of white Mazda lamps has
proved very- effective in enhancing the artistic
appearance. The low brightness of these
lamps has resulted in their being used in local
lighting units for the inspection of machined
interiors and similar places difficult to light
with any general lighting system.
The tendency to use the white Mazda
lamp without reflecting equipment because
of the softness of its light should be dis-
couraged. For most locations the bulb is still
too bright to be used alone, and in addition
to reducing glare, it is just as important, from
the standpoint of effective distribution of the
light generated, that a good reflector be used
with this as with any other incandescent
lamp.
M
M
xSX^
$
— r — f f-
— —
90'
"*V\
— tPr
^^5^7'
)-c^ter-
~-^5-
"'\\
v^
K/Yj 1 111 \ v(\/
nf
^* \ V'
Kj^
>-j 1 1 1 \^A^
v/ 7 — fo — r v
</"'
55-\ \
^^
y-T — Iro — X\
\/"'
" 25-^ 'i
"jT (OO 'T'
s- es-
015 •
•__ 5- \ 5- (.
C '' '' "'
■ X i*i-^\ \ leo- i
y^i*i-
/ '\
X \
//5*
^^<^WL^
^
S^
A
95'L f _/
QJI. t
T — f— 4 -W~WL
m
^
%
\
V:
es' — \
\ \.~\'\z5yiy'°^\^
^tJ
' ~
— ■
"■*5-
'-T — \
$v
ih
/-
\V\
\rv /X XNL / 11' \ J
^vO
/Vv
7
**'\ \ ^
\5<
/X^
b'
**' \X
■x v/^jC / — j^°rv
k\^
K//
ys,i
V \/\
/xJt^^T/^ "°t^
^^
■^/\
//
h=
3^
V^
yy
J5-
ej* /5- — s'
S'
IS-
25-
a-
__
White Mazda
Zone Lumens
Per Cent Tolal
Clear Lamp
0- 60 264
0- 90 .-iOl
61
Mazda B
Zone
0- 60
O- 90
232
2Sl
Per Cent Total
Clear Lamp
52
Fig. 4. Distribution Curves of White Mazda Lamp and Maz Ja B Lamp, Both with Bowl-shaped Enameled Steel Reflector
■15
The Reward for Efficiency*
By HcN. Edwin O. Edgerton
President, Railroad Commission of the State of California
Mr. Edgerton's belief is that the best of which man is capable is never produced by punishment. Man
exerts himself to the utmost only when he sees before him a reward proportionate to his efforts. Maximum
efficiency from man and machine is to the best interests of the public, and knowledge by the former that
merit will be quickly recognized will serve best to bring about this condition. What is true of the individual
worker with respect to reward is also true of the investor, the manufacturer, the public utility, and the public
service comrrission. — Editor.
Introduction
I want to define in my own way the subject
which has been given me. I want to broaden
it from a mere discussion of definite specific
reward and treat it from the standpoint of
inducement to do the best that is in each mian
who is connected with this great industry,
and also, finally, I want to suggest induce-
ment to the public utility commissioners to
do their best, and on that score I want to
say something along the line of what you
gentlemen owe these Commissions.
Let us take money. By the way, for some
reason that I have not been able to under-
stand— and I have made considerable inquiry
on the subject — finance is e.xchided from pub-
lic discussion, as an ordinary thing. I looked
over your program — and I speak in no sense
of criticism — I looked over your program
seeking the place where finance would be dis-
cussed, and I found nothing. Today, at
least in California — and I think it is true over
the nation — the greatest single job that we
have is to finance these pulDlic utility con-
cerns. And why that great subject should
not be pulled frankly out into the open so
that the most of intellect and ability can be
brought to bear, I don't understand.
If I dwell too much on California condi-
tions, it is not in any provincial spirit, it is
not in any way feeling that we out here
dominate the situation in any degree, but it
is first because I know California conditions
better, and, next, because of the situation we
are in; the problems perhaps are accentuated
out here as compared w^ith those of the East.
Now, the job immediately ahead of financ-
ing these ptiblic utilities is a huge one. In
California every time I have made an estimate
of the amount of money required in the next
few vears, I have found that somebody raises
that'estimate. I started in with $2.50,000,000
in the next ten vears, and I have been raised
now to .i;500,()()0",000; and I don't know where
it is going to end.
* An address before the Annual Convention of the N.E.L.A.,
Pasadena, Cal.. May, 1920.
The other day in Chicago a very famous
banker was on the witness stand before the
Public Utility Commission of Illinois, and he
made this statement, published largely over
the nation, that anybody who invested in
public utility stock needed a guardian. I
don't admit that the electric public utilities
of California are financially sound, basically
sound. I aflfirmatively assert, I insist that
it is true, and I say that that statement as a
generalization is unsound, and as applied ti
California is particularly tmsound; that sort
of statement has the effect of offsetting in
some degree at least the efforts that are being
made to build up the credits of these com-
panies in order that the investors may have
confidence and assurance. And a gentleman
of prominence who goes publicly on the wit-
ness stand and makes a statement of that
kind ought to be careful that by a glittering
generality he does not do serious dam_age,
probably not intended by him.
I am one of those who believe that you can-
not produce the best that is in men and women
by punishment. You cannot whip a man
into efficiency. In that belief I am convinced
you must proceed by inducement ; that every
one of us in some degree, in order that he may
do the best that is in him., requires that before
his eyes there be some reward.
Let us apply that to the fellow who has
money to invest. I recognize that we cannot
get money out here from the East into Cali-
fornia by threat, by argtiment, by any form
of punishment, by any suggestion that be-
cause of investment already made they can-
not quit; I recognize that the money must
come by inducement; and in my judgment it
can be itiduced without paying exorbitant and
unreasonable prices for it. We must bid for it,
yes; but must we bid for it against speculative
securities, against anything anybody is willing
to offer, bid for it against the man who offers
a chance for large reward? No; I say not. I
believe that we have assets in California
which, if properly used, will produce the
fundam^entals of inducement to investors, this
"16 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 8
being absolute security of the investment it-
self, the assurance that the dollar will not be-
come 90 cents or something less, coupled with
certainty and regularity of return. And if
we have these assets, why not make use of
them? Then if money is available on any
terms, we w ill get that money.
Service an Asset
Now, what are these assets ? Power houses
and transmission lines? Surely. But over
and above that we have the assets of a great
vital and essential servace — -a ser\-ice which
the people must have, a ser\nce which they
cannot get along without. And that is an
assurance to an investor that that asset can-
not disappear. My judgment is that the
public utility financiers and the Railroad
Commission have these assets in trust for the
people — -I think they can make bitter com-
plaint if we do not use them coupled with
others, so as to produce the necessary- money
to do the absolutely essential development
that must go forward in this state.
If it is necessar\- now to persuade invest-
ment, to insure regularity and certainty of
return, why not face that fact? Why con-
tinue to indulge in discussions of technical
methods of valuation never settled' Eight
years' experience now in the Railroad Com-
mission, with constant discussion of the
methods of valuation and proper rate bases
and reading of the decisions of courts, puts
me in a position to say to you that those
questions are no more certainly settled today
than they were eight years ago. Now, why
not face that as a fact, and why not seek some
other method of determining the rate of re-
turn the company should get? To face the
situation clearly and conscientiously, what
methods shall we now pursue?
My judgment is that the thing to do is to
start with sound capitalization — an admitted,
agreed, established capitalization, and there-
after fix rates based on getting the necessar\-
bond interest, dividends, and fixed charges to
support that capitalization. Then in order
to meet the contention of the investor that
perhaps later a different policy may be
adopted, set up a cash resers-e out of rates, to
be rigidly held for the purpose of insuring bond
interest and dividends, so that when you take
your securities to the money markets of the
country' the investor can be shown an actual
cash reserve as an insurance policy for his
bond interest and dividends. I realize that
this suggestion means almost a complete
reversal of the attitude of regulating bodies
towards this question. But why not? If it is
the sound thing to do, why not reverse the
attitude'
Degrees of Efficiency
By the way, I make these suggestions in no
spirit of finality. If there is one thing I have
learned, it is that final opinions are never
final. But I do make them for the purpose
of starting discussion on this important sub-
ject. If those suggestions are not sound,
weaknesses may be pointed out, and I for one
will welcome such suggestions. I welcome
criticism of regulation in California, only pro-
viding that the fellow who criticises accords to
m,e the same thing that I accord to him, and
that is sincerity of purpose.
Now, there is something else — rewards for
efficiency. Are all the electric public utilities
at the highest point of efficiency? Well, they
are not all here, so we can say they are not.
And I think we can conclude that there are
different degrees of efficiency among the
companies. I believe it is sound to suggest
that inducements definitely be held up to the
companies to become thoroughly efficient.
I think it would pay the public to hold up
definite rewards to that end . And incidentally
that suggestion has been made quite fre-
quently to our Commission and others.
There is a feeling and has been for years that
regulation has a tendency to hold the return
down to a dead level, that the inefficient
company enjoys the same rate of return as
the efficient company, no more, no less, and
initiative is destroyed. Why work, toil,
think, to produce efficiency if the regulating
body immediately appropriates the results to
reduce rates. A ven,- fair suggestion. But this
is to be thought of: Whom must you reward
to get efficiency ?
Now. if you have taken care of the investor,
if you have produced a situation where his
investment is safe and intact, his return is
regular and sure, must you stimulate him to
make him efficient? Well. I would suggest
that that is not quite what we arc thinking of.
Then who is it that we should stimulate by
offer of reward' In my judgment, it is the
organization of the company itself. Now
don't misunderstand me. When I say "or-
ganization" I mean from the jiresident to the
office boy. I don't believe it jiroduces ef-
ficiency merely to hand the management, as
such, the reward; in fact. I am inclined to
think that that would produce the ojiposite
effect, because all down through the organiza-
tion would go the feeling that the reward
THE REWARD FOR EFFICIENCY
717
earned by each individual's efforts v. as going
to somebody else, and there is not anything
in the American mind that produces more re-
sentment than that situation.
Employees Interested
Rewards for efficiency. Yes, the regulating
body, in my judgment — and I think the time
has come to consider that seriously — -ought to
give definite assurance that for increased effi-
ciency and economy, always coupled with good
service, a reward should be accorded by the
regulating body. But having said that, the
job thereafter, in my judgment, is distinctly
one of management to make it effective.
Management to make it effective must see to
it that every member of the organization, no
matter how humble, is made to understand
thoroughly that his increased efforts towards
efficiency and economy will be rewarded.
And I speak of reward not only in a money
sense, but I speak of recognition of service
well performed by men lower in the ranks; I
speak of certainty of advancement when
opportunity comes; I speak of the absolute
elimination of ptill or influence in promotion
in the organization; I speak of a situation
where the management studies personnel,
works at the problem, is constantly on the
job — first to know what its personnel really
is, to know when efficiency and economy are
being striven for by the individual in the
ranks, and then with absolutely dead cer-
tainty to reward that effort on the part of
the members of the organization.
Aids Labor Situation
I have had clerks telephone me from the
inside of public utility organizations, fearful
to give their names to rtie, complaining that
the entire class represented by the speaker
over the telephone had been overlooked in
the wage increase, and that the wage increase
had been given in large part to the organized
employees. And in that clerk's heart was the
sense that because of that organization, be-
cause of the threat held up to the manage-
ment, additional wages had been accorded
and merit had been forgotten.
I think that there is no better way to solve
the question often called the labor question —
there is no better way to take the body of
labor, working people, clerks, away from the
demagogue leader of the union than to accord
to every individual in an organization re-
wards for efficiency. It is the best safeguard
against the chap who comes about preaching
anarchy, preaching this proposition, and it is
always his fundamental proposition and he
is clever enough to understand it: "You are
not getting a square deal." What better
safeguard could there be against the agitator
than to have in the heart of the fellow he
comes to, the knowledge that he is getting a
square deal ? That is better, in my judgment,
than all the panaceas that have been sug-
gested; but it requires on the part of the
management, work, thought, study, effort.
I remember talking to one of the directors
of a company that is generally considered —
and I think properly so — one of the most
efficient in the United States in the matter of
organization. It is generally understood that
the employes of that company seldom quit. I
began to be interested to know why, and so I
talked to one of the directors. Here is the
system: He said, "It is the job of the highest
officers in our company, it is their principal
job to study, to watch, to investigate, to
come in contact with all of the employees of
the company. A rigid rule is adopted that
promotions are made upon merit, that in no
instance is an outside man called into the
organization unless there is in the organiza-
tion no one with the special knowledge re-
quired at the moment; that men starting with
the company have an assurance of a career
for life, and that they have the assurance
that they will be constantly promoted as
opportunity offers." And I said to him:
"Well, if this is the job of the principal officers
of the company, how do you find time for the
ordinary business of the concern?" "Well,"
he said, "if our organization is efficient, the
business takes care of itself. In other words,
we sharpen the tool, and we have the confi-
dence thereafter that it will cut."
Commissions Interested
The Railroad Commission — it, too, should
be efficient. We are the same kind of animals
as you fellows — no different. True, the job
is a little dift'erent; but in a real sense we are
wrapped up in your success, will suffer by
your failure. Remember, if the private
ownership and operation of public utilities
goes down in failure, regulation goes down
with it. No matter whether the Commis-
sioners can point the finger of blame to the
utility men, if the wreck occurs, regulation
has failed.
So you fellows owe a duty to the public
utility commissions. You should hold up the
reward for efficiency on their part. And the
only reward that you have available is af-
firmatively to make regulation as little necep-
"18 August, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. S
sary as possible; beat the regulating body to
it, and you will do the best thing you can do
for the regulating body. By the success of
your own efforts in satisfying the public will
you make the job of the utility commission a
job worth haA^ing.
It is a cheap and easy thing, and sometimes
results in temporary- gior\', for commissioners
to denounce public utilities; but it is a very
costlv thing for the companies and for the
commissioners, because finally a denunciation
mav go to the point of wreck, and regulation
and commissioners go down with the wreck.
And finally, I will say this from the stand-
point of California: There is a tremendous
job ahead — a job for the electric utilities, a
job which involves not only carr^'ing forward
the business which naturally accumulates,
which normally increases, but the unques-
tioned job of taking over the work now being
done out here by oil. There is not another
substitute in sight for oil except hydro-
electric energy; and it is a sobering thought
to think of the tremendous responsibility
this will place upon the electric utility men
of this state and the Commission working
with them. When you stop to think of a
state like California, with agriculture, in-
dustry, and the people absolutely'dependent
upon the ability of electrical men to produce
service, the responsibility that you not only
ought to take upon your shoulders, but that
you will have to take is staggering.
Industry Is a Unit
Now, realizing this fact, I make this sug-
gestion, that it is the wise thing to do to
approach this whole problem with an open
mind, without any fear whatever of disclosing
secrets, withholding information, striving for
advantage, this company against the other,
to solve the problem in a way that will be
sound, considering the whole power situation.
I believe the proper position for both the
companies and the Commission to take is
that it is one great problem, and that it is
not sound operation on the part of the com-
panies to allow one company to be seriously
injured, because that injures the industry- as
a whole.
If to produce the ser\'ice under the condi-
tions that must be accorded the people it is
necessary- to consider complete unification of
all plants, why not discuss it ? I say the man
who is afraid openly and frankly to discuss
that subject is fearful it may succeed. Why
doesn't the fellow who believes that there is an
inherent fundamental weakness in unification
gladly come forward and discuss it with any-
body, in the conviction that he can defeat the
suggestion ?
I only suggest it, because I know it touches
the heart of the fellow who has built up a
company, who has been with it in fair weather
and foui, and has a pride in that company in
giving sen-ice, and has a pride in its entity.
I know that, being a human being, he does
not like to merge that company with another.
I know that he feels he is giving good ser\-ice
and he is reluctant to consider the unification
with other companies that he thinks are below
him in the standard of ser\-ice.
But why not discuss it? If it is the answer,
then I say the personal feelings of each of us
must go by the board. The great thing to
accomplish is the doing of the job, and finally
the public will hold us responsible for having
the job done. It does not understand all the
angles of the problem. The public of Cali-
fornia today, if you suddenly said, "Let us
put all the companies together under one
great organization," might rebel, because it
would not understand what that meant; but
that does not exclude us from going fonvard
with the discussion to see whether it is sound.
My conception finally is that the electric
public utility men, with the utility commis-
sion, are in responsible charge of doing the
job. We must do it efficiently, and then we
are entitled to our rewards. If we don't do
it, we ought gracefully to take what undoubt-
edly will result.
719
IN MEMORIAM
George A. WooUey, Manager of the Denver
District of the General Electric Company,
died at his siunmer home in the mountains
near Evergreen, Colorado, July 3, 1920. The
immediate cause of his death was apo]3lexy
superinduced by a cerebral hemorrhage.
GEORGE A. WOOLLEY
When a young man Mr. Woolley decided
to make his home in the West, and in 1S9()
allied himself with the Edison Electric Com-
pany at Denver. He was soon made Manager
of the Supply Department and held this
position until the formation of the General
Electric Company in 1893, when he was made
District Manager of the Supply Department.
In December, 1913, he was appointed Dis-
trict Manager and retained this position until
his death.
On several occasions in early life oppor-
tunity to transfer his activities to other lines
of business promising cjuicker and perhaps
greater financial reward presented itself, but
in each instance was rejected. He had
unfailing faith in the ultimate success of the
electrical industry and no amount of per-
suasion could make him forsake his chosen
vocation.
Mr. Woolley's later years were saddened
by the death of his younger son, Frederic
H. Woolley, in October, 1918. Frederic was
residing in Schenectady at the time of his
death and was employed in the Testing
Department of the Company. Mr. Woolley
was deeply impressed by the great kindness
shown to his son by the officials of the Com-
pany and their families during his sickness.
Their sincere desire to comfort him as he was
leaving Schenectady on the saddest of all
his journeys was never forgotten. Many
times since he has stated that a corporation
is best judged by the personnel of its execu-
tives; that a company headed by officers
considerate and kind enough to ignore the
pressure of business to go to one in trouble,
to extend their sympathy and share another's
sorrow, cannot be classed as soulless.
The Company has lost an able, loyal and
highly esteemed leader, and we in the Denver
Office a revered friend and conscientious
adviser.
Mr. Woolley was married to Semira Hartzell
in Kearney, Nebraska, February 15, 1893.
He is survived by his wife and his son, George
Allan Woolley, who is at present in the
employ of the Great Western Sugar Company.
Mr. Woolley was a member of the Quarter
Century Club, Mohawk Club, Denver Motor
Club, Denver Athletic Club and the Rotary
Club. B. C. J. Wheatlake.
GENERAL ELECTRIC REVIEW
AUGUST. 1920
Where to Get G-E Service —
Quick service is best obtained from the nearest G-E
sales office, distributing jobber, or foreign representative
For Business in the United States
CJrwtUM. v..
G-E S*les Office CE DUlributing Jobber
Alabama. BlrmlnKham Matihe** En^ '^iiiifh < i-
Alabama. MobUeJ ... .MauhewB tilec Supply to.
Arltansaa. Ullle R<Kk
Caiirornla. Los Anneleat PartBc States Fl«irlp Co.
California. Oakland: Paclflc Slates Klerlrir Co.
California. San Franclarolf Pacirtc Siau-s Klwirlr Co.
CoI<K-ado. DcQVcrt Tbe Hemjrie A Boltboa MK A
Sup. Co
ConnecUcut, Hartford
Connecticut. New Haven
ConnecUrut. Waterbtiry: Ne« Enslaod Enf Co.
District of Columbia. Wastilnc-
lon XaOonal Elec'l Supply Co
Florida, Jacksonville Florida Elec J'uppir Co
Florida. Tampa: Florida Klec supply Co
Georgia. AtlanUlt Carter Elcctrlr < ompany
Georgia. Savannah! Carter Flecirlr ( ompany
Illinois. ChlcagoSt Central F.leririr (ompany
Commonwealth Kdlaon Co
Indiana. Fort Wayne
Indiana. Indianapolis IndlanapolL* Kirc Supply Co.
Indiana. rV>utl) Beodt .South liend KIrrirlr Co
Iowa. r*e3 Moln« Ml<I-Wcst ElcrtflrCo,
Kentucky. Louln'Ulc Belknap Hardware A Manu^ar
lurlne Co . Inc
Louisiana, New Orleans Gulf states F.lrrUlc Co.. Ine
Mar>land. BalUmore. Southern Flectrtc Co.
Massachusetta, Boatoot PetUncell-ADdrtwt Co.
G-E Sales Office
\>w Vork, Ai^anv;
Vpw York. RuriDo
New York, Elmlra.
New York City St ..
G-E Dbtribtitiiic Jobber
Havens FlecU'leCo., Ine
HoDcrtso[M.~ataraeC Elee Co.
E D. Latham A Conaaay
Ro>-al E;aatera Elfc'l Sup Co.
^bkr-Pttmaa Elee. Corp.
' n'bcckr-Grceo Elecl Sup. Co.
. Mohawk EIee"l Sup. Co,
Eire. Supply A EqutpmrDi Co.
The F D Lawreft*^ Eke Co.
Republic Electric Co
W C NacH EltelrtcCo.
Vew York. Ntacara Falb . . .
New York. Rochester
New York. iCcheoecUdr .. .
New York. Syracuse
North Carolina. Charlotte
Ohio. Cincinnati* -
Ohio. Cleveland
Ohio. Columbus
Ohio. Dayton .
Ohio. Toledo
Ohio. Younestown
Oklaboma. OkUft'ina Cltyt
'>recoo, Portlan-lt
PennsylranU. F>ie
Pennsylvania, Phlladelphla|t . Phlladetphia ElectrV- Company
.'tiipply nepartmcnt
Pennsylvania. Pittsburcht t'nioa Electric Company
Rhode Island , Provtdeore
South Carolina. I'olumbla:. . Ptrry-Mana Elee. Co.. Ine.
Tennessee. Chattanooo Iimw Supply Co"
Tenneaaee. Knotvllle
Mkhlcan. Detroit
Michigan. Grand Rapids .
M inneaota . Duluth
MtanesoU, Mlnneapollst.
M Idnesola . St . Paul;
Missouri, Jopllat
Mtasouri. Kansas Citrt...
.Frank C. Tnl CoiDpsny
^Norihwntero Elecuic F.culp
mcnt Company
Preflcaa Elrctriral Co
.Northwestern Eire Equip. Co.
!tbe B-R Elecirlr Co
Traarmee. Memphis
Tenneaaee. Nashville
Texas, Oaltaat
Teias. El Pasnt
Teiaa. Houston*
Itah. Salt Ijike Cltyt
Vtrelnla. Rlchmor>d
Washington. Seaitlet
Washlncton. Spokan<-
Wasblncton. Tacoma .
Wr« VtnctnU. Charl*«ion
Wlsronaln, Mllwaakee
For Hawaiian bualnen addras Ckttoa. Ndll *
t\Vareh"u«e (s*^*!*-* Shop.
Etectrtc Supply Company
Southwest t^E Co.
Soulhw«t V^E Co
Southwrst CE Co
CaplUI F-lrctrtc Compnny
Southern Electrte Compway
ParlV SUtM Elcctrto Co.
Montana. BullefTT Butte Hwirtc Supply Co,
Nebraska. OmalUl Mid-West Elecirlr Co
^'ew Jersey. NewarkI Trl-CH) Electrir ( o . Inc.
tNoG-E Offlre
Diatribulora for the General Electric Company Oulaide of lb* United Slate*
INTERNATIONAL GENERAL ELECTRIC COMPANY. INC.
120Bro«dwa* New York. N Y. Schenectady. N. Y 83 Cannon Street.
Foreign Offices and Rcprescniati
ArKciilliia: General KIrctrlc. S A , I'
Australia; Australian Grnrral Elcciri
Melbourne
Beiclum and Colonies: Soclete dElecirlclir el de Mecaolque
Proredea Tbomsoa-Houslon A Carels Soclete Anooyme.
Brussels.
Bolivia: Intrrnailonal Machinery Co.. I.a Pat and Oniro.
Braiil: General EIrctrk, S. A., Rlo de Janeiro and Sao Paulo.
Canada: Canaillao General Electric Co.. Ltd . Toronto,
Chile: Inirrnaiional Machinery Company, ^antlaco. Aaie-
tacaita and Valparaiso.
China: Andersen, Meyer A Company. Ltd . Shanfhai
International General ElecUlc Co. lor (General repreaeou-
tlvts of ibe Far East, esciudlns China and Japenj Shanxhal.
Colombia: Wesselhoeft A Poor. Medellin, Barranquilla and
Bocou.
Cuba: General Electric Company of Cuba. Havana.
Dutch East Indies: loteraatlonal Genera! Electrk Co.. Inc..
Soerabala. Java.
Ecuador: Carlo* Cordovet. Guayaquil and Quito.
Egypt: British Thomson- Houston Co., Ltd.. Cairo.
France and Colonies: Compac&le Fraocalse Thomson-Houston.
Paris
creal Britain and Ireland: Brlti-ihThomson-HoustonCo-, Ltd
TtiJBhvr Inirrn-illonsl CrnersI Flerlrlc Co . Inc . London
Coinpa<n>e Francalse Thomsoa-
Muuatou, Paris. France.
India- British T bom sop- Houston Co . Ltd . Calctiita and
Bombay: Intemail«nal General Eketrtc Company. Inc..
Cakutia
Italy and Colonies tranco Toil Socleia Aoonima. Milan
Japan: Shlbaura tncioeerinf Uorkt. Tokyo Tokyo Llertrte
Co.. Ltd . Kaaasalil
Meiko: Mexican Gcorral Electric Co . City of Meako and
Guadalajara.
New Zealand: Nattnoal Electrical A Enclneerlnt Co., Ltd.
Auckland. Dunedln. Chrutrhurch. and Welllixtoo
Paracua). (.General Llecirk. S A . Buenos Aires . Arfentlsa.
Peru: U . R (..race A Co . Lima
Philippine Islands. Pacific Commeeclal Company, MaalU
Porto Rko: Intemaikmal (ieorral Electrk Co.. Itte.. 8aA Juaa
?anuci
Psrb. France
Kuisia: U spobabtrhaU ElecirtcbeskaU Kempaaia. P««r«<r«d
and Vladlvustuk
South Africa, south African (ieneral Electric CcLtd.Jo-
batiDrsburc and i apetown
Spain and colonic* < ompacnle H
Acetkce d'b^pecnc Madrid. Spain
L'rucuay tieoeral Ekt'irlc, - A .Montevideo
Vcncjtt^lB- UMwift.wti A Poor, Caracas.
General Office
Schcnectadv: N.^'.
l^Elff'^tric
Company
Sales Olhces in
all large cities
Sov \ou snu' it iuhi-rlisid in lite Genkr.m. Electrk Rf.vifw
TWO DOLLARS PER YEAR
TWENTY CENTS PER COPY
GENER/a ELECTRIC
REVIEW
VOL. XXIII, No. 9
Published by
General Electric Company's Bureau,
Schenectady . N. Y.
SEPTEMBER, 1920
yo^
More Than Four Years of Electric Operation on the Mountain Divisions of the Chicago, Milwaukee & St. Paul Railway
Have Amply Demonstrated the Merits of Electric Traction for Main Line Service. The operating
results of this electrification for the year 1919 are published in the article, page 724
For
Fractional H. P. Motors
Ol'AI.ITY seeks qualit) — which explains why
"NORmfl Bearings are standard in hundreds of
of thousands of high-speed, high-duty electrical ma-
chines internationally known for their super-service
ahilitv. Fundamentally right in design and uni-
fonnly maintained at the highest standard-; nt
materials, workmanship and precision, "NORff^fl"
Bearings will make an\- machine a better machine,
capable of longer, better service.
See (hat your Motors
are "HORri\f\" Equipped
Ball, Rollei', Thrust and Combination Bearings
General Electric Review
A MONTHLY MAGAZINE FOR ENGINEERS
Associate Editors, B. M. EOFF and E. C. SANDERS
Manager, M. P. RICE Editor, JOHX R. HEWETT In Charge of Advertising, B. M. EOFF
Subscription Rates: United States and Mexico, $2.00 per year; Canada, $2.25 per year; Foreign, $2.50 per year; payable in
advance. Library and Student Rates: United States and Mexico, $1.50 per year; Canada, $1.75 per year; Foreign, $2.00 per year;
payable in advance.
Remit by post-office or express money orders, bank checks, or drafts, made payable to the General Electric Review, Sche-
nectady, N. Y.
Entered as second-class matter. March 26, 1912, at the post office at Schenectady, N. Y., under the Act of March, 1879.
Vol. XXIII. No. II Av Co,^rTgS//cT,„/>.,,. September. 1920
CONTENTS Page
Frontispiece: A Modern Electric Furnace and Automatic Temperature Control Equipment 722
Editorial: Some Results of the Chicago. Milwaukee & St. Paul Electrification. . . . 723
Electric Power Consumption on the Rocky Moimtain and Missoula Divisions of the
C, M. & St. Paul Rwy 724
By Reinier Beeuwkes
The Production and Measurement of High Vacua
Part IV. Manometers for Low Gas Pressures 731
By Saul Dushman
The Cooper Hewitt Mercury Vapor Lamp
Part I. Theory and Operation . . ■ 741
By L. J. BiTTOLPH
The Importance of the Electrical Industry in the Foreign Trade of the United States . . 752
By M. A. OuDiN
Power Control and Stability of Electric Generating Stations: Part II 756
By Charles P. Steinmetz
Relative Thermal Economy of Electric and Fuel-fired Furnaces 708
By E. F. Collins
Condenser-resistance Protective Device 774
By J. L. R. Hayden
Typical Installations of Electric ]\Iine Hoisting in South Africa 775
By E. B. Bell
Opportunities in Office Work 782
By Anna McCann
A New Co-operative Course in Electrical Engineering 784
Bv W. H. Timbie
I
-
L
V
,^
■ ■.- -
\
A
L
Metallic Resistor Electric Furnace and Automatic Temperature Control Equipment InstaUed in Building No. 17. Schenectady
Works of the General Electric Company. Used for hardening punches, dies and cutters. See article.
•Relative Thermal Economy of Electric and Fuel Fired Furnaces." page 764
m ElECTMC
SOME RESULTS OF THE CHICAGO, MILWAUKEE & ST. PAUL
ELECTRIFICATION
It is now more than four years since electri-
cal operation was be^^un on the C. M. & St. P.
Rwy., and the perfomiance for the year
1919 therefore covers a period of seasoned
operation; in other words, the equipment was
neither new nor worn out. The statistics for
the year 1919 compiled by Mr. R. Beeuwkes
and published in this issue may therefore be
taken fairly to represent the results of this
electrification and should be carefully ana-
lyzed by all engineers and executives inter-
ested in railway operation.
The statistics on power consumption of
34.9 kw-hrs. for freight and 39.7 kw-hrs. for
passenger service indicate that 40 watthours
per ton mile at the high tension bus (com-
monly used for estimating purposes) is a
conservative figure.
It should also be noted that the entire
amount of power charged to train haulage
(including switching) is apportioned between
the passenger and freight service. The
segregation of the small amount of energy
used by the four 70-ton switchers is unim-
portant because of the relatively small pro-
portion.
In the table showing power outputs at the
high tension bus and at the locomotives the
mistake should not be made of assuming that
the ratio derived is the efficiency for the sub-
station and overhead distribution. In this
case the introduction of the regeneration
must be taken into account; otherwise the
distribution efficiency for a road using regen-
eration would appear to be less than it would
be without this feature. It will be noted that
the greater the amount of regeneration the
lower will be the ratio shown in this table.
The gross figure before subtracting regener-
ated energy should, of course, be used to
compute actual distribution efficiency.
In selecting substation capacity provision
was made for future requirements, wh ch have
not yet been reached. That a great reserve
capacity is still available is shown by the low
average load on the substation. This accounts
for the rather high distributing losses. Under
the terms of the power contract, however, the
losses at light load do not aff'ect the power
cost when the (lO per cent load factor is not
exceeded.
The data show an interesting record of the
experiments made to determine the proper
limit setting for the power-limiting and
indicating system, and the percentage of
time during which the limiting feature was
operative for the different settings. In the
interest of rapid train movement it was
desirable to reduce this figure to as small a
value as possible without injuring the load
factor. It will be seen that the limit setting
of 12,000 kw. must have appreciably slowed
down the train schedule, while the 16,000-kw.
setting with a somewhat higher average load
affected the speed of trains only 2 per cent
of the time. Obviously the information con-
, stantly at hand with this indicating system
is invaluable and the apparatus has undoubt-
edly paid for its initial cost many times over.
In arriving at the figure for the total unit
cost of freight transportation in cents per
gross ton mile, items are included to cover
the cost of maintenance to the electrical dis-
tributing system, two of which (3S3 and 305)
represent principally the cost of substation
attendance. The cost of electric power is
charged against other accounts under the
heading of train power purchased. The fig-
ures showing the total expense of furnishing
energy for train operation are computed both
on a basis of trailing tonnage, and total train
weight including the locomotives.
724 September, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, No. 9
Electric Power Consumption on the Rocky-
Mountain and Missoula Divisions of
the C, M. CS, St. P. Rwy.*
By Reinier Beeuwkes
Electrical Engineer Chicago, Milwaukee & St. Paul Rwy.
The figures presented in this article covering the operation of the electrified divisions of the C, M. &
St. Paul Rwv. for the year 1919 should be carefully analyzed by railway executives and engineers. Volumes
have been written and published to show the economy and other advantages of electrical operation of main
line railways, specially over mountain grades, but no arguments for electrification on paper will carry the
weight of conviction of actual performance. If the operation of our main line railways by electric power is
ever to be an accomplished thing present experience would seem to indicate that it will be realized by means
of high voltage direct current, and this fact gives added significance to the results obtained by the C., M. &
St. Paul. — Editor.
Power for the electrical operation of the
Chicago, Milwaukee & St. Paul Railway
between Harlowaon, Montana, and Aver\-,
Idaho, is delivered to the transmission system
in the form of '100,000-volt, three-phase,
60-cycle current. The power is supplied under
two separate contracts, one for the Rocky
Mountain division, extending from Harlowton
to Deer Lodge, and the other for the Missoula
division, extending from Deer Lodge to Aver>-.
The power company's 100,0UU-volt trans-
mission lines are shown in the single line
layout of the system, as are also the points of
power deliven,- to the railway company and
the latter's lOO.OOO-volt transmission system.
* Paper presented at the Pacific Coast Convention of the
American Institute of Electrical Engineers. Portland, Ore.. July
21-2.3. 1920.
Burfte Lines 24 M.
The railwa\- transmission line of the Rocky
Mountain division extends from Two Dot
substation to the Morel substation, a distance
of 184 miles, the former point being 12 miles
from Harlowton, eastern terminus of the
division, and the latter point 17 miles from
Deer Lodge, the western terminus. Power is
delivered by the power company at the Two
Dot, Josephine. Piedmont and Morel sub-
stations. The railway transmission line of the
Missoula division extends from Gold Creek
.substation, IN^o miles from Deer Lodge, a
distance of ISO miles, to the substation at
Aver\-, the western terminus of the di\-ision.
Seven substations on each division are
used to convert the lOO.OOO-volt alternating
current of the transmission line to the .'JIIOO-
Horlowion
lowttft
TMoOot
Fig. 1
DitrLodqt \C-J
Dupatchtri
Office
^nacondd _
Jannty^
Transmission Lines of the Montana Power Co. and Substation Layout for the C.
Blectrificati on
ELECTRIC POWER CONSUMPTION ON THE C, M. & ST. P. RWY.
TABLE I
SUBSTATIONS AND THEIR EQUIPMENT
Substations
Transformers
Motor- Generators
Rocky Mountain Division
Two Dot...
Two 2500 kv-a.
Two
2000 kw.
Loweth ....
Two 2500 kv-a.
Two
2000 kw.
Josephine . .
Two 2500 kv-a.
Two
2000 kw.
Eustis
Two 2500 kv-a.
Two
2000 kw.
Piedmont. .
Three 1900 kv-a.
Three
1500 kw.
Janney. . . .
Three 1900 kv-a.
Three
1500 kw.
Morel
Two 2500 kv-a.
Two
2000 kw.
Missoula Division
Gold Creek .
Two
2500 kv-a.
Two 2000 kw.
Ravenna . . .
Two
2500 kv-a.
Two 2500 kw.
Primrose , . .
Two
2500 kv-a.
Two 2000 kw.
Tarkio
Two
2500 kv-a.
Two 2000 kw.
Drexel
Two
2500 kv-a.
Two 2000 kw.
East Portal .
Three
2500 kv-a.
Three 2000 kw.
Avery
Three
1900 kv-a.
Three 1500 kw.
volt direct current used for traction purposes.
Each motor generator consists of two 1500-
volt direct-current generators connected in
series and driven by a 2300-\'olt synchronous
motor supplied from the substation high
tension busses t hrough a three-ph ase , 1 00 , 000 /-
2300-volt transformer and is guaranteed for a
maximum five-minute overload of 200 per
cent. The rated capacities of these stations
are given in Table I.
The railway company's high-tension line,
arrangement of apparatus in the substations
and the general layout of the 8000-volt dis-
tribution or trolley system, are shown dia-
grammatically in Figs. 1 and 3.
The contact wires of the trolley system
consist for the main line of two No. 0000
B.&S. grooved trolley wires flexibly supported
side by side from a }/i-'m. steel catenary and
tapped at intervals of about every 1000 ft.
to a feeder which connects to the adjacent
substation busses through switches and auto-
matic circuit breakers. Over passing, indus-
trial and similar tracks only a single No. 0000
copper trolley wire is used. There is an
insulated air gap in the trolley in front of each
substation separating the trolley system west
of the substation from that east of the sub-
station; that is, portions east and west of the
substations are fed, respectively, through
separate feeder breakers. There is also an
insulated air gap at the beginning and end of
every passing track, so that by means of a
section switch installed in the feeder at the
gap the district between any two gaps may be
isolated in case of trouble so as to permit opera-
tion up to the location of the open switches.
10 eo 30 40 50 eo to eo 90 roo no i2o 150 i40 \5o 160 i7o 180 190 soo iio eso rao
Fig. 2. Graphic Train Sheet and Load Curve for the Rocky Mountain Division, February 19, 1920
72(i September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 9
if V-
,2£
11 tie^
•S3
?E°
<3 U V
'^ c u u y_
>t O c <S £9 ft>
o « i
'Pi
1 -s^ nm 111
5o £ii >. " u
•=Z §21 1*
aog u-o'' g S
-2 O-O-n^ c * i.
-^ a ^ c 0::; «
O *'Q (5 O ^^
•'S o «.- •- *'
u —
« O w
^ u a
O OfQ
£ ox
f u .
"" - o
ccoz
.2*
%^^
-. w « 2^ ..
ELECTRIC POWER CONSUMPTION ON THE C, M. & ST. P. RWY.
The return circuit consists of the 90-lb.
running rails and, in general, of a No. 0000
B.&S. copper supplementary negative wire
which is run along the trolley poles and
connected to the track at intervals averaging
about 8000 ft. through each alternate signal
system reactance bond. This supplementary
negative, however, is intended more as a
safety measure to bridge open rail bonds than
to increase the return circuit conductivity.
However, on various feeder cutoffs on the
mountain grades, where the conductivitv
of the positive circuit closely approaches that
of the return circuit, one of the two feeders
on the cutoff is in parallel with the running
rails and is provided for the purpose of
increasing the return circuit conductivity.
Power Demand Controlled by Train Dispatcher
The terms of the power contracts are similar
and each provides for a minimum payment
on basis of a 60 per cent load factor. Where
the load factor exceeds 60 per cent payment is
made on basis of the actual kilowatt-hours
consumed, the rate being 5.36 mills per
kilowatt-hour. The demand is controlled for
each division by means of a so-called power
indicating and limiting system,* which on
the Rocky Mountain division was put into
operation early in the year 19 IS and on the
Missoula division a few months ago. Briefly,
this system is so arranged as to indicate and
record at the dispatcher's office at Deer
Lodge the total kilowatts or dem.and being
supplied in any instant by the power company
to the railway company and to prevent the
maximum demand from exceeding a certain
amount as detennined by the "demand
setting made by the dispatcher," this limiting
action being secured by lowering of the sub-
station direct-current voltage and therefore
of the train speeds.
The effect of this limiting action is clearly
indicated on the graphic time-table (Fig. 2) of
train movements on the Rocky Mountain divi-
sion for February 19, 1920, and corresponding
load curve traced by tapalog meter of the
power indicating and limiting system with
the load limit set at 16,000 kw.
The percentage of time when the limiting
action will take place, for a given amount of
business, will depend on the demand setting
and on the possibilities of spacing the trains
so that as few as possible will at one time be
operating on the heavisr grades, the latter
matter, except as regards passenger trains
♦This system was described in the General Electric Review
for April. 1920. page 292.
TABLE II
ST. PAUL ELECTRIFICATION, LIMIT SET-
TING AND AVERAGE KILOWATT-
HOUR TAKEN MONTHLY
Month
Limit
Setting
Average
Monthly
Load in
Kilowatts
July,
Aug.,
Sept.,
May,
Aug.,
Sept.,
Oct.,
Nov.,
Feb.,
Mar.,
April,
1918
1918
1918
1919
1919
1919
1919
1920
1920
1920
,000
000
000
,000
000
1919 14,000
,000
000
,000
,000
,000
12,1
12,1
12,1
14,1
14,1
14,i
14,1
16,1
16,1
16,1
8020
7820
6675
7840
7650
8230
,8420
7115
8625
8680
8620
Per Cent
Time
Limiting
Action
Takes
Place
13.0
15.5
8.2
4.62
4.12
9.50
10.65
8.24
2.40
2.20
.90
and certain time freights, being to a consider-
able extent in the hands of the train dis-
patchers. The slowing up of the train speeds
of course results in increased train and engine-
men's expense and increased time in getting
freight over the road, and a proper balance
must be struck between this increased expense
and the saving in power cost and the limit
setting determined upon accordingly. Table
II gives an idea of the percentage of time the
limiting action takes place with average
kilowatt load and settings as indicated, this
percentage being based on the number of
hours the limiting system was actually in
service.
In arriving at the amounts chargeable for
power against the respective classes of train
service, the total kilowatt-hours to be paid
for — that is, the actual kilowatt-hours, or the
actual kilowatt-hours increased, if necessary,
to correspond to a minimum 60 per cent load
factor — is taken and from it is deducted
the kilowatt-hours m_etered against substation
lighting, auxiliar\- power, signal system
supply, etc., amounting to about 1 per
cent. The remaining kilowatt-hours is then
divided between the different classes of train
service, freight, passenger and non-revenue,
in proportion to the total net kilowatt-hour
readings obtained for these respective services
from wattmeters installed in the locomotives.
These readings are taken by the engine crew on
entering or leaving the engine on the form pro-
videdforthepurpose, andarecordof the power
consumption of each train is thus obtained.
The readings are referred to as "net" read-
ings, as they represent motored energy less
regenerated energy.
728 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 9
TABLE III
ST. PAUL ELECTRIFICATION-
AVERAGE INPUT
Month
January . . .
February . .
March ....
April
May
June
July
August ....
September .
October. . .
November .
December.
EOCKY
MOUNTAIN DIVISION
MlSSOt.X.\ DIVISION
Actual Kw-Hrs.
Net Kw-Hrs.
Actual Kw-Hrs.
Net Kw-Hrs.
System Input for
Input at
Ratio
System Input for
Input at
Ratio
Locomotives
Locomotive
Locomotives
Locomotives
6,381,233
4.838,480
75.9
5,540,581
3,753,430
67.6
4,610,607
2,921,840
63.3
4,107,960
2,702,710
65.8
5,795,859
4,351,126
75.2
5,412,048
3,469,120
64.2
5,949,840
3,962,650
66.6
5,429,932
3,574,080
65.8
5,803,455
4,146,517
71.4
5,745,397
3,795,770
66.2
5,662,650
4,100,810
72.3
5,697,785
3,853,590
67.6
5,744,738
3,794,940
66.2
5,318,692
3,505,630
65.8
5,648,815
3,755,280
66.5
5,133,008
3,255,820
63.4
5,892,430
3,799,830
64.5
5,102,562
3,434,010
67.3
6,222,486
3,971,149
63.8
5.389 ,88:j
3,654,955
67.8
5,095,937
3,425,458
67.2
4,879,130
3,181,456
65.2
5.809.976
3,830,870
65.8
4,971,601
3,382,700
67.9
68.618.026
46,898,850
68.3
62,728,579
41,563,271
66.3
The ratio of the total net locomotive watt-
meter readings, all services, to the total
actual kilowatt-hours input to the system
chargeable to locomotives for the various
months of 1919 is given in Table III.
As there are no wattmeters installed in the
direct-current side of the substations, a ratio
for net substation output to system input or to
locomotive is not obtainable. There are,
however, wattmeters in the circuits of the
individual motor-generator sets and Table IV
considered in connection with the profile of
the line will be of interest in showing the
manner in which the energy is distributed
among the respective substations, average
kilowatts being used for convenience instead
of total kilowatt-hours, and the whole of the
year 1919 being taken.
Operating Figures for 1919
The figures in Table V show forthe year 1919
the net kilowatt -hours per thousand gross ton-
miles for freight revenue sen-ice and passenger
service, respectively, and corresponding cost of
these kilowatt -hours at the high tension bus or
point of delivery- of the power to the railway
system. The lesser consumption of energy-
during the summer months as compared with
the winter months will be noted. The figures
for the passenger ser\-ice are approximate,
as the ton-mile data are based on the assump-
tion of an average weight per car, no record
of the particular cars handled in all the
separate trains being available.
The cost of maintaining and operating the
transmission lines, substations and trolley
system for the year 1919 is given in Table VI
TABLE IV
ST. PAUL ELECTRIFICATION— AVERAGE INPUT OF SUBSTATIONS
ROCKY MOUN'TAIN DIVISION
Substation
Two Dot .
Loweth . .
Josephine .
Eustis. . . ,
Piedmont.
Janney . . .
Morel ....
System Total.
Average Annual Kw. Input Net
to Motor Generators
•Total
•• Per Motor-
Generator
895
962
1014
1022
1218
1.390
1047
7548
813
783
1013
1016
617
559
1072
MISSOfLA DIVISION
Substation
Average Annua
to Motor
♦Total
Kw. Input Net
Generators
*• Per Motor-
Gaocrator
Gold Creek
1150
915
908
843
790
1390
812
1128
Ravenna
1115
925
Tarkio
Drexel
East Portal
803
778
778
Avery
523
System Total
6808
* Total kw-hrs. computed on the basis of 8856 hours in the year (four extra days in December being included).
•• Computed from total kw-hrs. and total running hours of motor g(
' generators.
ELECTRIC POWER CONSUMPTION ON THE C, M. & ST. P. RWY.
-29
TABLE V
ST. PAUL ELECTRIFICATION OPERATING STATISTICS FOR 1919
Net Kw-Hrs. per Thousand Gross Ton- Miles
Month
Thousand
Gross
Ton-Miles,
Trailing
At High
Tension
Bus
At
Locomo-
tive
At High
Tension
Bus
At
Locomo-
tive
Load
Factor
Cost of Kw-Hrs.
per M. Trailing.
Gross Ton-Miles,
Cents
Rocky Mountain Division
Freight Service:
January
February
March
April
May
June
July
August
September
October
November
January-November .
98,478
79,859
118,297
121,646
124,395
122,264
120,723
111,092
115,787
108,920
86,267
Averages
47.8
43.1
.39.0
38.5
36.5
36.7
36.7
40.9
39.7
45.8
44.0
40.5
36.3
27.3
29.3
25.6
26.1
26.2
24.3
27.2
25.6
29.2
29.6
27.7
41.2
37.3
33.9
33.1
31.7
31.7
31.6
34.9
34.1
39.4
37.7
34.8
31.3
23.6
25.5
22.0
22.6
22.9
20.9
23.2
22.0
25.1
25.3
23.8
63.7
57.7
65.3
61.1
56.0
56.4
55.4
54.6
58.8
60.0
50.9
57.3
25.7
24.0
20.9
20.7
20.9
20.9
21.3
22.4
21.7
23.6
27.8
22.5
Rocky Mountain and Missoula Division
January
February
March
April
May
June
July
August
September
October
November
January-November
87,958
73,481
103,613
109,133
118,331
116,660
106,045
101,017
99,578
100,504
78,459
Averages
44.3
39.8
40.3
38.5
37.9
37.8
38.1
.38.8
38.5
40.0
45.3
39.7
29.9
26.2
25.8
25.4
25.1
25.6
25.0
24.6
25.9
27.1
29.5
26.3
38.6
35.2
35.6
34.1
33.5
33.3
33.5
34.3
34.1
35.3
39.2
35.0
26.1
23.2
22.8
22.4
22!2
22.5
22.0
21.8
22.9
23.9
25.5
23.1
23.8
21.7
21.6
20.2
20.3
20.3
20.4
20.8
20.6
21.4
24.3
21.3
Rocky Mountain and Missoula Divisions Combined
January-November.
January-December .
Passenger Service:
January- November.
January- December.
2,302,507
2,476,085
340,480
378,080
40.1
56.8
27.1
38.7
34.9
39.7
23.5
27.1
21.9
22.3
38.4
38.1
ST.
TABLE VI
PAUL ELECTRIFICATION— DISTRIBUTION OF OPERATING COSTS FOR 1919
Account
Total All 1 p„ TT„;,.
Services , ^^"^ ^nit
255. Power substation buildings
257. Power transmission system
259. Power distribution system
261. Power line poles and fixtures
$ 8,487 $ 606.00 per building
1,773 4.87 per mile
78,461 1 179.00 per route-mile
24,299 ' 55.50 oer route-mile
306. Power substation apparatus
383 1
„„-■ > Train and yard power produced
40,224
102,152
2,870.00 per station
7,300.00 per station
Total
$255,396
730 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 9
TABLE VII
UNIT COSTS. INCLUDING COST OF POWER
1. Cost per thousand gross ton-miles trailing freight as actually distributed in accounts 28.8 cents
2. Cost per thousand gross ton-miles train freight as actually distributed in accounts 24.9 cents
3. Cost per thousand gross ton-miles trailing freight on basis distribution in proportion to freight
kilowatt-hours 30.2 cents
4. Cost per thousand gross ton-miles train freight on basis distribution in proportion to freight
kilowatt-hours ., 26.2 cents
5. Cost per actual kilowatt-hours, delivered to locomotives 1.1 cents
* Note. — The items in the table refer to the classification numbers prescribed by the Interstate Commerce Commission for steam
railroad accounting. The several groupings are defined as follows:
255. Power Substation Buildings.
This shall include the cost of repairing buildings of power substation * * » used to transform power for the operation of trains and
cars, and to furnish power, heat, and Ught for general purposes: * * *
257. Power Transmission System.
This account shall include the cost of repairing systems for transniitting high-tension power from power houses to the point where
transformed for use, including the cost of work-train service and special tools furnished for such work.
259. Power Distribution Systems.
This account shall include the cost of repairing electric distribution systems, whether overhead, surface, or underground, for conveying
low-tension power for propelling trains and cars, and for power, heat, light, and general purposes.
261. Power Line Poles and Fixtures.
This account shall include the cost of repairing and replacing electric line poles, cross arms, and insulating pins; brackets and other
pole fixtures; and braces and other supports for holding poles in position; also the cost of repairing structures primarily for supporting
the overhead electric construction.
306. Power Substation Apparatus.
This account shall include the cost of repairing machinery and other apparatus, including special foundations, for transforming or
storing power in power substations used for the operation of trains and cars and for power, heat, and light for general purposes.
Details of Power Substation Apparatus.
Rotary Converters Switchboards
Storage Batteries Transformers
383. Yard Switching Power Produced.
This account shall include the cost of the production and distribution of electric power used in operating locomotives and cars in
switching service in yards where regular switching service is maintained, and in terminal switching and transfer ser\*ice.
Employees. — The pay of employees engaged in operating electric-power stations and substations, such as engineers, firemen, elec-
tricians, dynamo men, oilers, cleaners, and coal passers.
Fuel. — The cost of coal, oil, gas and other fuel, including the cost of labor unloading or stocking fuel.
Water. — The cost of water used to produce steam or to operate water plants, including pumping, rent of ponds, streams, and pipe
lines, also water tests, boiler compounds, and other like supplies and expenses.
Other Supplies and Expenses. — The cost of lubricants, such as oil and grease used in lubricating engines, shafting, dynamos.
and pumps; cost of waste, carbon brushes, fuses, lamps, and other supphes; also the cost of heating and lighting power plants, and other
expenses not elsewhere specified in connection with operation of electric -power plants.
395. Train Power Produced.
This account shall include the cost of producing and distributing electric power for the propulsion of electric locomotives and cars
in transportation train service. Otherwise same as account No. 383.
and a final figure thus arrived at showing the
approximate total- operating costs involved
in the delivers^ of the electric energy to the
locomotives is given in Table VII.
Conclusion
The installation being comparatively new
it might naturally be assimied without
consideration of other facts that the figures
for the maintenance are considerably lower
than those which will eventually obtain, but
it should also be borne in mind that the
maintenance and operating costs given \y\\\,
except for power, remain more or less constant
so far as any consideration of their being
affected by the business handled is concerned,
so that the cost per thousand ton-miles would
be correspondingly reduced as business is
increased. It is also expected that consider-
able improvement will be effected in main-
tenance methods, which would again tend to
reduce costs. The figures are therefore given
merely to show the results which are at
present being obtained.
'31
The Production and Measurement of High Vacua
PART IV. MANOMETERS FOR LOW GAS PRESSURES
By Dr. Saul Dushmax
Rese.\rch Laboratory, General Electric Company
This installment and the next contain a description of different types of manometers used in connection
with high-vacuum technique. The present installment deals mainly with the McLeod gauge and different
forms of viscosity gauges. The next installment will discuss the Knudsen, Pirani-Hale, and ionization
gauges. — Editor.
For the measurement of pressures that lie
between one atmosphere and one cm. mer-
cur\', a standard form of mercun.^ barometer
is generally used. Such a method is obviously
ver\^ insensitive when it is necessary to
measure pressures below this range, and con-
sequently a number of types of manometers
have been developed by different investiga-
tors for this purpose.
In the simplest type of low-pressure gauge,
the difference between the actual pressure and
that in an extremely good vacuum is measured
by some very sensitive optical method. This
is the principle of Rayleigh's manometer.
On the other hand, the McLeod gauge repre-
sents an interesting application of Boyle's
law to very low pressure. By compressing a
given volume of the gas whose pressure is to
be measured to a very small known voltime,
the pressure is amplified several thousand-
fold and may read directly.
Again, instead of attempting to measure
the pressure directly, use may be made of the
fact that the amount of heat conducted from
a stirface varies with the gas pressure.
Similarly, the damping effect of gas on a body
set in vibration or rotation varies with the
pressure. In each case, however, it is neces-
sary to know the law of variation between the
observ'ed effect and the pressure.
In the following section are described some
of the different types of low-pressure gauges
that have been used by different investigators.
Only those forms are described in detail which
have proved to be most generally useful in the
present state of high-vacuum technique ; while
other forms, which are of more or less histori-
cal interest, are mentioned rather briefly.
MERCURY MANOMETERS
Rayleigh's Gauge^
The essential parts of this gauge (Fig. 32)
are two glass bulbs, one of which com-
1 Phil. Trans. 196. A. 205 (1901) Zeits. physikal. Chem. 37,
713 (1901).
' Zeits. f. Instrk. S9. 344-349 (1909) K. Jellinek. Lehrbuch
d. physikal. chem. I. 1, p. ,321. Ann. d. Phys. S9. 723 (1909).
» Zeits. (. Instrk. 6'. 89 (1886) and 2.',. 276 (1904).
' Ann. d. Phys. (4). gl. 320 (1906).
municates with a good vacuum by a tube C,
and the other with the system in which the
pressure is to be measured. Two glass point-
ers are sealed into the bulbs, and the latter
are connected to a T-connection which forms
the upper end of a barometric column A.
Fig. 32. Rayleigh's Gauge
Mercury can be raised and lowered in the
bulbs by means of the reservoir D and the
level thus brought up so as to be flush with
the ends of the pointers. Any difference in
pressure on the mercury in the two bulbs is
then measured by gradually tilting the frame-
work AK and observing the deflection on a
mirror which is fastened vertically on top of
the bulbs at /. According to Rayleigh this
gauge can be used to read pressures between
1.5 mm. and 1 X 10~' mm. of mercur^'.
A modified form of this gauge was used
by K. Scheel and W. Heuse^ for measuring
the vapor pressure of water at temperatures
below 0 deg. C, and similar manometers
have been constructed by M. Thiesen', and
E. Hering^.
More recently, an ingenious modification
of Rayleigh's method has been used by C. F.
'32 September, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, No. 9
MundeP for measuring vapor pressures at
very low temperatures. A verj' sensitive
optical method for measuring slight dif-
ferences in level of two mercury- surfaces,
developed bv K. Prvtz^ has been used
X Pl/mp
To source of
gases or vapors
Fig. 33. Optical Lever Manometer
extensively by different investigators in con-
nection with Rayleigh's method.'
In the optical lever manometer, described re-
cently by J. E. Shrader and H. M. Ryder,*
the same object is attained by a very simple
constuction. The following description is
quoted from the orignal paper;
"A mercun,' U-tube manometer (Fig. 33)
is formed in the usual manner, except that
the surfaces of the mercury- are so arranged as
to be of relatively large area. Above one of
the surfaces, within the tube, is arranged an
optical lever as shown in the illustration.
This lever is supported by two knife edges,
a-a. which rest on loops of wire, which in turn
are sealed into the glass walls of the tube; a
glass bead 6, fused to the end of the lever aim
acts as a float on the mercury surface, and in
this way transmits the motion of the mercury-
surface to the lever arm. A mirror M
attached at the position shown acts in the
usual manner to reflect a beam of light from a
lamp to a scale, if the gauge is to be arranged
as an indicating instrument. If the gauge is
to be used for recording vairiations in pressure,
» Zeits. f. Physikal. Chem. SS. 433 (1913).
« Ann. d. Phys. (4). 16. 73.'> (1<H).">).
' C. F. Mundcl loc. cit., and M. Knudsen. Ann. d. Phys. (4)
SS. 1435 (1910).
• Phys. Rev. tS, 321 (1919).
the scale may be replaced by a photographic
device such as is used in oscillographic work.
"The cross connection e provides an easy
means of evacuating the whole system with
one pump located as shown. With this stop-
cock or mercury cutoff open, a zero reading
can be easily obtained, after which this con-
nection may be closed and the gases or
vapors introduced for measurement. This
system provides also for the measurement of
Fig. 34. McLcod Gauge
small variations in pressure, with an original
pressure of any desired value, this value in no
way affecting the absolute sensibility of the
gauge."
A sensitivity of 10"' mm. of mercur>- is
claimed for the gauge, and it certainly ought
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
733
to prove useful in those cases where the
McLeod gauge is inapplicable.
McLeod Gauge
The principle of this gauge consists in com-
pressing a given volume V, of the gas whose
pressure P is to be measured, to a much
smaller volume v and observing the resultant
pressure p which in accordance with Boyle's
law is given by the relation
V
V
The greater the ratio — , the greater the sen-
sitivity of the gauge.
One of the simplest forms of McLeod
gauge is shown in Fig. 34. The bulb V, to
which is attached a capillary tube aa, is con-
nected to the low-pressure system at E and
also to the barometric column T. In order to
avoid errors due to the effect of capillarity, a
tube bb of the same diameter as aa is sealed on
as a by-path to the larger tube E. To operate
the gauge the reservoir B is raised, thus forc-
ing the mercury in the barometric column
upward until the gas in V is shut off from the
remainder of the system.
As the mercury is raised further, the
volume of gas V is compressed until finally
the mercury in the capillary bb is level with
the upper end of the capillary aa (correspond-
ing to the point O on the scale) . The pressure
on the gas in the capillary is then evidently
equal to that of the mercury column of height
/;. Now let a denote the volume of the capil-
lary per unit length, and P denote the pressure
in the system at E. Then it follows from
Boyle's law that
P = —h^
(22)
Since a and V are constant for any particular
gauge, it follows that the pressure is pro-
portional to the square of the observed value
of h. It also follows from this equation that
the smaller the ratio a/V the greater the
sensitivity of the gauge. Practical con-
siderations, however, make it impossible
to use either extremely fine capillaries or very
large volimies for V. The following data for
a gauge used by the writer are of interest in
this connection as an indication of the range
of pressures that can ordinarily be measured
with a McLeod gauge :
' Ber. d. deutsch. Physikal. Ges. 10. 783 (1908).
'» Phil. Trans. (A) 196. 20.5 (1901).
" Ber. d. deutsch. Phvsikal. Ges. ;/. 10 (1909).
" Ann. d. Phys. 41. 289 (191.3).
Gauge No. 1
V= 171 cm.'', a = 0.00407 cm.^ per cm. length
(diameter of capillary = 0.72 mm.).
Hence, measuring h in cm.,
p = ^M^ML^ = 2.38 X 10-' /z2(mm . of mercury)
= ^X2.38X10-Wi= = 0.317 /j-' (bar)
750
That is, for h = l cm., P = 0.317 bar; and
for h = l mm., P = 0.0032 bar, so that for a
10-cm. length of capillary aa, the range of
pressures that could be measured with this
gauge is from 0.003 to 32 bar.
Actually, it is impracticable to make V
larger than 500 cm.' and with capillaries
smaller than 0.5 mm. the mercury tends to
stick badly and the gauge is very sluggish in
operation. With V =500, and a = 2X10-*
(d = 0.5 mm.), 1 cm. on the capillary would
correspond to 4X10~* mm. of mercury, or
approximately 0.053 bar, and 1 mm. to
0.00053 bar. In general, the lower limit of
pressure that can be measured with a McLeod
gauge is about 0.01 bar.
It is evident that the McLeod gauge does
not indicate the pressure of mercury vapor
and condensible vapors such as those of oil,
water, and ammonia. Even in the case of
carbon dioxide the gauge is very inaccurate.
In using it to measure very low pressures, such
as those produced by a Gaede molecular or
Langmuir condensation pump, a liquid air trap
should be inserted between the gauge and the
remainder of the system.
Regarding the accuracy of the gauge for
indicating the pressure of the so-called
permanent gases {Hi, He, Ne, Ar, O2, N2 and
CO) a careful investigation carried out by
Scheel and Heuse" has shown that if the bulb
and tubing are carefully dried (to eliminate
the presence of a film of water) the results
obtained in the case of air are certainly
reliable down to pressures of 0.01 mm. of
mercury and are probably just as exact at
lower pressures.
Lord Rayleigh'" found by means of his
differential manometer that in the range of
pressures 0.001 mm. to 1.5 mm. Boyle's law
holds accurately for N2, H2, and On; and
Scheel and Heuse" observed the same result
with their membrane manometer. A very
careful investigation on this point was
carried out by W. Gaede '- in connection with
his work on the laws of flow of gases at low
pressures. He found that in the case of
nitrogen and hydrogen, the McLeod gauge.
734 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 9
when care is taken to dr\' the walls thor-
oughly, is \-ery accurate down to very low
pressures (below 0.0001 mm.), while in the
case of oxygen errors are liable to arise
because of the formation of an oxide scum on
the surface of the mercur\- which causes the
surface to wet the glass in the capillary. How-
ever, this sciun may be got rid of by heating
the capillary- carefully and the mercur\- then
becomes quite clean again.
There are certain features about the
McLeod gauge that must be carefully ob-
ser\-ed both in its construction and operation.
In sealing off the upper end of the capillan,-
aa (Fig. 34), care should be taken to have the
capillar^' bore terminate in as blunt a surface
as possible, so as to ensure a fair degree of
accuracy in reading the ver%- lowest pressures.
The rubber tubing connecting the reser-
voir B and the tube T should be thoroughly
cleaned and dried before use to get rid of
any loose particles and also to eliminate as
much as possible the injurious action of the
sulphur present in the rubber. Only the
cleanest mercun.' should be used and all
glass parts of the gauge should be dried thor-
oughly before filling with mercun,-. A new
McLeod gauge will be found to give very-
erratic results at the beginning until all the
condensible vapors adhering to the walls
have been removed by gentle heating with
simultaneous exhaustion.
For extremely sensitive gauges, where the
volume V is large, the mass of mercur\- to be
raised and lowered is so great that the design
shown in Fig. 34 becomes impracticable.
In these cases, the reser\-oir T may be re-
placed by a wide bore glass tube with snugly
fitting glass plunger. Where a rough vacuiun
line is available, the top of the reser\-oir can
be closed by a rubber stopper through which
passes a two-way stopcock; one way being
connected to the rough vacuum, and the other
to the atmosphere. The mercur\- in T can
then be raised or lowered by opening the
stopcock to the atmosphere or to the rough
vacuum respectively."
In order to a\'oid the error which arises
when reading the ver>' small volume of the
capillan.- at the upper end, it is often prefer-
able to compress the gas in the cap!llar>- aa
to a definite volume and then observe the height
» An excellent description of the construction of such a
McLeod gauge, sensitive to pressures as low as 10""* mm., is given
by Gaede in the article referred to in footnote (">. He used a
capillary tube 0.3.') mm. in diameter, while the volume of the
bulb was about 1 liter.
" Zeits. f. Instkunde. Si. 97 (1914).
>' Ber. d. deutsch. chem. Ges. 58, 4149 (1905).
« Zeits. f. Angew. Chem. 19. 755 (1906).
/; of the mercun,- in the capillar},' bb above
this level. Under these conditions, since
it is e'V'ident that h is directly proportional to
the pressure to be measured. The value of h
may then be obser\-ed very- accurately by
means of a cathetometer. This method of
using the McLeod gauge is, however, not as
sensitive at low pressures as is the preceding
method described. Again, in some cases,
where the range of pressures to be measured
is fairly large, the single capillan,- aa may
be replaced by two or more capillaries of
gradually increasing bore, the coarser bore
being sealed onto the bulb I' and the finer
are on top of this. The serious objection to
this construction, however, is the inaccuracy
of the measurements at the junction between
the two capillaries.
While the construction shown in Fig. 34 is
the usual form of McLeod gauge used in
exhaust work, a niunber of modifications
have been suggested which are more con-
venient in special cases. An interesting con-
struction is that designed by H. J. Reiff"
and shown in Figs. 3.5 and 36. The advan-
tages of this form are its compactness and
avoidance of the use of rubber tubing which,
as Reiff points out. sooner or later causes
the mercun.- to get dirty. The gauge is
mounted on a board which can be turned 90
deg. about the axis at C (Fig. 36). The
system in which the pressure is to be measured
is connected at R by rubber tubing. In the
position sho\\-n in Fig. 3.5, the reser\-oir G and
tube r' are filled with mercun.- up to the
stopcock H. To measure the pressure, the
board is turned into a vertical position and
H opened until the mercun.- rises in M
to the desired level. The bulb (7* prevents
any mercury- from overflowing into the tube
r-. After the measurement is completed the
board is again turned into the position shown
in Fig. 35 and the mercun,- returned into the
resen-oir </.
In the same paper ReifF has also described
a further modification of this construction
in which the readings are directly proportional
to the pressure. Other forms of the McLeod
gauge have been described bv A. Wohl and
M. S. Losantisch." and L. Ubbeholde.'*
MECHANICAL MANOMETERS
A number of attempts have been made to
construct low-pressure manometers indicat-
ing the mechanical deformation suffered hy
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
735
a surface under pressure. At ordinary pres-
sures this principle has been utilized in the
construction of the Bourdon Spiral. Laden-
burg and Lehmann/' and subsequently M. G.
Johnson and D. Mcintosh''*, have de-
scribed a low-pressure gauge consisting of
a flat tapered glass tube bent in the form of
a spiral. The walls are usually very thin, so
that the device may be sensitive to small
pressure differences. A glass mirror is
attached to the end of the spiral and the
latter is sealed into another chamber in which
the pressure may be varied. The system
whose pressure is to be measured is connected
to the spiral. In using the instrument, the
pressure outside the spiral is varied until it
Scheele and Heuse's membrane manom-
eter-" consists of a very shallow cylindrical
glass box separated into two compartments
(parallel to the flat sides) by a thin copper
membrane. One compartment is connected
to the system, while the other is connected
directly to a high-vacuum pump. The
deformation of the membrane, due to the
slight difference in pressure on the two
sides, is then measured by noting the
number of interference rings produced by
the pressure of the membrane against a
glass plate.
The instrument was found capable of
measuring pressures down to about 0.0001
mm. of mercury, but difficulties were en-
Fig. 35
Short Form of McLeod Gauge
Fig. 36
is equal to that in the spiral, as indicated by
the mirror, and the pressure outside is then
measured by an ordinary mercury manom-
eter. The device has been used for measur-
ing the pressure of corrosive gases like chlorine
and ammonium chloride vapor. A similar
type of manometer has also been used very
recently by C. G. Jackson for measuring the
dissociation pressure of cupric bromide.
These gauges are, however, not sensitive to
pressures below about 100 bars.''
" Verh. d. deutsch. Phys. Ges. S. 20 (1906).
18 J. Am. Chem. Soc. SI. 1138 (1909) ; Zeits. f. Physikal. Chem.
ei. 457 (1908).
1^ For full details regarding this type of manometer, refer to
K. Jellinek. Lehrb. d. Physikal. Chem. I, 1, p. 638, also to the
references given in footnotes (i^) and ("*).
» Zeits. f. Instrk. £9. 14 (1903).
Ber. d. deutsch. Phys. Ges, 1909, p. 1.
21 In this connection the author has quoted to a large extent
from his paper on the "Theory and Use of the Molecular (jauge,"
Phys. Rev. S. 212 (1913).
countered in using it because of the con-
tinual gas evolution from the walls of the
device.
VISCOSITY MANOMETERS
Theory'''^
If a plane is moving in a given direction
with velocity u relatively to another plane
situated parallel to it at a distance d, there
is exerted on the latter a dragging action
whose magnitude may be calculated from
considerations based on the kinetic theory of
gases.
At comparatively higher pressures where
the mean free path of the gas molecules is
considerably smaller than the distance be-
tween the plates, the rate of transference of
736 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 9
momentum across unit area is given by the
equation
(23)
5 = ^
where tj denotes the coefficient of viscosity."
According to the kinetic theorv' of gases,_
this coefficient ought to be independent of
pressure. The confirmation of this deduction
by Clerk Maxwell and others, over a large
range of pressures, has been justly regarded
as one of the most striking argiunents for the
validity of the assumptions on which the
kinetic theory is based.
It was found, however, by Kundt and
Warburg^' that at verv' low pressures, where
the mean free path of the molecule becomes
of the same order of magnitude as the dis-
tance between a moving and a stationary
surface placed in the gas, there is distinct
evidence of a slipping of gas molecules over
the planes, so that the apparent viscosity is
decreased. As the pressure is lowered the
amount of this slip is found to increase and at
very low pressures it varies inversely as the
pressure.^*
Denoting the coefficient of slip by 5, it can
be shown that the amount of momentum trans-
ferred per unit area from the mo\ing surface
to that at rest is
B= '^'^ (24)
d + 25 ^
Thus, owing to slip, there is an apparent
increase in the thickness of the gas layer
between the two surfaces, which amounts to
6 for each surface.
As has already been stated, Kundt and
Warburg found that at ven,- low pressures 6
is inversely jjroportional to the pressure, and
approximately of the same order of magni-
tude as the mean free path L, of the gas
molecules at the corresponding pressures.
More generallv, we can write
5 = aL
where a is a constant. It is evident that at
very low pressures, where d is small com-
pared to L, equation (24) reduces to
7JM
5 =
•2aL
»■ See Part I of this series of articles, June, 1920, p. 498. Poynt-
ing and Thomson Properties of Matter, pp. 218-220 give an
exceptionally clear explanation of the physical significance of »i
from the point of view of the kinetic theory.
" Pogg. Ann. lua. 340 (18".')).
2* ' 'The diminution of the viscosity at very low pressures is well
shown by an incandescent lamp with a broken filament. If this
be shaken while the lamp is exhausted it will be a long time
before the oscillations die away, if. however, air is admitted into
the lamp, through a crack made with a file, the oscillations when
started die away almost immediately." Poynting and Thom-
son, loc. cit. . , , .„■
« See equation (9). Part I. p. 498. and equation (41 p. 497.
» See Part I of this series of articles June, 1920. pp. 499-501,
and Part 11. July. 1920.
Since-
« ^ = o:.np^i
SM
L
^\1FRT
„ 2X0.31
B = p
\ M
''\2VRf
(25a)
That is, with a given gas at constant tem-
perature, the rate of transference of momen-
tum is directly proportional to the velocity
of the moving surface and also to the pressure.
It follows from this that given the value of a,
it would be possible from measurements on
the mutual effect of a moving surface and one
at rest, to measure the pressure of the gas.
The exact interpretation of a from the
kinetic theory point of view has, however.
proved to be rather a difficult matter. While
the further discussion of this subject must
be deferred for another section of this series
it may be obser\'ed that a relation of the
same form as (25a) may also be deduced by
considerations similar to those used by Knud-
sen in connection with his investigations on
the laws of molecular flow.-'
According to the kinetic theory-, the mass
of gas striking unit area of a surface per unit
time is equal to
}4pn=pJ^
V 27r
M
RT
where pfi = density
= average (arithmetical) velocity.
Assuming, as Knudsen does, that all the
molecules striking a surface are reflected in
directions which are absolutely independent
of the directions of incidence and that these
reflected molecules follow Maxwell's dis-
tribution law, it follows that the rate of trans-
ference of momentum per unit area from a
surface moving with velocity u is
B
\2irRT
(25b)
This relation will, of course, hold true only
at such low pressures that the molecules can
travel across the space between the two sur-
faces without suffering collisions with each
other.
It will be observed that equations (25a)
and (25b) agree in the conclusion that
at very low jjressures B is proportional to
P\ M (RT). so that we can express the rela-
tion in the general fonn
B
M RT
(25c)
where k is a constant, which may be slightly
different for different gases and i)robably
varies also with the nature of the surface.
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
737
In applying the relation between the coeffi-
cient of slip and the pressure to the construc-
tion of a gauge, two different methods have
been used. In the first of these, which we
may designate for reference as the "decre-
ment" type of gauges, a surface is set in
oscillation and the rate of decrease of the
amplitude of oscillation is taken as a measure
of the pressure. Physically, the damping may
be explained as due to the gradual equalization
of energy- between the moving surface and the
molecules of gas striking it.
In the second type of construction, a sur-
face is set in continuous rotation and the
amount of twist imparted to an adjacent
surface is used to measure the pressure.
The molecules striking the moving surface
acquire a momentum in the direction of
motion which they tend in turn to impart to
the other surface. If the latter is suspended
and free to turn about an axis which is per-
pendicular to the direction of motion of the
rotating surface, it will be twisted around
until the force due to the incident molecules
is just balanced by the torsion of the suspen-
sion. We may, therefore, designate this as the
"static " type of viscosity gauge, to emphasize
the fact that observations with this method
are taken under stationary conditions.
"Decrement" Type of Viscosity of Gauge
A gauge based on this principle was first
suggested by W. Sutherland" and subse-
quently a very careful investigation on the
same subject was carried out by J. L. Hogg.-^
The construction used by the latter, which
was essentially the same as that used by
Maxwell and Kundt and Warburg in their
determinations of the coefficient of viscosity
is shown in Fig. 37. A thin glass disc is
suspended by means of a wire between two
fixed horizontal plates N. The wire carries
a mirror which may be viewed through a
plate glass window D by means of a telescope
and scale. At the top, the wire is supported
by clamps and is connected to a soft iron
armature / which is supported by the swivel
head A'. By turning this armature by means
of an external magnet, the center disc can
be set in oscillation and the rate of decrease
of the amplitude of these oscillations is then
observed by means of the telescope pointed
at the window D.
Now let T denote the period of oscillation,
and Si and 52 two successive amplitudes of
" Phil. Mag. 43. S3 (1897).
M Proc. Am. Acad. 4^.115 (1906). and 45.3(1909). Contribu-
tions from the Jefferson Physical Lab.. 1906. N'o. 4, and 1909,
No. 4.
oscillation. Solving the differential equation
for the rate of damping of the central disc,
it can be shown that
aT
='' (26)
5i
Sl^''
r\
Fig. 37. Decrement Type of Gauge
where X is defined as the logarithmic decre-
ment. That is, the amplitude of oscillation
decreases in geometrical progression for
successive equal intervals of time. The
constant a depends upon the moment of
inertia of the vibrating disc and its dimen-
sions.
Thus X is a measure of the rate of transfer-
ence of momentum from the ^•ibrating plate
to the stationary plates.
At higher pressures, since the viscosity is
independent of pressure, the logarithmic
738 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 9
decrement has a constant value which may
be denoted by /. Denoting by M the decre-
ment due to the suspension itself, it was
shown by Sutherland that the following
relation ought to hold true:
Cr^-O-'
(27a)
where p is the pressure and C is a
constant for the particular arrange-
ment used.
The results obtained by Hogg were
found to be in satisfactory accord
with this equation down to pressures
of the order of 0.0004 mm. of mercury
in the case of hydrogen.
It is evident from the form of the
above equation that the gauge is un-
suitable for measuring very low pres-
sures (say below 0.0001 mm.) as the
value of X then becomes comparable
with that of fx, thus involving large
experimental errors. Furthermore, as
mentioned by Hogg, the construction
of the gauge and its actual manipula-
tion require extremely great care.
A verv recent contribution to the
theory of this type of viscosity meter
has also been published by P. E.
Shaw-^ He derives an equation of
the form
p = C-K (27b)
and records measurements of pres-
sures down to 0.35 by 10~^ mm. of
mercurv.
Quartz-Fibre Gauge
This method was originally sug-
gested by I. Langmuir'" for measuring
the residual gas pressure in a sealed
off incandescent lamp, and has been
used in this laboratory in a number
of investigations. It is specially use-
ful in measuring low pressures of
chemically active vapors such as
those of chlorine, iodine, and mercury
which are liable to attack metal parts.
A discussion of the theory of the
gauge and actual details as to its mani])ula-
tion have been published by F. Haber and
F. Kerschbaum''.
The construction of the gauge is shown in
Fig. 3S. It consists of a thin quartz fibre
sealed into the top of a glass ttibe. The fibre
is set in oscillation by gently tapping the glass
bulb and the rate of decrease of the amplitude
» Proc. Phys. Soc. London. S9. 171 (1917).
» J. Am. Chem. Soc. SB, 107 (1913).
" Zeits. f. Elektrochem. iO. 296 (1914).
V
Fig. 38. Quartz
Fibre Gauge
of vibration is then obser\-ed by means of a
telescope and lamp as shown in Fig. 39.
Let t denote the inter\-al of time required
for the amplitude to decrease to half value.
Then it has been shown by Haber that
p\/U = --a (28)
where p denotes the pressure, M is
the molecular weight of the gas,
and a and b are constants for the
particular quartz fibre. That is, for
any gas, the pressure varies linearly
with the reciprocal of t.
In the case where the gas to be
measured is a mixture of different
vapors, the sum of a nimiber of
terms py/^ must be taken corre-
sponding to the partial pressure of
each constituent.
The constant b in equation (28)
is proportional to the diameter
of the fibre, that is, the finer the
fibre the smaller the pressure at
which the amplitude will decay
to half value in a given time.
On the other hand, a is a func-
tion of the elastic properties of the
fibre.
It is evident from the form of
equation (28) that a b corresponds
to the value to at which the ampli-
tude would decrease to half-value
in a perfect vacuum. For calibra-
tion, it is necessary" to obtain only
two points, corresponding to the two
constants a and 6. One of these
may be determined by obser\ing the
value to in a \-er>- good vacuum,
while the other point may be ob-
tained by calibrating against a
McLeod gauge with some gas of
definite composition.
The following data are given
by Haber for a quartz fibre 7.0
cm. long and 0.013 cm. in di-
ameter. Air was used for calibra-
tion.
Pressure
in mm. Hg
0.00302
0.00494
0.00775
0.0117
0.01880
0.0260
0=0.0003
*%/.«
< (seconds)
»
0.01625
74
1.22
0.02654
46
1.23
0.0417
31
1.30
0.0630
oo
1.39
0.101
12
1.23
0.140
10
1.40
Avg. 1.28
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
739
Some measurements with air taken by Mr.
Huthsteiner in this laboratory, using a fibre
3.8 cm. long and 0.0043 cm. diameter, are
given for comparison.
Pressure in mm, Hg
t (seconds)
0.00058
0.00342
0.0080
0.0190
105
31
16
6.5
Plotting p against — gave a straight line
131
whose equation is 103 p = 0.655. For
air, .1/ = 28.96. Hence, for this particular fibre,
0.705
P\/m
t
-0.00353
Since ^0 = 200 in this case, it is evident that
this fibre could not be used for measuring
pressures below 0.0001 mm. of air It also
follows from the form of the above relation
that the heavier the gas the lower the range of
pressures over which the gauge may be used.
The optical arrangement suggested by
Haber (Fig. 39) may be varied in practice by
fastening a scale to the back of the gauge and
placing the lamp in such a position that the
light beam passes practically parallel to this
scale. The scale and tip of the quartz fibre
are then sighted by means of a cathetometer.
While Haber used tubes which are more or
less flattened on two sides, ordinary cylindri-
cally walled tubes are more convenient and
this can readily be accomplished. In view of
the simplicity of construction and relative
ease of manipulation, the quartz fibre gauge
ought to find a useful field of application in
low pressure technique, where the pressures
to be measured are not below about 0.05 bar.
Fig. 40. Molecular Gauge
Fig. 41.
Rotating Commutator Connection
for Molecular Gauge
Fig.
39. Optical Arrangement for
Quartz Fibre Gauge
almost as satisfactory. As observed by
Haber, care should be taken to tap the glass
in such a manner that the fibre vibrates in the
plane at right angles to the line of sight from
the cathetometer. With a little experience,
■2 Phys. Rev. /. 337 (1913).
" Phys. Rev. 6. 212 (1915).
Static Types of Viscosity Gauge
The molecular gauge suggested by I.
Langmuir'- represents a direct application of
equation (25c).
The construction and results obtained with
a gauge built on this principle were described
by the writer^' as follows :
"It consists of a glass bulb B (Fig. 39) in
which are contained a rotating disc A and,
740 September, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, No. 9
suspended above it, another disc C. The
disc A is made of thin aluminum and is
attached to a steel or tungsten shaft H
mounted on jewel bearings and carrying
a magnetic needle NS. Where the gauge is
to be used for measuring the pressure of
corrosive gases like chlorine, the shaft and
disc may be made of platinum. The disc C
is of very thin mica, about 0.0025 cm. thick
and 3 cm. in diameter. A small mirror M.
about 0.5 cm. square is attached to the mica
disc by a framework of thin akuninum.
This framework carries a hook with square
notch which fits into another hook similarly
shaped, so that there is no tendency for one
hook to turn on the other. The upper hook
is attached to a quartz fibre about 2 by 10~^
cm. diameter and 15 cm. long.
"The lower disc can be rotated by means
of a rotating magnetic field produced outside
the bulb. This field is most conveniently
obtained by a Gramme ring, GG, supplied
at six points with current from a commutating
device rotated by a motor (Fig. 40). In this
way the speed of the motor determines abso-
lutely the speed of the disc, and the speed of
the latter may thus be varied from a few revo-
lutions per minute up to 10,000 or more."
By applying equation (25c) it can be
shown that the angle of torque (a) on the
upper disc is given by the equation
:^ — tI/'WxItt^ (29)
where
of oscillation
of
t = natural period
mica disc,
^=moment of inertia of disc,
r = radius of rotating disc,
and CO = angular velocity of rotating disc.
Hence, for any one gauge, the torque on
the upper disc is proportional to the product
of the speed of rotation of the aluminum
" Ann. d. Physik. J,0. 971 (191.3).
disc and the quantity p\/ M/{RT). The
sensitivity of the gauge can thus be increased
by increasing the speed of rotation; also by
illiuninating the mirror and using a similar
arrangement to that used for galvanometers,
it is possible to use the gauge to measure pres-
sures of the order 10 "'to 10~* bar.
The gauge actually used for measuring ver}'
low pressures showed a deflection of 1100
mm. per bar of air, at 1000 r.p.m., with the
scale 50 cm. from the mirror. Up to a pres-
sure at which the mean free path of the gas
molecules becomes comparable with the
distance between the two discs, the deflec-
tions, at constant speed of rotation, were
found to be proportional to the pressure
as observed by a ^IcLeod gauge.
At extremely low pressures (below oXlO"*
bar) the indications of the gauge were found
to be inaccurate because of two sources of
error. First, the rotation of the magnetic
field produced by the Gramme ring tends to
induce eddy currents in the metal frame work
used to hold the mirror; and second, there is a
tendency for the ujjper disc to start swinging
especially at very high speeds of rotation of
the aluminum disc. As the damping at low
pressures is very feeble, it is very difficult to
stop this oscillation when once started.
Working independently of Langmuir, and
about the same time A. Timiriazefl" also
suggested the application of equation (25)
to the construction of a low-pressure gauge.
As he was primarily interested in determining
the laws of slip for different gases his actual
design is not suitable for a very sensitive
gauge. Instead of using a rotating disc with
a stationary disc situated svTnmetrically above
it, Timiriazeff used a rotating cylinder with
a stationary- cylinder ])laced s>Tnmetrically
inside it and suspended by a phosphor
bronze wire.
(7*0 he CoHlinued)
741
The Cooper Hewitt Mercury Vapor Lamp
PART I. THEORY AND OPERATION
By L. J. BuTTOLPH
Engineering Department, Cooper Hewitt Electric Company
This article is the first of a series of three on the theory and uses of the Cooper Hewitt Mercury Vapor
Lamps. The second article, which will appear in the October issue, will illustrate the advantages of these
lamps for industrial illumination. The third article will be descriptive of the Cooper Hewitt Quartz Lamp
and its characteristics. — Editor.
GENERAL PRINCIPLES OF THE
MERCURY ARC
The Cooper Hewitt lamp consists of a tube
of glass or of quartz containing mercury,
mercury vapor and wires sealed into the ends
of the tube to conduct electricity to and from
the current carrying vapor. In the manu-
facturing process all foreign gases are re-
moved and the tube closed vacuum tight. In
operation there is a direct current arc from
the cathode electrode of mercury to an anode
electrode of iron or of tungsten' (Fig. I).
^ 3^
Fig. 1.
Upper — Direct-current Cooper Hewitt Lamp
Lower — Alternating-current Cooper Hewitt Lamp
The wattage of a lamp of a given size is
limited by the heat resisting quality of the
glass used. Two types of lamps have there-
fore been developed, one of glass to operate
at relatively low temperatures, and one of
fused quartz to operate at relatively high tem-
peratures. The normal volt-ampere char-
acteristic of a lamp is determined primarily
as a very complex function of the mercury
vapor pressure and density and of the length
and cross section of the tube. With the tube
dimensions fixed the vapor pressure is deter-
mined largely by the mini-
mum temperature within
the tube, while the vapor
density varies according to
^"^^^^-^ the heat distribtition, be-
•^if^SeM^M^ ing in general a minimum
along the central axis of
the tube. In standard in-
dustrial units the normal
volt-amperage is then
finally detennined through
the tube temperature by a
condensing chamber in the
form of a bulb on the cath-
ode end of the lamp tube.
A condition of complete
equilibrium is reached when
the light and heat radiated
and conducted from the
tube equals the electrical
energy input. The effect
on the tube voltage and
current of the temperature
rise during starting is shown
in Fig. 2, where they are
plotted as functions of time.
In the actual design of a
lamp these several vari-
ables are so balanced as to
give at once that critical
vapor density at which the
light-giving efficiency is
742 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 9
120
110
100
90
80
70
60
90
«0
30
20
10
0
V
12
11
10
9
7
6
S
4
3
2
1
0
A
1
_
_
_
GE
r
t
^
-
-UC_ruaE_^
tlAgg
I
^ ^ / "-If,
h
^
i'SaT____
~
~
..^ — ,
\
— 1
^
1
c
r.
JRPENT
\
■n
nC LAMP cu
-\
=="
•e ID w 40 to ■> ^,!? is *°
«=»•• 1 2 « • -"-!- 10 11
MINUTES ' . .
Fig. 2. Volt-ampere Starting Characteristic of Standard Lamps
greatest and a volt-ampere characteristic
allowing maximmn current regulation with a
minimvun sacrifice of wattage for that purpose.
Modem theorA- gives a strikingly graphic
picture of the electrical condition in the arc
column of the Cooper Hewitt lamp. Accord-
ing to it the tube is filled, during operation,
with mercury molecules, mercury ions, and
electrons. The ions are molecules which
have gained or lost one or more electrons or
unit negative charges of electricity, thereby
being left charged either negatively or posi-
tively as the case may be. These molecules,
ions, and electrons move with various char-
acteristic velocities and in individual direc-
tions determined by their collisions with their
fellows according to the well known kinetic
molecular theory of gases. This commotion
characteristic of all gas molecules is further
complicated by the fact that a constant dif-
ference of potential of about one and one
third volts per inch of arc length is maintained
on the electrodes located in the ends of the
tube, and that because of the heat of the
cathode and the impact of the electrons, ions
and molecules on each other and on the elec-
trodes more electrons and ions arc produced
than are usually needed to carry the current.
The effect of the electromotive force on this
gas column is to produce an arc current which
may be described as a continuous drift of
electrons from the cathode to the anode and
a relatively much slower movement of posi-
tive ions towards the cathode. The excess of
ions and electrons produces the effect of a
partial short circuit with a continuous ten-
dency to become a more com-
plete short circuit. The resixlt
is a periodic increase of current
and fall of potential of a fre-
quency determined by the ca-
pacitative and inductive react-
ance of the arc column and of
the supply circuit.
For transient variations of the
current this inverse variation of
voltage is characteristic of the
mercun.- arc, a cathode phe-
nomenon apparently, for the
whole range of practical current
values and arc temperatures.
It is most pronounced for low
currents, but decreases rapidly
with increase of normal cur-
Fig. 3. Volt-ampen "St^tiooary" Charaeterittlc*
Regulation
THE COOPER HEWITT MERCURY VAPOR LAMP
743
rent. For slow changes of the current this
same volt-ampere relationship is character-
istic up to a certain critical current value.
With further increase of current from this
point the tube voltage passes through a
minimum and then rises rapidly as shown in
Fig. 3. For maximum light efficiency, the
Cooper Hewitt lamp is operated at the point
of minimum tube voltage, where, if unre-
stricted, the arc current will fluctuate over a
wide range on constant voltage. In order to
operate this unstable and essentially constant
current device on supposedly constant volt-
age power lines two forms of regulation are
necessary. The current is steadied by an
inductance coil, connected in series with the
arc and as directly as possible to the cathode
so as to oppose every transient action of the
current by an instantaneous induced reaction.
The falling voltage characteristic of the arc
as well as the voltage variations of the line
are compensated by an ohmic resistance in
series with the inductance coil and the arc as
shown in the wiring diagram, Fig. 4. This
resistance is so chosen that, for normal opera-
tion, with any increase of current the decrease
in arc voltage will be less than the mcrease
in resistance potential. In Fig. 3, curve
STABTVN<« BE3I6TAMCK
Fig. 4. Wiring Diagrams and the Detail of Direct -current Lamp Auxiliary
744 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 9
D 0 E V F is the volt-ampere characteristic
of a Cooper Hewitt arc showing inherent
stability above and instability below four
amperes. Line P U O R represents the line
voltage minus the resistance voltage for vari-
ous currents, or in other words, the voltage
available at any time for arc operation.
Point U is, therefore, one of arc instability
since any current increase is accelerated by
the resulting excess arc voltage. On the
other hand, point 0 is one of stability, a cur-
rent decrease being opposed by an excess of
arc voltage and an increase being limited
by the available arc voltage. In this case
the regulating series resistance is eleven ohms.
Supply
Inductance Cot.
wvTm I.;
ducfance Coil
^ C ::j_
Cover Fastener
Type "F"Tube-
Startlng Bdna
Curve C H 0" M, the volt-ampere character-
istic cur\^e of the whole lighting unit, is the
continuous sum of the resistance potentials
B O" R and the arc potentials. Point H there-
fore represents the minimum maintenance
current and voltage of the outfit for the
amount of regulation used. The regulation,
which is defined as the per cent fluctuation of
normal \'oltage producing a current change
from one half ampere below to one half am-
pere above normal current, is in this case S
per cent. In the Cooper Hewitt industrial
units the series resistance is adjustable to pro-
vide for operation on various and on var}.-ing
voltages.
Starting the Arc
To Start the Cooper Hewitt lamp
it is only necessary to start and main-
tain the formation of electrons in a
so-called '"hot spot" on the surface
of the mercun,- cathode. Collisions
with mercury- molecules immediately
result in the formation of more elec-
trons and ions than are needed to
form a current, with the results de-
tailed above. The temperature of
this spot, several thousand degrees
at least, may be accounted for by
the very small cross section of the
spot and the fact that some eighteen
watts of energy are converted into
heat in this small area of liquid
WIRING DIAGRAM EC AUXILIARY
POSITIVE
aeesTAMCE.
SHIFTER RESISTANCE
AUTO -TRANSFORMER SHIFTER 5TA.R-nM« BAND
suppi.y s/M/A/c porrs
POS/r/i^£ LfADf
POS/T/V£ Rer/s-TA/^cf
AurO TRANffORMER
/A/OacrANOE CO/IF
SHirr£R/iBS>STAA/C£ \
Fig. 5. Wiring DioKroms and the Dctoit of AUcrnutinR-currrnt Ljiinp Auxiliary
THE COOPER HEWITT MERCURY VAPOR LAMP
745
vapor inter-surface, the cathode drop in
potential being about 5.3 volts. There is
a difference of opinion as to whether ioniza-
tion at the cathode results from the direct
emission of electrons from mercury vapor
heated far above its boiling point or whether
it re.sults from the impact of positive ions
upon hot molecules. In either case the condi-
tion is easily produced by bringing the mer-
cury cathode into contact with the anode and
then breaking the circuit thus formed, as with
the ordinary carbon arc. This tilting method
is now used to start the relatively small
Cooper Hewitt quartz mercury lamps. An
alternative automatic starting method stand-
ard for the glass lamps consists in short
circuiting a small current through the arc
regulating inductance in series with the arc.
This current is broken by a. mercury switch
or "shifter" magnetically operated by the
inductance coil itself. The resulting induced
high potential is sufficient to start a localized
cathode discharge and the arc is formed. A
metallic coating placed on the outside of the
cathode end of the tube opposite the mercury
cathode and connected to the positive side
of the supply circuit serves to increase the
electrostatic capacity of the cathode and
hence to give a greater current density to the
induced high potential discharge when it is
localized to form an arc. See Fig. 4 for the
arrangement of the circuits.
The effectiveness of the "shifter" or mer-
cury switch as a quick acting cut-out switch
is worthy of note. It is itself a small glass
chamber evacuated except for mercury and
mercury vapor and is supplied with leading-in
wires for electrical connection. It is made
in the same manner as the regular lamps and
is itself essentially a small mercury vapor
arc. It is so mounted as to be easily rotated
by an armature actuated by the magnetic
field of the inductance coils. Its operation,
in detail, is therefore as follows: At the
moment the lamp is connected to its source
of electric supply a current is short circuited
past the lamp tube, through the arc regulat-
ing inductance coils and resistance, through
an additional shifter or starting resistance and
through the shifter itself (see Fig. 4). The
lightly mounted shifter rotates, the mercurv
pool connecting the two leading-in wires is
widely separated, and at the moment of
separation an induced electromotive force
reaching a verv high peak voltage appears
on the terminals of the shifter and therefore
on the terminals of the arc tube. This voltage
is sufficient to start the arc as outlined before.
but it will not form an arc in the shifter since
the total resistance of the shifter circuit is
such as to keep the shifter starting current
well below a minimimi arc maintenance value.
Since the inductance coils used have rela-
tively low self inductance the high induced
voltages obtained result from the extremely
rapid current decrease when the circuit is
broken in the shifter. The effectiveness of
this mercury-vacuum switch for this purpose
as compared with oil immersed or quick act-
ing circuit breakers is accounted for by the
uniquely rapid rate of deionization of cold
mercury vapor and the heat dissipating prop-
erty of volatile contact points. These con-
siderations afford an explanation of the ob-
served fact that the colder the "shifter"
the more effective is its operation.
Exhausting the Tubes
In the manufacture of Cooper Hewitt lamps
two features are of special interest, the method
of evacuation and the treating of the metal
anodes. When ready for evacuation the tube
containing about twice its final amount of
mercury is hung vertically in an upright gas
furnace or hot air oven and connected by a
tube at the upper end near the anode through
a mercury trap to an ordinary vacutun. As
the tube heats up to the boiling point of
mercury the relatively heavy mercury vapor
rises in it, displacing the remaining traces
of foreign gases and water vapor. This proc-
ess is continued until, with the mercury in
the tubes boiling vigorously and with the
glass walls of the tube nearly at their melting
temperature, the tube is acting as a highly
efficient mercury diffusion pump to produce
its own high vacutun with reference to all vol-
atile stibstances other than the mercury itself.
When this process has resulted in the distilla-
tionf rom the tube of ameasured amount of mer-
cury the process is stopped and the bulb sealed
off at the tttbe. Thereafter the"vacutmi" of
the tube is determined by the vapor pressure
of mercun,' at any gi\'en tube temperature.
To free the metal electrodes from occluded
gases they are heated to a white hot temperature
during the ptunping process. This treating is
done by operating the lamp on an alternating
current at some 4000 to 6000 volts.
The heat of the cathode hot spot is highly
localized so that in a glass Cooper Hewitt lamp
the arc column temperature varies from some
500 deg. C. in the center to about 125 deg.
C. at the surface of the ttibe. Therefore the
vapor pressure seldom, rises to over one milli-
meter. There is a potential drop at the anode
746 September, 192Q
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 9
of about 5.7 volts and the anode is so designed
that its temperature is normally about 350
deg C.
The Quartz Lamp
The Cooper Hewitt quartz lamp differs
from the glass lamp as follows ; The arc tem-
perature is much higher, var^'ing from some
1400 deg. C. in the center to about 450 deg.
C. at the surface of the tube. The vapor
pressure is therefore an atmosphere and over.
The potential drop is about 25 volts per inch.
To withstand the higher temperature a
tungsten anode is used, which is white hot
in normal operation. The quartz burner has
no condensing chamber, direct radiation and
the construction of the mercury' filled cathode
providing the required cooling. Fig. 6 shows
a cross section of a 220-volt quartz burner in
operation. The cathode surface is relatively
smaller than in the glass lamp to restrict the
fluctuations of the cathode spot. The arc
as is apparent from the oscillogaph cur\-es
of Fig. 7. The mercur}* arc is essentially
a unidirectional conductor because its main-
tenance is dependent upon the existence and
peculiar properties of the so-called cathode
"hot-spot." This can be formed and m.ain-
tained at a low voltage, 5.3 volts at ordinary'
temperatures onh^ on mercur\' and certain of
its alloys, and once formed is itself only main-
tained by continuous operation; and even
with a mercury cathode this discharge of mer-
cury vapor and electrons can only be started
by drawing an arc by contact or by a potential
of several thousand volts. These peculiarities
of the arc are utilized in the Cooper Hewitt
lamp as follows: The cathode of the lamp is
connected through inductance to the middle
point of the secondary of an auto transformer
(see Fig. 5), while, the anodes are connected
to the terminals. Therefore the cathode is
continuously negative with respect to one or
the other anode during operation. The arc
Fig. 6. Cooper Hewitt Quartz Lamp
Stream is further steadied by deflection from
the axis of the tube to the horizontal surface
of the mercury. When cold the mercur\' flows
down out of the cathode chamber and a slight
tilting of the burner permits starting by the
contact of the electrodes. The quartz mer-
cury lamp requires the same regulation as the
glass lamp, but a smaller per cent of energy is
required for the ])urpose. The quartz arc col-
umn appears to be constricted along the center
of the tube in contrast with the unifonn ap-
pearance of the arc in glass.
The Alternating-current Lamp
The Cooper Hewitt alternating-current
lamp is a highly specialized form of Cooper
Hewitt single-phase constant voltage alter-
nating-current rectifier. As shown in Fig. 1 ,
the construction is identical with that of the
direct current lam]) cxcejit that there are two
anode electrodes. The current in the lamp
tube is a pulsating direct current of a fre-
quency twice that of the alternating current,
is started by an induced voltage, the mercury
electrode becoming the cathode for the
reasons indicated above. Thereafter the
cathode functions as continuously negative
with respect to one or thef other anode. Thus
the two halves of the transformer secondary
and the anodes connected to them function
alternately, the arc shifting from one to the
other anode with the alternations of the
sup])ly current. The series inductance, in
addition to steadying the current for transient
variations, has the more important function
of sustaining the cathode spot and the arc
current during the time of zero voltage, or
in other words, of causing the current to a
given anode during a half cycle to lag its
voltage and overlaj) the current to the other
anode to such an extent that the resultant arc
current never falls below the minimimi main-
tenance value. Although the potential be-
tween the two anodes is obviously always
double that between the active anode and the
cathode, there is little or no leakage between
THE COOPER HEWITT MERCURY VAPOR LAMP
747
them. For an alternating current of a given
frequency the minimum sustaining inductance
is definitely detennined and this also fixes the
minimum practical power factor of the outfit.
Regulation such as that provided by series
resistance in the case of the direct-current
lamp could obviously be provided by in-
ductance or choke coils instead of ohmic re-
sistance at a slight gain in efficiency but with
th disadvantage of low power factor, viz.,
50 p r cent. In the Cooper Hewitt alternat-
ing-current lamp, ohmic resistance is placed
in the anode circuits. Fig. 5, and to secure a
aximum of regulating effect on fluctuating
voltage a special iron wire resistance unit is
used. It is so designed that because of the
high temperature co-efficient of resistance of
iron the voltage absorbed by the resistance
varies more rapidly than the current. The
volt-ampere characteristic of a certain iron
wire resistance is as indicated by the curve
BO"N" in contrast with a nearly straight
line for an ordinary resistance, Fig. 3 ; and the
effect of using such an iron wire resistance
with a direct-current lamp might be as indi-
cated by the dotted lines, / A''. In actual
practice the series inductance provides part
of the regulation, absorbing an appreciable
amount of the transformer voltage as shown
in Fig. 7, K, and helping to produce a power
factor of 85 per cent.
Fig. 7 shows some of the relationships be-
tween voltage, current and time in various
parts of a standard alternating current lamp.
A, the primary voltage, is approximately a
sine function as usual, but the current wave
form. B, is distorted by the reactance and the
arc characteristic of the secondary circuit.
D is the e.m.f. between the arc cathode and
the active anode, while H is the e.m.f. during
the succeeding half cycle when the other anode
becomes the active one. C is the anode current
corresponding to voltage D, while G is the
current in the other anode during the succeed-
ing half cycle. E is the voltage drop in the
anode resistance units during their current
carrying intervals. / is the superimposed
anode currents, while / is the resulting rec-
tified arc current. L shows the superimposed
arc voltages and their induced overlap which
causes the anode currents to overlap as in /.
Curve K showing the voltage drop in the direct-
current reactance coils is of unusual interest.
The inductive reactance of the arc circuit and
the arc characteristics cause the pulsating arc
current to rise more slowly than it decreases.
The point of anode current overlap also comes
during the time of arc current decrease. The
bearing of these facts upon the wave form of
the direct-current reactance voltage is evident
from / and A'. Thus points of zero voltage
correspond to zero time rate of current
change, maxima and minima current or to
momentarilv constant current; while the
Fig. 7. Oscillograph Record of Alternating-current
Lamp Characteristics
points of maximum voltage come when the
time rate of current change is a maximum.
The effect of the overlap discontinuities of
the arc current on the corresponding induced
voltage maximum is evident. During the
period of current overlay, current flows to
each anode and there is during that time no
potential difference between them, as shown
lay a prolonged interval of zero voltage on the
approximate sine curve of the voltage be-
tween the two anodes. The energy repre-
sented by this variation from the full sine
curve form of the transformer secondary'
e.m.f. is momentarily absorbed in the common
coils of the transformer which are constructed
for high self inductance against each other.
Ask evident from B and _/, Fig. 7, the tube
current fluctuates over a much smaller range
than does the usual alternating current.
This fact and the lower intrinsic brilliancv
748 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 9
account for the success of the lamp for high
intensity illumination on alternating currents
of frequencies as low as 25 cycles. On the
other hand, alternating-current lamps are
built for operation on frequencies as high as
133 cycles by modifications in the auto-
transformer design.
4
Fig. 8. Candle-power-life Characteristics
While the characteristics of individual
lamps var}^ over a rather wide range, ob-
ser\'ations for years and on lamps operating
under all types of adverse industrial condi-
tions give the basis for a candle-power life
of the nature shown in Fig. S. The regenera-
tive nature of the arc material gives a lamp
of theoretically indefinite life, but manu-
facturing difficulties in the line of impure
materials and contamination during the glass-
blowing operation seriously complicate the
situation. That improved manufacturing
methods will produce lamps of even better
candle-power life is inevitable.
Luminescence Color Sensations, Visibility and Visual
Acuity
The light of the Cooper Hewitt lamp may
be thought of as produced by electrical
forces acting directly upon the vapor par-
ticles in the arc stream. Specifically the
phenomenon is thought to be connected with
the ionization and deionization of the merctm-
molecules as outlined above. The result is a
relatively cold light since the temperature of
the luminescent vapor of a Cooper Hewitt
lamp is from 200 deg. C. to 500 deg. C, while
that of the filament incandescent lamp is
some 2800 deg. C.
The intensity of the light from any artificial
source varies for the difTerent wave lengths,
being in general greatest for the long wave
lengths in the infra-red, as shown on the insert
in Fig. 9. These intensities plotted against a
wave length scale form a relative spectral
w>«cuK«m
Fig. 9. Energy Distribution in Varioui llluminants
THE COOPER HEWITT MERCURY VAPOR LAMP
749
ULTRA-VIOLET,
distribution curve for any given light source.
With increase in the temperature of the
source this maximum in the infra-red moves
towards the visible part of the spectrum.
In so far as this change is not according to
Wein's displacement law for a perfect radiator
the radiation is said to be
selective, favorably so if
producing greater visi-
bility.
For the discontinuous
spectrtmi of luminescent
light the energy distribu-
tion may be located by
lines and bands along the
wave length scale and the
corresponding intensity by
their height, as for mer-
cury in Figs. 9 and 10.
The spectral distribution
of pure luminescence is
completely selective and
has not as yet been shown
to be a function of tem-
perature.
Light of each distinct wave length produces
its own characteristic effect of visibility, color,
visual acuity, photographic effect and psycho-
logic reaction. All these effects differ in
quality and intensity with the nature of the
light waves. Some of these complicated re-
lationships may be shown graphically by
plotting the relative intensities of these effects
against a wave length scale, as in Figs. 9 and
10. These effects also vary with different
eyes, photographic plates, and nervous tem-
peraments. The human visibility curve repre-
sents the average of a large number of eyes
studied by the Bureau of Standards. The
photographic sensitivity curve represents ap-
proximately the effect of white light on an
ordinary photographic plate.
Visible light of any given wave length
produces the sensation of a single color —
monochromatic light. Light of all wave
lengths and uniform intensity utilized in the
proportions indicated by the visibility curve
produces the sensation of white light. Color
or white light produced by any other than
these natural means is described as subjective.
White light may be produced by the proper
mixture of a series of complementary hues,
such as orange with blue, yellow with blue-
violet, or yellow -green with violet-purple.
The whiteness of the Cooper Hewitt light is
due to the combination of the nearly comple-
mentary hues of the yellow-green lines with
the blue and violet lines. The difference
between such a subjective white and true
white light is only apparent when examining
objects of colors other than those making up
the former, since colored objects have their
color by virtue of the colored light they are
able to reflect. One method of studying
Fig. 10. Color Sensibility Curves
light considers it as made up of combinations
of three primary colors, red, green and blue.
Ives has found that on the basis that white
light is one third each of red, green and blue
the mercury arc light gives the effect of being
20 per cent red, 30 per cent green and 41 per
cent blue. Green and red produce the sensa-
tion of yellow; therefore the mercury arc
light may be said to be 59 per cent yellow
and 41 per cent blue, there being an excess of
9 per cent green and 12 per cent of blue light
more than needed to produce the sensation
of pure white.
Analyzed in tenns of hue and saturation
the light of the Cooper Hewitt lamp may be
described as apparently of dominant hue
0.49 or blue with an admixture of 70 per cent
of white light. Other lights analyzed on the
same basis are :
PER
CEKT
White
Hue
Sunlight
100
0
Cooper Hewitt light
70
.490m
Average clear sky
60
.472
Mazda C
53
.584
Carbon glow lamp
3.8 w.p.c.
25
.592
Neon tube
6
.605
Transparent and solid objects are seen as
colored only when they select and absorb
750 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 9
from the light illuminating them all but some
characteristic color or colors which they
either transmit or reflect. Therefore any
change in the color of an illuminant by means
of colored glass globes or reflectors involves a
decrease in luminosity since the color is
produced by a process of subtraction from
the original light. When the Cooper Hewitt
light is produced in a glass tube of true spec-
tral red color the glass absorbs nearly all the
light and transmits little or none since there
is no objective red in the Cooper Hewitt
light. Similarly dull dark red objects may
appear nearly black because of maximum
absorption and minimum reflection.
As has already been said, the most unique
property of the Cooper Hewitt light is that
CS" 15° 25' 35'
Fig. 11. Candle-power Distribution About a
Bare Lamp
while it produces the sensation of white light
the independent investigations of Luckiesh
and Bell show that it is essentially a mono-
chromatic light giving a visual acuity some
50 per cent higher than white light. It is of
interest to note also that for equal illumina-
tion by monochromatic lights of various
colors visual acuity is a maximum for yellow
light of wave length 58 microns, which is also
nearly the color of maximvim visibility
(Fig. {)). High monochromatic visual acuity
and a white light containing a full range of
spectral colors are mutually exclusive.
While it is claimed by some that subjective
white light and colors should not be compared
with ordinary white light and the spectral
colors, yet there is no basis for defining the
difference, which appears to be one of degree
only. In fact, according to the most widely
accepted theorx' of color vision, all colors are
subjective in the sense that we never actually
see the true primary colors which are them-
selves excited not b\' narrow ranges of wave
lengths of light but in var\-ing degree by wide
ranges of light vibrations. Koenig's hue sen-
sation curves for true primary red, green and
blue are analogous to the visibility curve for
the hirman eye (Fig. 10). Therefore the yel-
low Cooper Hewitt light excites moderately
the sensations of primar\' red and green, and
feebly the blue, producing the subjective
sensation of yellow. The green Cooper
Hewitt light excites strongly the primarv"
green, moderately the primar%' red, and
feebly the blue, producing the subjective
Fig. 12. Representative Light Distribution
with a Standard Reflector
sensation of green. The blue light excites
strongly the primary blue and but slightly
the other two, producing a subjective sen-
sation of blue. The human eye apparently
integrates these three jirimary color sensations
as white — light visibility. The data for the
insert in Fig. 10 were obtained by Ives by
methods based on such a three color theon*-.
Photometery
The physical evaluation of the Cooper
Hewitt light has been a perj^lexing problem
for years. Added to the well known difficul-
ties of heterochromatic photometn*- is the
questionable process of comparing a light of
discontinuous spectrum with a standard light
of continuous spectrum. As yet direct com-
parisons with and without color corrective
screens have failed to give thoroughly con-
sistent results, nor does the flicker photo-
THE COOPER HEWITT MERCURY VAPOR LAMP
751
meter seem to solve the jjroblem for all its
effectiveness in general heterochromatic pho-
tometry. Integrating spheres and hemi-
spheres are limited by the marked effect of a
diffuse reflecting surface in increasing the
selectivity, by reflection, of a selective lijjht.
Direct comparison with calibrated color
filters to reduce the color difference on the
comparison field seems as yet to most nearly
approximate a physical valuation of use to
of a bare lamp, Fig. 11, is characteristic of
any line source. A standard reflector widely
used for industrial installations is so designed
as to give identical distribution curves, in
planes both parallel and perpendicular to
the tube, of the type shown in Fig. 12. In
the layout of an industrial installation these
curves and accompanying data are used
according to the principles fundamental in
all illuminating engineering practice.
TABLE I
umi-
3e in
>
S
Si
Bis
I-<
V
,„ 3
o."
p-m
Current
Length o
nous t
inches
3
Q.
6
<:
rt
^
So
S
■ci:0
0-5
« q
"So
Direct
50
110
3.5
385
550
17.9
875
.44
14.2
Photograph
Illumination
Direct
2-50
220
3.5
770
940
15.4
1500
.52
12.2
Photograph
Illumination
Alternating
50
110
or
220
110
3.8
85
430
615
17.9
975
.44
14.2
Photograph
Illumination
Direct*
67
7.
770
Blue Printing
Direct*
67
110
1,5.
1650
Blue Printing
Direct
3
110
4.
440
Quartz Arc
Direct
6
220
■
3.5
770
Quartz Arc
•Made also for alternating current,
standard lamp tubes.
Variations in length and shapes of above lamps provide some 25
the illuminating engineer and the candle-
power data of Table I were obtained by this
method. As the common form of Cooper
Hewitt lamp is distinctly a source of finite
area, and especially of finite length, the lamp
is photometered at such a distance as to re-
duce this error to less than one per cent
while in calculating the mean spherical candle-
power the usual spherical reduction factor is
used. The approximate distribution curve
Table I is a tabulation of some of the
characteristics of standard types of Cooper
Hewitt lamps. The larger tubes are used in
blue printing machines rather than for light-
ing, and illtmnination data are therefore
omitted. These straight tubes are modified
into specialized forms by variations in length
and by bending the standard 50 in. tubes into
U and M shapes for photographic enlarging
outfits.
752 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 9
The Importance of the Electrical Industry in the
Foreign Trade of the United States
By M. A. OvDix
Vice-President International General Electric Company
Ten per cent of the electrical apparatus manufactured in the United States during 1919 was exported, the
value of which was ninety million dollars or 50 per cent in excess of the value of these exports for 1918. This
high percentage of exports to total production is surprising in face of the exceedmgly unfavorable exchange
rates and is a true index of the importance of electric power in all countries of the globe. If this electncal
apparatus were not absolutely essential to the industrial rehabilitation of European countries it most certainly
wT)uld not have been purchased in such an unfavorable market. At home this foreign business has served to
maintain production in certain lines of electrical apparatus for which domestic demand has temporarily fallen
off- and as the manufacture of electrical goods involves the use of many other manufactured products, the
exploitation of foreign markets by American electrical manufacturers has greatly increased the business of
allied home industries. We should strive to maintain our present advantageous position in foreign trade by
improving our knowledge of international business methods. — Editor.
The participation of the electrical industry Europe took about as much as the Far
in the foreign trade of the United States is East. Our principal customers in Europe
were Great Britain, Nor«-av (which gave the
United States over S4,U0U.6c)0 worth of elec-
trical business, a surprisingly large amount for
so small a countr\-). France. Italy and Spain.
This high level of electrical exports during
1919 has been attained in the face of foreign
exchange rates, which have become increas-
ingly unfavorable. But the demand for
electrical products in Europe especially has
been urgent and insistent, so necessary- is the
utilization of electrical power regarded for
its industrial rehabilitation of Europe.
The fact that American electrical manu-
facturers have been able in general to maintain
prices, means that this export business has not
been handled at a loss, or in the way of dump-
ing surplus products. It has been developed
as a vital and necessan,* part of their whole
business.
How important this foreign business has
been in maintaining the production in
American factories in lines in which the
domestic demand has temiiorarily fallen off.
may be seen from the fact thai, while foreign
business as a whole was only about U) per
cent of that production, in many instances
it became a far greater jiercentage, offsetting
a decline in domestic demand.
For instance, in the case of certain material,
such as turbine sets, the foreign demand was
about 2."> per cent of the total; for street car
equipment, it was (10 per cent, and for electric
raihvav locomotives, still higher.
The falling off in domestic demand for
these items reflected the financial handicaps
of central stations and electric and trunk line
railways in the United States, which, because
of more or less stationan,- rates, had to face a
rapidly declining net income and in the tight
money market found it difficult to finance any
considerable purchase of new equipment.
of very considerable magnitude. Conse-
quently, electrical exports constitute a power-
ful factor in insuring the prosperity of the
industry- itself, in preventing the unemploy-
ment of labor and in contributing to the main-
tenance of wages. There are many allied
lines of machiner>^ which are essential ad-
juncts in the use of electrical products and
which are exported as a result of that associa-
tion. The prosperity of such industries is
closely bound up with that of the electrical
industry. Finally, the ramifications of the
electrical industry- and its dependence upon
other industries are such that its condition of
prosperity directly affects through these
contacts the welfare of a host of men, women
and children in this country.
As to the important part played by the
electrical industry in the foreign field, reliable
figures indicate that the production of the
electrical manufacturing industry in the
United States during 1919 was about .§900,-
000,000, of which over $90,000,000 or 10 per
cent was exported. This was an increase of
50 per cent over the amount exported in 1918.
As great a percentage of electrical products is
exported as in any other manufacturing
industn,- in this country, with the exception of
typewriters, cash registers, harvesting ma-
chinery and a few others.
The accompanying table shows the exports
of electrical material from the United States
in 1919 compared with 1918.
It is worth while to note the destination of
these electrical exports.
Over 40 per cent went to North and South
America, Canada, Brazil, Argentina and Cuba
being our best customers in the order named.
Over 25 per cent went to the Far East,
Jaoan, China, Australia and India being our
principal customers in that part of the world.
ELECTRICAL INDUSTRY IX THE FOREIGN TRADE OF UNITED STATES To.'!
The increased export demand for these
items of equipment meant that the manu-
facturers and their employees were relieved
of the necessity of extensive layoffs, or ex-
pensive shiftin}^ of labor.
Any factor like this, which helps to stabi-
lize demand and consequently production,
does a great deal to insvire a steady income to
the workmen and to prevent the distress and
complication attendant' upon unemployment.
The export of electrical manufactures aids
not only the electrical industry of the United
States, but also the manufacturers of allied
machinery and materials, the sale of which
j^'oes hand in hand with the sale of electrical
machinery.
Statistics show that where American-made
electrical m.achinery has successfully entered
foreign markets there has been extensively
sold allied machines, such as boilers, con-
densers, pumps, water wheels, hoists, mining
machinery, sugar mill machinery, rails and
accessories and a host of like material.
The exploitation of foreign markets by
large electrical manufacturing companies has
done much to open those markets for the prod-
ucts of other American manufacturers. This
intensive cultivation is largely responsible for
the wide prevalence of A.merican engineering
practice in m_any foreign countries. And this
again has reacted favorably upon the exporta-
tion of a wide variety of products of American
manufacturing industries.
At a recent meeting of the shareholders of
the Siemens & Halske Com-pany of Berlin,
Kark F. Von Siemens, Chairm.an of the
Company, emjjhasized the fact that the
manufacture of electrical goods belongs to
that class of industry which can do nothing
with purely raw m.aterials alone but also
depends on the articles produced by other
industries.
Indeed, it may be said that more than any
other industry, electrical manufacturing
draws upon the jjroducts of the soil and the
out])ut of mines and in no less a degree, the
manufactured articles of other industries in
a finished and semi-finished state.
The accompanying map, while showing the
location of plants manufacturing electrical
m.achinery and supplies, indicates onlv a few
of the materials entering into these m.anu-
factures and then only the raw materials.
It does not indicate the location of other
industries which contribute innumerable arti-
cles required in the manufacture of electrical
machinery and supplies.
In addition to the raw materials, the source
of which is approximately indicated on the
map, there are very important items such as
UNITED STATES EXPORTS OF ELECTRICAL MACHINERY AND APPLIANCES (INCLUDING
ELECTRIC LOCOMOTIVES)*
1919 and 1918 Compa"-ed
Year 1913
Year 1918
Per Cent
Chanctes
Batteries
Carbons
Dynamos and Generators .
Fans
Heating and Cooking Apparatus
Insulated Wire and Cables
Interior Wiring Supplies (incl. fixtures!
Lamps
Arc
Incandescent
Carbon filament
Metal filament
Locomotives (Electric)
Magnetos, Spark Plugs, etc
Meters and Measuring Instruments
Motors
Rheostats and Controllers
Switches and Accessories
Telegraph Apparatus (incl. Wireless) . .
Telephones . - .
Transformers. .
All Others
$5,998
1,.392
.■5,800
1,421
1,580
8,815
2,319
17
203
4,674
836
3,035
2,891
10,635
515
3,565
831
3,783
3,788
27,827
Total Electrical Machinerv, etc $89,925
$3,178
1,601
3,363
847
686
5,605
1,429
14
103
3,369
183
2,7,50
1,888
8,225
289
2,195
379
2,687
3,529
17,846
$60,166
+ 88.6
- 13.
-I- 72.5
-I- 67.8
-1-130.
-I- 57.4
-I- 62.3
-I- 21.5
-I- 97.
+ 38.7
-1-356.
+ 10.4
-I- 53.1
+ 29.5
-I- 78.3
4- 62.4
-f 119.2
-I- 40.9
-I- 7.3
H- 56.
-f- 49.5
(*) Expressed in thousarids of dollars.
754 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 9
iL
3E
• i 3 b
==.= 2
O O W C
-St. ^
y OX —
t' u u u
r
.-! - - O ?
i ^ E
-vis
c « E t c
T £ E F
'- ^ a m m
:^ E > > >
*j t* V
= JCJSJS
ELECTRICAL INDUSTRY IN THE FOREIGN TRADE OF UNITED STATES 755
cotton yarn, cotton cloth and cotton tapes
manufactured in the eastern and southern
]jart of the United States.
From the east also there come copper shapes
and brass manufactured in endless variety,
also such articles as dry goods, textiles and
paper products, and for construction pur-
poses almost every kind of machinery and
for production purposes machine tools, these
coming for the most part from the Middle
States. Other articles which lead in the
list of domestic commodities, and which are
required b\- the electrical manufacturing
industry, in addition to those already men-
tioned, are the following:
Vulcanized fiber, porcelain materials, steel
wire, metal alloys, petroleum wax from the
East; aluminum, hardware, coal, coke, pig
iron from the Middle West; asbestos, mer-
cury, rutile, turpentine, rosin, oils, pitch from
the South; mica and slate from the North,
while the West furnishes many of the articles
already mentioned and in addition, the raw
items of gilsonite, lead and spelter.
This list is sufficiently long to indicate that
the electrical industry is of vital importance
to every section of the United States. More-
over, the export of electrical goods contributes
to the commercial activities of the important
seaports of the Atlantic Coast, and of the
ports of New Orleans, San Francisco and
Seattle. The prosperity of the ports on the
Pacific and on the Gulf will be greatly en-
hanced if our electrical business with Latin
America and the Far East continues at its
present rate of growth.
With the certainty of the depreciation of
European currency continuing for a long time
to come and thus the maintenance of a barrier
against exports to Europe, more and more
does it behoove the American manufacturers
to look to those markets which are not af-
fected by unfavorable exchanges. With the
inevitable falling off of exports to Europe,
new business must be secured from the
markets of South America and particularly
from those of the Far East. New competi-
tion requires new business methods and the
demand for electrical material will not be
maintained at its present high rate unless
there is created a demand for new kinds and
new uses of electrical goods.
In the development of electrical projects , also,
American interests can do far more than they
have done in the past by aiding in the financing
of such developments without which many of
them will not be undertaken in the near future.
American trade has been given a tremen-
dous impetus during the years of the war and
the months which have followed the cessation
of hostilities. Whether we shall retain our
dominant position and not revert to our
restricted and provincial pre-war position
will be determined by the answer to this
question: Do American merchants, manu-
facturers and capitalists desire to play the
international game'
If this answer be in the afiirmative, American
enterprise no doubt will overcome the chief
handicap to attaining this position which is
found in our relative inexperience in the sphere
of international politics, finance and commerce.
756 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 9
Power Control and Stability of Electric
Generating Stations
PART II
By Charles P. Steixmetz
Chief Consulting Engineer, General Electric Company
The effect of power limitation upon the stable operation of large generating stations is of so great im-
portance that at the recent annual convention of the A.I.E.E. at White Sulphur Springs a session was set
apart for its discussion. Also, a session of the coming Edison Convention at New London will be devoted
to it. Our August issue contained an editorial and the first half of Dr. Steinmetz"s article on the subject.
The remaining half appears below. — Editor.
DISCUSSION OF E.M.FS.
The foremost difficulty, and uncertainty
in the application of the preceding equations,
is found in the selection of the proper values
of the machine e.m.f . E. E is not the terminal
voltage: by slipping past each other without
external impedance, the terminal voltage
of the alternators goes down to zero. Neither
is E the "nominal induced voltage," as this
has no actual existence, but is the voltage
which would be induced by the field excitation
if the saturation curve of the machine
continued as a straight line. It appears to
me that E must be considered as the "true
induced voltage, " or actual induced voltage,
that is. the voltage induced by the actual
field flux, that is, the field flux due to the
resultant field excitation and armature re-
action. The armature reaction, however,
fluctuates with the current between zero and
a maximum, while the actual field flux often
may be assumed as practically constant,
since the magnetic field cannot follow the
relatively rapid fluctuations of armature
reaction.
The magnetic efTect of the armature
reaction is represented electrically in the
synchronous reactance .t'o. The synchronous
reactance thus consists of a true self-inductive
reactance .Vi, which is instantaneous, and an
effective reactance of armature reaction .Vj,
which requires api^reciable time to develop,
and does not correspond to anv real magnetic
flux.
.Vo = .Yi-|-.v:
In turbo-alternators, a« usually is very
much larger than .Vi.
Electrically, the actual induced e.m.f. thus
should be the nominal induced voltage fo,
which corresponds to the field excitation, less
the reactance drop of the average current in
the elTective reactance of armature reaction.
If / equals the maximum efTective value
of the fluctuating current, the average current
is -, and the actual induced voltage thus is :
It is, however, in two alternators connected
together out of synchronism, through an ad-
ditional reactance :
2£ = /(2.vi-F.v)
where .v is the additional reactance through
which the alternators of actual induced
voltage E and true self-inductive reactance
.Vi are connected together, while running out
of s>"nchronism with each other.
From these two equations follows:
Maximtmi (effective) ^■alue of the fluctuat-
ing interchange current :
2x,+X2+x
and, actual induced voltage:
/ =
(47)
2xi-|-.v
2xi+Xi+x
(4«)
where Ca equals nominal induced voltage
effective value.
If the alternators are connected through
an impedance ;. z takes the place of .v, com-
bining vectorily with Xi and .vj.
In this calculation, the armature reaction
has been assumed as demagnetizing, and the
impedance voltage therefore subtracted from
the nominal induced voltage. This appears
correct, as the interchange current between
the alternators out of synchronism with
each other is essentially a lagging current,
throughout, as illustrated in Fig. \'.i.
If the two alternators are in synchroni^an,
but out of phase with each other by a maxi-
mtim phase displacement angle 2 Uo, it is :
2 £ 51 w Wo = / (,2 .» 1 -I- v )
and again assuming the armature reaction as
demagnetizing:
E = €o-
Ix,
POWER CONTROL AND STABILITY OF ELECTRIC GENERATING STATIONS 757
thus the maxmium (effective) value of the
fluctuating exchange current ;
J = - g" ^'" "» (49)
2 X1+X+X2 sin Wo
and, actual induced voltage;
_ 2xx-\-x ,.,..
E = eo::, 1 j : (oO)
2 X]+x+XiStn coo
where eo is the nominal induced voltage,
effective value.
However, in this case of alternators in
synchronism but oscillating against each
other, at least for small and moderate values
of Wo the interchange current I is essentially
an energy current with regard to the machine
voltage, and the reactive component of this
current alternately changes between lag and
lead, that is, between demagnetizing and
magnetizing. Therefore, the correctness is
doubtful of subtracting the impedance voltage
from the nominal induced voltage to get the
induced voltage, but it would be :
and as i varies between zero and /, the average
E would be the mean between
Co and V eo'^ — Pxi^, thus:
combining with the equation ;
2 E sin Wii = / (2 Xi+x)
gives :
2 eo(2 Xi-]-x)sin Wu
E = eo-
(2 Xi+xy-\-Xi^ sin^ Wo
(2.v-,+a-)=
(51)
\2x-,+xY^x.,}stn-0i„ ^'^"-*
It is probable that the true value of E lies
between those given by equations (50) and
(52),- but nearer to that of (52).
Substituting these values of equations (48),
(50), and (52) into the equations of sections
A, B, and C, and substituting
c = 2.vi+-v
in these equations, as the impedance of the
circuit between the two alternators, gives the
equations referred to the nominal induced
voltage, fo, that is, the field excitation.
The nominal induced e.m.f. et, is derived
by combining the terminal voltage e with iz,
where z is the total impedance inside of the
terminals, true reactance as well as effective
reactance of armature reaction. For non-
inductive load — and synchronous machine
load may be assumed as approximately
non-inductive — this gives :
C0=V'e2-|-(jA-j2
where ^ is the percentage reactance, and the
resistance is neglected as small compared
with the reactance.
However, this expression neglects the
change of reactance with increase of magnetic
saturation, increase of magnetic leakage be-
tween field poles, etc., and, therefore, espe-
cially in turbo-alternators with their enormous
magnetic fields, high saturation and high field
leakage, this expression is not very accurate,
and reasonably reliable only in the mean range
of current and voltage.
In sections C and D, the case of two
alternators or groups of alternators out of
synchronism with each other, the equations
of synchronizing power, energy, and critical
slip : P, Po, W, 5o contain the term :
2 X\-\-x
(2x,+x + x-^-
thus are a maximimi, if this term is a maxi-
mum. This is the case if :
or:
A-2 = 2 X\+X
x = Xi. — 1 X\
(54)
that is:
The synchronizing power between alter-
nators out of synchronism with each other is
a maximtmi, and the frequency difference
from which they pull each other into syn-
chronism is greatest, if the alternators or
groups of alternators are connected together
through a reactance which is equal to the
effective reactance of armature reaction, less
twice the self-inductive reactance of the
circuit between the alternators or groups of
alternators. With two alternators or groups
of alternators connected together without
any external reactance, this means if the
self-inductive reactance of the alternators or
groups of alternators is one-third the syn-
chronous reactance. With turbo-alternators,
the self-inductive reactance usually is much
less, and with such machines the synchroniz-
ing power is increased by the insertion of
external reactance.
Substituting above relation into the equa-
tions of sections C and D, gives as the
expression for the case of maximum syn-
chronizing power:
Actual machine e.m.f. :
= e\'\^-^-
(53)
E--
eo
'2
Resultant
e.m
.f.:
£» =
= eo
sin
-^ «
758 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 9
Resultant current:
J _eo sin s <f>
%■<.
Power fluctuation ofilow frequenc}-:
p _ gp- sin 2 <j> sir^OL
4 x-i
Energy transfer of low frequency:
W = -
-sin a
8 T sjxi
Continuous power transfer:
ce^ sin a cos (a+ff)
P —
Critical slip:
Sx.
^o = -
eo
iV-Zirfx.M
FEEDER REACTORS
A: General
Economy in cost and space makes it
desirable to use the smallest feeder reactor
which is reasonably safe, the more so as the
number of feeder reactors required to jjrotect
even,- feeder going out from the generating
station is usually much larger than that of
the generator and busbar reactors.
Any reactance inserted into the system
increases the reactive lagging volt-amperes
and therefore, if the load on the system is
lagging, lowers the power-factor, the more,
the greater the reactance of the feeder reactor.
In 2.5-cycle systems, this is of no moment, as
the load usually is almost exclusively syn-
chronous machines, equally operative with
leading as with lagging current, so that
even with large values of feeder reactors the
system operates at unity power-factor. In
()0-cycle systems, however, a considerable
part of the system usually comprises induction
machines and other apparatus which pro-
duce lagging current; the power-factor thus is
below unity, lagging, and much additional
reactance is therefore undesirable. An at
least approximate investigation of the
relations between the size of the feeder
reactance and the disturbance in the generat-
ing station caused by a short circuit at the
generating end of the feeder is thus desirable.
The function of a feeder reactor is three-
fold:
(1) It reduces the short-circuit current on
the generating station in case of a breakdown
of the feeder near the generating station, and
thereby reduces the shock on the system.
(2) It permits setting the feeder circuit
breakers for a much shorter time of ojjcning,
due to the lesser short-circuit current which
they have to open, and thereby reduces the
time during which the system is exposed to
the short-circuit stresses.
(3) It keeps at least partial voltage on the
busbars of the generating station during the
feeder short circuit, and thereby reduces the
liability of the generators, stations, and
substations falling out of s\-nchronism with
each other.
Without a power-limiting reactor in the
feeder, a short circuit in the feeder near the
generating station — which is much more
liable to occur than a short circuit on the
busbars — is practically a short circuit at the
busbars. The short-circuit current thus is the
maximum which the generators can give, and
its momentan,- or initial value (about eight
times the final value, with the usual amount
of generator reactors) is so great as to make it
necessary to set the circuit breakers for a
considerable time limit so as to allow the
momentarv- excess current to die out. During
the short circuit, the busbar voltage is zero or
])ractically so, thus there is no s]iTichronizing
l)ower between the generators of the affected
station section, between this station section
and the other station sections, and between
the synchronous converters of the substations
fed by the aflfected generating station sections
and as due to the time limit of the circuit
breakers the short circuit lasts an apjjrcciable
time, it is probable that during the short
circuit the synchronous machines in the
substations, and the generators have drifted
out of step with each other so much that at
the opening of the short circuit they do not
catch into synchronism any more, and a
shut down of a considerable part of the
system results.
At the moment when a short circuit begins,
the alternator field and thus the machine
voltage still has full value, and the inductive
short-circuit current thus is limited by the
true self-inductive reactance otily — the
external and internal reactance of the gener-
ators, and the reactance of the feeder reactor,
where such is used. At the moment when the
short circuit begins, the busbar voltage drops
from its normal jjrevious value, to zero if no
feeder reactor is used, or to the reactance
voltage of the feeder factor under the
nKmientary short-circuit current, which may
he a considerable jiart of the nonnal busbar
voltage. If then the short circuit could be
opened instantly before the altemator field
can build down under the demagnetizing
action of the inductive short-circuit current,
POWER CONTROL AND STABILITY OF ELECTRIC GENERATING STATIONS 7.59
the busbar voltage would recover instantly,
to its previous value. If, however, the short
circuit lasts any appreciable time, the alter-
nator fields gradually — and rather rapidly —
build down; the machine e.m.f. and the
short-circuit current decrease (and the busbar
voltage, with feeder reactor; without feeder
reactor the busbar voltage is zero, as stated) .
If now the short circuit is opened, the busbar
voltage does not instantly recover, but jiunps
up only to the value corresponding to the
then prevailing field flux, and then only
gradually — and rather slowly — -recovers by
the field flux again building up under the
effect of the exciter voltage.
In turbo-alternators, the rate at which the
machine fields build down under dead short
circuit, and at which the busbar voltage de-
creases which appears at the moment of open-
ing the short circuit, is very high, that is, the
field is demagnetized in about half a second,
so that with the delayed opening of the cir-
cuit breakers the field has practically been
demagnetized before the short circuit is
opened; but the rate at which the voltage
of the machine recovers after the opening of
the short circuit is rather slow, from three to
five seconds or more (depending on the
existing field exciting current and thus on the
load previously on the machine).
With a power-limiting feeder reactor,
however, of a reactance which though small
with regard to the rating of the feeder is
considerable compared with the reactance of
the generating station (internal and external
generator reactances), the rate of demagneti-
zation of the field flux is greatly slowed down,
due to the lesser demagnetizing action of the
smaller short-circuit current, that is, the time
required for the demagnetization of the
machine field is of the magnitude of one and
one half seconds. It is the larger, the higher
the feeder reactance and greater the num-
ber of generators connected to the busbars,
smaller with lower feeder reactance and fewer
generators on the busbars. If then the circuit
breakers can be adjusted to open quicker,
which appears feasible at the lesser short-
circuit current, most of the field flux will
still be there at the opening of the short
circuit, and the voltage thus would, at the
opening of the short circuit, jump back to
nearer full value. Considering that even
during the short circuit of the feeder cable,
considerable voltage remains on the busbars,
and that the duration of the short-circuit
period is greatly reduced by the permissible
quicker opening of the circuit breakers, it
appears feasible, with a moderate value of
feeder reactor, to limit the voltage drop and
its duration in the affected station so that all
or at least most of the synchronous apparatus
on this station section will remain in step.
B: Armature and Field Transients of Synchronous
Machines
(1) If
p = number of poles
Mo = number of field turns per pole
Zoo = exciting current at no load
and
$0 = magnetic flux per pole
then pni)i>o is the total number of inter-
linkages, and ^-— r — - the flux interlinkages per
Joo
unit current, that is, the inductance of the
field circuit. That is, in standard units;
Lo
p tlo^o
too
10-' /;
(55)
is the inductance of the field.
If
eo is the voltage of the exciting current
and
io the (permanent) field current,
the resistance of the field circuit is:
eo
ro = -
lo
(56)
This is the total resistance, field winding as
well as external rheostat, etc., as both have
the same action in the field transient. The
duration of the field transient then is:
^00 = ^ (57)
ra
that is:
ao = — = attenuation constant, and (58)
'00
n = /oe-*>'
E = Eoe-'^' (59)
is the field discharge, or the transient by
which the field current and thus the field
magnetism and the induced voltage decrease
on withdrawal of the exciting e.m.f., and
i2 = «o(l-e-*'0
E = Eoil-e-'^') (60)
is the charging transient of the field or the
starting current of the field, that is, the
transient by which field current and field
magnetism and thus the induced voltage rises
on the application of the exciting voltage or
recover after the demagnetizing action of an
excessive lagging current, such as a short
circuit.
76U September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 9
(2) On inductive load, the armature cur-
rent of an alternator demagnetizes, and to
give the same field flux the field exciting
current thus has to be increased to counteract
the demagnetizing armature reaction.
In a three-phase alternator;
If
n = nvmiber of annature series turns per
pole per phase
and
/ = amiature current per phase (effec-
tive), the armature reaction per pole is:
F= 1.5\ '"2 III ampere-turns,
and
iitsi= \..}\/"2 nl
thus gives the field current
1 - /— " T
? = l.av 2 — i
iln
where :
c= l.ov' 2
Ho
(«1)
(G2)
is the reduction factor from armature to field.
Thus
If Jo = field exciting current,
and an additional inductive load of / amperes
is put on the alternator, to keep the same
magnetic flux and thus the same voltage, the
field exciting current has to be increased from
ia to io+cl.
[This does not consider the change of the
magnetic flux distribution resulting from the
inductive armature current /, such as the
increase of leakage flux, corres])onding in-
crease of saturation, etc., which requires a
somewhat greater increase of field excitation.
That is, c is somewhat greater than gi\cn !)>•
equation (G2).]
(3) Let £o be the voltage produced by the
no-load field excitation im- An inductive load
of current / requires an increase of the field
excitation cl. This additional field current
cl would produce (assuming a straight-line
saturation curve, that is below saturation) a
voltage ;
Ei = ■:— £o
loo
thus gives an apparent internal reactance
(if the machine;
E.,
or;
/
cEa
loo
This is the effective or equivalent reactance
of armature reaction. It is not a true react-
ance, and differs from the true self-inducti\-e
reactance, that the latter is instantaneous,
while the effective reactance of armature
reaction x^ requires some time to develop.
Or, if 7o equal full load or rated armature
current, the effective reactance of armature
reaction, given as fraction (or in per cent), is
j=^lL» = £^ (64)
£■0 loo
that is, the ratio of the field equivalent of the
armature current, do, to the no-load field
current I'oo, is the effective reactance of arma-
ture reaction, as fraction.
Substituting (G2) into (03) gives:
.V-. = 1..)\ 2w-
= 1.5v
woioo
^ £o
o i;
- fo
(65)
(03)
where Fo is the no-load field ampere-turns per
pole, which give the voltage £o-
(4) Let Eo equal the voltage and I'o equal
the field exciting current of an alternator.
Let then an inductive load of current /o be
suddenly thrown on the alternator, for
instance by a short circuit beyond a feeder
reactance, or on the busbars. If then the
reactance (true self-inductive reactance) of
the circuit of this inductive load is .Vi, the
current is:
/o = - (66)
-Vl
This current U however demagnetizes with
the field equivalent c7o. and the magnetic
field flux of the machine, and thus the
voltage must therefore decrease. The field
flux however cannot change instantly, as in
changing it induces a voltage and therefore
produces a current in the field circuit, which
oiiqioses the change. That is, the field flux
begins to decrease at such a rate as to induce
in the first moment a voltage in the field
winding, increasing the field current by do.
the field equivalent of the armature current.
That is. in the moment when the inductive
load current Fa is thrown on the alternator
armature, the alternator field current iimij)s
from /o to iu+cln-
As, however, the exciting voltage fo can
maintain only the current lo in the field
circuit, the momentar>- excess field current
io + cia gradually decreases, down to the
permanent value I'o, and with it decreases the
field flux and the voltage of the machine.
POWER CONTROL AND STABILITY OF ELECTRIC GENERATING STATIONS 761
from the initial values $o respectively £u, to
the final values;
to:
to + CI 0
and
Er-
to
r- Ea = bEo
(67)
(68)
to + do
and with it decreases the current, from the
initial value /o, to
to
io+cio
(69)
where ;
(70)
b = -^- -
io + cio
At the first moment the field flux is still 4>o,
the field exciting current however is io+ch-
Field flux 4>o and no-load field exciting
current j'oo give the field inductance Lo. Field
$0 and field exciting current ^'o+c/o thus give
an apparent or equivalent or eflfective field
inductance:
L--
«00
to -\- do
Eo — boLo
where :
6o = -r
too
(71)
(72)
io+do
That is, when throwing an inductive load on
an alternator, field flux and voltage decrease by
the demagnetizing arm.ature reaction, and dur-
ing the field transient, the mutual inductance
of the armature current on the field reduces
the field self-inductance from the true self
inductance Lo to an apparent or eflfective
inductance L = boLo-
As the resistance of the field circuit rem.ains
the same, the duration of the transient
resulting from a sudden inductive load, such
as a short circuit, thus is given by:
L
to = -
ro
, Lo
= 0o —
ro
= bo too (73)
and the attenuation constant is:
and the equations of the transient thus are:
The armature current, changing from:
J En
io = —
/, = w„
or:
7=/i-f(/a-/i)e-"
/ = /„ b + (\-b)t-
(75)
(76)
and the voltage then is :
E = xJ {11)
and, if of the reactance Xi, the part .v is
external, Xi — x internal in the machine or
station, the terminal \-oltage is:
E'> = xl (7S)
(5) Equations and Denotations
ro = resistance of exciting circuit
Co ,
= — ohms
to
(56)
eo = exciter voltage
/oo = field current at no load
f'o = field current at full load
Ln = true inductance of field exciting
circuit
_ ptto *o
ioo
10-8 /j
ioo)
(57)
(58)
p = number of poles
«(, = number of field turns per po'e
<i>o = m.agnetic flux per field pole, produced
by exciting current ion
/oo = duration of field transient
To
Oo = attenuation constant of field tran-
sient
_J_
E = E„i~'"''
= field discharge transient (59)
£ = £;„(l-e-"°0
= starting transient of field (60)
xi = total self -inductive reactance of
alternator circuit
X = external part of this reactance
£o = machine voltage before closing the
circuit on reactance Xi
/o = initial (or momentary maximum)
value of the current in reactance
X (effective valuej
Xi
(66)
762 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 9
/] = final (or permanent) value of current
= W„ (69)
b = ^
io+cio
c = 1.5v/2
(70)
no
= reduction factor from armature to
field circuit
M = number of series armature turns per
pole per phase
I = h[b + {l-b)]e-<" = Ii + {Io-I,)t-'"
(75) (76)
= armature transient,
a = equivalent attenuation constant of
transient
_l
"to
to = botm (73)
200
(74)
bo = -. , J
to+cl
£ = total voltage
£" = terminal voltage
= xl
(72)
(77)
(78)
(6) From these equations, and the numeri-
cal constants of the alternators, it is possible
now to calculate the action of a short
circuit or similar disturbance on the system,
and the effect thereon of the reactance of the
feeder reactor, by calculating, and plotting
with the time as abscissae, the transients of
induced voltage, current, and terminal \-olt-
age resulting from the application of a short
circuit. This gives for the moment of the
opening of the circuit breaker the values of
current, terminal voltage, and induced volt-
age, and from the latter the value of the
terminal voltage immediately after the
moment of the opening of the circuit breaker.
Calculating then, and plotting, with the
latter as, initial values, the field transient
gives the voltage recovery curve of the
system. From the drop of voltage, and its
duration, then can be estimated whether any
synchronous apparatus such as converters,
operated from the affected generating station,
are liable to be thrown out of synchronism,
and whether by the voltage drop the syn-
chronizing power of the station against other
stations tied to it by busbar reactors is
sufficiently lowered to fail to kee]) in step,
♦ A reactance of n per cent means « per cent of the value of
rated voltage
rated current'
and whether in this case the duration of the
voltage drop is sufficient for the machines or
stations to drift far enough apart so that at
the voltage recovery they are not able any
more to pull each other into step.
C: Numerical Calculations
The constants of some typical steam
turbine alternators of large size, three-phase
machines of 25 and 60 cvcles, are given in
Table I.
Considering now as a numerical instance
the efl:ect of a feeder short circuit close to the
busbars, on a 25-cycle 9U00-volt generating
station of 60,000 kw. steam turbine alternator
capacity, without and with feeder reactors,
assuming that the short circuit is opened
after one second. Assuming as a fair average :
An equivalent effective reactance of arma-
ture reaction of 85 per cent.
A true self-inductive internal reactance of
the alternators of 6.S per cent, and
An external reactance, as power-limiting
generator reactor, of G per cent.*
Let the duration of the field transient (full-
load condition) be
<oo = 4.51 sec.
the field attenuation constant thus
ao = 0.222
The field transient then is given by :
e = ei+{eo-ei)t-''^ (79)
where :
ec = voltage of the machine at the moment
/ = 0, for instance, initial voltage in the
moment when a short circuit has been opened.
t'i= machine voltage corresponding to the
exciter voltage impressed upon the machine.
that is, final voltage of the machine.
Consider the three cases :
(a) No feeder reactor, thus dead short
circuit on the busbars.
(b) A feeder reactor of 0.5 ohms per phase,
or 2.9 times the true reactance of the generat-
ing station, or .S7 per cent.
(c) A feeder reactor of 0.7 ohms per phase,
or 4.05 times the true reactance of the
generating station, or 52 per cent.
(a) With 12..S per cent self-inductive react-
ance, the momentary or initial short-circuit
current, as fraction of the rated current of the
station, is given by:
'° = (U28 = "-^
From the machine constants, it follows:
b =0.172
^0 = 0.1 29
POWER CONTROL AND STABILITY OF ELECTRIC GENERATING STATIONS 763
TABLE I
CONSTANTS OF THREE-PHASE STEAM TURBINE ALTERNATORS
25 Cycles
Rating, kw i 12.000
Speed, r.p.m.
NIech. momentum, mega
joules 3 M
No. of poles, p
Volts, terminal
Amperes, full load
Synchronous reactance, Xo,
per cent i
Internal true reactance, .rn,
per cent
Eff. react, of armat. reac-
tion. .v«. per cent
External reactance, Xi; per
cent
Regulation, non-ind. load,
per cent
Field
No. of turns per pole no. • ■
Amp., no load, ioo
Amp., full non-ind. load, io
Exciter voltage, fo
Flux per pole, ml, *»
Armature . I
No. series turns per pole
per phase, n
n
c = 1.5v 2—
no
i„ = ^'L«^10-s(henrys),,.
Joo
No Load
ro = -.— = (ohms)
loo
/o(, = — = (seconds)
ro
1
Full Load
ro = ^ = (ohms) . . .
Jo
l^ = — = (seconds) .
fo
1
««=r
'ou
Short Circuit after Full Load
Initial current, I„ = — . ..
io+ch
Final current, /i
. 'oo
6o = ,
--bh.
id+ch
L = feoio = (henrys) .
/„ = — = (seconds) . .
Cn
a=l .
/o
50
150
4
9000
770
89 1
5.41
8.3.6;
21
95
250
.333
125
HI
12'
0.268
1.69
0.5
3.38
0.296'
0.375
4.51
0.222
6700
14,000
750
200
4
9000
900
85
4.6
80.4
22
96
307
388
125
135
10
0.221
1.69
0.408
4.14
0.241
0.322
5.25
0.191
8450
0.156
0.172
1050
1450
0.117
0.137
0.198
0.232
0.527
0.72
1.90; 1.401
20,000
750
233
4
9000
1280
129
6.8
122.2
4.13
29
78
328
522
125
137.5
10
0.271
1.30
0.385
3.38
0.296
0.24
5.42
0.185
11,700
0.141
1650
0.089
0.116
0.482
2.08
30,000
1500
262
2
9000
1925
108
9.0
99
6.25
24
198
226
335
250
276
12
0.129
4.84
1.11
4.3o
0.229
0.7471
6.48
0.154
12,700
0.170
2160
0.114
0.551
0.737
1.35
35,000
1500
204
2
9000
2250
60 Cycles
28 I
i
243!
185:
320
250;
252
12
0.105
6.62
1.35
4.91
0.204
0.781
8.47
0.118
147
120
11.4
12
135.6
108
0.167
2530
0.0965
0.638
0.8161
!
1.23 1
12,500 14,000
1800 720
97' 220
4i 10
12,000' 12,000
*750 1900
74
9
65
12
39
385
463
125
68
0.217
0.662
0.325
2.04
0.49
0.2
2.45
0.408
31
150
122
190
125
71.5
10
0.141
3.53
1.02
3.45
0.29
0.66
5.35
0.18
15,200 6200
0.180
1120
0.115
0.405
0.615
1.63
20.000
720
258,
lOl
12,000!
*1200
125
9
116
16
36
348
517
125
68
4
0.?36
:;.7>3
0.36
1.96
0.51
0.242
2.91
0.344
20,000
1200
30,000
1800
310 190
6 4
12,000, 12,000
*1200 tl600
10,000 13,.300
0.176
1760
0.146'
0.097
0.358
2.77
0.142
1880
0.095
0.067
0.276
3.63
73
13
60
15
84
286
367
230
130
15
115
25
95
235
368
230
95.5 117.5
5, 6
0.1261 0.134
1.80
1.90
0.8061 0.98
2.23
0.448
1.94
0.515
0.628 0.625
2.92j 3.05
0.342 0.328
1
9200 10,600
0.240j 0.206
2200 2180
0.187 0.137
0.337 0.260
0.547 0.418
1.83
2.39
* At 80 per cent power-factor.
t At 75 per cent power-factor.
♦ At 90 per cent power-factor.
764 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 9
Thus, final short-circuit current, as fraction
of rated current:
{i = bto= 1.34
and, duration of the short-circuit transient :
^0 = ^0^00 = 0.582 sec.
Thus, attenuation constant;
1
to
-=\.~,
Plotted as cur\-e " ei" in Fig. 15.
The terminal voltage is zero during the
short circuit ; at one second, with the opening
of the short circuit, the terminal voltage
jimips back to the same fraction of the
terminal voltage before the short circuit, to
which the induced voltage has dropped, that
is, to the point £>i of curve "^i, " 32.2 per
cent of the previous terminal voltage.*
J/77/7
~
«
n
-^
3.0
\
I
28
\
V
V
Z6
V
\
2.4
v
Z.2
\
1
Volt
1.0
1
s_
2.0
A
\^
';
I
s
^
>
/.8
0.9
\
^j
^
"S
V
f\
1 ' '
>0-
/"2-
1
1
k
S
^
kJ^'
v^
0_^
_4_
— j
0/^
^A
^
^r-
-s, ^
h
«^
1 — ' —
—
1
—\—
'
1.6
J4
07
^
-^
'^'
^s
^^
k
■*
— 1 -
cc;^
'
1
1
^
^^
^
■»,
^
H
%
-^
b!>
^
**
■»«
■"■r—
1
1
1.4
il!
^
> I
"*"
■«.
J
^^
?^-_
t~.
_^^_^
1
!■'
• -■ ^
— —
—
"
-"
—
—
-
^
■*/
06
\
-<
^
^
^
^
^^
*""
-—
— +^
^.J^— —
I.Z
\
,.
«*
»j
—
k^
-,«,^
:2
S
^^
r*
=-
Of
A
<
(
H
-n
^
1 ,
f-___,
' —
LS
c.
I.U
\
"~"
^
„,,— '
■* 1 \
— — ^
—^
H—i
.^::=a
0.4
'S
i.
^
fSTra
=^T1
-
Y~-. ■ i
^
^rt
o.s
V
s
.^
-t5
=^ir;
k===
^^^— ^-^
p'. , 3
,*
0.6
(71
b|
<
n
1
1
1
»«,
1
1
04
OZ
"^
■»•
^
1
1
1
"1
r
' \ —
^
^
-^
-t-
—
*,
OZ
0 1
a,
e"
c,
1
?
0
z
0
4
0
.6
0
<5
/
0
/
z
1
4
/
b
/
*
2
0
. 1
5
2
2
4
2
£
2
8
J
0
3.
z
J
4
S
6
-J
8
-«
(7.
t<c
and, equation of the short-circuit current
transient :
J=ii + («o-«i) *-"'
= 1.34-1-6.46 t-i'72( (80)
This current is plotted in Fig. 15, in dotted
line , "j,."
Proportional hereto is the induced generator
voltage, and thus is given, as fraction of the
induced voltage immediately before the short
circuit, by the transient :
e = h-\-{\-b) t-"
= 0.172-|-0.S2,Se-i-7-'' (81)
* Assuming that the conditions of the external load have not
materially changed or have no material effect, which latter may
be assumed approximately, since the short-circuit currents are
very large compared with the normal-load currents.
From this point, of 32.2 per cent voltage at
one second, the \oltage now gradually recov-
ers on the field transient, equation (73) for
f,= l, To = 0.322, 00 = 0.222, thus:
f»=l-0.67St-«^2/
During the short circuit, the terminal volt-
age thus traverses the values of .4 Hx C\ Di F\
in Fig. 15. As seen, the voltage recover)- is
ver\- slow, and it is not probable that any s>ti-
chronous apparatus will remain in step.
The short-circuit current after one second —
which the circuit breaker has to open — is 2.5
times the rateil current, or 150, 01)0 kv-a.
(b) With 12.8 ix-r ct^nt self-inductive gener-
ator reactance, and 37 per cent feetier
POWER CONTROL AND STABILITY OF ELECTRIC GENERATING STATIONS 765
reactance, the total reactance is 49 per cent,
the initial short-circuit current thus:
1
'0 =
:2.01
0.498
In this case, it is:
/) = 0.44G
do = 0.335
Thus, final short-circuit current :
n = fc^ = 0.894.
Duration of short-circuit transient:
^ = ^0^00= 1.51 sec.
Attenuation constant :
o = - = 0.6G2
Short-circuit ctirrent transient :
^■ = 0.894 + 1.107 e-0-662<
Induced voltage :
e = 0.446-1-0.554 e-o-sea'
The terminal voltage during the short
circuit now is not zero, but is the same
fraction of the induced voltage, as the
reactance of the feeder reactor is to the total
self-inductive reactance :
X + Xi
the
- = 0.744
37+12.8
Thus, the terminal voltage during
short circuit is:
e" = 0.744 e = 0.332+412 e-o-662i
The transients of short-circuit current,
induced voltage, and terminal voltage are
shown in Fig. 15 by the curves "i->," "^2" and
"e,o."
As seen, at the moment of short circuit, the
terminal voltage makes a sudden drop to
curve "e..", " to 0.744 of the previous value,
then follows the curve "^2°" for one second;
then on opening the short circuit suddenly
jumps, from 0.543 on "Ci"" to 0.732 on " e-i,"
and then gradually' recovers on the field
transient given by the equation :
e»= 1-0.268 €-0-222'
The terminal voltage thus traverses the
broken curve AB^CiDiF^.
As seen, while the voltage recovery after
the short circuit is slow, the terminal voltage
even during the short circuit remains above
half value.
The current after one second, when opening
the short circuit, is only 1.46 times full-load
current.
(c) In the same manner the curves are
calculated for the 52 per cent feeder reactor,
giving :
10=1.55
6 = 0.511
ho = 0.383
fi = 0.79
/o=1.73 sec.
a = 0.578
t = 0.79 + 0.76 e-"s78<
e = 0.51 1+0.489 €-0-"8'
- = 0.802
X+Xi
e» = 0.41 +0.392 e-0-578(
and the recovery transient :
^0=1 -0.215 €-0-222(
The values are shown in Fig. 15 as "js,"
"^3," "^3°, " giving for the terminal voltage
the broken curve ABzCsD^Fs.
The short-circuit current when the circuit
breaker opens, after one second, is only
9 per cent above full-load current.
Table II gives the numerical values of
voltage and current, at the beginning of short
circuit, after one second and final, as fractions
of rated voltage and current.
The question then arises of the bearing of
these voltage curves. Fig. 15, on synchronous
operation.
During the period of dead short circuit or
zero terminal voltage, BiCi, there is no
synchronizing power. There is no load on the
generators beyond the Pr and the load losses
which are moderate even at the initial high
momentary short-circuit current and rapidly
decrease with the decreasing short-circuit
current. Thus the alternators speed up, until
the governor shuts off steam or the emergency
governor trips. The former necessarily must
take an appreciable time to avoid trouble
from steam governor hunting. Thus usually
the speeding up will occur until the emergency
trips and cuts off steam, about 10 per cent
above normal speed. Then slowing down
occurs until the machines are again put on
their governors. The speeding up, however,
occurs at different rates, due to the differences
in the momentum of the different machines;
the speed of tripping cannot be exactly the
same, as absolute reliability rather than
exactness of speed is the main requirement of
the emergency cut off; and furthennore,
some speeding up will continue after the
closing of the governor, due to the steam
contained between the cut off and the turbine.
766 September, 1020
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 9
and in the turbine.* Thus, if during this
period the machines do not have sufficient
power to keep each other rigidly in step, at
the time when the short circuit is cleared
and the voltage returns , the machines probably
have drifted so far apart that they cannot
pull each other in step again but continue
slipping out of synchronism, short circuiting
each other and keeping zero terminal voltage
indefinitely.
Let P equal the load on the machine before
the short circuit. With the load taken off.
the power P then accelerates the momentum
M of the machine, until the steam is cut ofT.
This means ;
2sM = Pt (82)
where 5 is the increase of speed in fraction,
and t the time or more accurately :
M((l+5)=-l] = P/ ' (83)
however, (82) is sufficienth- accurate for our
purposes,
thus:
s = ^^t (84)
Substituting the values of P and M from
Table I, gives the acceleration curves. In
Fig. 16 are given such cur\-es for four 25-cvcle
machines, the 12,000, 14,000, 30,000, and
35,000-kw., as (1). (2). (3). and (4). As seen.
the acceleration is ver\- rapid, from 3.5 to 8.()
per cent per second.
* One cubic meter of steam at 14 atmospheres (200 Ibs.^
retained between the turbine and the steam cut off would speed
up a 35.000-kw. steam turbine alternator by more than one per
cent, after the cutting off of the steam.
The limits of synchronizing power, that is.
the maximum speed difference from which
the machines can pull each other into step
promptly, is given by equation (38) as:
2s^ = E
I sin a
\2irfzA.
fzM
Choosing the same values as in Fig. 15, that
is, per 10,000 kw. rated machine capacity:
s = 2.v,= 2.08 ohms
3 M= 125X10* joule
/ = 25 cycles
a = 90 deg.
^ 9000 ...,,„. ,
E = — 7= = o200 volts
V'3
gives :
2 5o= 1.4 per cent.
In the moment however when the short
circuit opens, the induced voltage of the
machine has dropped from the initial value,
due to the demagnetizing effect of the short-
circuit current, on the cur\-e ei of Fig. 15,
and the critical speed 2 5o has dropped pro-
portional thereto.
In Fig. Hi thus is given in dotted line the
cur\-e 2 So. as (0). As seen, even in a fraction
of a second, that is, in a time much shorter
than the circuit breaker can open the short
circuit, machines of different types have
drifted apart b>- greater speed differences
than those at which the machines can pull
each other in step again at their reduced
s\-nchronizing power.
However, even with identical machines,
especially if the speeding continues to the
TABLE II
SHORT CIRCUIT ON 60,000KW., 25-CYCLE, 9000-VOLT STATION
USISTANCB or FEEDEK (KACTOK
None
0,5 ohms
0.7 ohms
Duration of field transient, seconds /oo 4.
oo 0.
Duration of armature transient, seconds U 0.
a •. 1 .
b (I.
'J bo 0.
»o 7
i 2
•i ^ 1
«« 1
e 0
ti 0
f Before Ct 1
I After ro° . .
f Before ««. .
1 After e d
*■•
Short circuit
current
Induced
voltage
Terminal
voltage
f Initial
•I After 1 sec.
I Final
( Initial
< After 1 sec.
I Final
Initial
I After 1 sec.
L Final
222
.W2
172
12!t
8
34
(H)
.{22
172
(K)
4.51
4.51
(t.222
0.222
1.51
1.73
().6ti2
0.578
0.44ti
0.511
0.335
0.38:5
2.01
1.55
1.46
1.09
0.894
0.79
I.IH)
1.00
0.732
0.785
(».44«
0.5U
I.IX)
1.00
0.744
0.802
0.543
O.fi.3
0.732
0.785
0.332
0.41
POWER CONTROL AND STABILITY OF ELECTRIC GENERATING STATIONS 7(57
tripping of the emergency steam valves,
inevitable inequalities in the tripping speed
and in the time of restoring the machines on
steam governor control probably cause greater
speed differences than permissible by the syn-
chronizing power. Furthermore, even if the
short circuit is opened in a second or less, the
induced voltage has drojjped so considerably
(^1 in Fig. 15) and the recovery curve (ei° in
Fig. 15) is so slow that the machines cannot
immediately take load, and speeding up con-
tinues for some time.
Thus, it may be expected that with a dead
short circuit at or near the busbars of a
high-power steam turbine station, the gener-
ators drop out of synchronism and are not
able to pull back promptly into synchronism,
but begin to drift indefinitely, slipping past
each other at zero voltage.
For a machine to remain in synchronism
with other machines, with full-load steam
supply and the load thrown off by a short
circuit, etc., the machine must be able to
transfer full load to other machines, within
its limits of synchronizing power, that is,
with a phase displacement not exceeding
90 deg.
The maximum power transfer between
two machines is given by equation (11) as ;
P = — sin a
where s is the total impedance between two
machines, and a may be assumed as 90 deg.
This gives:
E = v/^ (85)
as the minimum voltage E at which the
machine will keep in synchronism at a ])ower
difference P between the load and the steam
supply.
Substituting thus for P the rating of the
machine per phase, and for z twice the self-
inductive reactance (external and internal)
per phase, gives the minimum voltages of
remaining in synchronism, that is, the
voltage limit of synchronizing power.
p
~
r
~
~
~
""
~
r-
/
r
/
i
/
i
f
/
J
/
/
/
/
i
/
f
/
/
/
(
/
»i.
/
/
f
/
t
/
/
/
f
ii
/
/
/
J
/
/
f
/
(
/
^
ii.
^
/
r
/
^
>
/
(
/
y
}
/
/
y
/
^
/
,
/
'
/
/
/
^
.
/
/
-^
f_
/
/
A
/
^
/
/
^
7
r
1*1
I
p;
/V
/
-H
t
f
i
-J
^
=
-
»-.
.
■—
-
'•
s
.4
_
,,
.
,,
/
10.0
a.s
9 0
»S
SO
&0
s.s
SO
4S
*.0
5.5
30
IS
zo
O 0.1 0.2 0.5 0.4 0.5 Oj6 0.7 0.8 0.9
Fig. 16
J 2 15 f4 IS Ac
This gives, for the machines in Table I, the
values recorded in Table III.
As seen from Table III, the voltage limit
of synchronizing power in most of these
TABLE III
VOLTAGE LIMIT OF SYNCHRONIZING POWER
Rating, kw
Speed, r.p.m
Volts per phase, e.
Rated current, i. .
- = (ohms)
X\ = (per cent)
s =2 I'l = (ohms)
P per phase, (watts) 10* .
E = V?z = (volts)
E in per cent
25 Cycles
60 Cycles
12,000
750
5200
770
14,000
750
5200
900
20,000
750
5200
1280
30,000
1500
5200
1925
35,000
1500
5200
2250
12,500
1800
6900
750
14,000
720
6900
900
20,000
720
6900
1200
20,000
1200
6900
1200
6.76
5.78
4.07
2.71
2 32
9.2
7.67
5.76
5.76
11.4
1.54
4.0
10.6
1.225
4.67
10.93
0.89
6.67
15.25
0.823
10.0
11.4
0.528
11.67
12
2.21
5.17
9
1.38
4.67
9
1.04
6.67
13
1.495
6.67
2480
47.8
2390
45.0
2430
46.9
2870
55.2
2480
47.8
3030
43.8
2530
36.6
2630
38.0
3150
45.6
30,000
1800
6900
1600
4.32
15
1.30
10.0
3610
62.1
76S September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 9
machines is a little below half voltage, and
the conclusion thus follows that:
As long as the machines do not drop below
half voltage, little danger exists of the ma-
chines breaking out of synchronism by the
sudden loss of load under short circuit or
other accidents, and:
If a feeder reactor limits the voltage drop of
the station, due to a feeder short circuit, to
50 per cent or less, the machines in the
station remain in synchronism, even when
speeding up due to the release of load,
when tripping their emergency steam cut
offs. etc., and the voltage thus recovers
immediately on the opening of the short
circuit.
As seen from Fig. 15, this is the case even
with the smaller feeder reactor, and under the
conditions of this instance the smaller feeder
reactor thus should offer complete protection
against loss of synchronism of the station as
result of feeder short circuit.
Similar relations then exist between gener-
ating station and synchronous machine loads,
such as converters and synchronous motors.
The s>'nchronous converter probably
represents by far the most frequent syn-
chronous machine load. Its internal character-
istics are somewhat similar to those of the
steam turbine alternator, that is, high
effective reactance of armature reaction and
low true self-inductive reactance, and it
therefore is probable that about the same
numerical relations pertain, that is:
Such values of feeder reactors, which are
sufficient to guard the generating stations
against loss of synchronism, by maintaining
the station under feeder short circuits above
the voltage limit of synchronizing power,
may in general be expected also to guard the
s>-nchronous converters in the substations
against being thrown out of step, that is,
shut down, by the shock on the system due
t(i feeder short circuit.
Relative Thermal Economy of Electric and
Fuel-fired Furnaces
By E. F. Collins
IXDUSTRl.AL HeATIM; DEPARTMENT, GEXER.\L ElECTRIC COMPANY
The pronounced advantages of the electric furnace in providing easy and positive temperature control
and uniform heat distribution account for the large number of these furnaces that have been placed in service
during the last few years. The extent of these apphcations is indicated by Mr. Collins' statement that nearly
all brass melting furnaces that are being installed are electric furnaces. In many cases the electric furnace was
chosen on the basis of performance rather than cost of operation: but today with the higher and still rising
costs o'' fuel even the cost of o]>eration may be decidedly in favor of the electric furnace, as is shown in this
article by the data on sp>-cific installations. It is not visionar\- to predict that electricity will ultimately be as
widelv used for industrial heating as it is now used for industrial power. — Editor.
It is well known that the thermal efficiency
of a furnace decreases rapidly with increasing
temperature. This decrease is due primarily
to the fact that the air required for com-
bustion must be heated to the temperature
of the furnace; the heat necessary- to raise
the air to this temperature being lost in the
products of combustion as they escape. A
secondary cause is the heat lost from the
walls of the furnace, which for convenience
is called the radiation.
It is not possible to state with exactness
the efficiency which may be realized in a
furnace unless all the conditions are known,
as there arc rrany factors which may influence
the amount of heat lost in the flue gases. For
instance, a part of the heat may be recovered
by ijrehcating the fuel or incoming air, or by
heating water or other materials for incidental
uses; or the furnace mav iit some ca.ses be
constructed on the compensating or counter-
flow principle, when a considerable amount
of the heat iti the flue gases will be given up
to the incoming charge.
These considerations apply in getieral only
to relatively large furnaces or installations,
in which cases the plant is carefully designed
to utilize the waste heat. In the ordinary
fumace. operating under ordinani' conditions,
with which we arc all familiar, none of these
considerations api)ly. and the thermal effi-
ciency can therefore be approximated by
calculations, since ])ractically all of the
theoretical heat represented in the tem-
perature of the flue gases is lost.
Data showing the theoretical los.scs in flue
gases for various fuels at various temperatures
have been submitted by earlier investigators,
but it is thought that it might be of interest
to complete these data and |)ut them in the
RELATIVE THERMAL ECONOMY OF ELECTRIC AND FUEL-FIRED FURNACES 769
Source of
Heat
Coke
Klectricity
City Gas
Fuel Oil
Anthracite
Bituminous
Coal
Natural
Gas
Per Cent
of Air foi
Perfect
Com-
bustion
Air at
TODeg.F.
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
100
150
Calorific Value
of Fuel or Power
B.T.U.
13,000
i:<,000
13,000
13,000
13,000
13,000
13,000
13,000
per lb.
per lb.
per lb.
per lb.
per lb.
per lb,
per lb.
per lb.
3415 perkw-hr.
3415 per kw-hr.
3415 perkw-hr
3415 per kwhr
590 per cu. ft.
590 per cu. ft.
.590 per cu. ft.
590 per cu. ft.
590 per cu. ft.
590 per cu. ft.
590 per cu. ft.
590 per cu. ft.
19,000 per lb.
19,000 per lb.
19,000 per lb.
19,000 per lb.
19,000 per lb.
19,000 per lb.
19,000 per lb.
19,000 per lb.
12,000
12,000
12,000
12,000
12,000
12,000
12,000
12,000
12,550
12,550
12,550
12,550
12,550
12,550
12,550
12,550
per lb.
per lb.
per lb.
per lb.
per lb.
])er lb.
per lb.
per lb.
per lb.
per lb.
per lb.
per lb.
per lb.
per lb.
per lb.
per lb.
1100 per cu. ft.
1100 per cu. ft.
1100 per cu. ft.
1100 per cu. ft.
1100 perc u. ft.
1100 per cu. ft.
1100 per cu. ft.
1100 per cu. ft.
Temperature
of Furnace
400 deg.
400 deg.
1600 deg.
1600 deg.
2300 deg.
2300 deg.
2800 deg.
2800 deg.
400 dep,. F
1600 deg. F
2300 deg. F
2800 deg. F.
400 deg. F
400 deg. F.
1600 deg. F
1600 deg. F
2300 deg. F
2300 deg. F.
2800 deg. F
2800 deg. F
400 deg.
400 deg.
1600 deg.
1600 deg.
2300 deg.
2300 deg.
2800 deg.
2800 deg.
400
400
1600
1600
2300
2300
2300
2800
deg. F
deg. F
deg. F
deg. F,
deg. F,
deg. F,
deg. F,
deg. F,
400 deg. F
400 deg. F,
1600 deg. F,
1600 deg. F.
2300 deg. F.
2300 deg. F.
2800 deg. F.
2800 deg. F.
400 deg. F.
400 deg. F.
1600 deg. F.
1600 deg. F.
2300 deg. F.
2300 deg. F.
2800 deg. F.
2800 deg. F.
B.TU
Heat Units
Supplied
100,000
100,000
100,000
100,000
100,000
100,000
U10,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,000
100,0t)0
100,000
Ic. per kw-hr.
Ic. per kw-hr.
Ic. per kw-hr.
Ic. per kw-hr.
$1.00 per M.
1.00 per M.
1.00 per M.
1.00 per M.
1.00 per M.
1.00 per M.
1.00 per M.
1.00 per M.
Rate Paid for
Fuel or Power
$10.00
10.00
10.00
10.00
10.00
10.00
10.00
10.00
per ton
per ton
per ton
per ton
per ton
per ton
per ton
(ler ton
per gal.
per gal.
per gal.
per gal.
per gal.
per gal.
per gal.
10c. per gal.
10c.
10c.
10c.
10c.
10c.
10c.
10c.
.f 10.00
10.00
10.00
10.00
10.00
10.00
10.00
10.00
per ton
per ton
per ton
per ton
per ton
per ton
per ton
per ton
$5.00 per ton
5.00 per ton
5.00 per ton
5.00 per ton
5.00 per ton
5.00 per ton
5.00 per ton
5.00 per ton
30c.
30c.
30c.
30c.
30c.
30c.
per M.
per M.
per M.
per M.
per M.
per M.
30c. per M
30c. per M
Cost per
100.000
B.T.U.
).0385
.0385
.0385
.0385
.0385
.0385
.0385
.0385
.293
.293
.293
.293
.17
.17
.17
.17
.17
.17
.17
.17
.0748
.0748
.0748
.0748
.0748
.0748
.0748
.0748
.042.'
.0425
.0425
.0425
.0425
.0425
.0425
.0425
.020
.020
.020
.020
.020
.020
.020
.020
.0273
.0273
.0273
.0273
.0273
.0273
.0273
.0273
PER CENT
HEAT LOST
Flue
(App.)
HEAT AVAIL-
ABLE FOR
USEFUL WORK
Per Costper
ICent of 100.000
Supplv' B.T.U
7.5
10.5
32.5
48.5
47.5
71.0
0
0
0
0
15
17.5
44
51
60
70
72.5
85
14
17.5
40
55
56
77.5
67.5
94
9
12.5
35
50
50
72.5
60
87.5
11
15
38.5
52.5
55
75
65
90
16.2
20
44
59
60
81
72.5
97.5
10
10
15
15
20
20
57.5 25
87 25
10
15
20
25
10
10
15
15
20
20
25
25
10
10
15
15
20
20
25
25
10
10
15
15
20
20
25
25
10
10
15
15
20
20
25
25
10
10
15
15
20
20
25
25
83.3
80.5
57.4
43.8
42.0
23.0
31.9 0.121
9.7 0.395
0.046
0.048
0.067
0.088
0.092
0.167
90
85
80
75
76.5
74.2
47.6
41.6
32.0
24.0
20.6
11.2
77.4
74.2
51.0
38.2
35.2
18.0
24.4
4.5
81.9
78.7
55.2
42.5
40.0
22.0
30.0
9.3
80.1
76.5
52.3
40.4
36.0
20.0
26.2
7.5
75.6
72.0
47.6
34.8
32
15.2
20.6
1.9
0.326
0.345
0.366
0.390
0.222
0.229
0.356
0.408
0.530
0.708
0.822
1.51
0.097
0.101
0.147
0.195
0.212
0.415
0.306
1.66
0.052
0.054
0.077
0.100
0.106
0.193
0.141
0.456
0.025
0.026
0.038
0.049
0.055
0.100
0.075
0.267
0.036
0.038
0.057
0.081
0.085
0.179
0.132
1.46
ro September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 9
form of tables and charts so as to make this
information available for ready reference,
and at the same time compare the vanous
fuels with electric heat on the basis of thermal
efficiency and the cost per heat unit.
90
m
70
(i, 10
r
"3
«>
wo
lb yo
5 so
;| X
zo
ij
o
400 Decrees Fahr.
1600 Degrees Fahr.
.■" ^^ 'TT ^^ ^^ '^'^ ^"^ i"^ "^"^ ^''i ^'^ '^■^ '^'^ "^^
ZSOODegrees Fahr.
1800 Degrees Fahr
Fig. 1. Curve Showing Heat Available for Useful
Work in Various Kinds of Furnaces
The applications of electric heating arc
being verv^ rapidly extended, and since the
cost of combustible fuels tends to increase
at a higher rate than the cost of electricity,
the point may soon be reached where the
comparative cost even on a B.t.u. basis will
not be unfavorable to electric heat; as a
matter of fact such a comparison even at
present prices is not unfavorable at the
higher temperatures, as will j^resently be
shown.
Table I has been compiled to show the heat
lost and the cost per lUO.OOO B.t.u. for various
fuels including elcctricitv, at temperatures of
400, lOOO, 2.S()0. and 2800 deg. F., the tem-
perature usually required for baking, heat
treating, forging and melting, respectively.
The values of flue losses have been calculated
for \Vi() per cent air, or the theoretical air
required for perfect combustion, and also for
1.50 per cent air, or 50 per cent in excess of
combustion requirements, which represents
more nearly the usual conditions.
Radiation losses of 10, 15, 20 and 25 per
cent respectively have been arbitrarily
assimied for the four temperatures mentioned,
and the same radiation loss has been asstmied
for all fuels, so that they are thus compared
on the same basis ; or it may be assumed that
all the fuels are burned in the same furnace.
Average calorific values of the fuels have
been assumed, and the costs per ton or per
gallon, etc., are in round figures for easy
calculation so that any other costs per ton
or per gallon may be readily compared.
It should be borne in mind that the heat
available for useful work given in the table
is the theoretical maximum for the conditions
stated, and that perfect combustion is
assumed in all cases. The values actually
realized in practice will in many cases
represent a much lower efficiency than the
tables and cun.-es show, except in the case of
electricity, which of course, is all converted
into heat and only the radiation loss escapes.
The values for the heat available as g ven
in Table I are arranged in the form of a chart
in Fig. 1, which shows in a rather striking
manner the relative efficiency of the various
fuels at the four temperatures chosen. The
rapid decrease of efficiency with rising
temperature and increased air supply is at
once apparent.
ix ■*)o 601 aoo looc iKo ><v ea <»w.«wi i.vo Ntioitixieaixoa
Fig. 2. Curve Showing Relative Co«t of Fueli at Various
Furnace Temperatures 100 Per Cent Air
Fig. 2 shows the relative cost of fuels at
various temperatures, with 100 per cent air
supply plotted as curves, and Fig. 3 shows
the corresponding cost for 150 per cent air
supply, the values having been plotted from
the last column of Talkie I. Fig. 1 therefore
RELATIVE THERMAL ECONOMY OF ELECTRIC AND FUEL-FIRED FURNACES 771
shows the high thermal efficiency of electricity
as compared with combustible fuels, while
Figs. 2 and '.i show its comparatively high
cost per heat unit.
For fuel costs other than those chosen,
curves may be plotted by multiplying the
values in the last column of the table by the
actual fuel cost. For instance, for oil at 14
cents ])er gallon, or electricity at 1.2,5 cents
per kilowatt-hotir, multiply the values in the
last column of the table by 1.4 or 1. 2.") and
plot a new curve.
The foregoing is, of course, on a strictly
B.t.u. basis, without regard to the expense of
handling or storing fuel, or to the cost of
repairs, convenience of manipulation, etc.
It will be attempted to show that even at
a higher cost per heat unit electric heat in
many cases is actually cheaper, and at the
same time it offers an opportunity to increase
output and improve the quality of the
product, so that the net result is a very
considerable reduction in the manufacturing
cost. Indeed it is hardly necessary to prove
this as a general proposition, because the
rapid increase in the use of electric heating
equipment is in itself proof that the
advantages far outweigh the additional cost
of heat units.
Coal and coke are relatively difficult and
expensive to store and fire, and natural gas is
restricted to a comparatively small area.
Further, the supply available at present is
such that it is not entirely dependable.
Artificial gas is rather expensive, and it is
not always available in quantities, so that
fuel oil has naturally become the chief source
of heat in many of our industries because
of the convenience with which it can be
stored, handled and distributed, its concen-
trated fuel value, and its formerly low cost
and abundant supply.
Oil has therefore been considered an ideal
fuel, and its use has become widespread.
However, the very considerable increase in
its cost recently and its apparent scarcity
have brought about a condition which is
vers- trying to industrial managers, many
of whom have become interested in electric
heating as a solution of the fuel problem and
of many other incidental problems.
It is proposed therefore to show very
briefly some of the later applications in
which electric heating has met with marked
success, and to indicate some of its possi-
bilities.
The advantages of electric heating are,
in general, well recognized, but the impression
seems to prevail that the cost of operation
is excessive. This is perhaps true in a few
instances, for example, in a case where the
nature of the process permits the recovery
of waste heat from the flue gases; but in
/Cl/
1 1 1 1
Cit,^Gas
:
-iiX-
i
T
t
I
t '
S -
_, ^
inz
Kj-^"
' i
J JZ
5 - J
*S?j
k
/
/ iVlTfl \ i
^rn
/
foSIHf
/I
7 IJU
jf \
t rw
■!; '
M
[
7 T"
' -40 ~
^'
1
.'II \l\Af-"'\
^ -
■-CVr^r'jyir^
A
"7^^5-
d^O -
:^ =
' "'-
-^
-
--^r'-lA-
^ — -
.t"" /coal
, ^-
-' r
L
.*!
^ ^
. — -
- — '
-,-.
: >P
^
^f
=:-
= a-
-^-
"
0 -
1
1 1
Fig. 3.
•ijo MO 600 :ooo iM) i-ioo lax 'fOO^Ja}^^oo^■*x^fOO^ea>Jm
Trmper<jturn-DsgreGS rohrenheit
Curve Showing Relative Cost of Fuels at Various
Furnace Temperatures 150 Per Cent Air
general it is entirely erroneous, as a few
typical examples will show.
Electric heating is also being successfully
applied over the entire range of temperature,
from low temperature ovens for drying and
baking to high temperature furnaces for
melting and refining, so that equipments are
available for practically every application.
Electric ovens for baking japan and enamel
are so widely used and their advantages so
well known that it is hardly necessary to do
more than refer to them here. Their extensive
use for enameling automobile bodies and
parts and in other work where a superior
finish is required is a sufficient argument for
their net economy.
Ovens for baking cores are also yielding
excellent results from the standpoint of
efficiency and net cost, particularly for small
cores in which breakage and loss from uneven
baking usually is a very considerable item.
The uniform distribution and automatic
temperature control possible with electric
heat considerably shortens the time required
for baking and produces perfectly baked
cores, so that the losses are reduced to zero.
Several installations have been made in
wire mills, all of which have shown a
n:
September, 1920 GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 9
substantial saving. The following data are
typical and show the actual cost of baking
steel wire to remove grease, and drying it
after pickling. These are operations which
require no refinement, and it would not be
expected that electric heat would have any
advantages whatever.
The installation referred to has both
electric ovens and coke fired ovens of the
same dimensions. Electricity is by far the
most expensive fuel on a B.t.u. basis at
drying and baking temperatures, as shown m
Figs. 2 and 3, while coke is among the
cheapest fuels. Therefore a direct comparison
in the same plant is interesting.
The normal output of these ovens is
considerablv less than the maximum capacity,
so that "normal" results, or results from
observation are given, as well as results which
might be expected if they were operated at
full capacity.
Baking at 525 Deg. F.
Electric
Normal
Output
Cost of power per net ton of steel $5,087
Annual charges per net ton 7.839
Maximum
Output
$1.19
0.727
Total cost per net ton $12,926 $1,917
Coke
Cost of coke per net ton of steel. $0,921
Annual charges per net ton 21.368
Total cost per net ton $22,289
Baking at 330 Dec. F.
Electric
Cost of power per net ton of steel $0,787
Annual charges per net ton 0.648
Total cost per net ton $1,435
Coke
Cost of coke per net ton of steel. $0,253
Annual charges per net ton 1.766
SO. 171
1.983
$2,154
$0,515
0.187
$0,702
$0,146
0.510
Total cost per net ton
S2.019
$0,656
These figures show, as might reasonably
be expected, that the cost of fuel in the coke
fired ovens is practically negligible in com-
parison with the cost of handling, rcjiairs and
other charges. In the electric ovens the cost
of power is the principal item.
Perhaps a more familiar example is an
ordinary hearth type furnace as used for
hardening tools, dies, cutters, etc. Tests
were rvm on an oil furnace and on an electric
furnace, both having about the same hearth
area and doing the same kind of work at the
same temperature.
Oil Furnace
Inside dimensions . 46 in. long, 24 in. wide, 20 in. high
Average temperature held 760 deg. C.
Specific gravity of oil at 60 deg. F 0.86
Degrees Baume ^^'I?
Weight, pounds per gallon /
Total time test was run. .46 hours
Total weight of oil used. 603. o lbs.
Total gallons of oil used 84 gaUons
Weight of steel heated .26.0 lbs.
Total time heating steel 31.75 hour?
Pounds oil used while heating steel 433
Gallons of oil used while heating steel 60.o
Average gallons per hour heating steel 1.9
Average gallons per 100 lbs. of steel 2.26
Time furnace was empty (holding tem-
perature only) 14.25 hours
GaUons oil used for holdmg temperature
only 23.5 gaUons
Average gallons per hour holding temperature
only .1.65 gaUons
Fuel cos't per hour at 14 cents per gallon .23.1 cents
Electric Furnace
Inside dimensions. 30 in. wide, 36 in. long, 22 in. high
Kw. capacity of furnace 20 kw.
Davs run covered by test 19 days
Total working hours 10. hours
Total weight of steel heated 14ol lbs.
Temperature }_*^ deg. b.
Average kw-hr. per hour in working hours
^ 8.96 kw-hr.
Average kw-hr. per hour empty (radiation
loss only) 8.04 kw-hrs
Fuel cost per hour at 1.25 cents per kw-hr. . . .10 cents
It mav be said that the electric furnace is
provided with'automatic temperature control
and a time switch which throws off the power
at the end of the working day and throws it
on in the carlv moniing so that the furnace
is always ready for use. The furnace and
control equipment are shown in Fig. 4.
It is obvious from these data that the fuel
used in actually heating steel is almost
negligible in either furnace, which of course
is well known. The data show that the cost
of maintaining temperature in the oil furnace
is more than double the cost for the electric
furnace.
This mav perhaps be stin>rising to many
who base their calculations entirely on the
relative cost of B.t.u. It would be still more
surprising if the cost of repairs and fixed
charges could be included in the comparison.
This'cost for the electric fumacc is practically
zero. The figures for this size of oil furnace
are not available to the writer, although they
are probablv available from other sources.
The fact which it is desired to emphasize
is that the oil fumace refemxl to is typical
of thousands in regtilar use for heat treating,
which could be heated by electricity for less
than half the fuel cost alone, to say nothing
of the saving in repairs.
RELATIVE THERMAL ECONOMY OF ELECTRIC AND FUEL-FIRED FURNACES 773
This represents a considerable loss to the
individual manufacturer, and in the aggregate
it is an enormous economic waste.
An interior view of this type of electric
furnace is shown in Fig. o, which however is a
somewhat larger furnace
than that shown in Fig. 4.
A large number of these
furnaces are in daily opera-
tion for such work as an-
nealing, hardening and car-
bonizing, for which they
are particularly well suited.
The furnace shown in Fig.
5 is one of a pair which are
used for carbonizing. All of
these furnaces are equip])ed
with automatic tempera-
ture control.
The furnaces referred to
are of the metallic resistor
type and are suitable for
operation to 1000 deg. C.
(ISOOdeg. F.). For higher
temperatures, such as are
required for forging and
melting, some fonn of arc
furnace is required.
There are a very large
number of electric brass melting furnaces in
operation, most of which have been built in
the last few years; in fact, nearly all the fur-
naces being installed today for melting brass
are electric furnaces.
This is an index of the rate at which
electric heating is being adojrted, and it seems
safe to predict that it will supplant fuel oil
to a very considerable extent for many other
purposes within the next decade.
^^^- }.: '-i
.:ss:2S^
Fig. 4. Box Type Electric Furnace and Control Panel Used for Hardening
Punches, Dies and Cutters
Enonnous quantities of oil are used in the
operations of annealing, heat treating and
forging. The furnaces previously referred to
are ideal for annealing and heat treating, and
operate at maximum economy, and electric
forging furnaces will soon be in
operation, with corresponding
results.
It is unnecessary to dwell
upon the many advantages of
electric furnaces, such as easy
and positive temperature con-
trol, and absence of noise, prod-
ucts of combustion and excessive
heat, as these features are well
known. It is desired rather to
emphasize the fact that, con-
trary to the general impression,
the cost of operation is not ex-
cessive, but in most cases is not
greater and in many cases con-
siderably less than that of fuel-
fired furnaces.
The era of electric heating is
at hand, and it may well be
expected that electricity will
ultimately be as widely used
for an industrial fuel as it is for
indtistrial power at the present
'^^^jtvr ■
1*1^4, 5 Interior View of Box Furnace Used for Carbonizing
tim-e.
774 September, 1920
GENERAL ELECTRIC REVIEW
Vo . XXIII, No. 9
Condenser-resistance Protective Device
By J. L. R. Haydex
General Engineering Laboratory, General Electric Company
The protection of a transmission system from the effects of high-frequency disturbances may be accom-
plished by the use of a condenser and resistance in series connected from line to ground. With a suitable
value of reactance and resistance, the condenser will hold back the low-frequency machine current but will
allow high-frequency current to pass freely into the resistance, which in turn will dissipate the energy of the
disturbance. — Editor.
In the transmission of electric energy the
circuits must be able to operate under normal
and abnormal conditions. One of the abnor-
mal conditions which occur in transmission
systems of any length are high frequency
oscillations created as a result of the read-
justment of the stored energy of the circuit
when connecting and disconnecting circuits,
when lightning strikes a line, or when change
of load occurs, etc.
In general these high-frequency disturb-
ances, or more correctly speaking abrupt dis-
turbances, produced by atmospheric lightning
and by circuit operation, such as switching,
may be divided into three classes:
(1) Impulses: that is, sudden waves of
voltage or current, which are not oscillatory.
(2) Oscillations: that is. periodic disturb-
ances which gradually die out. more or less
rapidly, depending on the dampening effect
of the circuit resistance.
(3) Cumulative oscillations or surges: that
is, oscillations which gradually increase in
amplitude, until destruction of the circuit
occurs, or which are finally limited by increas-
ing energy losses.
To guard against these high frequency
disturbances the use of a condenser-resistance
is of value, serving to shunt the disturbance
and as a high frequency absorber. The
operation of the condenser-resistance pro-
tective device absorbs the energy of the high
frequency disturbances, and thereby keeps
them from building up ti> dangerous voltages.
It consists of a condenser and a resistance
in series, shunted across the circuit which is
to be protected, from line to ground.
The part which does the protecting is the
resistance; the condenser is provided merely to
keep the low frequency machine current out
of the resistance.
If a fairly low noninductive resistance is
shunted across a circuit any high frequency
disturbance entering the circuit is shunted
into the resistance and there rapidly dis-
sipated, so that a building up of the high fre-
quency to higher voltages (such as occur in
high voltage power transformers) is made im-
possible, the disturbance is rapidly absorbed,
and effective protection is afforded.
However, such a relatively low resistance —
a few hundred ohms — permanently shunted
across the high voltage circuit would con-
tinuously absorb a ven,- large amount of
energv', and would have to be very large,
indeed impracticable. For instance, assum-
ing 40U ohms in a (iO.OOO-volt circuit the
resistance would continuously consume l.'jO
amperes at ()0,OU(J volts, or 9000 kw., which
obviously is out of the question. Therefore
a condenser is connected in series with the
resistance.
This condenser practically obstructs the low
frequency machine current. At high frequency
however, the higher the frequency the larger the
current through the condenser and if the con-
denser is large enough and the frequency high
enough, the condenser affords no obstruction.
Thus the condenser is a means of cutting
off the low frequency machine current, with-
out interfering with the high frequency dis-
turbance, and the latter is absorbed by the
resistance.
As regards numerical values, the larger the
condenser (provided it is not so large as to
pass considerable low frequency current) the
more high frequency current it will pass, and
the greater will be the protection. With a
given size of condenser, that value of resist-
ance will obviously be the best which dis-
sipates the energy most rapidly, that is, where
the resistance equals the condenser reactance
at the danger frequency.
For illustration, to protect a l.{,200-volt
alternator by condenser resistance: Experi-
ence shows the danger frequencies in electri-
cal apparatus, that is, frequencies most
liable to build up to high voltages, to be 2(t
to 100 kilocycles, or an average of (U) kilocycles.
Thus a condenser which at (iO kilocycles j)asses
the rated machine current at rated machine
voltage should be able to absorb such high
frequency as may be anticipated; this gives
the capacity reactance, and equal thereto
should be chosen the resistance for maximum
dissipation.
Typical Installations of Electric Mine Hoisting in
South Africa
ByE. B.Bell
Engineer, South Africa General Electric Company, Limited
The successful operation of electric mine hoists in South Africa and the rapidly increasing number of such
installations amply demonstrate the superiority of electric drive for this purpose. The mines are deep, are
located thousands of miles from the plants of hoists manufacturers, and the gold content of the ore is compara-
tively low which factors require high speed, great reliability, and highly efficient operation. In the following
article Mr. Bell describes the mines, shafts, hoists, electrical equipment, and control devices of the latest and
largest installations of electric mine hoists both of the direct-current motor and the induction motor types. —
Editor.
The problem of hoisting from great depths
has received very serious consideration from
engineers on the Witwatersrand in South
Africa.
Since the gold content of the ore raised is
com.paratively low (from 5 to 1 3 pennyweights
per ton) it has been necessary to study very
carefully the questions of efficiency, speed,
and reliability.
Until comparatively recent years steam
engines were used entirely for this class of
service; and it was only after the advent of
cheap power, made available by the inaugura-
tion of the Victoria Falls & Transvaal Power
Company in 190S, that the electrification of
hoists began to be seriously considered.
Since 1908 the great advantages to be ob-
tained from electric hoists have been thor-
oughly appreciated and today South Africa
probably leads the world in their use, includ-
ing among many large installations two
practically identical sets, the largest in the
world, one of which is in operation at the New
Modderfontein Gold Mining Co. and will be
described later.
Although over 75 per cent of the electrified
mine hoists on the Rand are induction motor
driven, due chiefly to the low first cost and
simplicity of installation and operation, the
Ward Leonard system, has been adopted for
the larger sizes not only because of the accu-
racy and simplicity of control but also because
of the increased safety and higher efficiency
on the heavy duty cycles encountered.
NEW MODDERFONTEIN HOISTING
EQUIPMENT
In m.any respects the hoisting equipment
at the circular shaft of the New Modderfon-
tein gold mine will be of interest not only to
electrical but also to mining engineers.
The mine, one of the largest in the world,
is situated about 25 miles from Johannesburg,
Transvaal. For details as to the mine and
shaft the author is indebted to a paper pre-
sented to the South African Institute of
Engineers, by H. Stuart Martin, Consulting
Engineer, and for details of the electrical
equipment to the Consulting Electrical Engi-
neers of the Central Mining and Investm,ent
Corporation.
Mine and Shaft
Since comm.encing milling operations in
1S92 up to the end of 191S, the mine has pro-
duced gold to the value of over $66, 000, 000
from 7,225,480 tons of rock crushed. It
has paid in dividends over $21,000,000, the
disbursements on this account in 1918 having
amounted to 51 J^ per cent on the capital.
The mine was originally opened up by
means of several inclined shafts sunk near the
reef outcrop. As mining progressed it became
advisable to provide additional means for
reaching the deeper levels, consequently an
IS-foot diam.eter vertical circular shaft was
sunk to a depth of 2258 feet, at which point it
intersects the gold bearing stratum, or reef.
This shaft was the first of its kind to be sunk
on the Witwatersrand, but of late a number
of sim.ilar shafts, including some of much
greater depth, have been undertaken. The
ore delivered by several electrically driven
endless rope haulages is collected and stored
in two large bins of 1000 tons capacity, each
having discharge doors which open onto a
level near the bottom of the circular shaft.
The rock discharged from bins is conveyed
to the surface in steel trucks each of six tons
capacity. These travel by gravity to the shaft
where their motion is arrested by means of me-
chanically operated brakes. From this point
they are pushed one at a time by air operated
appliances onto the cage which raises them to
the surface where similar braking and pushing
arrangem.ents are installed. As a loaded
cage is raised another carrying an emtpy
truck is lowered in the adjacent compartment.
776 September. 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 9
Each cage is kept from, swinging in the shaft
by means of four guide ropes 1^4 in. in diame-
ter, suspended from the head gear, and to pre-
vent the two cages from touching as they pass
each other, two division ropes 2 in. in diameter
are suspended from, a girder at the surface.
These guide ropes are held in place by massive
cast-iron weights at the bottom, of the shaft.
The chief advantage claimed for this system
is that when raising men the same platforms
which carry the wheeled trucks can be used
to their fullest capacity, whereas, when skips
are employed which dump ore onto grizzly
bars it is necessarv- to disconnect these skips
and substitute men cages, an operation neces-
sitating large delays in spite of ingenious
mechanical arrangements. It might be in-
teresting to state that even when hoisting at
3500 ft. per min. the absence of vibration
in hoist and guide ropes is little less than
phenomenal.
The load on the hoisting rope totals .33,300
lb. m.ade up of rock 12,000 lb., cage 15,300
lb., and truck GOOO lb., while the rope fully
extended weighs approxim.ately 22,000 lb.
The headgear is of steel N4 ft. high to the
center line of the sheaves which are 18 ft. in
diameter. Fig. 1 shows a view of the head
gear and engine house.
Hoist
The hoist is located in a modern brick and
stone engine room approximately 140 ft. back
from the shaft and consists of a single built
up cylindrical-conical-cylindrical drum car-
ried by a continuous steel shaft 24 in. in diam-
eter coupled at each end to the motor shafts,
as shown in Fig. 2. The smaller cylindrical
portions of the drum are 15 ft. in diameter and
are made of cast-iron, while the large cylin-
drical and conical portions are made up of two
parts each consisting of four cast steel seg-
ments securely bolted together, the diameter
of the former being 24 ft.
In climbing from the smaller to the
larger cylinders, the rope is carried by five
complete machine cut grooved spirals, the
grooves extending the complete length of the
cylindrical portions as well. For adjusting
the length of the ro])es, two cast steel reels are
fitted to the smaller parallel parts of the
drum; displacements relative to the m.ain
hoist being obtained by means of small hand-
operated gear wheels. The total weight of
the drum, shaft and bearings is approximately
180 tons.
The post brakes, 17 ft. in diameter by 14 in.
broad, are built up of m.ild steel plates and
rolled steel channels and are lined with poplar
wood. They are both weight loaded and air
operated with oil cataract devices, either
being sufficient to safely hold the maximum
unbalanced load on one rope.
Compressed air for operating these brakes
is normally drawn from the main supply, but
there is in addition an aiixiliar\- air compressor
of 100 cu. ft. per min. capacity driven by a
25-h.p., three-phase, squirrel-cage motor with
a 250-cu. ft. receiver, and automatic starting
and stopping control apparatus.
The control gear of the hoist is mounted on
an elevated platform sufficiently high to
enable the dri\-er to obtain a clear view of the
drums, ropes and cages as they come to the
siuface.
Below- this platform, which is shown in
Figs. 2 and 3, are mounted retarding and over-
winding de\-ices consisting of large slowly
revolving cams, dri%-en from the main shaft
by substantial gear and shafting, which reg-
ulate the movement of the driver's lever and
consequently the controller, thus preventing
the driver from accelerating the winder in
either direction beyond a predetermined
value and gradualh- bringing the operating
le^'er back to the neutral position sufficiently
early in the event of forgetfulness on his part.
There are also stops on an adjacent revoh-ing
disk which trip the brake operating mechan-
ism in the e^'ent of an overwind.
Electrical Equipment
To the dnmi shaft are coupled two 2000-
h.p., .53..5-r.p.m.. .)()()- volt, direct-current motors
with commutating poles, mounted on a steel
shaft 22 in. in diameter. The armatures are
10ft. loin, in diameter while the outside diam-
eter of the magnet frames are 15 ft. 5 in.
Each motor has 22 main field poles wound for
220 volts excitation. The bearings in addi-
tion to being ring oiled are piped for forced
lubrication, but experience has proved this
to be an unnecessary adjunct.
The motors are connected in series and
recei\-e their power from a motor-generator
set consisting of two I().">()-kw. generators,
a (iO-kw. exciter, and a .^(HMI-h.p. induc-
tion motor. A view of this set is shown in
Power is supplied to this set by the Victoria
Falls and Transvaal Power Co., through the
necessary transformers, oil switches, etc., at
2000 volts, .")0 cycles. The motor is started by
means of a liquid rheostat in the basement
below, operated through a handwheel and
shaft extended to the ground flot^ir.
INSTALLATIONS OF ELECTRIC MINE HOISTING IN SOUTH AFRICA-, 777
in u
u <
§1
■ ■£ e
•a .
ab *o
K E =
778 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 9
As the set approaches synchronous speed
the starter is entirely cut out by means of
a hand-operated brush-lifting and short-cir-
cuiting gear.
The generators are each wound for .500
volts, and in order to successfullv commutate
Fig. 5.
Drum, One Brake, and One of the Two 2000-h.p.
of the Crown Mines' Hoists
the hea\-3' currents at the var\'ing voltages
em_ployed are equipped with commutating
poles and, in addition, with compensating or
pole-faced windings.
The exciter is a standard compound-wound,
self-excited m,achine of fiO-kw. capacity at
220 volts feeding directly onto two main
busses which supply the motor and generator
fields, the brake magnets, and all control
devices.
Control and Protective Devices
For controlling the speed and direction of
rotation of the hoist motors the well known
Ward Leonard system of control is used; or
in other words, the control is effected by
var\-ing the strength and reversing the polar-
ity of the generator fields.
The motor fields are excited in parallel
through equalizing resistors from the 220-volt
busses ; while in series with each field is a field
relay and one of the brake coils, ensuring the
application of the brakes and the opening of
the main circuit breaker connected between
motors and generators in the event of reduc-
tion of the motor field strength beyond a i)rc-
determined value. In addition a motor field
econom}' resistor is connected in the circuit to
permit of long periods at standstill without
injurious overheating of the fields and inci-
dentally^ to reduce the standby losses. The
resistor is short circuited by means of a con-
tactor operated from segments on the con-
troller as soon as the driver's lever is moved
from the neutral position.
The generators receive their excita-
tion indirectly from the exciter busses.
A large resistor, having several tapping
points brought out to two rows of studs
over which fingers of the controller
slide, is connected across these busses.
In this manner a large number of dif-
ferent ^■oltage values are obtainable in
either direction for field excitation.
The generator fields, like those of the
motors, are also connected in multiple
through equalizing resistors, and to-
gether are in series with a weight oper-
ated emergency rheostat normally short
circuited b>- a contact arm and clutch,
which is solenoid controlled. In the
event of the tripping of the main alter-
nating-current oil switch through over-
load or failure of supply, or through the
opening of the main circuit breaker
Motor? through any cause, the brakes are ap-
plied through the operation of brakecut-
out relays and the clutch of the emer-
gency rheostat is released, whereby at first an
increasing resistance is inserted inthegenerator
fields which are finally disconnected from the
supply altogether. It may be interesting to
state in this respect that there is located
within reach of the driver a small emergency
switch which is connected in series with the
aforementioned solenoid, the operation of
which has already been described.
In general the direct-current breaker will
open under any of the following emergenc\-
conditions and in turn will operate the brakes
and emergency generator field rheostat:
(1) Extreme overload.
(2) Loss of exciter voltage.
(3) Overspeed of motor-generator set.
(4) Loss of motor field excitation.
To prevent the generators from building
up when the motors are at rest, a connection
commonly known as the "suicide connection"
is made. As soon as the controller is throwni
into the neutral position, the generator fields
are disconnected from the exciter busses and
contacts made whereby the fields are thrown
across the armatures in such a direction that
the residual and generated voltages buck or
kill each other.
The calculated duty cycle curves of this
hoist are shown in Fig. 0.
INSTALLATIONS OF ELECTRIC MINE HOISTING IN SOUTH AFRICA
Unfortunately there are no test figures
available as to the operating efficiency of the
complete unit, although without doubt ex-
ceptionally good results would have been
obtained when judged from its excellent
operating record.
CROWN MINES' HOIST
The Crown Mines' equipment was ordered
and installed a few months prior to that of
the New Modderfontein gold mine. As the
two mines are in the same group, or in other
words, under the same control, it was thought
advisable when ordering the latter to dupli-
cate the Crown equipment.
Due to certain inherent differences in the
mine and shaft, the mechanical portions
differ widely; even so it was still possible to
order identical electrical and control equip-
ments, a description of which would be only a
repetition of that of the New Modderfontein.
The South Rand shaft of the Crown Mines,
situated on the outskirts of Johannesburg, is
a six-compartment rectangular shaft 3540 ft.
deep.
Double winding drums of the conical-
cylindrical type are used, one of which is
arranged for clutching at will to the driving
shaft, thus m.aking it possible to hoist from
several different levels. These drums and one
of the 2000-h.p. driving motors are shown in
Fig. 5. The diameter of the rope centers at
the small end of the drums is 12 ft., and at
the larger end, 29 ft. S in. The drums are
arranged for 25 complete turns on the conical
portion, 21 turns on the first layer of the
Fig. 7 shows the calculated duty cycle and
rope speeds when hoisting rock from a depth
of 3540 ft.
An interesting feature relating to both the
New Modderfontein and Crown equipments
is that due to the loading limits prevailing
7000
6000
nZ" _L1_U 1 1 11 1J_
,1 Output, of Hoist Motor
,'• — Input to MO Set
J
17 - 'X-^
^t It t
! 1 •' / h : /
tt t/s-- j/
■7. Z "^^x ~,l
-il It"^ J-
i XT' _ _ 2
1 ^ ^-L r
O lO 20 30 ao y^/o 60 70
seconds .'^
- ,'2 1
. ^
?coo
~~
-
~
—r
T1
—
—
—
—
—
n
1
-
—
-
— -
-
-
-
^6000
f
[
QuLpuL Hojst Motors
,
Input from Line
1*°°°
~J
Velocity Ascending Rop
e
• 1
Velocity Descending Rope
-
-
J
i-
-
(O -3000
-
J
7
-
-
.=
'^
-
-
-
-
-
' 1
1 .
._
\
^30GC
1
1
H _
^
-•
"1
*,
'
-N
^
/
/
"2 2000
' ■
'
.
•
■•^.
'
i
f
/
/
-^
,
■
-'
-
f
L nnr
'
'/
/
-
—
—
^
_
'
5 yoo
I
\'
/
£
V
■;
-
-
-y
9,
)
10
20
30
AO
50
60
70
•/
ao
90
njo
/
.
7
,;'''
3000
L
u
L
1 1
_ J
L
L
u
_
_
-\
_
_
Fig. 7. Calculated Duty Cyrle Curves of the Hoist Shown in Fig. 5
cylindrical portion, and 14 turns on the
second.
The weight of ore raised per trip is 1(1,000
lb., that of the skip 8700 lb., and the rope fully
extended weighs 22,300 lb. The total weight
of the drum.s is approximately 270,000 lb.
3000 ■-
Fig. 6. Calculated Duty Cycle Curves of the Hoist
Shown in Fig. 2
on the railways in South Africa; the hoist
armatures had to be completely built up on
site, including core building, winding and shaft
pressing.
That the hoist has given complete satisfac-
tion since its installation in 1917 is fully rec-
ognized on the Witwatersrand; and, although
approximate, results obtained when hoisting
water gave an overall efficiency of 57 per cent
calculated from the amount of power con-
sum.ed at the switchboard and the foot pounds
of water raised.
RANDFONTEIN CENTRAL HOISTS
During the latter part of 1919 an
order was placed with the South
African General Electric Co., Ltd., for
two hoisting equipments for the Rand-
fontein Central Gold Mining Co.
These will be of interest in that when
installed they will constitute, as re-
gards power capacity, the largest
hoists in the world; and as such will
be briefly described, particularly with
respect to the novel features involved.
The m.ine is about 25 miles west
of Johannesburg and the hoists will
be installed at two different shafts
which are situated about two miles apart.
The duty cycles and general characteristics of
the two shafts are similar and consequently
identical sets will be installed. Fig. 8 shows the
calculated duty cycle for balanced hoisting on
which the design of the equipments was based.
780 September, 1921)
GENER.\L ELECTRIC REVIEW
Vol. XXIII. \o. 9
The load to be hoisted froin a depth of 5000
ft. is made up of rock 10,000 lb,, skip 7500 lb.
and rope fully extended '27,500 lb.
Due to the great depth, conical-cylindrical
drums were considered impracticable and
therefore double cylindrical drums are to be
used. Each drum will be 12 ft. in diameter
and 6 ft. wide, necessitating the winding on of
four layers of the 1 ^4'-in. rope. A drum speed
of 100 r.p.m. is to be used, giving a winding
speed of appro.ximately 4000 ft. per min.
The electrical and control equipments will
be somewhat different from those at the New
Modderfontein and Crown Mines. For each
hoist there will be two 16-pole, 2500-h.p., 106-
r.p.m.., 600-volt, direct-current, separately ex-
cited motors coupled direct to the drum shaft,
one on each side. These will be connected in
series with two generators each 2000-kw.
capacity, driven by a .>00()-h.p. induction
m.otor with liquid starter and brush-raising
and short-circuiting gear.
The speed and direction of rotation of the
hoist motors will be regulated by means of a
master controller governing the action of
magnetically operated contactors which will
cut resistance into or out of the generator
fields. Special reversing contactors will be
used for changing the polarity.
The control equipment will provide 15 steps
for both forward and reverse direction of motor
rotation, all of which will be automatically
controlled by adjustable current-limit relays.
The driver can, if he wishes, regulate the
speed by hand control for purposes such as
shaft ins])ection where very low speeds arc
necessan.'.
In addition to the ordinary mechanical
overwind and deceleration devices, similar to
those .described under the New Modder-
fontein equipment, two electrical limit
switches will be arranged for gearing one to
each hoist dnun. These will shut off power
from the hoist motors and make an emergency
application of the brakes in case the skip
travels past the landing platfonn by any pre-
determined amount.
In each shaft there will also be located
switches actuated by the skips themselves to
form a further protection against overwind.
Re-establishment of the generator field after
an emergenc\' shut down will be prevented,
by means of an under-voltagc contactor relay,
until the controller has first been brought to
the off position.
To partly relieve the brakes under emer-
gency conditions a de-ice will probably he
included, whereby, should the direct-current
circuit breaker trip, the armatures of the
hoist motors will be short circuited on suitable
resistances.
Due to the enormous peak loads which the
m.ine power station would have to carry% a
system of relays and interlocks is now being
1
I
1
<
1 i
' '
3i
i
1
1
1
s
1
1
,
1
1
'
1
1
1
dooo
J
1
1
1
-/
;/
1
)
6000
X-Lpot to HoiSt Motors »
D_ _ C 1 1
i;
1,
"^ /
4/-
^
/ /
%
( 1
H* r
\'
r"
u
2k
k
1
' n
i
*/ *' "
s
vj
\
"7
% 200O
^
r,
V-
/
ij
s
\
ZTx.
,w
k
1'
N
X
It 3;
0
1
1
1
^, 1^^
/
f
I
1
T -\
_/
0 lO 20 1 30 40 60X N OO l.'/l
100
peconda
'
V>3(
I
1
t
1
^ (
iL
I
r"
1
jt
1
'*
1
-. it
'
H 5
I
; t-
- h
-
l_
~ it"
SOOO
_
_
_
_
_
'1
Fig. 8. Calculated Duty Cycle Curves for Balanced
Hoisting at the RandTontcin Central Gold Mine
worked out which will prevent the two hoists
from being started simultaneously or ha\'ing
their peak loads ()\erlai>.
For the medium and smaller hoist equii)-
ments, induction motor dri\-e with secondary
control is specified in most instances, chiefly
because of its low initial cost, simplicity of
operation, and installation, and the general
ruggedness and dependability of the motor.
ROSE DEEP HOIST
The drums which are 10 ft. in diameter by
5 ft. face and are connected to the shaft by
means of friction clutches, were formerly
driv'en by a Vales and Thorn douhlo-landem-
compound steam engine with Corliss gear;
and it was only a short while after the instal-
lation of the hoist that it was decided to
change over to electric drive. The conversion
was accomplished in IDIO by removing the
connecting rods and supplying a new disk to
the dnma shaft equipped with an Oldham
coupling. The other hall of this coupling is
carried by an intem\ediate shaft upon which
INSTALLATIONS OF ELECTRIC MINE HOISTING IN SOUTH AFRICA 7.S1
is mounted a double helical Citroen gear S ft.
4 in. in diameter by 15 in. face. Geared to
this bv means of a Citroen pinion 18 in. bv
15 in. is the 900-h.p., 375-r.p.m., 2000-volt, 50-
cycle induction motor.
The mine shaft has a length of 2880 feet
and is at an inclination of 37 deg. to the hori-
zontal. The amount of rock hoisted per trip
is SOOO lb., while the empty skip weighs 630(1
lb., and the T^g-in. rope 2.3 lb. per foot.
Primary reversing contactors, mechanically
and electrically interlocked, are installed
between the line and motor and are operated
from the driver's platform by a lever actuat-
ing a small master controller.
Speed control is obtained by means of a
liquid rheostat in the rotor circuit, which pro-
vides for gradual acceleration. Mine hoisting
requires rather exacting characteristics in a
rheostat of this kind. A high resistance must
be provided for starting and low-speed run-
ning, and a low minim.um resistance is essen-
tial for high-speed operation to prevent
excessive slip and loss in the rotor circuit.
The rheostat under consideration is pro-
vided with two separate sections of electrodes,
one consisting of widely spaced pipes and the
other of a nest of closely spaced plates. At
starting the high resistance section only is
connected in circuit, but as the level of the
liquid rises and the motor speeds up, with a
consequent large decrease in secondary volt-
age, the low resistance section is cut into the
circuit in multiple with the pipes and the
motor acceleration is completed. A small
motor-driven centrifugal pump forces the
electrolyte into the electrode chamber from a
storage tank formed by the lower portion of
the rheostat, where the electrolyte is cooled
by means of coils through which water circu-
lates. The speed and acceleration of the
motor is varied bv means of a lever situated
TABLE I
MOTOR-DRIVEN MINE HOISTS IN SOUTH AFRICA ABOVE 250 HORSE POWER
BY GENERAL ELECTRIC COMPANY
SUPPLIED
H.P.
275
275
320
350
400
400
400
4.50/225
450 22.")
4.50
500
550
550
550
600
835
835
700
900
900
*2000 1
2000 J
»2000 1
2000 /
*t2500
2500
*t2.500 1
2500 J
R.P.M.
1
VoUape
300
500
300
2000
375
2000
273
2000
187 'i
2100
375
2000
375
2000
360/180
2100
360/180
2100
500
500
375
2000
187 >.;
2100
375
2000
375
2000
375
2000
375
2000
.375
2000
.1.1
300
375
300
375
2000
.53.5
1000
.53.5
1000
106
1200
106
1200
Type
Induc-
tion or
Ward
Leon-
ard
Ind.
Ind.
Ind.
Ind.
Ind.
Ind.
Ind.
Ind.
Ind.
Ind.
Ind.
Ind.
Ind.
Ind.
Ind.
Ind.
Ind.
W.L.
Ind.
Ind.
W.L.
W.L.
Wt.
of
)epth
Ore
n Ft.
Raised
per
Trip
1.500
4000
4500
6000
6000
24000
4500
8000
3300
5700
3800
12000
3800
12000
2500
6000
2500
6000
1850
10000
2240
6000
2720
uioon
2800
6000
2800
6000
2.500
6000
2708
8000
.5000
6000
3600
5400
2500
10000
2880
8000
2258
12000
3540
16000
5000
10000
5000
1 0000
Wt.
of
Skip
and
Ore
Rope Rope
Speed > Dia.
Ft/mn.| in In.
8000 1200 , 1
10000 1440 ;
37200 750 1
14000
11270
20000
20000
10000
10000
17000
10800
17000
lOSOO
10800
10800
1.3325
10000
13000
17000
14300
32000
1.500
1600
1500
1.500
2000
2000
1500 I
2000 '
1500
3000
3000
2000
2000 ,
2000 '
2056
2000 I
2000 ;
1
1
1
1
1
1
1
I '4
40500 2
24700 3500
17.500 4000
17500 I 4000
1?4
Type of
Drum
Cyl.
Cyl.
Cy!.
Cy!.
Cyl.
Cyl.
Cvl.
Cyl.
Cyl.
Cyl.
Cvl.
Cyl.
Cyl.
Cyl.
Cyl.
Cyl.
Cvl.
Cyl.
Cvl.
Cyl.
Cyl. con. cyl.
Cyl. con. cyl.
Cyl.
Cyl.
Surface
or
Under-
ground
Surface
Under
Under
Under
Under
Under
Under
Under
Under
Under
Under
Surface
Under
Under
Surface
Under
Under
Surface
Surface
Surface
Surface
Surface
■ Surface
Surface
Where Installed
Falcon Mines. Rhodesia
Knight Central G.M. Co.
Modderfontein "B" G.M.
Co.
Wit waters rand Deep G.M.
Simmer Deet) G.M. Co.
Government G.M. Areas
Government G.M. Areas
Crown Mines. Ltd.
Crown Mines, Ltd.
Cinderella Consd. G.M.
Co.
Xourse Mines, Ltd.
Van Ryn G.M. Estates
Crown Mines. Ltd.
Crown Mines, Ltd.
Bantjes Consd. Mines
Durban Roodepoort Deep
Durban Roodepoort Deep
Cinderella Mine
Bantjes Consd. Mines
Rose Deep, Ltd.
\ew Modderfontein G.M.
Crown Mines. Ltd.
Randfontein Central G.M.
Randfontein Central G.M.
* Motors and generators in series.
t On order.
SUMMARY
No. of
Equipments
19
H.P.
9990
11700
Ward Leonard, above 250 h.p.
5
Totals
24
24690
782 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 9
on the driver's platform operating a hinged
weir in the electrode chamber, and the posi-
tion of this weir determines the level of the
liquid. Due to the limited capacity of the
pump, acceleration beyond a predetermined
rate is prevented, even should the driver
throw his lever suddenly in the fullspeed
position. The liberal dimensions of the weir,
however, permit practically instantaneous
insertion of maximum resistance should
he throw his lever suddenly into the off
position.
The clutches and post brakes are operated
by compressed air. The latter are auto-
matically set, by controlling solenoids wired
across a phase of the main supply, should any
emergency causing the opening of the main
circuit breaker arise.
The installation thus briefly described is
typical of induction-motor-driven hoists on
the Rand. The majority of these have been
operating since their installation under ver\-
hea^n' duty cycles with unqualified satisfac-
tion and exceptionally low maintenance costs.
Opportunities in Office Work
By Anna McCann
Alternating Current Engineering Department, Gener.\l Electric Comp.\nv
In these days, when science is making rapid
strides, when men are successfully accomplish-
ing what the great thinkers of a centur\- ago
conceived as mere dreams, when we look about
us and see what has been achieved by man
on the land, in the water, and in the air.
little do we realize the many details that have
had to enter into the processes of these inven-
tions. Do we, for instance, stop to think of
the part played by the office force of a great
concern ?
During the last quarter of a century the
business of the General Electric Company
has grown extensively, and its office force has
kept pace, offering great opportunity in its
clerical work. Nowhere is the change in the
employee's position better illustrated than
in his conquest of office work. Unquestion-
ably, this is the opirortunc time for the faith-
ful and interested employee.
Of the factors that lead to a successful
office career, those to be first considered are of
a personal nature. This is a day of exacting
requirements and consequently a contempla-
tive employee would do well to complete at
least a high school course, so that upon such
a foundation a thorough knowledge of office
work may be built.
Many, however, who ha\'e risen to respon-
sible positions have not had the benefit of
a high school training; a large number ha\c
had merel}' an elcmentarv' training but pos-
sess the natural ability and encrg\- to become
successful. By persistent study of English,
spelling, and punctuation, and of the duties
of an office employee, they acquire a good
all-round education.
Another requirement of the employee is
trustworthiness. Many persons hold excel-
lent positions for the reason that they are
dependable; they are often more valuable
to the firm than one more brilliant and
clever but less punctual in his duties. They
have the welfare of their employer at heart;
in other words, they are loyal to his interests.
They work not merely for material returns
but because they realize that they are filling
their niche in the great industrial world.
Still another powcrt'ul asset to a successful
office career is a pleasing personality. This
can best be cultivated by reflecting sincerity
and good feeling toward others. The employee
should so conduct himself that a favorable
impression will be made upon those with whom
he comes in contact. A dignified, courteous
bearing is sure to win the respect of all.
The efficiency of an office is dependent
upon each individual employee. In the suc-
cessful pursuance of any worthy project, the
panacea for troubles is co-operation. For
business success, the manager and his subor-
dinates must have at heart the true interests
of the finn by which they are employed. The
manager who j)laces certain responsibilities
upon his subtmlinales. who is not afraid to
intrust them with that portion of the work
which they are capable of pcrt"onning. who
will lead but not drive, is the one who will
obtain best results. He will recognize those
who possess initiative and will be glad to
OPPORTUNITIES IN OFFICE WORK
783
assist in the development of ideas that will
make for the success of all concerned.
Again, in the division of labor, the mana-
ger of an office should show tact; he should
so apportion the work that the brunt of it will
not fall on the most willing, but that each
individual will be allotted his just share. If,
as in other lines of activity, there be one or
two who, on account of superior qualifications,
are more valuable than the others, the wise
manager will do all in his power to have
merit recognized in a substantial way followed
by words of personal appreciation.
In co-operative relationship the manager
and his subordinates will be able to work out
several problems that will make for the effi-
ciency of the office. The exchange of con-
structive criticisms, the keeping abreast of the
times by reading helpful business methods,
and the consideration of the hixman element —
that the individual is subject to human feel-
ings and shortcomings — are forces which help
to promote harmony.
Another factor that must not be lost sight
of in considering this subject is the office
equipment. With the best possible manager
and office force but without proper equip-
ment, no office can be efficiently conducted.
Office ventilation and lighting, attractive and
suitable office furnishings, convenient ar-
rangement of desks, files, and other equip-
ment, including cleanliness, all tend toward
ideal conditions. In bettering the physical
conditions of the office, much time may be
saved and more work accomplished if the
manager and his workers co-operate in the
study of conditions and in the planning of
improvements. They surely are the ones
who know to what extent they are handi-
capped by unfavorable physical conditions
and should, when possible, confer to remedy
them.
With the expansion of the Company's busi-
ness office conditions have not always been
ideal, in fact, they are at present not all that
could be desired; but, when we look back
upon the conditions of twenty-five years
ago and compare them with those of today,
we must admit that the business methods
have been progressive, that the morale of the
office force has, in general, been directed
toward the Company's interests and that
the Company in turn has not failed to show
its appreciation of services rendered. That
the General Electric Company's office condi-
tions are favorable, for the most part, may
be evidenced by the fact that the employee
is made to feel that he is a useful cog in this
great electrical industry to which every em-
ployee owes loyalty.
ERRATA
A number of typographical errors occur in
Dr. Tolman's article, "Relativity Theories in
Physics," published in the June number of
the General Electric Review. Of these
the most important are;
Equations (1) should read
l'=xl i' = xt ;«'=- e' = e S'=S
X
Equations (2) should read
E
v' = v E'-
;V x'
Equation (14) should read
-I — \/-{dx'+ dy' + dr^ -c'-dn=0
Equation (10) should read
3/^
£„
2 — \'' gndxi} + giidxidxi +gudxidX3 +... +etidxt' = 0
or
!/-=f\/J
S,j dxj dXj =0
784 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. <)
A New Co-operative Course in Electrical
Engineering
By W. H. TiMBiE
Associate Professor of Electrical Engineering at the Massachusetts Institute of Technology
The co-operative electrical engineering course described in this article is an attempt on the part of the
General Electric Company and the Massachusetts Institute of Technology to solve the problem of supplying
each year to the manufacturing industry a number of highly trained electncal engmeers who can, in a minimum
length of time after graduation, take responsible positions in the manufacture of electrical appliances. In
this plan the students have the advantage which the Institute offers in the way of theoretical and technical
training combined with the enormous resources which the General Electric Company offers for practical ex-
perience in the manufacture of electric appliances. Most of the theoretical training is given at Carabndge;
the greater part of the practical training is given at Lynn, a distance of about ten miles from Cambridge, in
six thirteen-week periods during the last three years of the five-year course. — Editor.
For the past year the General Electric
Company in conjtinction with Massachusetts
Institute of Technology has conducted a
co-operative course in Electrical Engineering.
The co-operative plan is not new in itself.
The decided advantage of making immediate
connection between the theor\^ as studied
in school, and practice as it exists in the
engineering field, early led to the establish-
ment of co-operative courses abroad. In this
country Dr. Herman Schneider years ago
inaugurated this type of education at the
University of Cincinnati. Similar courses
were put in operation at the University of
Pittsburgh and other technical schools. The
cooperative plan that was started last year
between the General Electric Company and
the Massachusetts Institute of Technology
presents a wide divergence from other plans
both as to the educational principles upon
which the work is founded and in the work-
ing out of the various details of operation.
In the first place, this co-operative course
does not take the place of the old and well-
established Electrical Engineering Course at
the Institute, but is operated in addition to
this course and for a specific and collateral
object. It is avowedly an effort at intensive
training of engineers to meet a specific de-
mand. Perhaps the easiest way to describe
just the field that these men arc being trained
for, is to consider for a moment the various
fields in which electrical engineers are needed.
We can roughly divide electrical engineers
into three rather arbitrary' divisions:
First, the consulting engineer, who is
usually attached to one or inore electrical
com])anies or users of electrical power, to
advise them in cases where expert electrical
knowledge is needed.
Second, the administrative engineer who
may have a large financial responsibility in
addition to his duties as electrical engineer.
Such a man would be called upon to take the
responsibility for the electrical end of any
project in the development and utilization of
electrical power. In fact, both of these first
two types are more closely connected with the
development and administration of projects
for using electric power than with the manu-
facture of the machinery involved in such
projects, and are in fields which are of them-
selves of tremendous magnitude, breadth
and importance.
In the third division, then, belongs the
engineer who is intimately connected with the
design and manufacture of electric machinen,-
and accessories. He superintends the design
and manufacture of most of the apparatus
used by the other two types. His qualifi-
cations call for an intimate knowledge of the
best manufacturing processes and a thorough
training in modem research methods — to
which must often be added the ability for
creative design. This is the engineer that
the General Electric Company and the
Massachusetts Institute of Technology are
endeavoring to train by means of the <-o
operative course in electrical engineering.
He is not confined to manufacturing elec-
trical appliances. This is the engineer that
will be needed in ever increasing numbers as
the country turns more and more to the manu-
facturing industries in order to sustain itself.
The alanuing rate at which the natural re-
sources of the country are being depicted has
made it imperative that the countn,; at large
eventually rely almost entirely upon its manu-
factures. No longer can we depend upon mir
exports of raw materials to pay our bills.
These raw materials, lumber, ores, and oil,
must be manufactured into finished prod-
ucts if the living exjienses of the population
are to be met. Our water ]xiwers mu.st be
utilized and new methods of using our oil
and coal more efficicnth- must be devised.
A NEW CO-OPER.\TIVE COURSE IN ELECTRICAL ENGINEERING
r85
In all this work manufacturing engineers of
the highest type are needed, and become the
most valuable asset of the country. It is these
men who in the last analysis must direct the
operation of the nation's industries; for our
industries cannot compete with those of other
countries unless they are conducted by men
who have large vision, intimate knowledge
of manufacturing details, and a thorough
training in science and scientific methods.
Alanufacturing must be conducted on a
sound financial basis, which means that
processes of production must be so managed
that the total cost of the finished article will
be low enough to compete with the products
of foreign factories. For this task the serv-
ices of an engineer who has a thorough
knowledge of manufacturing processes are
invaluable and his duties multifarious. He
must not only be familiar with the best
methods of production, but he must thor-
oughly understand scientific research, in
order that he may take advantage of dis-
coveries and continually better his methods
of production. This cannot be stated better
than in the words of the Governor of Mass-
achusetts, Calvin Coolidge: "Our pros-
perity comes from our industry and our
industry cannot flourish unless it is directed
with the highest intelligence. Far more in
the future than in the past will this intel-
ligence call for sound training in science and
in its innumerable applications to industry."
The co-operative electrical engineering
course covers a period of five years, the first
two years being identical with the regular
course in electrical engineering at the Insti-
tute; the last three years being divided be-
tween the instruction in theory at the Insti-
tute and training in manufacturing methods
at the Lynn Works of the General Electric
Company. The co-operative features thus
occupy only the last three years, starting in
the summer after the sophomore year. The
course is supervised by a joint committee of
the Institute and the Company. A professor at
the Institute is associated with an officer of the
company in the duty of supervising the prog-
ress of the students while at the Lynn Works.
While at the Works the students are given
a fixed pa^Tnent per week as employees of the
Company, which is the same for all depart-
ments to which the students are assigned.
At the completion of the five year course
the students receive the Master of Science'
degree and the Bachelor of Science degree,
their graduation taking place at the regular
commencement time at the Institute.
The first class was limited to thirtv mem-
bers. The class which entered July Gth this
year consisted of sixty. The size of future
classes may be still greater.
The first week in July the members of the
entire class who have just completed their
sophomore year at the Institute are sent to
the General Electric Company's Works at
Lynn and placed in various shops. Here they
remain for thirteen weeks. At the opening
of the fall term at the Institute, one half of
the students return to Cambridge and pursue
for one term what is practically the regular
course in electrical engineering. At the end
of this term they have a vacation of two
weeks and then go back to the Works of the
General Electric Company for further experi-
ence. The other group then returns to the
Institute for further theoretical instruction.
This schedule is carried out for three years,
each group spending alternately thirteen
weeks at the General Electric Company's
plant and eleven weeks at the Institute. The
vacation of two weeks given the students at
the end of their period at the Institute
divides the year into four equal periods of
thirteen weeks each. The last period of the
fifth year is spent by both groups at the
Institute, so that the two groups graduate
together at the regular commencement time;
yet each group has spent an equal number of
weeks in theoretical instruction and practical
application.
Perhaps the most striking feature in which
this arrangement differs from the former
co-operative plans is in the length of the
periods. This, however, is perhaps the least
important difference. It was endeavored
to arrange the details of the course so that
they would fit into a system of education
which the founders believe is basic. This
system combines the rudiments of S]:)encer's
theory of education with the central idea of
Josiah Royce's. It is an endeavor to develop
all the desirable sides of a student's mind,
character, and body, and at the same time
inculcate in him the spirit of loyalty to his
life work. The course had to be planned so
that these several activities would be carried
on uninterruptedly throughout the periods
which the student spends at the Institute
and at the Works. You will note that in the
scheme as outlined, the following activities
are carried on continuously throughout the
course: Instruction is given in theory,
classes are conducted in some humanistic
study, time is given and facilities provided
for collateral reading, and arrangements are
786 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 9
made for physical exercise and recreation.
The change at the end of each period there-
fore does not mean so much a change in occu-
pation as a change in emphasis, and the length
of the periods thus becomes a rather unim-
portant detail.
The period of thirteen weeks at the shop
and eleven weeks at the Institute followed
by a vacation of two weeks was decided upon
for the following reasons: It was believed
that the period at the Works should be long
enough for the student to spend in each
department an uninterrupted period of suffi-
cient length to become thoroughly familiar
with the men, methods, materials and spirit
of that department. In some departments
the time required for this is practically three
months, and in others it may be as low as one
month. The thirteen weeks period will
therefore meet the conditions required for
those departments in which he must spend
the longest time and does not prevent him
from dividing his time among two or three
departments, in case he is able to master
the details in a shorter time. The same is
true concerning the length of the period at
the Institute. The shortest course at Tech-
nology is ten weeks in length, and all longer
courses are some multiple of ten weeks. The
student is thus able to pursue his studies at
the Institute in units of standard length.
Furthermore, the fact was not lost sight of
that at each change some time was lost by the
student in getting started on the new work.
Therefore, the periods were made of sufficient
length to keep the number of changes as low
as practicable. Finally it was hoped that
the length of the period had l)een so chosen
that the sojourn at the Works would come as
a sort of mental relief and recreation from the
term's work at the Institute. In fact it was
hoped that toward the end of the term's
work the student would begin to look forward
to the change as a welcome break in the
routine of study, and on the other hand, that
the length of the period at the Works would
be .sufficient to quicken his desire and apjxnitc
for further mental concentration and study.
The fact is, the thirteen weeks' period has
pro\'ed that these results have been accom-
plished. Whether a somewhat shorter or
longer period would produce the same results
has not been experimented with, because the
period of thirteen weeks fits into the Institute
calendar in such a way that the periods s])cnt
at the Institute are jiracticably coincident
with the regular Institute terms. So much for
the length of period.
The real vital difference between this
course and other co-operative courses con-
ducted in this country, is the fact that the
co-operating company recognizes that for
three years these students are placed in its
plant for the particular purpose of being
educated and trained as electrical engineers
of a particularly high grade. There is not the
slightest effort or inclination on the part of
this Company to use these students for the
purpose of getting out greater immediate pro-
duction. It is clearly understood that these
students are in the shops and offices to leam
manufacturing methods, and the best relations
of labor, mechanism, and materials in high-
grade production, and to leam them thor-
oughly. Because he can best obtain this
knowledge by actually doing the work himself,
and because the skill which he attains in any
process is the only fair indication of his knowl-
edge of that process, the student is put on
the company's pay roll and becomes part of
its organization. The length of time spent in
each department is regulated not by the needs
of that department but by the value of the
experience to the student. As soon as it is
deemed that he has all the knowledge of the
details of the department that a manufactur-
ing engineer should have, he is immediately
changed to another department. This change
is made upon consultation between the fore-
man of the shop, the officer of the Company,
and the professor of the Institute, who are
associated in conducting the course.
It is not to be inferred from this statement
that the co-operative students do not work
as earnestly and as consistently as the other
men in the various departments. The
co-operative students are graded on the
amount and the quality of the work which
they do in the various shops, and as strong
inducements to do good work are put before
them as are put before the regular workmen.
The only difference between their work and
that of the other employees is that the
students' work is so laid out that they receive
a maximtun amount of experience from each
job and they are kept at it just long enough
to enable them to become fairly proficient in
the necessary- operations. In this way the
minimimi amount of time is spent in learning
the details of manufacture in the different
shops, testing departments, drafting rooms
and engineering offices.
This spirit on the part of the co-operating
company is the fundamental contribution
which this co-operative scheme offers to
engineering education. All the other points
A NEW CO-OPERATIVE COURSE IN ELECTRICAL ENGINEERING
787
of difference between this and other co-oper-
ative courses are made possible and have
their origin in this spirit of the General
Electric Company, which, we believe, is the
true CO operative spirit. It is the one factor
which has allowed us to carry out the plans
of the originators and to make such innova-
tions and experiments as we believe will
improve the curriculum. I will explain a
few of these innovations more in detail later,
but I want it clearly understood that it is
not these changes and departures from the
ordinary curricula which are the important
things in this course, but rather the real co-
operation which the General Electric Com-
pany has offered.
Do not think, however, that this Company
has an entirely unselfish motive in this work.
The officials of the Company frankly confess
that they are pursuing this work because
they believe that by this method they can
procure the future engineers who will be so
badly needed by the Company and by other
industrial concerns in the near future. It was
because they believe that this is the best way
to secure these men that they have entered
into this scheme and after one year's trial
they report that they are more convinced than
ever of the value of co-operative education
conducted along these lines.
Perhaps the spirit in which the work is
being done is most clearly manifested in the
attitude which the ofificials of the Company
are showing in their lectures on manufactur-
ing methods. Once a week one of the super-
intendents has the students come to his
office, and in an informal way talks to them
for an hour concerning the details of the work
of which he is in charge. This feature was
introduced into the course at the suggestion
of one of the superintendents and is given
entirely upon the superintendent's own time.
Some of these men have already prepared
six or seven talks on their work, many of
them illustrated with lantern slides; some
have prepared exhibits of the work in the
shops, showing the material in the different
stages of manufacture, and arranged in order
of the processes. Others have arranged to
have the students come to their shops in small
groups in order to follow through the manu-
facture of some typical article so that they
may become familiar with the output and the
processes before the lecture is given. When
we think of the amount of labor and time that
is involved for the superintendent who does
these things, we can appreciate how the
spirit of the originators of the course has
permeated the personnel of the Works. It
has entered into the attitude of the workmen
themselves, who at all times have shown the
finest spirit of helpfulness and of real co-
operation. Twenty-eight students mingled
for a whole year with the other employees of
the shops and offices and experienced nothing
but extreme courtesy and eagerness on the
part of the men to show them all about the
work and to demonstrate and explain the
details of particular processes. This, it seems
to me, shows the degree to which the Company
has entered into this scheme as a purely co-
operative project.
Of course, you must also appreciate that
the students have done their share in co-
operating. They have in every case entered
into the spirit and work of the shops and
offices and have quickly become a part of the
company. Particular attention was paid to
impressing upon the students the great
factor which human engineering plays in their
chances for success. They are impressed
with the fact that nothing is of more im-
portance than to understand the sterling
qualities of the men with whom they are
working, to study and learn how to adapt
themselves to the personal characteristics
and eccentricities of the various foremen
under whom they are working— that these
things will be of the utmost importance when
they are in a position to direct the work
of others. Therefore during their sojourn at
the shop they get experience not only in
electrical engineering but also in human
engineering, and each man's progress along
this line is followed carefully by those in
charge of the course. Do not get the idea
from this, however, that the students are in
any way coddled. Here is an excellent
chance for them to learn to stand on their
own feet with all kinds of fellow workmen
and all kinds of foremen, and they are com-
pelled to do so. Of course, they make
mistakes and occasionally get into trouble,
but it is better for them to make these
mistakes and get into these few troubles while
they are still students under the supervision
of the instructing staff of the plant and school
rather than later. Each mistake is used as
material to impress upon them the value of
human engineering. They are thus able to
learn valuable lessons without having to
suffer too severely from the mistakes.
Great credit belongs to Magnus W.
Alexander, whose initiative was a principal
force in the origination of the plan, and to
C. K. Tripp, Superintendent of Appren-
788 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 9
tices, of the General Electric Company for the
excellent work they have done along these
lines. They have worked out highly success-
ful methods for utilizing in the most practical
way all the opportunities which the shop work
affords for education and training in the
human element of the job. To be sure, they
have thrown themselves enthusiastically into
all activities of the work and have brought
to the task experience of twenty years in
training executives. But in the development
of the human side, I believe they have
contributed a particularly valuable feature
to the educational program of engineers.
This discussion is a little from the point I
was trying to bring out, but I believe it shows
how the far-sighted policy of the co-operating
company allows us to broaden the training
in every desirable way.
As the third point of difference from other
co-operative courses, should be mentioned the
continuity of the theoietical studies and
humanistic subjects. All through the course,
both while at the Institute and during his
sojourn at Lynn, the co-operative student is
pursuing the study of electrical engineering
theory. At the same time he is taking courses
in the study of English. While the main
purpose of the latter is to train the engineer
in more effective speaking and writing, it
also affords opportunities for enlarging his
vision and creating new interests. Accord-
ingly when a student goes to the Works he
continues the study of electrical engineering
just as though he were at the Institute.
During this period, however, we have found
that he can comfortably cover only about half
as much ground as he would in a like period at
the Institute. This schedule calls for six
hours of study per week and three recitation
hours for the two subjects. Electrical Engineer-
ing and English. Thus, including the one hour
lecture given by the sho]) superintendent each
week, the students spend ioxir hours per week
in recitations or lectures. Their schedule at
Lynn, therefore, comprises:
48 hours per week in sho]) or offices
4 hours ])cr week in class room work
() hours ])er week in jireparation for class
room work.
This schedule allows the student to do all
his study in three evenings a week and yet
get to bed at half-past nine. There is still
left for him three week-day evenings, Satur-
day afternoon, Sunday and Sunday evening
for collateral, reading and recreation. At
the Institute his schedule calls for a total of
forty-eight hours per week in class room and
preparation. Thus while he is at the Works
the student's weekly schedule is increased L
from forty-eight hours of classroom and study f
to fifty-eight hours of a combination of shop
work and mental work.
Inasmuch as the material used in the )■
study of English and the method of con-
ducting the English classes is unique. I |j
believe a word concerning this work would
be interesting. It is a well known fact that
engineering students as a class have an
aversion to the study of English for its own
sake. So it was felt necessar\' to arouse aii
interest in this work before starting it. A
plan which was conceived by Professor H. G.
Pearson, head of the English and History
department at the Institute, was adopted.
At the first session, letters from successful
graduates of the Institute were read to the
class. These letters all brought out the fact
that the higher the engineer rises in his
profession, the greater is his need to be able
to speak well and to write well. Several
instances were cited where promising engi-
neering projects were turned down by com-
mittees or boards of directors because the
engineers back of the schemes were unable to
present their side of the case effectively, while
a lawyerwho knewnothing about the engineer-
ing features was able to talk effectively and
persuasively. The class was then formed into
a committee or a board of conferees and the
session, instead of being called a "recitation
in English," was called a "meeting of the
board." At each of its sittings one of the
students presided, and two or three members.
acting as a subcommittee, presented a report
to the board and advocated its adoption.
This report generally consisted of some
engineering project. For instance, at a
typical session, the class was formed into a
committee from a manufacturing company
about to build a machine shop of a given size,
and requiring a definite amount of energy for
lighting, heating and power. Two members
of the class, acting as engineers of a company
dealing in power plant equipment and sup-
plies. ])ut before the board the advantages of
the comi)an\- owning its own jjower plant.
Two other members of the class representing
the local electric jjower company advocated
that the board purchase central station power.
After the presentation of each side, the class
discussed the matter and finally voted upon
the question. The presentation and dis-
cussion were made without notes, cxcei^t
for numerical data, etc At the close of the
discussion an instructor in Englisli criticized
the session, taking up such jxtints as the work
A NEW CO-OPERATIVE COURSE IN ELECTRICAL ENGINEERING
789
of the presiding officer, showing how he might
have avoided some of the difficulties he
encountered, and how he might have more
easily extricated himself from those he did get
into. The effectiveness of the presentation
of the subject was taken up from the gram-
matical, literary, and psychological stand-
points, the discussion of the class was com-
mented upon from the point of view of its
relevancy, and the vote of the class was
criticized as to whether the class had really
voted upon the merits of the question. A
member of the engineering faculty usually
discussed the whole subject from an engi-
neering standpoint, generally as to whether a
fair statement of the facts had been made. At
the succeeding session of the class, a written
report was always handed into the English in-
structor by the men presenting the projects to
the class and by the secretary of the board.
During the first period at the Works every
man in the class had an opportunity to serve
on two subcom:nittees, to preside over the
meetings twice, and act as secretary twice.
During the second period at the Works, the
sessions in English took a different tack.
On the previous occasion, emphasis had been
placed on effective presentation of engineering
projects in the interest of some company for
which the student was supposed to be work-
ing. Dtiring this term's work the emphasis
was laid upon the effective selling of one's
own service or project. The instructor took
pains to explain the purpose of this term's
work before asking the students to talk or to
write. He showed them that each letter and
each article that they wrote was written for
the purpose of producing a certain effect and
everything in the letter or article should add
to this effect; that any piece of writing was
effective only in so far as it produced the
result that was desired. A letter written for
the ]3uq30se of obtaining the writer a job is
effective if it lands the job. A prospectus
written for the purpose of selling goods is
effective if it sells the goods. In keeping with
this idea the instructor put forward as the aim
of this term's work, effectiveness in writing.
Every bit of writing done that term was to
have a definite purpose and be written to
produce a certain effect. Such exercises as
these were used: Write a description of a
dusty room which will make the reader
sneeze. Write a letter asking for an appoint-
ment that will make the reader really desirous
of seeing you.
When the idea and aim of this course was
understood, and the purpose of these themes
explained, interesting writing competitons
arose and work which formerly was looked
upon as drudgery became an exciting contest.
The class used as a textbook examples of
forceful writing contained in a volume of
short articles by a well known reporter.
I have dwelt at considerable length upon
this first year's work in English because I am
convinced of its importance to engineers. We
still have two more years with this class in
which to continue the work, and plans are
being formulated to develop courses which
combine industrial psychology and instruction
in English in such a way as to appeal to the
engineering student and induce him to put
sufficient effort into the work to make it
effective.
To go back to the differences between this
co-operative course and others, I should like
to mention as a fourth point the provision
which is made for further liberalizijig the
engineering student's education by means of
collateral reading. So important do we con-
sider this side of the engineer's education that
his program has been laid out with the defi-
nite purpose of giving him an opportunity for
reading outside of the prescribed courses.
We thoroughly believe in the desirability of
creating the habit of general reading on the
part of the engineer. This is in line with our
conviction that if the men with engineering
training can be induced to bring this training
to bear on public questions and civic affairs,
a great dynamic force for good will be put into
public life, a force which has behind it all the
power of a highly trained mind.
The fifth point of difference which this
course offers is the intense spirit of loyalty
which has been inculcated in the members
of this course; a loyalty to one another, to the
Institute, and to the co-operating company.
Several things have contributed to bring
about these conditions. First, the closest
connection has at all times been maintained
between the Institute and the group at the
plant of the co-operating company. On three
or four days each week a member of the
electrical engineering department of the
Institute spends a half day at the plant visit-
ing the various shops and offices to which the
students are assigned, or to which they are to
be assigned. In this way, as well as by the
direct contact which is maintained in the two
sessions a week in the class room, the student
is made conscious of the supervision which
the Institute exercises over his work in the
shop. Another feature which has not only
kept the men closer to the Institute, but has
made a closer bond between the m-embers of
the two groups of co-operative students, has
790 September, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 9
been the practice of having the group at the
Institute visit the Lynn Works on evenings
when the officials of the company deliver their
lectures on manufacturing methods. During
good weather the department of electrical
engineering furnished automobiles for the
transportation of the men. The members
of the group at Lynn were also encouraged
to make frequent trips to the Institute and to
enter into the student activities there when-
ever it was possible.
The fact that these students are all taking
thes^meco-operativecourseat the same institu-
tion, and areworking for the same co-operating
company, and finally are living together under
the very pleasantest conditions, quickly
develops this three-fold loyalty. It is this
loj-alty to the Institute and to the Company,
founded upon the student's conviction that
both the Institute and the Company are
planning the work for his highest education
and best welfare, that in the last analysis
must be depended upon to produce results in
the way of conscientious and intelligent effort
in delivering an honest day's work in the
plant and in doing the full quota of study.
Finally the co-operating company relies upon
this loyalty to influence some of the men in
each class to remain with the company after
completing the course.
As the sixth point of difTerence in conduct-
ing this work should be mentioned the fact
that throughout the three years the students
are kept in the plant of the same co-operating
company. Under the right conditions we feel
that this has its decided advantages, because
once the officials of a company have deter-
mined upon a policy, this policy can be
maintained and pursued in every department
through which the students pass. Of course
the company must be of such a size that it
has all the departments in which an electrical
engineer needs experience. The General
Electric Company admirably meets these
requirements. It designs and manufactures
electrical and mechanical machinen,- and
apparatus of nearly every description and of
the widest range of capacities, and many
mechanical devices large and small, some
of the most intricate design. Thus we are
able to offer the student a wide choice in the
departments in which he is to get his experi-
ence as well as in the particular branch he
desires to specialize in. This arrangement
has the advantage of offering the student
the same diversity of work which a large
ntmiber of smaller companies might offer him,
without the disadvantage of a lack of co-
ordination in details and in educational ideals.
The only remaining point of difference I
wish to call attention to is the unusual amount
of theoretical work in the course and the fact
that a Master's degree is awarded by the
Institute upon the completion of the five
years' work. From the beginning of his fresh-
man year to the end of his postgraduate year,
the student pursues one course after another
in mathematical physics. In the middle of his
sophomore year he starts his work in the
principles of electrical engineering and con-
tinues it without a break four terms a year
for the remaining three and one half years
of his course. During the last year the work
at the Institute is composed of advanced
research and creative design, while at the
Works the student is given experience in the
research laboratories of the company, or
upon important work in the engineering and
manufacturing offices. For a successful
completion of this course the institute con-
fers the degree of Master of Science. The
degree of Bachelor of Science, conferred as of
the year preceding the conferring of the
Master's degree, is associated with the
Master's degree. By this the Institute shows
its appreciation of the value of advanced
theoretical training combined with practical
experience, which has been intelligently
planned and carefully supervised.
In conclusion it may be fairly stated that
the one year's experience with this course
has demonstrated that it represents a work-
able program in electrical engineering that
includes the following three fundamental
elements which it has become recognized
should be a part of the training of even.-
modem engineer.
1. It includes eighteen months practical
experience in the industry-; experience which
is not gained hit-or-miss. but experience
which has been carefully jilanned and thor-
oughly supervised. Evcr\- engineer must
sooner or later obtain such experience before
he can fill any responsible ixisition.
2. It provides a greater amount of the-
oretical training than is usually given in a
course of electrical engineering. The practi-
cal experience, therefore, is not gained at the
expense of the theoretical instruction which,
above cven.-thing else, it is the function of the
school to provide.
:?. It is enriched with a wealth of luiman-
istic studies and experience in human engineer-
ing which it is believed is adequate to enable
the young engineer early to lake his rightful
place among his fellow workers in the industry
and among his fellow citizens in social activi-
ties.
A GROUP OF ARTICLES ON
RADIO COMMUNICATION
TWO DOLLARS PER YEAR
TWENTY CENTS PER COPY
GENERAL ELECTRIC
REVIEW
VOL. XXIII, No. 10
Published by
General Electric Company's Publication Bureau,
Schenectady. N. Y.
OCTOBER, 1920
RADIO FREQUENCY ALTERNATORS IN THE RADIO TEST BUILDING OF THE
GENERAL ELECTRIC COMPANY, SCHENECTADY, N. Y.
For
Fractional H. P. Motors
SOME things are so unalterably good, so inescapa-
bly essential, as successfully to resist even the
cheapening effect of low-priced competition. With
the passing of years, the demand steadily increases
for "NORffifl Precision Bearings — an inevitable result
of the steadiU' growing appreciation ot quality as a
factor in serviceability. A "NORfflfl" equipped elec-
trical machine is a better, more serviceable, more
dependable, more protitable machine — tor builder
and buyer.
See that your Motors
are "NORfflfl" Equipped
THE M^HM/^ (g^MF/^MY
l^u§ IglaniKol €S^
Ball, Roller, Thrust and Combination Bearings
General Electric Review
A MONTHLY MAGAZINE FOR ENGINEERS
,„„.-„ „^„,^^^ Associate Editors. B. M. EOFF and E. C. SANDERS
Manager, M. P. RICE Ed.tor, JOHN R. HE^\ ETT . ^^ ^^^^^^ ^^ Adverfsing, B. M. EOFF
Subscription Rates: United States and Mexico. $2.00 per year; Canada. $2.25 per year; Foreign. $2.50 per year; payable in
advance. library and Student Rates: United States and Mexico, $1.50 per year; Canada. $175 per year; Foreign. $2.00 per year;
payable in advance.
Remit by post-office or express money orders, bank checks, or drafts, made payable to the General Electixc Review. Sche-
nectady. N. Y.
Entered as second-class matter. March 26. 1912. at the post office at Schenectady. N. Y.. under the Act of March. 1879.
Vol. XXIII, XO. 1(1 ,y Cn^j7E^r^^,.l,o„r OcTOBER, 1920
CONTENTS Page
Frontispiece; Charles E. Patterson 792
Editorial :
The Development and Commercial Application of Radio Communication
Charles E. Patterson Elected a Vice-president of the General Electric Company . 793
Transoceanic Radio Communication 794
By E. F. W. Ale.xanderson
Radiophone Transmitter on the U.S.S. George Washington S04
By John H. Payne
Duplex Radiophone Receiver on U.S.S. George ]]'asli!ngton S07
By Harold H. Beverage
The .■\lexanderson System for Radio Communication 813
By Elmer E. Bucher
Some Practical Operating Features of Tungsten Filament Electron Tubes .... 840
By W. C. White
The Production and Measurement of High \'acua
Part \'. Manometers for Low Gas Pressures 847
By Saul Dush.man
A Special Form of Phosphoroscope 856
By W. S. Andrews
The Cooper Hewitt Lamp
Part II. Development and Application 858
By L. J. BuTTOLPH
CHARLES E. PATTERSON
Elected a Vice-President of the General Electric Ccmpany Stptember 10. 1920
General Electric Review
THE DEVELOPMENT AND COMMERCIAL
APPLICATION OF RADIO
COMMUNICATION
To one who does not follow the art of
radio communication by a study of its tech-
nical literature, information as to its progress
comes chiefly from newspaper items and
articles of a more or less popular nature. To
the engineer, however, such references are
often not very satisfactory.
This condition is largely due to the fact
that no part of the electrical apparatus em-
ployed is operated, or in most cases is even
seen, by the person sending or receiving the
message. With the telephone, electric light,
and most of the applications of electric power
a part at least of the apparatus is in plain
view, and changes and improvements are
quickly noted and interest aroused to learn
further of the nature of the changes.
To a person just becoming familiar with
this art it would also seem that the amount
of time, effort, and expense devoted to re-
search and development work is out of all
proportion to the relatively small amount
of apparatus built and actual!}' used com-
mercially. Nevertheless while radio com-
munication has been in real practical use for
over 15 years, many new developments are
usually made even today between the iimc
of the preliminary design of a radio set and
manufacture in commercial quantities. This
state of affairs is typical of a new art.
The everyday applications of radio arc
far behind its laboratory accomplishments.
One of the principal reasons for this condition
is the fact that the various phases of the art
have been developed by many widely scat-
tered investigators. To utilize and co-ordi-
nate these many developments and discover-
ies for the manufacture of practical radio
apparatus, a large group of engineers and
ample facilities are necessary.
In the January, 1913, issue of this magazine
an editorial pointed out the advantages of
continuous-amplitude high-frequency alter-
nating currents over the spark system pro-
ducing groups of damped currents for radio
communication and predicted the increasing
use of the former. This prediction is being
strikingly fulfilled, and it is interesting to
note that the articles of this issue on widely
different phases of radio deal, as a basis,
entirely with the production and utilization
of continuous-amplitude high-frequencv cur-
rents. W. C. W.
CHARLES E. PATTERSON ELECTED A VICE-
PRESIDENT OF THE GENERAL
ELECTRIC COMPANY
Charles E. Patterson, comptroller of the
General Electric Company, was made a vice-
president on September 10, 1920.
Mr. Patterson was born in New York City
in 1S66. He entered Princeton University
with the class of '86, but soon after was
obliged to leave college on account of his
father's death. He secured a position in
New York and during his leisure hours con-
tinued the college studies.
In 1885 he entered the employ of the New
York Central Railroad and for 15 years sacri-
ficed his vacation periods in striving for knowl-
edge and advancement. In 1899 he took up
residence in Princeton, N. J.; and by crowd-
ing his New York Central work into three
and one-half days and many nights, and
spending the remaining two and one-half
days each week at the University, he com-
pleted the interrupted college course and
received his degree in 1901.
At this time Mr. Patterson had risen in
the ranks of the New York Central to assist-
ant comptroller. On the same day that he
received his diploma at Princeton University
he was elected comptroller of the American
Locomotive Company, which position he held
for eight years.
Mr. Patterson has had wide experience in
accounting, in fact has been engaged in this
line of work for the past 25 years. Even
while preparing for and completing the inter-
rupted college course, he was studying higher
accountancy and corporation finance. In
1909 he became associated with the General
Electric Company to study its organization
and methods with a view to introducing a
more comprehensive system of accounting
and statistics. His untiring efforts and ag-
gressive business qualifications resulted in
promotion in 1913 to the position of comp-
troller, filling the vacancy caused by the
death of R. E. Steele.
That Mr. Patterson's advancement to a
vice-presidency is well earned is best ex-
pressed by the brief statement made by one
of the Company's officials when informed
of this elevation ; " He is a tremendous worker
and has earned a just reward. It is not
unusual to find him busy at his desk at ten
or eleven o'clock at night. "
A. E. T.
79-i October. 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 10
Transoceanic Radio Communication
By E. F. W. Alexaxdersox
Chief Engineer, Radio Corporation of America
A certain spirit of romance has been directed in turn toward the initial feats of spanning the oceans by
the sailing vessel, steamship, cable, radio, submarine, airplane, and airship. The passing of the romance
attached to the earlier of these means has revealed us in possession of another thoroughly practical and
established transoceanic type of communication. In the line of succession, radio now stands in midst of its
transition stage. Skillful developmental work is hastening the process. The following article briefly reviews
the highly successful Alexanderson system of telegraphic and telephonic radio. Each component piece of
apparatus is described, its function outlined, and the operation of the whole equipment explained. — Editor.
During the last few years a system of
transoceanic radio communication which has
been developed by the General Electric Com-
pany under the direction of the author has
come into use in the United States. This sys-
tem has been adopted by the Radio Corpora-
tion of America which recently absorbed the
interests of the American Marconi Company.
The system has been adopted for future in-
stallations by the British Alarconi Company.
The object of this article is to describe the
principal features of the system.
Historical
The continuous wave system of radio com-
mtinication which is now exclusively used
over long distances was foreshadowed by the
early work of Tesla and Fessenden. In order
to find means for putting his ideas in practice,
Fessenden turned to the General Electric Com-
])any with the request for de\-elopment of an
alternatorwith freqtiencies from .")0,000 to 100,-
OOO cycles, which to that time had
been considered impractical. The
result of this was the develojimcnt of
a 2-kw., lUO.UOO-cycle alternator.*
A number of these 100,000-cycle
alternators were built and one of
these found its way to the labora-
tory of Mr. Marconi who took
personal interest in this develop-
ment. In 191.), Mr. Marconi
made a visit to Schenectady in
order to witness the tests of a
.")0-kw., 50,000-cycle alternator, and
on his invitation this alternator
was installed experimentally in the
transoceanic radio station of the
American Marconi Company in
Xew Brunswick, N. J., which was
not then in use. This provided
the opportunity not only to test the alternator
and other feattires which ha\-e been dex-eloj^cd
in connection with it. such as th? magnetic
amplifier and speed regulator, but gave the
author the opportunity to demonstrate on a
* Alexanderson. General Electric Rsvitw, January, Ifll.'i.
large scale his theory for radiation and
improvements of antenna design.
The experimental demonstrations of teleg-
raphy and telephony which were made dur-
ing 1917 with this installation attracted the
attention of the United States Government
and scientific commissions that were sent to
the United States on account of the war.
A circumstance which particularly brought
the new system into prominence during the
war was the partial failure of the cable system
and the urgent demands for transoceanic
radio communication that developed in con-
nection with American militan,- operations
in France. The .iO-kw. alternator set in Xew
Brunswick, though installed in an experi-
mental wa>", was commandeered for official
transoceanic ser\'ice by the United States
Xavy in Januar\-, 191S, and was operated
until it was replaced by the 200-kw.
alternator set which is now in use in that
station.
Fig. 1. 50 kw. Hi^h Frequency Alternator Installed at the
New Brunswick Radio Station
Radio Transmitting System
Several types of radio transmitting systems
are at present in use with a high degree of
success. The descriptive matter in this ar-
ticle will, however, be confined to the sys-
tem for which the author is responsible, as
TRANSOCEANIC RADIO COMMUNICATION
79.5
represented by the Naval Radio Station at
New Brunswick, N. J.
Generally speaking, any radio transmit-
ting system consists of three essential ele-
ments:
1 . The generator of radio frequency energy.
2. The modulating system whereby the
energy is controlled so as to produce
the dots and dashes of the telegraph
code or the modulations of the human
voice.
3. The antenna or radiating system.
tween these poles are radial with the a.xis of
the disk and are filled with non-magnetic
material so as to present a smooth surface
and thereby reduce air friction to a minimum.
The disk runs between the two laminated
armatures which are cooled by water pipes,
as shown in the photograph. The armature
winding which consists of wire back and forth
in straight open slots, is divided in 64 sections,
each section generating about 100 volts and
carrying 30 amperes. The current generated
by these 64 windings is collected in the air-core
Fig. 2.
200-kw. High Frequency Alternator. Another view of this machine with air-core transformer mounted
over the machine is shown on page 814
Generating System
There are four types of generating systems
of radio frequency energy in use at the present
time.
1 . The spark or impulse generator.
2. The Poulsen arc generator.
3. The radio frequency alternator.
4. The Aacumn tube oscillator.
The system which will be described is of
the type employing a radio frequency al-
ternator. The installation in New Brunswick
contains a .')0-kilowatt alternator shown in
Fig. 1 , which was operated for some time for
experimental purposes with radio telephone
at a wave length of SOOO meters, and later in
transatlantic telegraph service at 9300 meters.
A larger equipment, which has been in
continuous service, consists of a 200-kilo-
watt alternator shown in Fig. 2. The poles
consist of projections on each face of the
disk near the periphery. The slots be-
transformer mounted on the top of the ma-
chine (see page 814). This transformer has 64
independent primary windings corresponding
to the armature windings. The single second-
ary winding of the transformer delivers the
complete output of the alternator. This col-
lecting transformer is thus to be considered
as an integral part of the generating unit;
and for all purposes of calculation the charac-
teristics of the generating unit, such as elec-
tromotive force and current, are given as
delivered from this secondary winding. At
full output the alternator delivers 100 amperes
at an electromotive force of 2000 volts. It
can thus be seen that the alternator is de-
signed for a load resistance of 20 ohms. How-
ever, the same machine might be adapted
for any other load resistance by selecting a
different number of turns in the secondary
of the collecting transformer. The reason
why this particular m_achine was designed for
a high voltage and low current will be given
796 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 10
later in the discussion of the new type of an-
tenna with which it is used.
The 200-kw. alternator when operated at
the New Brunswick wave length of 13,600
meters runs at a speed of 2170 r.p.m. It is
driven by an induction motor through a gear
HIGH FBEQ. SET
MOTOR ALTER.
PDWfff GO
TO Ar^TCf^^A
Fig. 3. Diagram of Speed Regulator for
High Frequency Alternator
having a ratio of 2.97 ;1. When the radio
frequency alternator is used as a source of
radiation the wave length is determined di-
rectly by the rotative speed of the machine.
Thus obviously it is important that the rota-
tive speed should be as nearly absolutely
constant as it is possible to make it. An im-
portant accessory of the alternator set is
therefore the speed regulator. The 5()-kw.
alternator set shown in Fig. 1 is driven by a
direct-current motor, whereas the 2t)0-kw.
set is driven by an induction motor of the
slip-ring type. The oO-kw. set was equipped
with a direct-current motor because the prob-
lem of speed regulation of that type of motor
is somewhat easier. Induction motors were,
however, decided upon for the later types
because alternating-current power is more
easily available in most localities.
Speed Regulator
The sjieed regulator consists of a speed-
determining element and a power-controlling
element. The speed-determining element is
a resonant radio frequency circuit fed by one
of the 64 alternator windings which is set aside
for that purpose. The oscillating energy of
this radio frequency circuit is associated by
magnetic couplings with a rectifying circuit in
which the radio frequency energy is changed
into direct current. This rectified current
in turn actuates the controlling magnet of
a vibrating regulator of the tyi)e that is gen-
erally used for voltage regulation in power
stations. When the driving motor is a direct-
current motor it is easy to see how this vibrat-
ing regulator may be made to control the
speed by regulating the voltage of the power
supply to the motor. In order to accomplish
the same object with an induction motor some
new features have been introduced.
An ordinary induction motor is operated
at constant potential. When the motor runs
light it draws from the line a magnetizing
current which is almost wattless. Thus it
operates at a low power factor. When the
motor is fully loaded, it draws power at a high
power factor, the motor used having a power
factor of 90 per cent.
When the New Brunswick station was ad-
justed for operation, it was found that a wave
length was desired which required the induc-
tion motor to work at 19 per cent slip. The
rheostat in the secondary of the motor could
easily be adjusted so that the motor woiild
deliver the desired power with full load at
19 per cent slip. However, inasmuch as the
output of the alternator varies continually
with the making of dots and dashes of the tele-
graph code, the motor is alternately loaded
and not loaded, therefore, the tendency
would be for the motor to speed up during the
intervals. If the potential of the power
supply to an induction motor is varied the
motor torque varies by the square of the
voltage. It is easy to show, by the theory-
of the induction motor, that if a motor con-
Fig. 4. Vibrator Regulator for High Frequency Alternator
sumes jiower at 90 per cent power factor at
full load and the load is reduced to '4 by the
reduction of voltage to ' o, the power factor
will remain 90 per cent. In fact, it will al-
ways consume power at 90 per cent power
factor regardless of its load if the voltage
TRANSOCEANIC RADIO COMMUNICATION
797
supply is adjusted accordingly, and so long
as the secondary resistance remains constant
and the speed remains constant.
Thus it may be said that the standard
method of operating an induction motor is
at constant potential and variable power fac-
tor. The method of operating the driving
motor of the radio set on the other hand may
be characterized as variable potential and
constant power factor.
The problem which thus presented itself
was to find means for varying the applied
voltage in accordance with the action of the
speed-determining element, and this has been
done in the following way:
Between the motor and the power supply is
introduced a choke coil with an iron core, the
permeability of which can be varied by satu-
ration. The change in permeability is pro-
duced by a direct current which is control-
led by a vibrating regulator. When the motor
carries full load the iron core is saturated
so that the choking effect is practically zero.
At fractional load, the choking effect is auto-
matically adjusted by the regulator so that
the motor delivers at all times the power re-
quired to hold constant speed. The motor
itself operates at all times at its maximum
efficiency and power factor, but the power
factor of the current drawn from the lines
varies with the load. Thus when the motor
operates at J4 load, the power factor of the
line is 45 per cent, while the power factor of
the motor is 90 per cent. The circuits of the
regulator are shown in Fig. 3 and the photo-
graph of the vibrator regulator in Fig. 4.
Modulating System
The method of controlling radio frequency
energy involves an apparatus which has be-
come known as the ' ' magnetic amplifier.
This device is described in a paper by the au-
thor in the Proceedings of the Institute of
Radio Engineers, January, 1916, and there-
fore needs to be referred to only briefly. The
magnetic amplifier is a device which is phys-
ically of the nature of an oil-cooled trans-
former. The iron core which is made of fine
laminations, is designed in such a way that
the magnetic permeability of the iron core
can be varied by magnetic saturation. By a
special combination of tuned circuits, as
shown in Fig. 5, it has become possible to
separate the controlling current from the radio
frequency current so that a comparatively
weak current of a few amperes controls as
many hundreds of amperes in the antenna.
When the transmitting station is used for
telegraphy, the magnetic amplifier is con-
trolled by the telegraph relays which are a
part of the wire telegraph system. During
the war service the telegraph key was oper-
ated in the centralized operating room of the
Naval Communication Department in Wash-
ington. When the station is used for tele-
Fig. 5. Diagram Showing Method of Controlling Heavy
Antenna Current by Means of a Combination
of Tuned Circuits
phony the controlling current is an amplified
telephone current.
While the magnetic amplifier has proved
to be a very satisfactory and reliable con-
trolling device for ordinan,^ telegraphy, its
particular advantages are most prominent in
high speed telegraphic transmission and
telephonic transmission, on account of its
instantaneous magnetic action without any
arcing contacts. Fig. 6 shows an oscillogram
of radiation at 100 words per minute and a
photographic record of reception at the same
speed. Fig. 7 shows the telephone modula-
tion of the antenna current when Secretary
Daniels was speaking over the telephone line
from Washington, controlling the output from
the New Brunswick station, thereby transmit-
ting his voice to President Wilson 's ship at sea.
The Multiple Antenna
The antenna of the New Brunswick station
represents a new departure in the method
of radiation. The old antenna structure was
originally one of the horizontal Marconi
antennae, .5000 feet (1500 meters) long, 600
feet (180 meters) wide, supported on towers
400 feet (120 meters) high. The original
antenna had a resistance of 3.8 ohms.
The antenna as now operated has a resist-
ance of 0.5 ohm, distributed approximately
as follows:
Ohm
Radiation resistance 0.07
Tuning coils and insulation 0.10
Ground resistance 0.3^
Total multiple resistance 0.5
79S October, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, No. 10
The reduction in total resistance of the
antenna is due to the reduction of the ground
resistance. While the old antenna had one
tuning coil located in one end, the new antenna
has six tuning coils as shown in Figs. 5 and 8.
<x3C/u.oai^AM or
M 0 K//Vg p /I R r s
N£iv eaiNswick' AT 100 nvoKPs pee MtNtm.
Fig. 6. Oscillograph Record of Transmission and Photographic
Record of Reception at a Speed of 100 Words Per Minute
Theory of the Multiple Antenna
The im.provements in radiation efficiency
which have been demonstrated by the use of
the multiple antenna can be explained in terms
of the Hertzian equation for radiated energy-:
Watts radiated = 1(100
m
This equation takes into account only the
oscillating current /, the effective height /i,
and the wave length X, but not the horizon-
tal dimensions nor the capacity of the antenna.
This explanation is accurate and convenient
and reduces the radiation efficiency into terms
of ground resistance. The im.provem.ents
of radiation efficiency by multiple tuning are
thus indicated by the measured reduction of
ground resistance as stated above. In ac-
cordance with this explanation the im.proved
efficiency is gained by spreading the antenna
over a large ground area and reducing the
ground resistance by leading the charging
current of the antenna to ground through a
multiplicity of tuning coils located far apart.
The minimum, ground resistance in any one
point is of the order of magnitude of 2 ohms,
and thus by utilizing several of these ground
connections in multiple the total resistance
of the antenna can be reduced.
The above explanation is convenient to
use in practical calculations, but the author
has found radio engineers and scientists not
always ready to accept this explanation with-
out further proofs, and quite justly so. be-
cause the Hertzian equation is only a con-
densed mathematical formula in which such
essential physical facts as horizontal dimen-
sions and capacity are apparently disregarded.
It shall therefore be attempted to present
the physical conception of radiation which
led the author to the development of the
multiple antenna.
There are two forms of radiators known
and in use at the present time; viz., the
electrostatic radiator and the electromag-
netic radiator. Of these the electrostatic
radiator is used m.uch more extensively.
Combination forms of radiators are some-
times used, such as the Man-nni <lirfcti\o
rSB. tt, /9/9
Fig. 7. Oscillograph Record of Antrnna Current Modulated by Radio Telephone
TRANSOCEANIC RADIO COINIMUNICATION
799
antenna and Bellini-Tosi antenna. The or-
dinary antenna with large capacity and
moderate height compared with the wave
length and large loading coils is almost ex-
clusively an electrostatic radiator. The purely
magnetic radiator is the closed magnetic
loop. While it is true that in any oscillator
circuit the energ\- alternately appears in
electrostatic and electromagnetic form, the
electrostatic radiator is characterized by the
fact that the energy when appearing in elec-
trostatic form is spread out over a large volume
of space, whereas the energy when appearing
in electromagnetic fonn is confined to a
tuning coil of small dimensions which does
not spread the magnetic lines to any appre-
ciable distance. The m.agnetic radiator is
characterized by the fact that energy when
appearing in magnetic form spreads over a
large volume of space, whereas the energy
in electrostatic form appears in an artificial
condenser. The radiation by an electro-
static antenna is prodiiced by the electrostatic
lines of force which reach as far away as one-
quarter wave length and there produce a
secondary electromagnetic field, thus throw-
ing off energy in the form of electromagnetic
waves; similarly, radiation takes place from an
electromagnetic radiator by the lines of force
which reach to a distance of one-quarter wave
length and produce an electrostatic field.
In accordance with the author's concep-
tion of the electrostatic radiator, it is suf-
ficient to create an electrostatic field which
has lines of force reaching into distance. It
is conceivable that an insulated plate may be
laid directly on the ground, but have such
dimensions and such potential charge that
it will throw electrostatic lines far into space,
and thus will become an effective radiator if
charged with a highfrequency oscillating poten-
tial. The distant effect is obviously propor-
tional to the size of the plate and to the poten-
tial applied . The height of the plate overground
and the charging current between plate and
ground on the other hand would appear to be
immaterial; thus it would appear that the two
quantities, height and charging current, which
exclusively determine the radiation efficiency
in accordance with the Hertzian equation,
are non-essential, while the potential and the
hoiizontal dimensions which do not appear
in the Hertzian equation are essential. This
apparent contradiction can, however, be
easily explained. It is only two methods
of stating the same fact, but these two state-
ments represent dift'erent points of view which
are apt to lead to different developments of
the technique. i \
Returning to the large plate laid on the
ground : it is evident that the closer the plate
is to the ground, the greater is the charging
current at a given potential. Furthermore,
the larger the plate the greater is the charg-
ing current required to maintain the same po-
SINGLE TUr^ED EQUIVAl.£l\IT CIRCUIT
MUL.TIPi.£ ANTCt^MA
f^ULTIPLE TUNeO EOUIVALerJT CIRCUIT
2 fa ?<^ / S_ 1-0 _ 0-5^
Fig. 8. Diagram of Maltiple-tuned Antenna Circuit.
This circuit has six tuning coils instead of one
as in older antennas
tential. Thus the horizontal dimensions of the
plate are in the Hertzian equation expressed
by the charging current. The voltage of
the plate is in the Hertzian equation expressed
by the height over ground, ina,smuch as a
greater height corresponds to a lower ca-
pacity and consequently to a higher voltage.
Thus the height in the Hertzian equation
corresponds to the charging voltage in the
electrostatic conception of radiation from a
low ground plate. So long as we are satis-
fied that the Hertzian equation is a general
and true expression, though it represents a
translation into a mathematical language
which does not directly correspond to the
physical phenomena, we can continue to use
the Hertzian equation for the purpose of
calculation without losing sight of the phe-
nomena which really take place.
800 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 10
The process of reasoning which led to the
development of the multiple antenna was
briefly :
The tendency in long distance communi-
cation has been towards longer wave length.
The Hertzian equation indicates only one
way of adapting the antenna to longer waves
and that is by increasing the height. This
has been carried to the extrem.e by building
towers approaching the height of the Eiffel
tower. The practical limit for height of tower
has thus been reached and the economical
limit m.uch exceeded. If we look at the an-
tenna as an electrostatic radiator, the ques-
tion presents itself: Why go so high in the
air? The only object of so doing is to throw
lines of force to great distance. Why not
accom.plish the sam.e result by mounting
the antenna wires at a moderate height, say
100 meters, but covering sufficient ground
area for the purpose' The Hertzian equation
confirm.s the correctness of this reasoning,
inasm_uch as two unit charges at a height of
100 meters will produce a radiation equal
to one unit charge at a height of 200 meters.
The cost of mounting an aerial capable of
taking two unit charges at a height of 100
meters is very much lower than the cost of
motmting an aerial for one unit charge at
200 meters. But if we are aiming at a
strength of signals corresponding not to one
unit charge but to ten units of charge, we
have the privilege of extending the antenna
indefinitely in horizontal dimensions, whereas
further increase of height would be impossible.
The antenna in New Brunswick on which
this theor\- has been dem.onstrated is 1.6
kilometers (one mile) long and the new an-
tennae which have been designed for future
and more powerful stations are (j.4 kilometers
(four miles long).
The reason why the ground resistance of
the original antenna at New Brunswick was
as high as 3.8 ohms was the fact that charging
currents passing between the aerial and ground
at a point one mile away from the station
had to be carried over one m.ile of ground
and back again through a mile of antenna
wires.
Multiple tuning consists in connecting
inductances between the antenna and ground
in various places. The current in the induc-
tance lags 90 deg. behind the antenna poten-
tial, whereas the electrostatic charging cur-
rent leads by 90 dcg. These two currents
which are equal and of opposite phase thus
neutralize each other and it is possible to
maintain a high antenna potential without
carr>'ing high charging currents from the
transm_itting set to the various distant points
of the antenna. In accordance with the
electrostatic theory of radiation given above
it is only necessary- to m.aintain the antenna
potential because it is only the lines of force
which reach a distance of a quarter wave
length that produce radiation. Thus by
using the expedient of neutralizing the elec-
trostatic charging current under the antenna
by corresponding inductance currents the
energ\- losses are avoided, which are otherwise
incident to carr\-ing currents long distances
through the antenna wires and back again
through the ground. The only current that
it is thus necessan,' to distribute through the
antenna wires is the energy- current, which
is about one half of one per cent of the charg-
ing current. This is the reason for the
measured reduction of energy- consumption
of the New Brunswick antenna at the ratio
of 3.8:0.5 ohms at the same average charging
potential and the sam.e radiation. The dis-
tribution of currents on this multiple timed
antenna at New Brunswick is shown in Fig. S.
What actually takes place is: The tuning
coil to which the alternator is connected
transforms the energy- of the alternator into
a power supply at a potential of (iO.OOO volts,
and each of the oscillating circuits draws
energfy from this power supply at that volt-
age. Thus the energy- current consumed by
each oscillating circuit is only 0.5 ampere. It
can thus be seen that while the total oscillat-
ing current of the antenna is (iOO amperes,
the energy- current which flows horizontally
from the power source to the multiple oscil-
lating circuits is only a total of 2.5 amperes.
In other words, the energ\- which is delivered
by the first tuning coil in the form of 100 am-
peres at 1800 volts is transformed by the first
oscillating circuit and distributed as in a
transmission line from which 0.5 ampere at
60,000 volts is drawn in five places. The
analog>- between the m.ultiple antenna and a
high tension power distribution system is
thus apparent.
This point of \-iew is a departure from the
conventional theoni- of radiation; but it must
be remembered that there was a time in the
de\'elopment of electric power technique when
the introduction of the high tension multiple
distribution system was a radical departure.
When an antenna is built which is four miles
long, it may be considered as four antennae,
such as a New Bnmswick antenna connected
in multiple. The ground resistance will then
be reduced again to one quarter of the mul-
TRANSOCEANIC RADIO COMMUNICATION
SOI
tiple tuned resistance of the New Brunswick
antenna. Calculations of the radiation ef-
ficiency of such large multiple antennae indi-
cate that it will be practical in the future
to construct radiators with a radiation ef-
ficiency of as much as 50 per cent or more,
instead of the radiation efficiency of a few
per cent that has been com.m.on up to the
present tim.e.
Directive Radiation
The multiple antenna as described in its
sim.plest form is adjusted so that the radia-
tion from, each of the individual oscillators
is in phase. If, however, the antenna dimen-
sions are so chosen that the phase displace-
m.ent of the travelling wave between the dif-
ferent radiators becom.es an essential factor,
it is possible to obtain directive radiation.
The radiated wave will then not be a simple
circular wave, but an interference pattern
which may be treated like the corresponding
phenom.ena in light and sound waves. Fur-
therm.ore, the phase displacem.ent of the os-
cillations of the individual radiators m.ay be
regulated by tuning. Thus a variety of
interference patterns m.ay be created and
analysis of these possibilities shows that
an efficient unidirectional radiation by such
methods should be possible.
Methods for unidirectional radiation have
been established through the well-known
work of Bellini and Tosi. Through the cour-
tesy of Mr. Bouthillon, of the French post
office, results of tests m.ade in France have
been placed at the disposal of the author
which show conclusively directive radiation
by the Bellini and Tosi antenna.
With the dimensions of antenna used up to
the present time efficient directive radiation
has not been practical. It has. however, been
proved by various tests that the system of a
central power source and a distribution sys-
tem of energy to a large number of multiple
radiators place means at our disposal for
constructing radiators of dimensions of one
wave length or more. The New Brunswick
antenna (1500 meters or 5000 feet long) has
a minimum wave length of 8000 meters as a
single antenna, whereas it can be operated
as a multiple antenna at 2000 meters wave
length. A detailed analysis of the possibili-
ties of multiple radiation would fall outside
of the scope of this article, but the author is
in position to predict with confidence that
directive radiation on a large scale will not
only prove practical but will be the most
effective method of radiation.
To add directive radiation to the proposed
program for increasing the capacity of radio
traffic would perhaps be premature until it
has been demonstrated on a large scale. How-
ever, it deserves mention in order to show
that new principles which may be utilized for
still greater expansion of the radio technique
ha\-e not yet been exhausted.
The Receiving System
The principal problems of the present day
reception of radio signals are the avoid-
ance of disturbances due to atmospheric con-
ditions and other radio stations. The solu-
tion to both of these problems appears to lie
in the development of unidirectional recep-
tion. The old type of static receiving an-
tenna receives signals and atmospheric dis-
turbances equally from all directions. The
magnetic loop antenna has a bidirectional
characteristic and is somewhat of an im-
provement over the static antenna. The
investigations undertaken by the author
during the war period and after on selective
reception have led to a type of receiver which
was adopted by the United States Navy for
transoceanic reception and has become known
as the "barrage receiver," because it was
developed to meet certain military require-
ments in France.
The Barrage Receiver
The barrage receiver is fundamentally a
unidirectional receiver. The principle of
tmidirectional reception was first developed
by Bellini and Tosi. While the unidirec-
tional Bellini-Tosi receiver has been used as a
direction finder, it has, to the knowledge of
the author, not been used to any extent for
reception of long distance signals. The
Bellini-Tosi receiver is based on the prin-
ciple of receiving the signal through two
antennas of different characteristics and neu-
tralizes the signals received from one direc-
tion by a system of balancing.
The principle followed b}^ the author in
devising the barrage receiver was :
1. That the antennas or energy collect-
ors should be aperiodic, because the
balance of two tuned circuits is fun-
damentally very delicate and difficult
to adjust for a perfect balance.
2. That the balancing should consist in
neutralizing the electrom.otive forces
in the aperiodic antennae before those
electrom.otive forces have had a chance
to create oscillating currents. The
phase shifting device should therefore
be aperiodic.
S02 October, iy2U
GENERAL ELECTRIC REVIEW
Vol XXIII. Xo. 10
3. The two or more antennae should be
of the same character: in other words,
it is preferable to balance a magnetic
exposure against another magnetic ex-
posure rather than against an electro-
static exposure.
^ <^ ® O
Fig. 9. Barrage Receiving Set Assembled in Carrying Case
The unidirectional Bellini-Tosi receiver
works on the principle that the electromag-
netic and electrostatic exposures are 90 deg.
out of phase. The barrage receiver takes
advantage of the geographic phase displace-
ment in the wave as it travels over thi' sur-
face of the earth. In the first barrage receivers
which were installed, the antennae consist of
two insulated wires laid on the ground a
distance of two miles (3.2 km.) in each direc-
tion from the receiving station. It was
originally intended by the author to mount
the wires on poles, but the easier procedure
of laying the wires on the ground was adopted
at the suggestion of Comm.ander A. Ho>'t
Taylor, and the arrangement has proved
entirely satisfacton,'. The barrage recei\-ing
set, photographs of which are shown in Figs.
9 and 10, consists of a standard receiving
set, combined with a phase rotator set. Fig.
10 shows the receiving set proper lifted out
of the box. This part of the set is arranged
so that it can be used as an ordinan.- recei\-ing
set. When used as a barrage receiver, a
condenser is used in place of the antenna and
the set is coupled to the aperiodic antenna by
the phase rotator set. The diagram, of the
phase rotator set is shown in Fig. 1 1 . Each
antenna is connected to ground through an
intensity coupler, the secondaries of the in-
tensity couplers being connected to the pri-
mar\' of the phase rotators. Each phase rotator
is built on the principle of a split phase in-
duction motor or induction regulator. A
single-phase current introduced in the prim.ar}-
is split into a quarter-phase current which
produces the equivalent of a rotating mag-
netic field inducti^-ely related to the secondan.-.
By adjusting the position of the secondarv-
ceil the electrom.otive force induced in it
may be made to assum.e any desired phase
relation to the primary,' voltage. The re-
ceiving set proper when used with the barrage
receiver has all the normal characteristics of
a standard receiving set. A signal originating
in any direction whatever may be neutralized
by adjustment of the intensity couplers and
phase rotators. This adjustm.ent is ver>- easy
to perform., even by an inexperienced opera-
tor, and is perfectly stable after it has been
made.
An experim.ental barrage receiving set was
operated for several m.onths of the sum.m.er
and fall of 191S. about three miles from
the New Brunswick, X. J., radio station.
Records were kept on the reception of Euro-
pean stations during the operation of the
New Brunswack station. As the New Bruns-
Fig. 10. Barrage Receiving Set Assembled on
Operating Bench
wick wave is 13,()00 meters and|[the Carnar-
von, Wales, wave is 14,200 meters, the re-
ception of Carnarvon was the hardest test
to which the set could be put. It was found
that in spite of the overwhelming intensity
of the Xew Brunswick signals on an unbal-
TRANSOCEANIC RADIO COMMUNICATION
803
anct'd receiver, the barrage receiver could be
adjusted so that the transmitted wave not
only did not interfere with the Carnarvon
signals, but the New Bnmswick signals could
be rr.ade entirely inaudible. During these
tests it was found that the directive charac-
teristics of the barrage receiver was a ip.aterial
help in reduction of inteference by static and
strays, as it was found very frequently that
solid copy could be obtained by proper
directive adjustm.ent, while the signals were
practicalh' unreadable with ordinary m.ethods.
The im.i^rovement in reception of signals by
the use of the barrage receiver depends upon
the highly directive qualities of this receiv-
ing system..
A rather surprising characteristic was dis-
covered by the use of the barrage receiver. It
was expected that this receiver could be used
to neutralize signals from, all directions except
the direction clo.se to the signal to be received.
As a matter of fact, it was found that inter-
ference originating from the same direction
as the signal could be neutralized. This was
first discovered in the New Brunswick instal-
lation. Signals from. San Diego, Calif.,
right in line with the transm.itting station,
could be received without great reduction in
intensity, while the set was adjusted so as to
neutralize the transm.itting station. The ex-
jjlanation for this is the fact that the -wave
front of the nearby station is curved and the
radiation diverging, whereas in the case of
the far away .station the radiation is parallel.
The receiving antenna covers a space of four
m.iles (6.4 km..) and in this space there is
sufficient divergence of the radiation from
the nearby station so that an adjustment can
be made whereby the diverging and parallel
radiation have different effects upon the re-
Fig. 12.
Diagram Showing the Horizontal Intensity
of Various Forms of Antennse
HO^tzOfrj^t M/freMf*^
MOrttOMT^tt. A/*T£M/^J^
Fig. 11. Wiring Diagram of Phase Rotator Set
ceiving set. The phenom.enon is com. parable
to the focussing of a field glass on nearb}' and
distant objects. In this case we have a radio
field glass of four m.iles (6.4 km.) diam.eter;
and for such dim.ensions, the focussing effect
is sufficient even at considerable distances to
produce an effective discrim.ination.
While the barrage receiver was worked out
prim.arily to avoid interference in trans-
oceanic communication, it may also be
found useful for simultaneous sending
receiving from, sm.all shore stations or
ship stations. In such cases it may be
u.sed to neutralize interference from
any other ship or shore station. By
the use of a double set of phase
rotators the barrage receiver may be
used to neutralize two stations in
different directions simultaneously,
and this principle m.ay be carried still
further if desired. It is thus hoped
that this developm.ent will open up
new possibilities in dealing with a
problem, which is perhaps the most
im.portant in the immediate future;
that is, to meet the demands of radio
technique for a rapidly increasing
number of svstems of communication.
804 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 10
Radiophone Transmitter on the U.S.S.
George JVashitigton
By JoHx H. Payne
Research L.^boratory, General Electric Company
This article and the one following describe the radio equipment installed on the U.S.S. George Washing-
Ion to enable President Wilson to engage in direct telephonic communication with his oflScials in Washington
while on his homeward trip from the Peace Conference in Paris. The makeup of the transmitting apparatus
is described below and interesting details of its operation are furnished. — Editor.
During the first part of March, 1919, the
Navy Department asked the Research Labo-
rator\- of the General Electric Company to
install a radio telephone transmitter on the
U.S.S. George ]]'ash!iigton, to work in connec-
tion with the New Brunswick station, so that
the President would be able to get into
telephonic communication with Washington
while still on the high seas. It will be remem-
bered that at that time President Wilson was
in Paris, attending the Peace Conference.
generators designed to operate from the ship's
mains and to supply these voltages, together
with their control and starting panels, were
hurriedly put together and the whole appara-
tus shipped to Hoboken by auto truck. There
the apparatus was assem.bled and installed on
the boat, and although the time was so short
that it could not be tested until after leaving
the dock, it perform-ed rem-arkably well and
gave practically no trouble during the three
months it rem.ained abroad.
[7v-|.H-'|«((ihM'lV
K>J
is ^^ii
'tv
Fig. 1. Connections of the Radiophone Transmitter Installed on the
U.S.S. George Washington
The General Electric Company had then no
finished equi])ment that would be suitable for
such purpose and it was necessary to design
and build a special set for this particular work.
It was decided to build a set employing a
ntimber of large pliotrons as generators of the
high -frequency- current required. There was
in the Laboratory at that time a panel
arranged to hold twelve of these large tubes
and to this was added another section con-
taining the necessary modulating apparatus.
These tubes required a source of LjOO to
2000 volts direct current and a separate source
of 20 volts direct current. Special motor-
The connections of the radiophone trans-
mitter are shown in Fig. 1. In this diagram
the actual number of tubes used in each stage
are not shown: also the details of the control
circuit, by which the operator of the trans-
mitter was able to supenise conversations
and to connect the set with the receiving
apparatus in the receiving room and to the
President's suite, are omitted for the sake of
simplicity.
The action was as follows : The microphone
transmitter M. a standard telephone desk set.
was used and the ctirrents generated by the
voice were stcp])cd tip in voltage by the trans-
RADIOPHONE TRANSMITTER ON THE U.S.S. GEORGE WASHINGTON 805
former T, and amplified by the pliotron tube
.4i. In this set this was a single small tube
having a rated output of about 50 watts.
The output circuit of this tube was connected
at B through a cajjacity to the grids of two
larger tubes of 200-watt capacity each, where
the voice currents were still further amplified.
Two similar tubes, C, were connected as
oscillators to generate an alternating current
of about 170,000 cycles (ISOO meters). The
high-frequency output (in watts) of these
tubes is proportional to the voltage supplied to
their plates. The plates of the tubes C and
B are connected to the direct-current source
through the high impedance /, and therefore
any change in the current flowing through
the tubes B will cause a voltage to be set up
across this impedance and a resulting change
in the output of C. In this way the high-
frequency output of C is made to correspond
with the voice currents supplied by the micro-
phone M.
Risa. bank of twelve 200-watt tubes having
their grids connected in multiple and coupled
inductively to the oscillating circuit of C.
The plates of these tubes were also connected
in multiple and inductively coupled to the
antennae. The tubes R therefore acted
simply as an amplifier for the fluctuating out-
put of C.
With an input to the plate circuits of all
the tubes of 1600 volts and between two and
three amperes, an antenna current having a
steady value of from 30 to 33 amperes was
obtained. An oscillogram taken when speak-
ing into the microphone showed that the cur-
rent in the antenna then fluctuated from 3 to
35 or more amperes. The wave length was at
all times maintained at 1800 meters.
After the installation of the set was com-
pleted on April 12th, a great many interesting
tests were made before the President boarded
the ship almost three months later.
On April 14th, we talked to the U.S.S.
Frederick, at that time about 150 miles ahead
of us, and they rejjorted: "Phone loud and
strong, easily understood." On April Ifith,
the log reads: "Before beginning the 3:00
p.m. schedule a broadcast message was sent
on the George Washington's spark transmitter
at 600 meters and at 952 meters asking all
ships to listen for our radiophone on 1800
meters and report how they received us and
giving their position." About a dozen ships
sent in reports. The ship farthest away that
reported was about 320 miles from us. They
reported " Phone fine on crystal with Marconi
type receiver." The U.S.S. President Grant,
about 150 miles from us, reported hearing our
radiophone 75 feet from the head phones using
a four stage amplifier.
Fig. 2 is a chart showing the names and
positions of the ships which reported that
they had received on the test. This was one
9 jff a
e -J s. __j (>_ J J, ^:
J O J^ 2
6'.Jj tj^of?GC ivAsf*f^i:> re -^
7-r^ Crt J- A/='/?'i. /6^
Fig. 2. Chart Showing the Names and Positions of Ships
which Reported Hearing Radiophone Messages
from the George Washington
of the first tests of this sort and from then on
we often entertained the operators of other
ships by phonograph concerts transmitted
via our radiophone. The log of the Navy
receiving station at Otter Cliffs, Maine, shows
that at one time, when we were 1000 miles
away, the music of one of these concerts came
in so loud that the sailors there danced to the
tunes we played.
On April 17th, when we were 2184 miles
from Ambrose Light, Otter Cliffs reported
hearing our tests but that the speech at that
tim.e was not clear.
Up to this time the full output of the set had
never been obtained because of some trouble
with the power circuit supplying our motor-
generator sets. In Brest this trouble was
remedied and thereafter all our tests were
made at about full output.
On April 27th, we began our first return
trip, Secretary of War Baker and several
806 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 10
thousand soldiers being on board. The static
most of the time was verv^ bad and we did not
get into touch with Otter Cliffs until May 4th.
when they reported : ' ' Your telegraph signals
excellent ; speech at first half of schedule loud
but not clear and on the last half ver\- loud
and clear." We had made an adjustment in
the middle of the schedule when we had dis-
covered that the ciuality of the speech was
poor. Four hours later they reported that
they were copying our speech on a typewriter.
Later in the day a number of commercial
messages were transmitted b}' the radio-
phone set. The first one read as follows:
"From U.S.S. George Washington via Otter
Cliffs, Maine, to Perkins Street, New York
City. 'Expect to see you Monday night.
Love. (Signed) Ted. 3:30 p.m.'" As far as
the writer is aware this was the first actual
paid commercial message ever transmitted
by radiophone from ship to shore. Later that
da\' Secretary Baker spoke a few sentences
over the phone which were received at Otter
Cliffs.
On the following day the set was used to
transmit to New Brunswick, where the
speech was automatically relayed over the
wires to Washington. Several persons talked
over the phone on the ship to people in
Washington, Secretary of War Baker talking
for some time with Assistant Secretan,^ of
Navy Roosevelt, and making arrangements
to meet some relatives in New York upon his
arrival there.
On the occasion of one of the coricerts which
we gave on this trijj the radiomen on the U.S.S.
Pastores. then (iOO miles distant, connected
the loud speaking telephone on the bridge of
the ship with the receiver and tuned in the
George Washington's wave. One of the radio
men, telling the writer of the incident, said
that some of the f)fFicers and men were so
sur])rised at the loudness and clearness of the
music and voices that they at first refused to
believe that the sounds were actually coming
from the George Wcisliingtov. and it was only
with difficulty that they were convinced that
the radio men were not n\i to some tnckery.
On May 10th the George Washington again
started for France and on this trip nimierous
tests were made. We were in communication
with New Brunswick until we were some S()(l
miles awav, when the interference and static
became so heavy that we gave up and all tests
were discontinued until our arrival at Brest.
While lying in Brest harbor we tested out
the set and representatives from Admiral
Sims' office in London listened for vis there.
We received the following report : ' ' Your
schedule received, signal strength ten, modula-
tion good . . . ." London is about 30U
miles from Brest and "strength ten" seems
ver\' loud.
While we were in Brest the XC-4, enroute
from Portugal to England; passed over and
circled the ship. An effort was made to talk
with her by the radiophone. The operator
of the NC-4 in conversation with the writer
later said that he heard our signals but that
they were so loud with the amplifier that he
was using as to be uncomfortable and scarcely
understandable. If more time had been
available it is probable that good com-
munication could have been established with
the XC-4 before she landed in England.
On June 29th, we started on our return trip.
President Wilson and party being aboard.
When we had approached to within 130U miles
from Ambrose Light we picked up a message
from Otter Cliffs to New Brunswick saying:
"For information, can hear George Washing-
ton's wireless telephone fine; can copy solid
but at present Glace Bay causing interfer-
ence."
On July 4th an effort was made to transmit
the President's speech to the troops by radio-
])hone. A telephone microphone was con-
cealed on the stand where he was scheduled to
speak, but due to a misunderstanding the
President spoke on a lower deck some 2tl feet
from the microphone. All ships had been
notified to listen for the President's speech,
but only an occasional word could be heard.
This was ver\- much to be regretted, as the
atmos])heric conditions were s])lendid at the
time. The writer read the President's speech
in the jjhone a few hours afterward. Colonel
Carr, Department Signal Oflicer of the South-
western Department of the Signal Corps, has
since told the writer that he heard portions of
the speech on a small antenna in San Antonio,
Texas. _ This distance was roughly .'>(i(Mi miles
and almost entirely over lanil.
On Jtily .'ith antl (ith. the static conditions
were so bad that we had difficulty in getting
into good communication with either New
Brunswick or t)tter Cliffs, hut on the 7th we
got several messages through, though it was
not at all satisfactory-. The ship was then
only about 37.i miles from Ambrose Light.
Later in the dav the conditions grew worse
and it was not until the following morning that
really satisfactor\' two-way communication
was obtained and the President was able to
send a message over the radiophone to Secro-
tar\- Roosevelt in Washington.
SOT
Duplex Radiophone Receiver on U.S.S.
George JVashington
B>- Harold H. Beverage
Radio Exgineerixg Department, General Electric Company
In arranging for duplex or two-way communication between the George IVashinglon and land stations,
one of the major diflficulties encountered was the prevention of each receiver from being affected by the pow-
erful interference of its own transmitter. This problem as applying to the land station at New Brunswick
was solved by separating the transmitting and receiving stations a distance of four miles. The solution em-
ployed on the George Washington was of necessity totally different. A description of the schemes used, with
particular reference to receiving, is given in the following article. E.xtracts from the ship's log are included
to show the successful performance of the equipment. — Editor.
New Brunswick's radiophone while the ship
was lying at anchor in Brest Harbor, France.
The results of these tests were so encourag-
ing that the Navy Department decided to
install a powerful radiophone on the George
Washington, to enable the ship to talk back
to the shore. The General Electric Company
was asked to furnish the radiophone equip-
m.ent, both transmitting and recei\-ing.
Introduction
On February 22, 1919, Secretary of Navy
Daniels, sitting at his desk in Washington,
picked up his telephone and spoke a few words
of greeting to President Wilson, then SOO miles
at sea oia the U.S.S. George ]]'ashingto)i.
Secretars- Daniels' voice was carried over
the 'regular toll line from Washington to the
Naval Radio Station at New Brunswick,
Fig. I. Receiving Set and Telephone Communication Station of the
U.S.S. George Washington's Radio Telephone Pliotron Equipment
N. J., where the voice currents were amplified
to such an extent as to moditlate the output
of an Alexanderson alternator. This modu-
lated energy was radiated from the New
Brunswick antenna and was picked up on the
George Waslirngtoii.
Previous to this demonstration, the oper-
ators on the George ]]'ash!)igto>i had heard
Requirements for Two-way Radiophone Conversa-
tion
In order to make a two-way conversation
possible over a radiophone, it is necessary
either to shut off the transmitting set when
receiving, or to so arrange the receiving appa-
ratus that it will be unaffected by the powerful
interference from the local transmitter.
808 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 10
On small radiophone sets the first method is
often used, the transmitter being started and
stopped by a convenient push button located
on the microphone support. More or less
confusion is likely to result from this method
of control, as it is impossible for the party
^
Detector
Rccieving and
Set Amplifier
Bg-imorTcla-phone Line
^^1^" ^ Miles "cjllis To New Brunswick Telephone E>chonqc
Recievmg Station
Transmitting 5totion
Smoli
Ptiotron
Amplifier
n.
Multiple Tuned Antenna
Large
Pliotron
Amplifier
Alexanderson
Alternator
IP
Maqnetic
Amplifier
Fig. 2. Connection Used for Duplex Operation at New Brunswick Station
talking to hear the other party until he
releases the push button, thereby shutting oE
his transmitter. This method of operation
is not readily adapted to remote control over
a long-distance toll line.
On high-power radiophones, particularly
where an alternator is used to supply the
radio frequency energA% the push button
method of control is impractical, if not impos-
sible in many cases. For these reasons, the
second or duplex method for two-way con-
versation was chosen for both New Brunswick
and the George Washington.
Duplex Arrangements at New Brunswick
At New Brunswick, Mr. Burke Bradbury
made arrangements for duplex operation by
setting up the receiving apparatus at a point
about four miles from the radio station, and
sending the amplified received currents back
to the radio station over an existing telci^honc
line connecting New Brunswick with the Mar-
coni receiving station at Belar.
Fig. 2 shows the connections used for du-
plex operation at New Brunswick. It will be
noted that the received currents are intro-
duced in series with the toll line leading
from New Bnmswick, enabling the party
talking from Washington or any other point
to hear the incoming radiophone speech
over the same wires which transmit his
own speech to New Brunswick, as in an
ordinary land wire connection. It is also
evident that, with this connection, the speech
and signals picked up by the recei\-ing
apparatus at New Brunswick
will modulate the alternator out-
put and be re-radiated again at
New Brunswick's wave length.
Anyone listening on New Bruns-
wick's wave length would, there-
fore, hear both sides of the con-
versation. This explains a point
which puzzled many amateur
operators, who reported that
thev heard the George Washing-
io« radiophone on SOOO or 13,600
meters, whereas the wave length
of the radiophone on the George
Washington was 1800 meters.
The wave lengths of SOOO and
13,000 meters were both used at
New Brunswick for radiophone
tests at various times. The
writer often heard short wave
spark signals while listening to
the New Brunswick radiophone
on the George Washington, the
signals being picked up by the receiving appa-
ratus at New Brunswick and being re-radiated
in the manner described.
Duplex Arrangements on U.S.S. George Washington
The solution for duplex operation on the
George Washington was necessarily dilTerent
than for New Bnmswick, as the receiving
and transmitting apparatus could not be
located at difTerent points as at New Bruns-
wick. The problem which presented itself,
therefore, was to provide receiving apparatus
sensitive enough to respond loudh- to received
currents of a few millionths of an ampere on
SOOO meters, and yet be practically unre-
sponsive to a radiation of thirty or more
amperes at ISOO meters, radiating on an
antenna stretched from the same masts as the
receiving antenna.
As a solution for this problem, Mr. E. F.
W. Alexanderson, Chief Engineer of the Radio
Engineering Deiiartmcnt, suggested the cir-
cuit shown in Fig. 3. This circuit was first
tried out in Schenectady, using the same
antenna for both receiving and sending. It
was foiuid possible to receive signals from
Europe on long wave lengths and at the same
time radiate ten amperes at 4000 meters on
the same antenna, using either an .Mexantler-
DUPLEX RADIOPHONE RECEIVER ON U.S.S. GEORGE WASHINGTON 809
son alternator or a pliotron oscillator as the
source of energy.
On the George Washington, however, sepa-
rate receiving and transmitting antennas were
used, arranged as shown in Fig. 4.
Fig. 3 shows the connections used in the
duplex receiver on the George Washington.
set is in series with the frequency trap, the
interference from the local transmitter is
reduced in the ratio of the impedances of the
two branches to ISOO meters, or 1(30/250,000,
so that the interference is reduced to 0.6 of one
per cent of the interference that would be
experienced without the filter circuit. The
W lOOft. —
174 Meter Set
Fig. 4. Arrangement of the Separate Receiving and Transmitting Antennae on the
U.S.S. George Washington
The current from the receiving antenna
divides through two parallel branches; Ci
being one branch, and FC-iL\, the second
branch. The "frequency trap" F is tuned to
the transmitting jWave length of 1800 meters.
pnonc Pluq For
SE 1000 Amplifier
To Radio Room
I- 50V linv
Fig. 3. Connections Used for the Duplex Receiver
on the U.S.S. George Washington
and offers an impedance of about 250,000
ohms at 1800 meters, but a very low imped-
ance to long wave lengths. The condenser
Ci has a capacity of 0.006 microfarads, and
offers an impedance of about 1 60 ohms at 1 800
meters. As the primary L\ of the receiving
remaining interference is so small that it is
easily taken care of by tuning alone.
The variable condenser C-y tunes the
primary of the receiving set to the long wave
length which it is desired to receive. The
inductance of the frequency trap F enters into
the tuning of the primary circuit, and there-
fore the frequency trap offers practically no
impedance to the long waves. The branch
FC2L1 offers only a few ohms effective resist-
ance to the long wave length to which it is
tuned, while the branch Ci offers a com-
paratively high impedance, about 1200 ohms
for 13,600 meters. It is, therefore, possible
to receive the long waves at practically full
intensity, and yet render the receiver very
insensitive to the effects of the powerful radia-
tion from the local transmitter.
The remainder of the apparatus is the same
as is used for receiving telegraph signals,
excepting that it is adjusted to receive a wider
band of frequencies than a telegraph receiver,
in order to receive all components of the
telephone wave necessary for clear speech.
If the receiver is too sharply tuned, the quality
of the telephone speech on long waves is very
poor, because only one frequency is received
strongly and other frequencies are suppressed.
For long wave telephony, it is very essential
to tune the receiver in such a manner that
it is capable of receiving a band of frequencies
within about 1 000 cycles on either side of the
carrier wave. On short wave lengths, 1000
cycles is a very small per cent of the carrier
frequency and an ordinary sharply tuned tele-
graph receiver will receive telephone speech
clearly. On wave lengths above 10,000 meters
810 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 10
lUOU cycles is several per cent of the carrier
frequency, and the average telegraph receiver
is tuned too sharply to receive all of the fre-
quencies necessary- for clear speech, the speech
sounding muffled and being ver\- difficult to
understand. With broad tuning, however, the
long wave length speech may be recei\-ed
clearly with good quality.
The adjustment of the duplex feature of
this receiver is very simple. First, the receiv-
ing set is tuned to the wave length it is
desired to receive. Then, the frequency trap
F is adjusted until the interference and noise
produced by the local transmitter disappear
and the distant signals are heard clearly. The
frequency trap adjustment is ver^i- sharp, and
in the case of the George Washington the radi-
ation from the transmitter was so powerful
that the detector bulb was rendered in-
operative until the frequency trap condenser
was within a ver\- few degrees of the correct
position.
Some idea of the effectiveness of the
duplex feature of the receiving set may be
gained from the following demonstrations:
When the pliotron transmitter on the ship
was in operation, radiating about 30 amp.
on the main antenna at ISOO meters, the
receiving antenna was sufficiently exposed
to light a 4()-watt lamp to full brilliancy
when the lamp was placed in series ^\ith
the receiving antenna. The intensity of the
interfering current reaching the receiving
set secondary' was so small, however, that
no interference whatever was experienced
above 6000 meters when the pliotron trans-
mitter was being used for continuous wave
telegraphy with 30 amp. radiation. When
the pliotron transmitter was modulated by
voice or buzzer, the modulation could be heard
weakly on 13.()00 meters, and a little stronger
on ,S000 meters. The interference from the
modulation was never strong enough to inter-
fere with telegraph reception from New Bruns-
wick on either 13.(>00 meters or 8000 meters,
even in Brest Harbor, but it was strong
enough to interfere slightly with the New
Brunswick radiophone when the shii^ was
several hundreil miles away from Xew Bruns-
wick, as the radiophone intensity was ver\-
much less than the intensity when Xew
Brunswick was telegraj^hing. The interfer-
ence or "side tone." however, was not loud
enough or clear enough to understand: and
it was necessary for intercommunication jjur-
poses to introduce an artificial side tone
enabling the transmitter operator to sjicak to
the receiving operator for changes in control.
etc. For duplex telephony the side tone is not
only unobjectionable, but is more or less
desirable, as it more nearly approxim.ates
the conditions in the ordinary land wire
telephone.
It was noted that in very wet weather the
side tone was appreciabh- stronger than in dr\-
weather, probably due to leakage currents
between the transmitting and receiving ante-
ras over the wet insulators. However, duplex
operation was satisfactory- in ver>- wet
weather, with a few exceptions. On one
or two occasions, it was found that wet
halyards swinging against the receiving an-
tenna produced more or less inductive dis-
turbance when the transmitter was working.
During rough weather, when the ship was
rolling and pitching badly, it was sometimes
ciuite difficult to keep the receiver quiet when
the transmitter was in operation. Loud induc-
tive disturbances, s\-nchronous with the
rolling or pitching of the ship, were obser\-ed
several times. In each case, the source of dis-
turbance was found to be an antenna lead-in
grounding somewhere on a metal stay or some
other grounded metal object. -As there were
about twelve antennae on the ship, it was
sometimes difficult to locate the source of dis-
turbance immediately. It was found neces-
sar\- to insulate each antenna lead-in care-
fully in such a manner that it could not swing
against a grounded object when the ship was
rolling and pitching in a storm.
Extracts from Log
Mr. John Pa>ne and the writer m.ade two
trips to France on the George U'ashitigtott.
On the first trij). the shi]) .mailed from Hobokcn
on April 11th. Due to the short time avail-
able for installing the radiophone, the ship
was out of range before arrangements were
completed for duplex operation and no duplex
conversations were tried.
During this trip, the New Brunswick radio-
phone was operated at a wave length of l.'?.((0(l
meters with an average antenna current of
120 amps. The New Brunswick radiophone
was o])erated on definite schedules (hiring the
day and evening, and was received consist-
ently U]) to about 1200 miles, with the excep-
tion of a few schedules when the ship was in the
gulf stream, and heavy static interfered with
reception. The speech was partially under-
stood up to 2.")0{) miles when the static con-
ditions were favorable, and was heard but not
imderstood at still greater distances.
On the retuni trip from Br-si. the New
Brunswick radio))hone was heard, but not
DUPLEX RADIOPHONE RECEIVER ON U.S.S. GEORGE WASHINGTON Nil
understood, as soon as the ship left Brest
Harbor. At a distance of 2UU() miles from
New Brunswick, complete sentences and
orchestra selections were recognized on the
shi]). The orchestra music was obtained by
placing a telephone near the orchestra at the
New Bninswick Opera House, and also at the
Hotel Klein. At a distance of 120U miles
practically all of New Brunswick's speech was
understandable under nomial static con-
ditions.
Due to very unfavorable static conditions,
duplex conversations were not satisfactor>-
until the ship was about 200 miles from New
York, on the morning of May oth. After
establishing satisfactory duplex conversation
with the engineers at New Brtmswick and
Navy officials in Washington, Secretary of
War Baker on the George Washington held a
conversation with Assistant .Secretary of Navy
Roosevelt in Washington. Secretary Baker
remarked that "the connection was as good
as over an ordinary toll line."
The George Washington sailed from Hobo-
ken again on May 10th. The first duplex
conversations were held while a heavy sea was
nmning and in a driving rain, and consider-
able difficulty was experienced from inductive
disturbances caused by the lead-in on unused
antennas swinging against grounded objects.
Later, fairly satisfactory conversations were
held with New Brunswick up to a distance
of about 400 miles from New York. At that
distance, very bad static was experienced
which made the reception tinsatisfactory on
both ends, particularly on the George Wash-
ington. After the ship was about SOO miles out
the static conditions on the ship end were
much improved, so that New Brtmswick was
easily understood again. New Brunswick
could hear the George Washington at this
distance, but could not understand the speech
well enough to work duplex satisfactorily.
While the George Washington was lying at
anchor in Brest Harbor, some radiophone tests
were made at New Brunswick, using a wave
length of 8000 meters. During these tests,
the average antenna current at New Bruns-
wick was about 50 amp., as compared with
an antenna current of 120 amp. at 13,600
meters. At night, the SOOO-meter wave length
radiophone was received much clearer and
stronger than the 13,600-meter wave length,
but very little was understood due to inter-
ference and static. However, it was possible
to recognize selections sung at New Bruns-
wick. On one occasion, Mr. W. W. Brown's
voice was clearly recognized singing "Amer-
ica." Sometimes, at night, the SOOO-meter
wave with 50 amp. average antenna current
was received by heterodyne note as strong or
stronger than the 13,600-meter telegraph
wave with an antenna current of 350 amp.
In the daytime, the conditions were reversed,
and the 13,600-meterwave length was received
much stronger than the SOOO-meter wave
length. In fact, it was often impossible even
to pick up the SOOO-meter wave length in
Brest Harbor in the daytime, although at
night the same radiation was very strong.
The George Washington sailed from Brest
again on June 2()th with President Wilson and
party aboard. When the ship was 2400 miles
from New York, New Bnmswick began to
make tests comparing the SOOO-meter and the
13, 600 -meter wave lengths with 50 and 120
amp. average antenna current respectively.
The first test was run at night, and the SOOO-
meter radiophone was understood fairly well
at this distance of 2400 miles, while the 13,600-
meter wave length could not be understood
due to static, although the radiation in
amperes was over twice as great. The next
test was made on the following day, with day-
light. The distance between New Bnmswick
and the George Washington was about 2000
miles. Both wave lengths were received about
the same on this occasion. From this point on
the SOOO-meter wave length was received
better than the 13,600-meterwave length, day
or night. From a distance of 1600 miles until
the ship docked, the SOOO-meter radiophone
was understandable at most schedules except-
ing when there was interference from Glace
Bay.
The first duplex conversation on this trip
was held on July 7th, when the George Wash-
ington was about 375 miles from New York.
The first conversation was not very satis-
factory, due to static and interference at New
Brunswick. Later in the day, fairly satisfac-
tory duplex conversations were held between
New Brunswick and the George Washington.
but it was necessary to repeat some of the
sentences several times before they could be
understood. It was decided that the con-
ditions were not quite favorable enough for
President Wilson to talk, and it was hoped
that the static conditions would improve
later in the day. The static conditions be-
came worse, however, so the presidential con-
versation was deferred until the following day.
On the following morning, satisfactory com-
munication was established, and the President
was able to send a message over the radio-
phone to Assistant Secretary Roosevelt via
812 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 10
radiophone to New Brunswick and thence to
Washington over the toll line.
Conclusion
On all trips, New Brunswick was received
well up to about 300 or 400 miles. Between
300 and 500 miles from New York, the
reception of New Brunswick radiophone was
generally comparatively poor on evers^ trip,
due to very strong static. Beyond 600 miles
from New York, the reception on the George
Washington was generally good again on all
trips. The static conditions seemed to be
most unfavorable on the western edge of the
Gulf Stream, and all signals appeared to be
comparatively weak there.
During all of these tests, the New Bruns-
wick radiophone was received with the
detector oscillating at zero beat. Non-
oscillating detector reception was also tried,
using a multistage radio frequency amplifier.
Good reception was obtained, but it was
found more practical for duplex operation to
use the oscillating detector, as it was not as
subject to interference from the transmitter
as the radio frequency amplifier system, due
to the fact that the radio frequency amplifier
tended to amplify the high-frequency harmon-
ics from the transmitter to a great extent.
The duplex reception with the zero beat
method of reception with oscillating detector
was very satisfactor\-. Under normal con-
ditions, the static produced a much louder
sound than the side tone from the local
transmitter, so that the distance that it
was possible to receive New Brunswick's
radiophone successfully was limited by the
static rather than the side tone from the
local radiophone transmitter. The master
oscillator method of generating the radio
frequency energy- used by Mr. Payne on
the George Washington radiophone produces
a wave which is remarkably free from
harmonics.
When the pliotron set was being used for
continuous wave or buzzer modulation teleg-
raphy, sending commercial messages to Otter
Cliffs, Maine, it was found possible to stand
watch on New Brunswick without the slight-
est interference from the pliotron transmitter.
This feature may be found useful in the future
on large ships with heavy traffic, as it enables
them to send and receive simultaneously.
With this type of duplex, it is possible to
work down to within about twice the wave
length of the transmitter when the trans-
mitter is being used for continuous wave
telegraphy.
Relaying Messages from Washington Through the New Brunswick Radio Station
813
The Alexanderson System for Radio
Communication
By Elmer E. Bucher
Commercial Department, Radio Corporation of America
The attainment of a system of radio communication free from the objectionable features and limitations
of the arc and spark gap types has been realized in the Alexanderson system. It has introduced into the
realm of radio the same degree of certainty of expectations and reliability of results, in generation, control,
and distribution, as has for a long time been characteristic of operations in the field of commercial light and
power. The generator, although necessarily of special construction, is designed in conformity with estabhshed
electric, magnetic, and mechanical laws. Its speed is held constant within one-tenth of one per cent by an
especially developed speed regulator. The output of the station, which fluctuates with the message being
sent, is controlled by an ingenious and extremely sensitive non-arcing device called a magnetic amplifier.
The multiple tuning of the antenna permits of radiating the energy at a vastly improved efficiency. The
following article thoroughly describes the complete system in great detail and in a manner that is easily
understood. — Editor.
General
Radio engineers early foresaw that the
ultimate generator of oscillations for radio-
telegraphy and telephony would be one of
a type providing more efficient and reliable
operation than the systems utilizing the "arc"
and "spark." In fact the literature of the
past makes frequent references to the desir-
ability of an oscillation generator constructed
along the lines of an ordinar\^ power-house
alternator; but because such alternators were
required to provide frequencies a thousand
times or more in excess of those used in power
engineering, new problems of designs were
encountered which were declared by many
to be well-nigh insurmountable. For a time
the development of the art seemed to follow
the line of least resistance, and it resulted in
the evolution of several systems utilizing the
"arc," the "spark gap," and the type of radio
frequency alternator which generates at a
comparatively low frequency, the necessary
increase of frequency- being obtained either by
groups of mono-inductive transformers exter-
nal to the alternator, or by tuned "reflector"
circuits associated with the alternator. None
of these systems, however, can be said to have
satisfied fully the exacting requirements of
commercial operation.
An oscillation generator suitable for com-
mercial radio service over great distances
should possess the following q^ialifications :
(1) It should generate a steady stream of
oscillations of constant amplitude.
(2) It should generate a so-called "pure"
wave; that is, a fundamental wave in
which the radiation incurred by super-
imposed harmonics is negligible.
(.3) It should provide a performance as
reliable as the ordinary power dynamo.
(4) It should operate economically and
efficiently.
(5) It should permit manufacture of units
in any desired power.
(6) The design of the whole system should
be such as to permit telegraphic signal-
ling at ver\' high speeds.
The above specifications are met fully and
fairly in the Alexanderson system.
As is well known, the design of radio fre-
quency alternators has occupied the attention
of Mr. Ernst F. W. Alexanderson of the
General Electric Company (U. S. A.), and his
staff for a number of years, and the pioneer
work of these men in that branch of radio
research is now a matter of common knowl-
edge. Starting with the development of sev-
eral experimental types of alternators, they
have steadily progressed toward the designs
of more powerful machines which are now
available for commercial use. Standardized
alternator sets for transmission at wave
lengths between 6000 and 10,000 meters and
between 10,500 and 25,000 meters are now in
production. This description is devoted
principally to the discussion of a 200-kilowatt
set, although sets of other powers are now
under construction.
The typical Alexanderson high-power
station may be said to represent a radical
departure from current ideas regarding radio
design. In fact, at first glance, it seems to
possess little in common with the apparatus
of other systems. These features will pres-
ently be described in greater detail.
The Radio Corporation, after an extensive
test of the Alexanderson system at its
high power station at New Brunswick, N. J.,
has acquired the rights to the Alexanderson
system, and it will be employed at all its
814 October, 1920
GENER.\L ELECTRIC REVIEW
Vol. XXIII. Xo. 10
stations devoted to long-distance signalling.
A 200-kilowatt alternator set was installed
at Xew Brunswick in September, 191S, and
from that time it has provided continuous
and most satisfactory service in continent-
to-continent communication. Normal trans-
mission is at present conducted at the wave
length of 13,600 meters, with antenna current
of 400 amperes corresponding to an alternator
output of approximately SO kilowatts. With
this fractional value of the available output
of the alternator, transoceanic communica-
tion has been maintained with European
stations throughout the twenty-four hours of
the day. The alternator is capable of supply-
ing 600 amperes to the New Brunswick
antenna, but its full output of 200 kilowatts
is not at present utilized, due to the lack of
adequate power supph' at that point. The
alternator, as installed at the New Brunswick
station, is shown in Fig. 1 .
With this brief disclosure of progress to
date, there will follow an explanation of the
basic principles of the Alexanderson system
and the fundamental circuits of a typical
station. This may be accepted as indica-
tive of a standard 200-kilowatt installation,
although largely based upon the apparatus
at the New Brunswick station.
Standard Equipment
A high power radio station of the Alexander-
son type contains three important develop-
ments :
(1) An Alterx.\tor — which generates cur-
rents directly at the frequencies which
are required for the radio circuits with
which it is associated. The frequency
of these currents is solely dependent
upon the nirmber of field poles on the
machine, and upon the speed at which
the rotating member is driven. This
is in distinct contrast to certain other
systems in which the radio frequency
currents are obtained indirectly by
means of "reflector circuits" or fre-
quency raising transformers electrically
associated with the alternator.
(2) A Magnetic Amplifier — which pro-
vides a non-arcing control of the alter-
nator output for radio telegraphy, and is
equally applicable to radio telephony.
(3) A Multiple TixedAntenx.x — adevel-
op»ment which has markedly reduced
the wasteful resistance of the flat-top
antenna, and has therefore increased
the transmitter overall efficiencv many
fold.
Fig. 1. 200-kw. Alexanderson Radio Frequency Alternator Installed at the Radio
Corporation's Transoceanic Station, New Brunswick. N. J.
THE ALEXANDKRSOX SVSTI- M FOR RADIO COMMUNICATION
S1.3
Alternator Development
To date the development in radio fre-
quency alternators has included the following
types :
(1) 2-kw., 1 00, OUO-cycle alternators.
(2) 50-kw., 50,000-cycle alternators.
(3) 200-kw., 25,000-cycle alternators.
The characteristics of several alternators
of other power outpvits have been investigated
from time to time. A standard 2.3-kw. and a
5-kw. alternator are now under construction
and will be shortly put into commercial pro-
duction.
With the object of providing a distinct
range of frequencies, both the 2r)-kw. and the
200-kw. alternators are manufactured with
armatures and rotors with different numbers
of poles; also with gears of different ratio for
different driving motor speeds. Thus the 2.3-
kw. machine can be assembled for any wave
length from (i.OOO to 10,000 meters, and the
200-kw. machine for any wave length from
10,500 to 25,000 meters. Frequencies lower
than these for which the machine has been
assembled can be obtained by running the
alternator at a reduced speed.
The standard drive for the 200-kw. Alex-
anderson alternator is two-phase, GO-cycles,
2.300-volt alternating current. By the use
of suitable transfomiers, the voltage of the
power supply can readily be transformed
to the value for which the motor was
designed. Special driving motors and con-
trol equipment can be supplied for other
frequencies.
The Alternator
The Ale.xanderson alternator is an inductor
type of generator with a solid steel rotor hav-
ing several hundred slots milled radially on
each side of the rim. The slots are filled in
with non-magnetic material, with the object
of reducing wind friction to a minimum.
The fillers are brazed into the disk in order
that they may withstand the centrifugal
strain of rotation. The rotor is designed for
maximum mechanical strength by providing
it with a thin rim and a much thicker hub.
With this construction the strain on the
material due to centrifugal force is the same
from the shaft to the outer rim.
The rotor of the 200-kw. alternator (with
half of the field frame removed) is shown in
Fig. 2. This also shows the collars of the
thrust bearings and a partial sectional view of
the main bearing housings.
Fig. 2. 200-kw. Alexanderson Alternator, with Top Half Removed
816 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 10
An assembled 200-kw. alternator with its
driving motor is shown on page 814, Fig. 1.
The alternator is driven by a 600-h.p. induc-
tion motor of the wound-rotor type, which is
operated from a 60-cycle, 2.300-volt, quarter-
phase source of supply. The motor is con-
nected to the alternator through a double
helical gear (with a speed step-up ratio of
1 :2.97), which operates in a container partial!}-
filled with oil.
The main bearings and the thrust bearings
of the alternator are oil-lubricated bv force
tVWDMSS
L^mnmrmp
MPunO!
CC/L
ff/MAfiy
mrfOiHGi
rdMMlMJU
COIL
J \secoKMfy
Fig. 3.
Schematic Diagram of Alexandenon Radio Frequency
Alternator Circuits
feed at pressures varsnng from o to 1 5 pounds
according to the demand on the bearing.
During the periods of stopping and starting,
and in possible emergencies, oil is supplied
by a special motor-driven pump mounted on
the alternator base. When the alternator
is working under normal operating conditions,
a separate pump geared to the main driving
shaft feeds the bearings, and the motor-
driven pump is automatically cut out of
ser^'ice. The oil-supply tank is located in the
base of the alternator, to which the oil
returns after being pumped through the
bearings. The oil gauge on the main feed
pipe is fitted with a signalling circuit to call
the attention of the operator in case the oil
supply fails. The main bearings of the
alternator, which are self-aligning, are also
water-cooled by a series of copper pipes
which run through the bearings near to
the friction surface. The armatures of the
alternator are also water-cooled from the
same pumping source by a series of parallel
copper tubes cemented in the frame along-
side the laminations.
In order to avoid large losses through
magnetic leakage, the air gap between
the rotor and the stator frame is main-
tained at a spacing of 1 millimeter.
It is important that the rotor be kept
accurately centered, for otherwise the
armature coils on one side of the rotor
will become overloaded. This is
accomplished by the use of specially
designed thrust bearings which are
inter-connected by a set of equalizing
levers with an adjustable controlling
leaf between them. These prevent
the possibility of binding between the
thrusts, due to expansion of the shaft
from heating, and they also take up
automatically all slack in the bearings
as they become worn. Any tendenc\-
towards a change in the air gap is thus
counteracted by the action of the
levers. The equalizers are, in part,
the hea\'>' vertical column shown at
the end of the alternator in Fig. 1 on
page SI 4. Should the air gap on
either side tend to get smaller, the pull
of the field on that side would cause
an excessive strain on the thrust at
that end and cause heating. This,
however, is pre\-ented by the leverage
system, which automatically corrects
this and holds the rotor in a central
position at all times.
In regard to some of the electrical
features of the alternator it will be noted from
Fig. .3 that the armature and field coils are
stationar>-, the requisite flux variations for the
generation of radio frequency currents being
obtained from the slots cut in the rotor. The
diagram points out the fundamental con-
struction of the alternator and the general
mode of winding the armature. The rotor
disk revolves between the two faces of the
field yokes. The direct current supplied to
the field coils produces a magnetic field
flux which passes between the field yoke
Au^tirrflt
THE ALEXANDERSON SYSTEM FOR RADIO COMMUNICATION
Si:
faces and through the rotor as shown by the
arrows.
The armature coils, which are placed in
slots cut in the two faces of the field frames,
are shown in the sketch as tipped away from
the rotor, although in the actual machine the
spacing between the rotor and the frame is
but 1 millimeter. Two distinct armature wind-
ings are thus provided, one on each side of
the rotor. There is but one conductor in each
slot and two of these slots make a complete
loop, and comprise a pole in the armature
windings. One slot in the rotor is therefore
provided for each loop in the winding. The
armature windings on each side of the rotor
are divided into thirty-two independent
sections, the circuits of which are completed
through transformer primary coils as shown
in Fig. 'A. Each primary consists of two turns
with sixteen separate wires in each turn.
There is no direct connection between the
individual armature sections, but through
the two-turn primaries they combine to act
upon the secondary coils of the transformers.
It is obvious that with this division of arma-
ture circuits the potential on any armature
coil (or on the corresponding transformer
primary) is very low, and as such, it permits
a grounded or open-circuit armature coil to be
cut out of the circuit and the operation of the
alternator to be continued with but a slight
decrease in its output — an obvious advantage.
A detailed view of a portion of the alternator
armature windings is given in Fig. 4, and of
the preliminary stages of assembly in Fig. 5.
Fig. 4.
Detail View of Section of Armature, Alexanderson
200-kw. Radio Frequency Alternator
Fig. (i shows the laminated armature, which
is wound with 0.037 millimeter steel ribbon.
The completed rotor and its shaft appears
in Fig. 7
Alternator-Antenna Transformer
It is to be noted that a transformer is pro-
vided for the armature coils on either side of
the rotor. There are therefore two trans-
formers, and they each contain the three coils
Pi, vSj, Si, and Ps, Ss, Ss, shown in the funda-
mental station diagram. Fig. 14. The primary
of each transformer contains two turns of six-
teen wires each, as mentioned above. The
intermediate coils Ss have twelve turns on each
transformer. The two intermediate coils are
connected in parallel, and are shunted by the
magnetic amplifier. The coils S.^ are also
connected in series with the secondary proper,
and the antenna system.
Fig. 5. A Section of the 200.kw. Alternator Armature
Fig. 6. Completed Armature Ready for Sawing Into
Two Sections
The secondary coils, which consist of seventy-
four turns on each transformer, are wound so
that their high potential ends are at the center
in order to provide a uniform potential gra-
dient. The two secondaries are connected in
parallel and their final terminals are in series
with the antenna circuit. More in detail, the
low potential terminals of the intermediate
coils are connected to the ground, the other
terminals of the intennediate coils are con-
nected to the low potential tenninals of the
secondary coils, and the high potential ter-
minals of the secondary coils to the antenna
loading coil. The intennediate coils Ss are
placed between the primary and secondary
of each transformer in order to obtain a close
coupling with the alternator. One unit of
the high frequencv transformer is shown in
Fig. S.
The voltage at the terminals of the second-
ary winding of the transformer when the
81S October. 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 10
alternator is operated at normal speed is
about 2000. The normal output current is
100 amperes. It is thus seen that the alter-
nator is designed for a load resistance of
20 ohms.
Speed Regulator
Since the antenna circuit is directly
associated with the alternator circuit, anv
Fig. 7. Typical Rotor Construction of Alexanderson Alternators
change in the rotative speed of this machine
would throw the alternator circuit out of
resonance with the antenna circuit; con-
sequently it is easily seen that the speed
variation of a radio frequency alternator for
substantially constant output must be held
within very close limits. The variable load
imposed by telegraphic signalling has a
tendency to cause a variation of speed that
must be compensated for by some de\-ice
which operates more critically than any of
the mechanical and electrical methods of speed
control de\-ised for ordinan.- power use. The
characteristics of any satisfactory- governor
must be such that a small variation of speed
will effect a maximum change in power input
to the device under control. To accomplish
this, some mechanism must come into such
a critical state at the speed to be maintained
that a low percentage of change in speed
causes a high jsercentage of change in itself.
It can be shown that a change in speed of
one-quarter of one per cent from that neces-
sary to maintain resonance will reduce the
antenna current in a station utilizing the
wave length of New Brunswick — l.'^.fiOO
meters — to one half its full value. This
clearly infers that tlie speed variation must
be much less than one fourth of one per
cent to maintain a constant output at the
alternator. As a matter of fact, a regulation
within one tenth of one per cent is obtained by
the Alexanderson speed regulator.
The necessity for close speed regulation
becomes equally important when considered
from the standpoint of the receiving station.
With a modern receiving apparatus of low
decrement, a very slight change in the wave
length of the incoming signal will materially
decrease the received current. A
change of wave length or frequency
is likewise detrimental when recep-
tion is obtained by the heterod\Tie or
beat principle, for should the speed
of the alternator \-ar\- markedly
while signalling, the beat note may
vary to the degree that will render
-^_^=: it objectionable for ear reception.
A variation, for instance, of 50 cycles
in the alternator will cause the beat
note at the recei\er to van^- by 50
cycles, which is the equivalent of a
speed variation of 0.2.3 per cent a-
the wave length of 13,(500 meters.
A solution of the problem of speed
regulation with a-c. motor drive
was found by Mr. Alexanderson
in the use of a resonance circuit,
which is tuned to a frequency slightly above
the frequency to be maintained at the
alternator. This circuit is supplied with cur-
rent from one of the armature coils on the
alternator. The current in this circuit in-
creases with the alternator speed and. through
the agenc\- of a rectifier, a d-c. component
FiR. 8. One Unit of the High Frequency Transformer
operates on a voltage regulator connected in
the circuit of the dynamo which supplies the
saturation current for a sol of variable imped-
ances in the two phases of the motor supply
circuit. The function of the regulator is to
prevent, within established limits, either an
THE ALEXANDERSON SYSTEM FOR RADIO COMMUNICATION
S19
increase or decrease of alternator speed.
Additional compensation for the load imposed
when signalling is provided by a relay which
also operates through the d-c. control circuits
to vary the line impedances mentioned above.
A detailed diagram of the speed control
system is shown in Fig. 32, and the theory
of operation is disclosed in greater detail on
pages 833 to 836.
The panel board of the voltage regulator
system is shown on page 796, Fig. 4.
Multiple Tuned Antenna
This may be said to establish a radical
departure froni the types of antennae fomierly
used for high-power radio transmission.
The immediate object of the multiple antenna
is to reduce the wasteful resistance of the
long, low, flat-top aerials formerly used and
to pennit the length of such aerials to be
increased indefinitely for the use of greater
powers. In the case of the New Brunswick
antenna, its resistance as a flat-toi^ aerial —
.'5.7 ohms — was reduced by multiple tuning
to 0.5 ohm. The radiation qualities of the
flat top are not impaired by multiple tuning,
as a series of tests have shown that with an
equal number of amperes in either type, the
same signal audibility is obtained at a receiving
station, but there is an enormous saving of
]5ower in the case of the multiple antenna,
as will be presently pointed out.
As shown in the station diagram. Fig. 1-1,
the multiple antenna has, instead of the single
ground wire usually employed, a number of
ground leads which are brought down from
the flat top at equally spaced intervals, and
connected to earth through appropriate
tuning coils.
The capacitive reactance of the flat top is
thus neutralized by inductive reactance at six
points to earth, instead of but one point as in
the ordinary system. The inductive react-
ance in each down lead is therefore made six
times the capacitive reactance at a given
frequency. The multiple antenna is thus
the equivalent of six independent radiators, all
in ])arallel and resonant to the same waA'e
length. Their joint wasteful resistance obvi-
ously is much less than that of an antenna
with a single ground, and herein lies the saving
of power which the Alexanderson antenna
brings about.
The relative power inputs required by
both types of antennae for the same value
of antenna current will be seen from the
following illustration: To maintain (i(l()
amperes in the multiple-tuned antenna at
New Brunswick, at a resistance of J4 ohm,
the power required is 600- X 0.5, or 180 kw.
To maintain the same antenna current in a
flat-top antenna with resistance of 3.7 ohms
requires 600- X 3. 7, or 13S0 kw. The economy
of power secured in the case of the multiple-
tuned antenna is an important consideration
from the standpoint of the cost of daily
operation.
Prior to the advent of the Alexanderson
antenna, theory and practice pointed to the
desirability of a very high antenna structure
Fig. 9. Schematic Diagram of Earth-wire System at the
New Brunswick Station
for long distance communication at high
powers, but as is well known, the cost of
erecting an antenna increases very rapidly
with the effective height. The multiple-tuned
antenna, however, pemiits the use of a less
expensive antenna structure, and gives the
same signal audibility at a given receiving
station as a high antenna of the old type with
less power. The example given above demon-
strates quite conclusively that the multiple
antenna will provide the same antenna current
as the flat-top type antenna, but only one
seventh of the power. The multiple-tuned
antenna is treated more comprehensively
on pages 823 to 833.
820 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 10
Earth System
The earth-wire system at the New Bruns-
wick station is a combination of a buried
metallic and a capacitive ground. Sixteen
parallel copper conductors are laid underneath
the antenna and buried one foot in the ground.
ax/^mpo^c- - »
-v tx/rixxff
.in,
equalizing coils is to increase the impedance
of the wires near the center and hence force
current in the outside wires. Since the coils
are wound in opposite directions they add
no appreciable inductive reactance to the
tuning circuits. In one instance, the use of
these coils reduced the multiple resistance of
the antenna system from 0.9 to 0.7 ohm.
A still better distribution of the earth
currents at New Brunswick was obtained
by using a capacitive ground commonly
known as a counterpoise, which is erected
underneath the antenna and a few feet above
the earth. A plan view of the counterpoise
is shown in Fig. 10. The capacitive ground
may be considered as a combination of a
tuned and a forced oscillation circuit, and
it has the effect of drawing the current from
the ground circuit more uniformly than with
wires hang on the ground or buried beneath
the surface. In practice the total current
in the down lead may be distributed between
the capacitive ground and the wire ground in
any desired ratio. The effect of adding this
unit to the system at New Brunswick was
to decrease the multiple antenna resistance
from 0.7 to 0.5 ohm. The capacitive ground
may be divided into separate units for each
tuning down lead or the units may be con-
nected together as shown. A schemaiii
diagram of the connections between the fla:
top and the capacitive and earth-wire ground.--
is shown in Fig. 1 1 , The equivalent circuit
is given at the right of the drawing. The
construction of the outdoor inductances for
multiple tuning is shown in Fig. 12.
Fig. 10. Plan View of Counterpoise at New
Brunswick Station
They extend the entire length of the
antenna and are spaced between
towers somewhat as shown in Fig. 9.
A network of wires and zinc plates
are also buried in the ground around
the station. At each of the five
tuning points outside the station,
connection is made from the antenna
flat top to the sixteen underground
wires.
In order to secure equal distri-
bution of current through the buried
ground conductors, equalizing coils
are inserted between the tap on the
down lead coil and the earth wires
at each of the five tuning points
outside the station, as shown in
detail in Fig. 9. The function of the
ax/yjiLcivr
o
— amur^
frc
P7*
AAAAAAA^A/\A
'WlVH.-VVmni Ml -•
Fig. 11,
Schematic Diagram of Antenna to Earth Connections of the
Multiple-tuned Antenna
THE ALEXANDERSON SYSTEM FOR RADIO COMMUNICATION
S21
Magnetic Amplifier
Telegraphic control of the large antenna
currents involved in high-power radio trans-
mitters has ever presented a difilcult problem.
Particularly has this been true when signalling
at high s]3eeds. Rapid signalling obviously
requires some device that will not cause
destructive arcs and will provide the desired
modu ation of antenna power without taking
upon itself the burden of carrjdng the full
power of the system, during the intervals
between signalling.
The magnetic amplifier is a device which
meets these exacting requirements, for it
provides a non-arcing control with a minimum
current in the key circuit, and it takes within
itself only a small proportion of the total
alternator output. A photograph of the
amplifier, removed from its container, is
shown in Fig. 13.
The magnetic amplifier in general may be
described as a variable impedance which is
connected in shunt with the external circuit
of the radio frequency alternator. Its func-
tion is to reduce the voltage of the alternator
and to detune the antenna system when the
sending key proper is open, and to perform
the opposite functions when it is closed. Thus
when the sending key is open the amplifier
short circuits the alternator and detunes the
antenna system, thereby reducing the antenna
current to a negligible figure. When it is
closed the output of the alternator is fed to
the antenna system.
A general idea of the operation of the
amplifier can be obtained from the funda-
^ 0*, 14
Fig. 12. Tuning Inductance for MuUiple-tuned Antenna
Fig. 13. Magnetic Amplifier Removed from Containing Case
mental circuit, Fig. 14, where it will be noted
that the radio frequency coils A and a control
coil B are mounted on a common iron
structure, and are so disposed that the effect
of the control coil upon the radio frequency
coils is obtained solely through the agency
of flux variations within the core. The im-
pedance of the amplifier is dependent upon
the degree to which the iron core is saturated
by the control winding. The saturation in
turn varies as the current is fed into the con-
trol circuit. When the control circuit is closed
the alternator is short circuited; when it is
open, the alternator assumes normal volt-
age and its output flows into the antenna
system.
The magnetic amplifier has been employed
in experimental telegraphic signalling at
speeds above 500 ivords per minute, at which
rates it functions without lag. It is equally
applicable as a modulator of antenna power
in radio telephony, in which case the control
current of the amplifier is modulated at
speech frequencies by a bank of pliotron
(vacuirm valve) amplifiers, which in turn are
controlled by an ordinary speech microphone.
The characteristics of the amplifier are
treated in greater detail on pages S3(i to 838.
822 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 10
Fundamental Station Circuit
The fundamental circuits of a typical
Alexanderson alternator station are shown in
Fig. 14. Beginning at the left of the drawing,
it is to be noted that a source of two-phase.
60-cycle alternating current drives an induc-
tion motor M, ha\-ing a wound rotor, the
circuits of which include a liquid rheostat R3.
The motor is connected to the radio frequency
alternator through a helical step-up gear.
The alternator armature coils are indicated
at A3, A4, the field coils at Fi, and the rotor at
A2. There are two sets of armature coils, one
on each side of the rotor, which as already
mentioned, are divided into 32 sections on
each side. The windings on each side con-
nect to the primaries of two transformers
shown at Pi, P2. The primary' of each trans-
former (see Fig. 3) contains two complete
turns of 16 wires in each turn, which carr\' the
current developed in the 32 sections of the
armature coils on each side of the rotor. As
can be seen from the diagram, there is no
direct electrical connection between the
armature circuits leading to the transformer
primar}-, but the individual primary- circuits
are disposed so that their magnetic fields at
any instant are in the same direction, that
is, their fields combine to operate on the
secondaries Si, So. In addition to the primarv-
and secondan,' coils, the two transformers
have intermediate coils S5 which are connected
in parallel and shunted by the magnetic
amplifier coils A. The coils Sj are connected
in series with the antenna system, and are
also closely coupled to the priman.- and
secondary-.
The multiple -tuned antenna, shown in the
upper part of Fig. 14, is a long, low, hori-
zontal aerial of the Marconi type, from
which are brought down leads to earth,
which include the tuning inductances Li, L;,
L3, L4, L5, Le. For any given wave length
ihe joint inductive reactance of the down lead
circuits Li . . . . Ls is made equal to the capacitivc
reactance of the entire flat top at the operating
frequency or wave length. The multiple
antenna is therefore the equivalent of six
independent radiating systems resonant to the
same wave length, and for all practical
purposes, the oscillating currents in them
flow in phase.
uuniPU rvNU juitiiima
nr'S; 'Hi'unnf
Fig. 14. Fundamental Station DiaKram of 200-kw. Alexanderson Alternator Set
Radio Corporation's Transoceanic Station. New Brunswick. N. J.
THE ALEXANDERSON SYSTEM FOR RADIO COMMUNICATION
823
The magnetic amplifier, shown to the right
of the diagram, comprises the parallel-con-
nected impedance coils A, which are connected
in series with the condenser Ci and the trans-
fonner amplifier coils S5. B is the control coil,
wound to include both branches of the wind-
ings A, which is fed with direct current,
regulated by the rheostat Rg. When the
control circuit is closed the impedance of the
amplifier coils A becomes a minimum; when
it is open the impedance is a maximum. In
the former case the alternator is placed on
short circuit and the antenna is detuned; in
the latter case the alternator assumes normal
voltage and its output flows into the antenna
system. In jjractice the capacity of Ci is
selected to neutralize the inductance of
windings A for some value of current in the
control coil.
The circuits of the speed regulator appear
in the lower left hand part of the drawing.
Note is to be made first of the variable im-
pedances N and O in the motor supply line
with their d-c. control coils P,; and the
variable impedance coils S7.
The extremely close speed regulation
essential to alternator operation is obtained
from the resonance circuit Lio, C4, P5, the coil
Lio being one of the alternator armature coils.
This circuit is made resonant to a frequency
slightly above the normal frequency at which
the alternator is to be operated and the cur-
rent developed therein acts inductively on
the circuit vSe, E, Mi (E being a rectifier).
The latter rectifies the radio frequency cur-
rent and sends a d-c. component through
Ml, which arts with an increase of speed to
decrease the voltage held by the voltage
regulator M2, Ti on the generator Ki. This
increases the impedance of the coils S7 and
therefore tends to reduce the speed of the
driving motor. As the speed now falls the
current in the resonant circuit falls off and
likewise that in the coil Mi. This permits
the voltage held by the voltage regulator to
increase, and therefore acts to reduce the
motor supply line impedance and thus increase
the speed. A given mean voltage is thus
maintained in the control circuit by generator
Ki, which depends upon the magnitude of
the control current in Mi. This keeps the
speed variation within exceedingly close limits.
Antenna Support
A standard tower for high-power stations
is shown in Fig. 15. This is of the self-support-
ing type erected on a suitable concrete base.
The antenna wires are suspended from the
steel cross arm at the top. This method of-
antenna suspension lends itself admirably to
the long narrow antenna which has been
found most suitable for the Alexanderson
system.
The antenna layout for a two-alternator
unit high-power station using these towers is
shown in Fig. 10 where two antenna wings
of any desired length extend in opposite
directions from the station house which is
located at the center. With this construction
the wings may be tuned to different wave
lengths and each energized by a single alter-
nator, thus permitting simultaneous trans-
mission at two different wave lengths; or the
two alternators may be joined in parallel to
energize both wings at some selected wave
length.
Fig. 15.
Section Standard Tower to Support Alexanderson
Multiple-taned Antenna
PERFORMANCE AND OPERATION OF
THE ALEXANDERSON SYSTEM
Multiple Tuned Antenna
The antenna? commonly used at high-power
radio stations may be broadly classified into
two types, viz., the long horizontal aerials
which are suspended on comparatively low
towers, and the vertical, fan or umbrella
aerials which are generally supported at great
heights. The flat-top antenna was adopted
for long distance transmission because it was
believed to have marked directional properties
and would therefore ]jrov-ide maximum radia-
tion in the direction desired and lesser degrees
of signal intensity in all other directions.
Experiment has indicated, however, that
this directional effect disappears at distances
beyond 300 miles or so from the transmitter
and thus the benefits of directional radiation
are realized only in a limited area. Beyond
this the flat-top antenna has been found to
have comparatively high resistance. This
may be said to be due to the long path
through which part of the ground current has
to pass to the far end of the antenna, which
is a path of relativ^ely high resistance. This
resistance cannot be materially decreased by
laving wires in the ground, for because of the
824 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 10
inductive impedance of such long wires (at
radio frequencies) a large percentage of the
ground current will still pass through the
earth. It is therefore evident that if the
length of the ground path in a radiating
system could be reduced, a considerable
sa\'ing of power would be effected.
At any given wave length the radiation
from an antenna has been found to be pro-
portional to the square of the effective height
and the square of the antennse current. The
, . . „. KiOO h-i- _, . . ^ ^
exact relation is 1 1 = ri; . This points to
A"
the desirability of a high antenna, but since
the cost of building such a radiating system
increases ver\^ rapidly with its height, the
factor of economy requires that the money
expended on a station be apportioned between
which is great compared with their horizontal
dimensions. It follows from simple electrical
principles that several antennae in parallel
will possess a lower joint resistance than a
long antenna of the same radiating capacity.
The result may be obtained from the ^Iarconi
flat antenna by bringing down leads from the
flat top, at regular inter\-als, to the ground
through appropriate tuning inductances. With
this construction it will be seen that the
antenna charging current has a much shorter
path through the down leads than it had with
the former design.
The improved efficiency of the multiple-
tuned antenna has been amply demonstrated
at the New Brunswick station where the
resistance of the Marconi flat top has been
reduced from o.7 ohms to 0.3 ohm with the
iTrnM
-izsort
-5000 ft-
Locatlon of Inductance Coll
II Z \ 3 4
Power House
Plan of Capacity Ground
1250 ft r
Plan of Antenna
5000 ft.
3
. ik
, laft.
. . . u . .
InbuctanceCoil ' Height of Capacity Ground
Elevation of Antenna
Fi3. 16. Antenna Construction and Counterpoise for T.vpical JOO-kw. Altcrnotor Installation
the cost of the antenna, power apparatus, and
m.aintenance in order to arrive at the lowest
total cost for transmission over a given
distance. It is obvious that if. by any means.
the wasteful resistance of the long, low, flat-
top antenna, that is. conductive losses, leak-
age through insulation, etc.. could be reduced.
and if its radiation jjroperties still could be
maintained, then assuming equal jjower
inputs into the two systems, a station using
a long, low and relatively cheap antenna
could produce the same signal strength as
that from a high and costly antenna.
The multiple-tuned antenna devised by
Mr. Alexanderson brings about a marked
decrease in the ground resistance of a flat-top
aerial. His antenna can be compared to a
station using a number of small antenna
connected in iiarallel. the height of each of
consequent saving of power pointed out on
page Slit.
Comparison of Radiating Qualities
The curves of Fig. 1 7 show the results of a
series of ex])eriments conducted between New
Brunswick, X. j., and Sdiencctady, N. V..
with the object of comparing the relati\-e
signal audibilities ampere for ampere in the
old aiitciiim u-itit ii single i^roiitiJ and the
Alexanderson iiiitcnna 'a-illi multiple grounds.
The results show quite conclusively that with
the same current in a flat-top antenna and
in a multiple-tuned antenna, substantially
equal audibilities are obtained at the receiving
station. However, the power required by the
jilain antenna for a given number of amixjres
is \ery much in excess of that fed to the
multiple-tuned antenna for the same total
THE ALEXANDERSON SYSTEM FOR RADIO COMMUNICATION
S2.-
current. Thus as the curve shows, to ]nit a
total of 70 amperes in the branches of the
multiple-tuned antenna with six grounds,
requires but 3 kw., whereas with the flat-top
antenna and a single ground, IS} 2 kw. are
required. This is, of course, a very small
proportion of the total output available at
New Brtmswick. The values shown in the
curve should not be taken as indicative of
those used in daily operation.
Theoretical Comparisons
The points of distinction between the two
types of antenna; may become evident from
the following comparative analysis. Thus the
flat-top antenna with single ground is shown
in Fig. IS. The equivalent circuit resolved
into lumped or concentrated values of
inductance and capacitance is shown in
Fig. 19. The schematic circuit of the
Alexanderson antenna is that of Fig. 20
where L\, L«, L3, L4, L5, Lg, are current paths
between the flat-top and the earth. The
inductance of each down lead is made six
times the capacitive reactance of the flat top
at the frequency of operation selected. The
capacitive reactance of the flat top is thus
neutralized at six places. The circuit is
alternator A'. The branches L, t'l, Lo C2,
L3 C3, etc., which are in shunt to one another,
are fed by the alternator. When each branch
is tuned to the frequency of the alternator it
will follow the well known laws for parallel
resonance. A large current will flow back and
nuiyALefJTOMjrr
SmSL£ /fNTf/ffM
Figs 18 and 19.
Fundamental and Equivalent Circuits of
Flat Top Antenna
^
1
I
1
?
Ml
VBIt
ry
ANO
AV
Ft
\a //
V/t/7
■CL
9m
/
1
M
W 6
fUVi
W/CA
■ y4A
■TTN
em
V4
'ML
Wirk
1
sm
■iu
HM.
IUI^
V
in
b
S
f
M
^
I
V
^
i
V
ff
m
SIX t
mu\
VJHO
i
^/
-c
jaoo
f>
»
too
/
f
r S"
^
^
€
ajo
/
/
Mll$.
tfS.
^
i»
4
ten
^
■^
->
p''
im
?
wo
— -
[^
^
fHA
(71*
•AjS.
/rj.
tiS
0
0
0
"
X)
"RFfi
40
A
'rerli
so
lO
70
1
Fig. 17. Curves Showing Comparative Signal Audibilities Obtained frcm
Alexanderson Multiple-tuned Antenna and the Open-ended
Fla'--top Antenna
therefore the equivalent of six independent
radiators operating in parallel.
The equivalent circuit of Fig. 20 is that of
Fig. 21, which is an artificial circuit com-
prising a number of parallel resonance
circuits adjusted to the frequency of the
forth between the inductance and the con-
denser, and the alternator will simply supply
power to compensate the resistance losses of
the circuits. These large currents
are directly due to the high voltages
maintained across the inductance and
the capacity, when the circuit is
tuned for resonance. These voltages
may be calculated when the value of
inductance or capacitance and the
current flowing therein are known.
If a parallel resonance circuit had
no resistance, the conditions for
parallel resonance would be strictly
the same as for series resonance.
These conditions are, however, very
closely realized in the parallel circuit.
In series resonance the e.m.f. on the
condenser is equal and opposite to
that of the coil and thus there is a
large flow of current between the
condenser and coil. There is also a
large current flowing between the
condenser and the coil in parallel
resonance, but viewed from the stand-
point of the feed or power supply
circuit, the feed current is simply the
difference of the currents in the condenser
and the coil.
The resistance of a parallel resonance cir-
cuit, in radio, is often treated as a negligible
quantity. This resistance, however, assumes
considerable importance in the multiple
826 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 10
antenna as it detennines the power taken
from the alternator. Thus if the wasteful
resistance of each branch in a multiple-tuned
arttenna of six branches is 2.7 ohms, their
joint resistance is 2.7/0 = 0.45 ohm (assuming
equality) and it is this resistance plus the
MULTrftr AVr£/Jf^/^
The oscillation frequency,
A' =
30,000,000
15,000
= 20,000 cycles
The capacitive reactance of 0.0G6 mfd. at
20,000 cycles
1
2 7rA^C
L--
1
(i,2,s;i2X 20.000 X0.000,000,0(>()
eooiyALfffr a/fcufr Mi/ir/Pis -f/VTT/V/W
= 120.5 ohms.
The inductance required to neu-
tralize the capacitiye reactance is
found from the relation
A'
L =
2 7r.V
120.5
Figs. 20 and 21.
^CUPPFf/r VALUCS ABCVF
Mrs /tssoMFo fW OfT/ur/av
AT /svotuTins
Fundamental and Equivalent Circuits of Alexanderson
Multiple-tuned Antenna
radiaiion resistance of the entire anteiuia
system through which the alternator works.
It is obvious that the alternator can be
connected as in Figs. 22, 2'.i and 24 with the
same effect as shown in Fig. 21. Thus in
Fig. 22 the alternator terminals are connected
in shunt to the parallel resonance circuits.
In Fig. 23 the alternator output is fed to the
antenna through the inductive transformer
P S. In Fig. 24 an auto-transformcr con-
nection is employed.
Multiple Tuning
In order to obtain resonance between the
alternator and the several radiators of the
multiple antenna; of Figs. 20 to 24, the joint
reactance or impedance of the down leads
Li, L«. L.-i, Li, Ls, Ae. must be chosen to equal
the capacitive reactance of the flat top at
some particular frequency. Hence with
multijjle tuning at six ]X)ints the reactance
of each down lead, for a given wave length
(or frequency), must be six times the capaci-
tive reactance of the whole antenna.
The method of computing the inductance
in the down leads for a given wave length is
as follows: AVe may take as a representative
examjile the capacitance of the New Bruns-
wick flat-top antenna, which is a long low
aerial of the Marconi type. Its capacitance
as measured is ().()()() mfd. Assume that
operation is desired al 15,01)0 meters.
6.2S32X 20,000
= 0.0()0,!).>S = 0.95S millihenry-
The total inductance of each down
lead should then be (5X0.958 = 5.74
millihenry; and the reactance of each
down lead, (iX 120.5 = 723 ohms.
Curves may be prepared to give the values
of inductance required to tune the multiple
antenna with various numbers of grounds at
different wave lengths. If then the line coils
be calibrated for dilTerent numbers of turns
at different frequencies, it is a relatively
simple matter to set these inductances to
Figs. 22, 23. and 24. Equivnlent Circuits of Alexanderson
Multiple-tuned Antenna
the correct value for any wave length. A
series of cur\-es showing the inductance
required to oi)erate the New Bnuiswick
antenna at various wave lengths are given
in Fig. 25. These are cited merely as illus-
trative examples.
THE ALEXAXDERSON SYSTEM FOR RADIO COMMUNICATION
827
Feed Ratio
The term ''feed ratio," for convenience,
has been applied to express the ratio of the
total current in the six radiators of the
multiple antenna to that flowing in the down
lead of the branch to which the alternator is
coupled. Assume that equal induct-
ances are inserted in each down lead.
With all other conditions equal, the
same current will flow in each of the
six circuits when supplied with energy
at the frequency which produces reso-
nance. *
Thus if the ammeter A, when con-
nected in series with the station down
lead. Fig. 21, indicates lUU amperes (at
resonance) . and the same current is ob-
tained in each branch, the total antenna
current is 6 X lOU = (iOU amperes.
The feed ratio is then equal to
Total Current
Current in the station down lead
which m ihis case = y— = b;l.
frequency 20,000 cycles, and the inductive
reactance at each down lead 728 ohms. If
now the wave length is reduced to 14,500
meters, the frequency increases to 20,700
cycles. This represents an increase of 700
cycles, which is 3 J4% of the original frequency
a 1
Wi
n
I !
¥///
t-
1
ii
'I
///
/
r^^
1
— 1
Ni
-f—f- —
//'; /
III' 1
1
///// ,■
1
//////
10000 UeOO 30000
WAVl IINGTH
Fig. 25. Graphs Showing Inductance Required to Tune the
Multiple-t jned Antenna at New Brunswick
to Different Wave Lengths
It is of interest to note that the above feed
ratio is only maintained when the inductance
in all the down leads is equal. Assume for
example, that the inductive reactance in the
branch through which the energy is supplied
is decreased and the frequency of the alter-
nator is raised for resonance. Assume also
that the feed ratio previous to this change
is 6:1, the wave length 1.5,000 meters, the
Fig.
26. Equivalent Circuit of Multiple-tuned Antenna for
Computation of Phase Difference
of 20, ()()(). It may be shown that 1% change
in frequency requires a 2% change of induc-
tance for resonance. Hence the inductive
reactance in the circuit for 20,700 cvcles is
100% -7% or 93% of the value at' 20,000
cycles; that is, 93% X 723 = 672 ohms.
Now if the five line coils to earth are left
unchanged and since each has an impedance
of 723 ohms at 20,000 cycles, or multiple
impedance of 723/5= 144.6 ohms, the imped-
ance at 20,700 cvcles obviously is 20,700/
20,000X144.6=149.6 ohms. The new feed
ratio is evidently proportional to the two
impedances or 672 149.6 = 4.49:1.
The value of this determination lies in the
fact that upon changing the wave length by
tuning at the station down lead only, the
new feed ratio can be computed, thus enabling
the operator to ascertain the correct feed
current necessary to maintain a given total
value of antenna current.
Phase Difference
After viewing the physical aspects of the
antenna layout in Fig. 19 it might appear
that a disturbing phase angle would exist
between the currents in the radiating circuit
embracing the alternator, and those in the
radiators placed at increasing distances from
the power source. It can be shown, however,
that for all practical purposes the currents in
all of the down leads are substantially in
phase. Thus in Fig. 26, the branch Le Ce.
since it is a tuned circuit, operates at unity
power-factor and therefore may be treated as
a non-inductive resistance of a value equal to
C7?(°''/?(2 7r.V=r=)
82S October, Uli'd
GENERAL ELECTRIC REVIEW
VoL XXIII. Xo. 10
If (at X= 15,000 m.) rB = ().011 mfd., Lf,=
(l.n(l.")74 henry and /('6 = 2.71 ohms, then the
impedance of any single branch to the e.m.f.
impressed thereon is equal to
O^OO^Ii = 192,500
0.000,000,011X2.71
ohms approximately.
Since the circuit Ls Ce Re is in resonance
with the e.m.f. impressed at Ti To, the current
in it is also in phase with the impressed e.m.f.,
which rra)- be considered to operate through
a non-inductive resistance of approximately
192,500 ohms.
Let now the inductance of the flat-top
between the fifth and sixth branches be
represented by L. The value of L is one-fifth
of the total flat-top inductance without
loading and in the case of the New Brunswick
antenna is approximately 0.00013 henry.
We then have in the last branch (Le Ce) a
current which lags behind the current flowing
in Li C:-, bv the angle 9 where
tan 9 = 5 —
_ 6.2632 X20.00()X 0.000 13
192.500
(which is negligibly small)
1
11,780
The phase difference between the sixth and
fifth radiator is thus negligible. The phase
difference between the currents in branch
Li Ci and branch Le, Cn is five times as great,
but it is still of negligible im])ortance. The
currents in the six radiators are therefore in
substantial phase, the effect of the inductance
between branches is negligible, and the
charging currents which are measured currents
in the down leads can be considered to tjc in
phase. Since the length of the antenna is but
a fraction of the wave length employed and
the phase difference is slight compared with
the wave length, no api)reciable directive
effects will be obtained.
Antenna Voltage
The antenna \H)ltage may be computed
when the equi\-alcnt capacitance of one
section and the current in the station down
lead, or the total antenna capacitance and
total antenna current arc known. This is
obtained from the relation.
c- 1 ,- 1
h = ^TT-p, or /i = . .
2 T A C A
where A' is the capacitive reactance of the
antenna at some frequency.
Using the values in the foregoing discussion.
assume that / as measured by an ammeter
in the station down lead is 100 amperes.
Then since the capacity reactance to be
neutralized by the down lead is one-sixth of
the whole capacity or 0.01 1 mfd., then
100
~ 6.2832 X 20,000 X 0.000,000,01 1
= 72,300 volts.
A current of 100 amperes performs the same
functions in each of the remaining branches,
so that the whole antenna is maintained at a
voltage of 72,300 volts by six separate cur-
rents, all in phase, of 100 amperes each. Since
the multiple impedance of the six branches,
as shown above, is 120.5 ohms, the total
antenna current is 72,300 120.5 = 600 am-
peres. This is merely a further proof of the
assiunption made at the outset.
As previously cited the branches of the
multiple antenna follow (except in one respect
explained below) the laws of parallel reso-
nance circuits with lumped inductance and
capacitance, and the current supplied to any
branch by the main or power supply circuit
is at any instant the algebraic sum of the
currents in the capacity and the inductance.
If there were no resistance in the branch
antenna it would have infinite impedance
to the power supply at resonance and no
current would flow in the feed circuit after
the initial e.m.f. has been applied. In the
actual circuit there must, however, be some
resistance and the energy for heating this
resistance must be supplied by the alternator,
that is, the alternator makes good this loss
of energy-.
The branch circuit of Fig. 26 at A' = 20.000
cycles, r = 0.01 1 mfd., /, = 0.00574 henry and
A' = 2.71 ohms, was shown to have an imped-
ance of apjiroximately 192.500 ohms. The
antenna charging voltage at 100 amperes is
approximately 72,300 volts. The energ>'
current supplied bv the power source to one
branch is therefore 72,300 192,500 = 0.375
ampere. The power supplied to each branch
is 72,300X0.375 = 27.1 kilowatts and to the
six branches (assuming equality throtighout)
6X27.1 = 162.6 kilowatts.
The foregoing method of computation while
correct for jjarallel rest)nance circuits with
lum])ed inductance and capacitance from
which feeble radiation lakes place, requires
some modification when the phenomena of
radiation from the multiple antenna is con-
sidered. Thus, in the multiple antenna, the
radiation resistance, whatever its value, may
THE ALEXANDERSON SYSTEM FOR RADIO COMMUNICATION
S2y
be said to be common to all six antennae,
whereas, the ground and coil resistances belonja;
to the different antenna; individually. The
combined circuit of the multiple antenna can
therefore be represented by a radiation
resistance common to all antenna; which is in
series with a group of six wasteful resistances
connected in multiple.
Thus assume now that the radiation
resistance of the individual radiators in the
multijjle antenna (at X=l.>.()()ll meters) is
O.Oti ohm and the ground and coil resistance
of each antenna individually, 2.()l] ohms. A
current of (U)0 amperes works through (J.Od
ohm radiation resistance, while 100 amperes
flow through each of the 2.().> ohm resistances.
The consumption of power in radiation is
()00=X0. 00 = 21. () kw., and in each branch
100-'X2.(i.")=20..> kw., or ()X20.o = 15!) kw.,
in the six branches. The total consumption
is therefore 1S().() kw.
The point to be brought out is that if the
radiation resistance of 0.00 ohms was added
to the wasteful resistance in each radiator,
and the energy consimiption computed there-
from, the result would be too small. Thus
assuming that the total resistance of each
antenna was taken as 2.().') + 0.()(> or 2.7!
ohms, the power in each radiator would be
27.1 kw. and in the six branches, 1()2 kw.,
but, as just shown, the correct value, when
the radiation resistance is treated properly,
is INO.O kw.
The multiple antenna rray be treated in
another way. With a total power consum]3tion
of ISO kw., the power supplied to each antenna
is 30 kw. and the energy current consumed
by each oscillating circuit at 72,300 volts is
0.4 1 5 ampere. Thus while the total oscillating
current is (iOO amperes the energy current
which flows horizontally from the power
source is 2.073 amperes. This distribution is
shown by the arrows. Fig. 21. In other words,
the energy fed to the system by the first
tuning coil in the form of 100 amperes at say
ISOO volts is transformed in the first oscillat-
ing circuit to 72,300 volts (in the case of the
particular problem cited) and distributed as
in a transmission line from which 0.415
ampere at 72,300 volts is drawn at five
places.
Multiple Resistance
When the inductance in each of the down
leads has been adjusted to provide resonance
with the alternator and the feed ratio has been
determined, the multiple resistance of the
Alexanderson antenna can be computed from
sim])le measurements taken within the station
house.
The process is as follows: Measure the
current in the station down lead at resonance
and then measure the open circuit voltage of
•the alternator (at the transformer secondary).
The voltage divided by the current gives the
"series" resistance of the antenna from the
standpoint of a load on the alternator. This
resistance is evidently the combined resist-
ance of the alternator and the "series"
resistance of the antenna system. The
resistance of the alternator must be obtained
from a separate measurement and subtracted
from this value to give the "series" or load
resistance of the antenna system.
Thus if the open circuit voltage of the
alternator transformer is 2000 and the current
in the down lead is 100 am])eres, the resistance
of the alternator plus the "series" antenna
resistance is obtained from R = E!I or R =
2000; 100 = 20 ohms.
Assume that the alternator resistance
(from the standpoint of the transformer
secondary) as obtained from previous meas-
urements is 2 ohms; then the series antenna
resistance (considered as a load on the
alternator) is 20 — 2=18 ohms. The mul-
tiple resistance of the antenna is then
equal to
Series Resistance
Square of the Feed Ratio
which in the problem above = -t^ = 0.5 ohm.
Proof of this formula is given below.
A set of curves showing the comparative
values of these two resistances at the New
Brunswick station for wave lengths between
2500 and 9000 meters are shown in Fig. 27.
Thus at X = S600 meters, the series resistance
is 32.5 ohms and the multiple resistance 0.9
ohm. It is the latter value that must be used
to compare the multiple tuned antenna with
the common antenna with single ground.
Curves showing the decrease of multiple
resistance at New Brunswick with increase of
the niunber of tuning points are given in
Fig. 28. It is to be noted that the data for
these curves and also that of Fig. 27 was
taken without the capacitive ground and the
current equalizers described on page S20.
In making measurements as above the
transformer must be regarded in all respects
as a part of the alternator, that is, the open
circuit voltage of the transformer secondary,
and the resistance of the alternator from the
standpoint of the transformer secondary must
S30 October, 192U
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 10
be treated as the voltage and the resistance
respectively of the alternator.
A proof of the formula Multiple Resistance =
Antenna Series Resistance
(Feed Ratio)-
MO
\
Y~
\
\
A
urr
«f5
if
WAfCC- Of- .
enaHoev -*nn
IM
SO
?4
\
^
TV ,
nriavictf /\ u
\ '
A
fit
puj^^a^^MrvanaMVuia^gnmB?
m
ze
\
\
?FSS
Mcr inKltrajrfV faff ami] V
\
\
1
Vl£
OfSfTAMCf SVC
VJ
II*
'1"
\
\
5r*/5 p** Tn? ivss
rrr
eO JfAT/Ol'
1
1 1
>
?
\
V
s^
1
\
\
s
mjtr
■PtF
«
J
^
^^
—
—
__
1
»
4
xm
»■-
"
xa
foa.
xoo
4«n
UH
aoo
1 Miims
JOOO
so,
Fig. 27.
Comparison of Multiple and "Series" Resistance i
Alexanderson Multiple-tuned Antenna
may be had from the following simple analy-
sis. Reference should be made to the equiva-
lent circuit, Fig. '29, which is assumed to be
inade up of a number of radiating systems in
Iiarallel. all tuned to resonance with the
alternator N.
Let E = open circuit voltage of transformer
secondan,'.
/ = current in the station down lead at
resonance.
Ra =the effective alternator resistance
from the standpoint of the
secondary.
/• =the "series" resistance of the
external or antenna circuit con-
sidered as a load on the alter-
nator.
Then
£ = / (R, + r)
from which
r=j-R.,
(Ra is obtained from
mcnt ) .
The i)ower consumed in the "series
circuit external to the alternator is
ir=/v.
Consider now the resistance of the complete
antenna from the standpoint of several
radiators in parallel :
Let F = feed ratio.
Then F/ = total antenna current in the
several radiators.
Also let 7?„ = multiple resistance
of the several radiators in parallel.
Then, the total energy in the
several radiators is equal to the
product of the multiple antenna
resistance and the square of the
total antenna current, or.
]V=iFir-R^.
This energA' obviously is the
same as that consumed in the cir-
cuit external to the alternator,
which as shown before =I-r.
Hence
(Fr)-R„ = r-r
from which
r
That is. the antenna multiple
resistance is equal to the "series"
or "alternator load" resistance
divided by the square of the feed
ratio. Expressed in terms of all the factors
involved
-Ra
Ra
se]iarate measurc-
or load
then,
»
1
Ml
It
^
n
p^-^JSKe^ \/\
If
■.IMS
^ ^^/
w
to
1
t
4
f \
m
/
1
^
/
1
'
m
/
/
1
s
./\
/
«
<
^-
m
nfu
_«c
?t_
~-
i 1
'
i
-HP'
flr^«
4
1
^ 1 \
c
L
Pig 26. Graphs Showing "Scries" and Multiple Resistance,
New Brunswick Antenna with Different Numbers
of Down Leads
THE ALEXANDERSON SYSTEM FOR RADIO COMMUNICATION
831
It is thus possible to compute the multiple
resistance of the Alexanderson antenna from
a few measurements made within the station
with instruments used in ordinary i^owcr
work.
Accurate measurement of the current in
each down lead is essential, prior to makin?
the foregoing measurements, as equal
divisions of current, due to physical
factors surrounding the station, can-
not always be obtained. Only in this
way can the true feed ratio be de-
termined.
of course, must be computed for each wave
length. The capacitive impedance for any
other wave length can be obtained from this
value, since impedance is directly proportional
to wave length. The inductive impedance of
each down lead should then be adjusted,
previous to tuning of the alternator, to a value
General
The muUiijle antenna can, under
some conditions, be used to advan-
tage with unequal currents through
the down leads although, in general. Fig. 29.
equality of currents gives the lowest the
resistance. This is apparent from
the fact that with unequal division some of
the current has a longer path to travel than
with equal division, making that jjarticular
branch of higher resistance. This also is
obvious from the fact that if a given amount
of current is to be passed through j^arallel
conductors their joint resistance will be less
if the division of current is in inverse pro-
portion to each path.
Unequal division of current is an ad\-antage
under two conditions. First, the "series" or
"load" resistance of the antenna can be
adapted to the voltage of the alternator, if the
alternator voltage cannot be adapted to the
antenna resistance. Second, by allowing
unequal division of current the wave length
of the system can be changed in a much
simpler manner than when equal division
is maintained. Each change of wave length
clearly requires a change in the inductance
of all the down leads to maintain equal current
division. If the inductance in all down leads
is not the same, the current will di\'ide itself
in inverse proportion to the inductance of
each path.
Further consideration will reveal that for
wide changes of wave length it may be
advisable to disconnect some of the down
leads.
Calibration of Ground Inductances
In order to compute the amount of induc-
tance tliat is necessary in each down lead
at some given wave length, the capacity of
the antenna must be measured by the
ordinary processes and its capacitive imped-,
ance calculated as shown on page 826. This,
Fundamental Circuit of Multiple-taned Antenna for Determinins
Distinction Between "Series" and Multiple Antenna Resistance
six times the capacitive impedance of the
antenna, if six tuning points are used. The
inductance of the down leads to the tuning
coils can be estimated roughly and the value
allowed for when placing tlie tap on the
ground coil.
The inductance of the tuning coils should
be computed for different numbers of turns at
different wave lengths and plotted in a series
of graphs as in Fig. 25. This will simplify the
operation of obtaining the correct inductance
for any wave length. In case there are no
means at hand of calibrating the tuning coils,
the required number of turns may be selected
by trial. The supposed num.ber of turns
required can be estimated roughly and con-
nected in all six down leads, but an allowance
must be made in the case of the station down
lead for the inductance of the alternator (or
for the inductance of the secondary coil of the
transformer). The speed of the alternator
may then be varied until resonance is found.
If the ntunber of turns selected tune at too
long a wave length too much inductance has
been inserted in the down leads, and if it
tunes at too short a wave length not enough
inductance has been added.
Capacitive and Wire Ground System
A general description of the earthing system
at the New Brunswick station has been given
on page 820. In the early experiments it was
found that when connection was made from
the tap on the down lead inductance to the
wire ground, the inner wires carried the
greater proportion of current, due to the fact
that they offered less impedance than the
832 October, li)21)
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 10
outer wires. A more equal current distribution
was obtained by inserting the equalizing coils
between the line inductances and the earth
wires as shown in detail, Fig. 9. These coils
are in inductive relation and are connected to
pairs of the buried wires as there shown. The
I
" TiTiTiT
^ARTHI , CAP/>CITY GWUNDS
.04 MF
Fig. 30. Equivalent Circuit Multiple-tuned Antenna.
New Brunswick Transoceanic Radio Station
effect was to increase the impedance of the
wires nearest the center and therefore to force
practically the same amount of current in the
outside wires as in the center wires. This
lowered the antenna resistance from (1.9 to
0.7 ohm.
A still better distribution of the earth
currents was obtained by installing the
counterpoise already shown in Fig. 10. As is
shown schematically in detail A, Fig. 1 1. the
section of the coil above the ground con-
nection may, for purposes of illustration, be
considered as positive with respect to the
ground and the section below the point at
which the ground is connected may be
considered as negative with respect to the
ground. The capacitive ground may therefore
be considered as a combination of a forced
and a tuned oscillation circuit. It has the
effect of drawing the current from the
ground more uniformly than with the wires
lying on the ground or buried beneath the sur-
face. The addition of the counterpoise in the
case of the New Brunswick station reduced the
antenna resistance from 0.7 to O.o ohm.
By suitable tuning, the total current
through the down leads may be distriliuted
between the capacitive ground and the wire
ground in any desired ratio. If the wire
ground is disconnected and the capacitive
ground is tuned to take all the antenna
current, the ca]jacitive ground then takes on
the characteristics of a tuned circuit. In this
case the wire ground may be connected to
the zero potential point on the coil (which
may be found by experiment), under which
condition it forms a path to earth for the
lightning discharges with no other appreciable
effect upon the system. An efficient ratio of
current in the wire and capacitive ground is
half of the total in each. The capacitive
ground may be installed in separate units
at each tuning point or may be connected
together as a single unit as shown in Fig. 10.
Taking into consideration the coun-
terpoise and buried wire ground, the
I equivalent circuit becomes that of
^ Fig. 30.
I It may be well to point out here
that the design and construction of
the grounding system for the multiple
_ antenna may undergo considerable
~ modifications in future high-power
installations. It is probable that the
system can be considerably simplified
and yet provide a lower total antenna
resistance than that obtained at the
New Brunswick station.
Radiation Efficiency
An antenna with a single ground and
effective height equal to that of the New
Brunswick aerial can be assigned at the
wave length of l.i.OOO meters, a radiation
resistance of 0.0(i ohm and a total resistance
of 2.71 ohms. This is, in fact, about the values
that would be obtained in practice. The
radiation efficiencv is therefore ().0(j 2.71 or
2.219c-
As a multiple tuned antenna the resistance
of the New Brunswick aerial is slightly under
0..") ohm, and the radiation efficiency is
0.0() ()..') or 12' (. The radiation efficiency of
the multiple antenna at this wav-e length is
therefore 129f against 2.21% in the indi\idual
antenna.
The radiation efficiency of the multiple
antenna is \cry much higher at the wave
length of NOOO meters which has been found
the most suitable for radio telephony. Thus
the radiation resistance of the New Brunswick
antenna at SOOO meters is 0.2 ohm and the
multiple resistance O.ti ohm. The radiation
efficiency is 0.2 '().() or 33%.
It is important to note that the New
Bnmswick antenna may be operated at the
wave length of 2.")()0 meters, although its
natural wave length as a flat-top antenna is
NOOO meters. Operation at such short wave
lengths obviously would not be ])ossibie with
the antenna in its old fonn. The multiple
resistance of the New Brunswick antenna at
2.")(I0 meters is 3 ohms, and the radiation
resistance is 2. 1 ohms. The radiation efficiency
is therefore 2.1 3 or ~V\^\\ whereas with a
single ground antenna the resistance at the
THE ALEXANDERSOX SYSTEM FOR RADIO COMMUNICATIOX
S.33
same wave length would be about 'iA ohms,
and the radiation efficiency, 2.1 .5.4 or 40'; f.
A cur\-e showing the computed values of the
radiation resistance of the New Brunswick
antenna, at various wave lengths, is given in
Fig. -i 1 . The multiple resistance as actually
measured at the wave lengths of "ioOO, ,S(JO()
and 13.()()(J meters is pointed out. The
radiation efficiency at these three wave
lengths should be noted, and also the com-
parative efficiencies of the common antenna
with the single ground and the Alexanderson
antenna with multiple grounds, at the wave
length of 13,()()U meters.
Although the radiation efficiency of all
tvpes of antenna; decreases with increases of
wave length the smaller absorption obtained
at the longer wave lengths offsets this
decrease. Efficient wave lengths for trans-
„,„ „j \C0Ml^[O PAiflATliti fffs/sKwcK
Fig. 31. Computed Radiation Resistance Multiple-taned
Antenna, New Brunswick Transoceanic Radio Station
oceanic communication have been found to
lie between 1(1,000 and 20.0(10 meters.
Alexanderson Speed Regulator
As pointed out on page SIS, in order to
secure a constant output at the alternator and
to prevent a diminution of the received cur-
rent at the receiving station, the speed varia-
tion of the radio frequency alternator, when
signalling, must be maintained within one
tenth of one per cent. It is evident that the
governing mechanism to maintain such con-
stant speeds must come into such a critical
state, at the motor speed to be maintained,
as to cause a high jjercentage of change in
itself for a low percentage change in speed.
The circuits of the Alexanderson speed
regulator have been shown in the fundamental
station circuit. Fig. 14. They are shown
separately in Fig. 32. Lio is an armature coil
which supplies a constant voltage at the
frequency of the alternator, d and P-, are a
capacity and an inductance which are tuned
to a frequency slightly above that at which
the alternator is to be worked. The coil 56 is
coiipled closeh" to Ph, but not so closely as to
affect appreciably the tuning of the resonant
circuit. £ is a rectifier (of the G-E Tungar or
Mercur\- Arc type) which is shunted by a
condenser C\ of 0.16 mfd. capacity.
M\ is an auxiliary control coil of the voltage
regulator. The latter through the contacts
7"i acts to control the voltage of a generator
K\. d is a condenser of 1 mfd. shunting the
coil Ml. Care is taken that the circuit Se, C4,
Co, is considerably off resonance with the
frequency of the circuit Lio, C4, Pi, in order
that the speed held by the regulator may be
changed with the greatest simplicity.
A' and 0 are variable impedances connected
in the two phases of the power supply lines.
They contain the d-c. control coils Pe and the
variable impedance coils S-,. R3 is a liquid
rheostat connected in the circuits of the
rotor.
The generator A'l, which is driven by the
motor Mi. is pro^•ided with field current from
a d-c. source of constant voltage which is
varied by the rheostat Rt.
In regard to the functions of the impedances
A' and 0. it may be said, in general, that with
zero current in the control coils Pe. their
impedance becomes a maximum. If on the
other hand the current through Pe is such as
to saturate the cores, their impedance
becomes a minimum. Any intermediate value
of d-c. control current will van,' the a-c.
impedance of the coils 5? accordingly.
It will now be shown how the motor input
may be varied inversely as the current fed into
the coil Ml from the resonance circuit brought
from a coil in the armature. Since the circuit
Lio, C4, Pi is resonant to a frequency slightly
above that of the alternator, it will develop
an increased current as the motor M speeds
S34 October, 192U
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 10
up. This will send a d-c. component through
the coil Ml which assists that flowing in coil
Ms; this causes the voltage regulator proper
to maintain a lower voltage at generator A'l.
This in turn decreases the current through the
coils Pe and therefore increases the impedance
in the power supply circuit, tending to
decrease the speed of the motor. When the
speed falls slightly the rectified component
Theory of the Speed Control Regulator
A series of graphs showing the phenomena
involved in the action of the speed regulator
are shown in Figs. 33 and 34.
In cur\-e A, Fig. 33, the "motor input" is
plotted against "per cent variation of normal
speed" with the normal line voltage and
frequency and with the resistance Rz (in the
rotor circuit of the motor) proper!}' adiusted
Fig. 32. Fundamental Circuits of Speed Regulator of the Alexandcrson
Radio Frequency Alternator System
through the coil Mi decreases, thus causing
the voltage regulator to maintain a higher
voltage on the generator /\'i and therefore
increase the control current through Pi. and
thus again decrease the impedance in the
power supply circuit. A given mean current is
thus maintained through the control coils Pe.
the value of which is determined by the value
of the current through Mi. The speed of the
driving motor is thus held constant.
to pro\ide the required power. The flat part
of the curve to d, indicates the motor input
with maximum field on generator K\. Fig. 32,
which is the result obtained with zerf> current
in the coil Mi of the voltage regulator. It
should be noted that the motor input with the
speed less than !»!t.*t.>'7 normal is well alxive
that required to drive the alternator with the
sending key closed. The motor will therefore
increase its speed up to point d. where the
THE ALEXANDERSON SYSTEM FOR RADIO COMMUNICATION
835
speed regulator takes hold. From here the
motor input drops off rapidly because of the
increasing current in coil Mi, (of the voltage
regulator) until its curve intersects curve B
which represents the power required to drive
the alternator at point e. Here the motor
input and the power required to drive the
alternator are equal and the speed will remain
constant.
When the key is opened, the power required
to drive the alternator drops off to that
indicated by the dotted line and the surplus
of power supplied to the motor speeds up
the alternator until the motor input has
dropped off to a value equal to that required
to run the alternator light. This condition is
represented at the interesction / at 100.05%
normal speed.
Point g represents the point at which the
speed regulator has decreased the motor
input the maximum amount possible, with
minimmn field on generator A'l; and for any
small increase in speed above this point, the
input will be the same as at g. Since here the
power required to drive the alternator is
greater than that sup])lied to the motor, the
motor will slow down until equality is
obtained as at point / with the kc}- open, or
as at point e with the key closed. With the
speed at point / when the key is closed, the
speed will decrease to point e, and when the
key is opened again, it will increase again to
that represented by /. This speed variation
being less than 0.1%, no inconvenience is
suffered.
MOTOR /MPUT A y
\
popvf/t Kisumeo to /e \ /^
DfflVe MTfRMTO/f / V/ . JVv
\
\/ TWtPTV(>C://T co/fPEnr
^^y A \ / '■;
...
-~_ \
7 A 9 ■■'■
PomiR \
c - — c' --''
9'
MS lOO.O >00 5
Fig. 33. Graphs Showing Certain Characteristics of the
Alexanderson Speed Control System
If, however, the characteristics of the speed
regulator are such that it lags in action, the
speed may fall below e, before the regulator
can effectively increase the power input. This
will cause a greater variation of speed than
would otherwise obtain. "Hunting" may
then take place and result in a speed variation
greatly in excess of the allowable variation
for constant alternator output.
The speed held by the regulator at a given
alternator frequency may be changed to some
other value by retuning the circuit Lio, C4 P5
through variation of its capacity or induc-
OTO 100 S
Fig. 34. Graphs Showing Certain Characteristics of the
Alexanderson Speed Control System
tance. This will change curve A. Fig. 33,
which will then maintain the same relation
to the curve C, thus providing a different
speed at which the power requii^ed to drive
the alternator will equal the motor input.
These conditions are represented in dotted
lines in Fig. 33, e' and /' representing the
speeds held with the key closed and open
respectively, and d' the point at which the
speed regulator takes hold.
To obtain proper regulation the speed
regulator must be adjusted so the point e will
be on the left or lower side of cur\'e B, for on
that side of the curve an increase in speed
will incur an increase in load (as resonance
in the alternator antenna circuit is approached) ,
which automatically will tend to keep the
speed down. On the other hand, if the point e
lies on the high side of the curve B an increase
in speed will decrease the load which will tend
to cause still further increase of speed. This
is prevented only by the fact that the speed
regulator causes the motor input to fall off
faster than the load falls off. Because of the
fact that better regulation is secured on the
low side of the curve, it is called the stable side,
and the high side the unstable side.
If the power supplied to the driving motor
is increased, such as by an increase in line
voltage or frequency, or by a change in the
setting of the motor circuits (such as a
decrease in the rotor resistance of an induc-
tion motor) the curve for motor input will rise
as to A", Fig. 34. If the power supplied to the
motor is decreased the ctirve of motor input
will fall as to A'".
S;5fi October. 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. X<
The motor adjustment must be maintained
so that point g on the motor input cur^-e will
be kept well below the power required to run
the machine light (as shown by the dotted
lines), and also point.t/ must be kept well above
the power required to drive the alternator at
system detuned. The joint effect of these two
phenomena is a reduction in antenna current
to 9% of its normal value. When the sending
key is closed, the alternator assumes sub-
stantially its normal voltage, the antenna
system returns to a state of resonance and the
Fig. 35. Magnetic Amplifier in Si-npli.^ei Form
maximum tune of the antennae. In case point
g is not well below the power required to run
the machine light a surge in line voltage or
frequency might increase it to g" where it
would be greater and thus cause the alternator
to run away when the sending key is left open
a short interval. Also if point d is not well
above the power required to drive the alter-
nator at maximum tune a slump in line
voltage or frequency might decrease it to d'"
and thus cause the machine to slow down to
e'" when the key is closed, with a consequent
falling off in signal strength and a swing in the
pitch of the received note.
If adjustments are made so that the condi-
tions outlined above are realized, no diffi-
culties are encountered in maintaining a
uniform speed at any desired alternator
frequency.
Magnetic Amplifier
This device already has been described as a
variable imjjedance connected across the
terminals of the radio frequency alternator
for the purpose of controlling the ]:)owcr input
to the antenna circuit. Its characteristics arc
alternator out])ut flows into the antenna
system.
The great advantage of the amplifier over
other methods of modulation is that it gives a
non-arcing control of the large currents
required in high-power radio transmission and
therefore permits rapid telegraph signalling.
In fact, the amplifier has been operated
experimentally at speeds in excess of .)(H(
words per minute with jierfect success.
An idea of the fundamental actions of the
amplifier can be gained from the circuit.
Fig. .'5.5, where the two windings designated
by .4 and R arc wound on a common iron core.
The windings .4 are connected in parallel
and shunted across the radio frequency
alternator .V. The coil B is an excitation
winding which includes both the positive and
the negative branches of the flux produced
in the windings .4. and hence, no voltages
are induced in B by the radio frequency
currents flowing in .4. This is illustrated by
the reference arrows in Fig. 'M\, which show
the direction of flux in the amplifier coil at a
l)articular half-cycle of the impressed current.
It is clear that the tcndencv to induce an
S
nmtnE
Fig. 36. Diagram Showing Inductive Action of Amplifier
Windings Upon the Control Winding
such that a relatively small current in an
excitation winding is enabled to control many
hundreds of ami)ercs in the antenna system.
The amplifier jjcrfonns two functions: When
the sending key jjroper is ojjen the alternator
is placed on short circuit and the antenna
e.m.f. in one side of the control coil by one
branch of .4 is counteracted by an opposing
e.m.f. in the other branch.
It is apjiarent that should the flux produced
in the core by the coil /> be sufficient to
saturate it fully, the impedance of windings
THE ALEXANDERSON SYSTEM FOR RADIO COMMUNICATION
S3 7
,4 would liecome that of a coil without an
iron core. On the other hand, with zero
current in the winding B. the core will be
magnetized by the windings A and the
impedance of ^4 will thus become a maximum.
In general, in order to obtain large flux
variations in the windings .4, the opposing
ampere-turns in B must be api^roximately
equal to those in .4. Utilizing the alternator
control circuit of Fig. .3."), the jjroblem is to
obtain a minimum impedance in the windings
.4 when the circuit to the excitation or con-
trol winding is closed and thus short circuit
the alternator; and to obtain a maximum
impedance when the control circuit is open,
so that the alternator may assume within
reasonable limitations its normal voltage.
In this way the necessary variation of the
antenna current for telegraphic signalling is
secured.
The characteristics of a magnetic amplifier
operated in a given instance as in F'ig. .'5o are
shown in the curve A, Fig. 37, where antenna
amperes are plotted against different currents
V
V
\
\
\\
1
\
X
\^4
rja/rsi
wnsr/!
\ .1
\
-~^
o.asM
<■ sairs
cmanvsa
^
a
uMrsfx
or anar
i«w\
/
8
y
Fig. 37. Control Characteristics of the
Magnetic Amplifier
in the excitation or control coil. The curve .4
shows incomplete modulation of the antenna
current, but it should be mentioned that with
this circuit it is possible to secure more
complete modulation with stronger currents
in the control winding.
A more sensitive control of the alternator
output to the antenna system can be secured
by the series condenser C\ of Fig. 38, for by
the use of this condenser a much smaller
control current is required to effect a given
^•ariation in antenna current. If the capaci-
A B
Mh^S
Fig. 38. Magnetic Amplifier with Series Condenser
tance of Ci is chosen to netitralizc the induc-
tance of the windings .4 for some definite
value of excitation current in the control coil
B, the impedance of the circuit Ci, .4, becomes
a minimum. The impedance at any lower
excitation is determined by the difference
between the inductive reactance of the
amplifier coil and the capacitive reactance
of the series condenser. However, the smaller
this difference the lower will be the amplifier
excitation which gives minimum impedance
and therefore minimum alternator voltage.
The increase in sensitiveness obtained from
the series condenser is well shown by the
curves B and C of Fig. 37. The curve A, as
already mentioned, shows the antenna cur-
rents for different control currents, without
the series condenser C\. The curve B shows
the control obtained with a series condenser
of 0.33 mfd. and the curve C with O.I2o mfd.
The curve B shows almost complete modu-
lation of the antenna current. Although it is a
matter of principal importance in radio
telephony it is pointed out here that the curve
B indicates a linear proportionality between
control and antenna currents almost through-
out its range. This is an essential require-
ment for satisfactory speech reproduction in
telephony. The excessive control indicated
at the right of point B with the larger values
of control current is a condition easily avoided
in practice.
In the final form of the magnetic amplifier,
the condensers Co and Cs are inserted in the
amplifier windings .4, as shown in Fig. 39.
Their function is as follows: If telegraphic
currents were introduced into the control
coil B with the condenser C-i and C,s absent, a
short circuit current would flow between the
branches of .4 without producing any flux
variations to the radio frequency current.
This, however, is prevented by choosing
838 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 1(7
values of C2 and C3 to have a low reactance to
the radio frequency currents and a high
reactance to the audio frequency currents.
These condensers have no appreciable affect
upon the tuning of the amplifier circuit.
In the commercial set the constants of Ci
are selected for the particular frequency at
which operation is to take place, and it is
therefore only necessary to vary the control
current in the coil B until the most complete
modulation of the antenna current is obtained.
In the event that the alternator is worked at
some frequency different from that originalh*
contemplated, a value of Ci can be found for
some definite value of control current in B,
at which a' minimum impedance in the ampli-
fier coils is'obtained.
Theoretical considerations of the circuits
involved and actual test show that this drop
in alternator impedance reduces the alternator
voltage and detunes the antenna system to the
extent that no more than 9% of the total
normal current flows in the antenna system
(when the current in the control winding is
zero).
In explanation of the control current of
IS amperes (fed by a 25(J-volt source) in the
case of a 2(J0 kw. installation, it may be said
that the same variation of alternator output
might be obtained with much smaller values
of control current. The larger value is
purposely used to permit rapid signalling,
that is, it permits the magnetic amplifier to
function without lag.
CO/L
V5 ''
L|ptf
Fig. 39. Magnetic Amplifier with Scries end
Short-circuiting Condensers
?S0/ D(l
^wnr-
or /^MPunen -
u
Fig. 40. Equivalent Circuit of Fig. 39
In summar\- of the foregoing the equi\-alent
circuit of Fig. 39 will be seen to be that of
Fig. 40 where the telegraphic key K when
closed reduces the impedance of the amplifier
and therefore the impedance of the amplifier-
alternator circuit. This simultaneously
detunes the antenna circuit and reduces the
alternator voltage.
Characteristic curves showing the ^■ariation
of alternator voltage, and change of alter-
nator-amplifier impedance with different
values of current in the excitation winding
(for the standard 200 kw. set) are presented
in Fig. 41. Thus with zero current in the
control circuit the alternator open circuit
voltage is 2000, and approximately oOO volts
with IS amperes in the control coil. Similarly
with zero current in tlie control coil the
altemator impedance is G7 ohms and it drops
to 37 olims with IS amperes in the control coil.
Radio Telephony
Since the magnetic amplifier provides a
linear control of the antenna current antl
functions with small values of control current,
it is applicable as a modulation device in
radio telei^hony. When telephonic currents
of suitalilc ami^litude arc ])asscd through the
control coil B. Fig. 30. similar variations of the
antenna current will be obtained, jirovided
the am])lifier characteristics are selected to
give linear proportionality; otherwise inac-
curate speech rej^roduction will result. It has
been amply demon.strated in practice that
such characteristics are readily obtainetl from
the amplifier. Thus the curves B and (". Fig.
37, both show the desired linear jiropor-
tionality between control currents and an-
tenna currents, but the curre B shtiws the
most complete modulation of the antenna
input.
THE ALEXAXDERSOX SYSTEM FOR RADIO COMMUXICATION
S39
The perfection of control provided by the
magnetic ampUfier has been well demon-
strated in a series of tests made on the 50-kw.
Alexanderson alternator. With a telephonic
control current varying in amplitude bv 0.2
ampere, the antenna current was changed
by the amplifier is here again well demon-
strated.
When the Alexanderson system is used in
radio telephony, the control circuit of the
amplifier is placed in the output circuit of a
bank of vacuum valve amplifiers. The input
CHA.^ACra?f5f/C5
1
i
\
Airt
X0>
rr
Tf»?M
PW
U7
TVS)
ceyti
vyvsa
MISM
COV.
rff£Q
■POL
cotis
COILS
WALL
pfAU.
a
L
1
mo
MOO
¥OLn
«■ n
w*
tSW7
C M
•PLin
■P Of
yf
i
??00
1
woo
mx)
\
J
^
xoo
\
V"
■ALTl
■JNATL
Q a
9V Ca
pcurr
kZX.1
#ar
7P
toe
\
to
/TOO
X
\
SH
vao
^
\
40
too
^
iLTfA
W7Z»
/MPi
■DAffC
-
X
soo
X
X
a?
400
^
lO
!00
2
4
i
ac'^
OfJTA
7.?
€P£5
ifffi£j
Fig. 41 Characteristic of Alternator-amplifier Circuits,
200-kw. Alexanderson Radio Frequency Alternator
from o.S to 42.7 kw., a variation of almost 37
kilowatts.
An oscillographic record taken on the
200-kw. set at Xew Brunswick, X. J., with
Secretarv- Daniels, of the U. S. Navy Depart-
ment, at Washington, D. C, speaking to
President Wilson aboard the U. S. S. George
Washington at sea is shown on page 798,
Fig. 7. The satisfactory operation provided
circuits of the amplifier bank are controlled
by three preceding steps of vacuum tube
amplifiers, which in turn are actuated by the
microphone.
In a number of experimental tests made
with the telephone set at New Brunswick,
the voice was projected to European stations.
At distances up to 2.500 miles very satis-
facton,- results were obtained.
S40 October, 1920
GEXER^JiL ELECTRIC REVIEW
Vol. XXIII. Xo. 10
Some Practical Operating Features of Tungsten
Filament Electron Tubes
By W. C. White
Rese.^rch L.\bor.\tory, Gexer.^l Electric Comp.\xy
In describing the operating features of tungsten filament electron tubes, as affecting the limitations and
possibilities of these tubts, the author analyzes the subject with respect to the filament, grid, plate or anole,
bulb and glass, vacuum conditions, tube circuits and their operation, and power supply. The information
contained is of value to all who have to d<-al with electron tubes, especially to those who are interested in
making experimental set-ups or in changing in some way apparatus in which the tubes are a part. — Editor.
General
During; the past few years a great deal has
been published upon the theon.- and methods
of using electron tubes, or vacuum tubes as
they are often called, particularly in the field
of radio communication.
In this article it is assumed that the reader
has some knowledge about the principles and
functions of vacutmi tubes. It is written for
those who have, or may have, occasion to
operate them.
When these tubes are an integral part of a
piece of apparatus, the designer has taken into
account many of the limitations and possi-
bilities of the tubes, so that some of their
characteristics are not obser\-able. However.
when the tubes are used in some experimental
set-up, or the apparatus of which they are a
part is changed or used in som-? special way,
unusual or unlooked for effect, often occur.
In this article some of these effects will be
discussed and some of the more unusual
characteristics of the tubes given.
The Filament
Since in most tubes of the kenutron and
pliotron type a tungsten filament is the
electron emitting cathode, some of the char-
acteristics of a tungsten filan ent as they
apply to these tubes will first be gi\'en
As used in vacuum tubes the useful range
of temperature for the filament is 2300 deg
K to 2700 deg. K.* At lower temperatures
the electron emission is very low and at
higher temperatures the life of the filament
very short. In this range of temperature the
resistance of the filament is approximately
\:i to 1() times as high as at room temperature.
Owing to this changing resistance the re-
lation between voltage and current is not
linear. In the operating range given above, a
1 per cent change in current causes approxi-
* These temperatures are expressed in the absolute or Kelvin
scale which is degrees centigrade plus 273.
mately a 1^4 per cent corresponding change in
voltage, this change in voltage being slightly
higher at the higher temperatures and lower at
the lower temperatures. In vacuum tube work
the two factors we are most interested in are
electron emission and life to filament bum-
out.
The electron emission varies rapidly with
the temperature and in amount follows the
curve shown in Fig. 1 . where for different
temperatures it is plotted as milliamperes
emission per watt of energ^• used to heat the
filament. This function is independent of fil-
ament length and diameter.
In the operating range a 1 per cent change
of filament current makes approximately a
20 per cent change in electron emission. A
1 per cent change in filament voltage makes
approximately an 1 1 per cent change in
emission. These changes are slightly less at
the higher temperatures and slightly greater
at the lower temperatures.
The filaments of vacuum tubes are usually
operated at ap])roximately constant current
or constant voltage. At constant current the
emission increases considerably during life: it
may reach double the initial value just before
burn out. At constant voltage the emission
may drop slightly or remain practically con-
stant during life.
The life of a filament in a vacuum tube can-
not he accurately ])redetermined by calcula-
tions based only on its dimensions and operat-
ing temperature.
If careful tests on a considerable nimiber of
tubes of a particular type have been made and
an average life determined then the life of a
tube of similar construction but with its fil-
ament of a different size, or operating at a
different voltage, may be calculated with suf-
ficient accurac>' for practical purposes.
Tests on a large number of filaments ha\c
shown that on the average a filament burns
out when a certain proportion of its mass has
TUNGSTEN FILAMENT ELECTRON TUBES
S41
been evaporated. Therefore, for the same
emission and temperature a larger diameter
filament will have a longer life.
A .'3 per cent increase or decrease in filament
current will respectively halve or double the
life of a filament. This shows the great gain
in tube operating cost that may be effected b\-
careful regulation of temperature. The cor-
responding figure for filament voltage change
to halve or double life is approximatelv .1 ]jer
cent.
With a filament operating at constant
voltage, the life is approximately three times
longer than with operation at constant
current. For this reason adjustment and
maintenance of filament temperature hx
voltage is to be preferred. It has usually
lieen the custom to adjust vacuum tube
filaments by current readings on an ammeter.
This was because in the early manufacture of
tubes a more uniform emission was obtained
by a current rating. However, modern
methods of tungsten tube manufacture insure
sufficient uniformity to allow voltage rating.
With operation at constant voltage the cur-
rent will drop .") to 10 per cent before burnout
t
,5
^
/
^•"'
/
^ "
/
>J
1
•4
J
^ ,.
/
^
1
f
1
1
f
1. ,|3
/
S
/
vJ
/
1 .
/
/
y
^
^
--
^
2100 2ZOO Z3O0 2^00 ZSOO 2600
Filament Temperature, Degrees Absolute
Fig. 1. Curve Showing Milliammeters of Electron
Emission Per Watt of Energy Used to
Heat the Pliotron Filament
occurs. However, as has been pointed out,
the emission varies only slightly and therefore
the operating features vary but little.
There is a factor which is negligible on
receiving and other low power tubes but
which is an im]X)rtant factor on high voltage
power tubes. This is the effect of the combi-
nation of the electron current of the plate
circuit with the current in the filament causing
a change in filament temperature and there-
fore a change in electron emission and life.
If the filament is operated from a 110 or 220
volt direct-current source with a series resist-
ance in the circuit, the electron current will
add to the filament current and increase its
temperature if the negativ'e of the plate volt-
age source is connected to the negative
filament terminal. It will subtract from it if
connected to the positive filament terminal.
This effect will not be uniform over the entire
length of the filament but will be variable
being a maximum or minimum at the end of
the filament to which the negative terminal
of the plate source is connected.
In case the filament is operated from a few
cells of storage battery or directly from a low
voltage direct-current generator so that little
or no series resistance is used, it is immaterial
whether the return from the plate circuit is
made to the positive or negative terminal of
the filament ; the heating current in the nega-
tive side of the filament is augmented by the
same amoimt wherever the return connection
is made.
The series filament resistance is essential
to any alteration in the distribution of the
flow of plate current through the filament
circuit as a safety precaution. In any case
the filament regulating resistance should be
so connected as to cause the additive currents
to flow through it while the differential
current flows through the low resistance side
of the heating circuit.
This result is always accomplished by con-
necting the return to the positive side of the
filament, regardless of whether the regulating
resistance is connected to the positive or to
the negative terminal of the filament voltage
source. It might also be accomplished by
connecting the return between the negative
terminal of the filament batterv and the reg-
idating resistance in the case where this rheo-
stat is on the negative side of this battery.
When it is remembered that the effective
plate current of a tube in the oscillating con-
dition usually has a value between 2 per cent
and 7 per cent of the filament current and
that, as previously stated, a 3 per cent increase
or decrease in filament current halves or dou-
bles the life, the importance of this effect is
apparent.
This filament heating effect of the plate
current also has another important aspect.
S-t2 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 10
If for any reason a high voltage, high power
tube stops oscillating the plate current will
usually rise to a value limited only by the
filament emission. If the connection, as ex-
plained above, is such that the plate and fila-
ment currents are additive this abnonnally
large plate current will increase the tempera-
ture and therefore the emission, which in
turn increases the plate current, this effect
being accmnulative, often destroying the fil-
ament in a few seconds.
If possible, alternating current should be
used for filament excitation with the regulat-
ing resistance in the power side of the trans-
former and the return of the grid and plate
circuits made to a center tap of the coil
supplying the filament. This connection
assures minimum disturbance in the plate and
grid circuits from the frequency of the fila-
ment source.
The effective plate circuit current in this
case divides evenly between the two filament
legs. Also the direct-current electron current
and the alternating-current filament current
add at a 90 deg. displacement to give the
combined heating current so the additive
effect is much smaller. Take, for instance, a
1 ampere electron current to the plate and
a 4 ampere filament current. If the latter
were direct current the current at the hottest
part of the filament would be o amperes, while
if it were alternating current it would be only
4.13 amperes and equally distributed.
In operating tungsten filament pliotron
tubes the following ])oints should be kejit in
mind to insure the maximimi possible life :
(1) Do not exceed the maximum rated
filament current or voltage, and in all cases
reduce the filament current to as low a value
as possible consistent with satisfactory o]-)cra-
tion.
(2) In order to make jjossible the reduction
of filament current mentioned, the best adjust-
ment of the set is the one giving the desired
result with the lowest value of ])late current.
(.3) It is poor policy to materially raise
the filament temperature to obtain a slight
increase of output.
Pliotron tubes are usually designed to oper-
ate in a certain position, that is, horizontally
or vertically, with a certain side or end up.
This preferred mounting position is usually
specified for each type of tube. This is
necessary because a hot tungsten filament has
a tendency to sag very slowly and unless the
tube is correctly mounted the filament may be
so displaced as to change the electrical
characteristics or even cause it to short circuit
against the grid.
Severe vibration, particularly rapid sus-
tained vibration, as from a high speed engine,
greatly accelerates filament sagging. Under
such conditions tubes should be spring sus-
pended.
In a detector or amplifier circuit in which
there is a telephone receiver directly in the
plate circuit it will be noted that, with all
other circuits closed, when the filament is
switched on there will be no click in the tele-
phone; when the filament is switched ofT there
will be a decided click. This is accounted for
by the fact that when cold there is no emission
from the filament so that the current through
the telephone rises slowly with the filament
temperature. However, when the filament
current is turned off^ the voltage drop along
the filament changes to zero instantly before
the filament has started to cool. This sudden
disappearance of voltage drop along the fil-
ament changes its average voltage with re-
spect to the grid, which of course, in turn
causes a change in ])late current and this
produces the click in the telephone.
Assume a filament taking 2(> volts to
o]jerate at normal temperature: suppose
also that the plate voltage is low. about .)(•
volts, and the grid is connected to the negative
end of the filament. The grid is therefore
negative to the entire filament length with the
exception of the extreme negative end. Under
these conditions of an effective negative grid
the plate current is greatly reduced. Now if
the filament current is suddenly switched oft"
the entire grid and filament are at the same
])otential so that there will be a sudden in-
crease of plate current. This is demonstrated
in a long filament tube operated as described,
with the filament above rated temperature
(so that emission will continue for a longer
time after current is switched off i and a cur-
rent indicating instnmicnt in the plate circuit.
If the grid connection is changed to the posi-
tive end of the filament the plate current will
make a sudden decrease on opening the fil-
ament circuit.
Occasionally while the filamcTit is cold
unusually severe vibration or mechanical
shock may loosen a filament weld or break a
lead or soldered connection in the base or
stem of the tube, the resulting contact re-
sistance being so high that no current flows
at the low voltage of the filament supply.
Occasionally in such cases a temporary emer-
gency rei)air can be elTected by connecting
the filament in circuit from a 110 or 220-volt
TUNGSTEN FILAMENT ELECTRON TUBES
S43
direct-current source with sufficient resistance
to bring the current to about rated vahie and
also inchiding an inductance such as a trans-
former coil or field winding. If, then, the fila-
ment terminals are short circuited and the cir-
cuit suddenly opened at this point, the voltage
set up by the inductive effect will often cause
a slight welding together of the loose contact
surfaces.
The Grid
In the construction of a transmitting tube
and its base, the grid and filament elements
are insulated from one another sufficiently for
all normal conditions of operation. However,
in experimental work unusual conditions may
arise which will set up voltages between the
grid and filament as high as ten times that
normally present. It is impractical to build
a tube to take care of this very abnonnal
voltage which only rarely occurs due to in-
correct adjustment.
A spark gap should therefore be provided
between the grid and filament terminals at
or near the base of the tube. This gap should
be set at y2 in. to }i in. depending on the
voltage employed and the number and type
of tubes used. This precaution should be
taken on any tube or group of tubes delivering
over 50 watts of alternating-current energy
or operating at a plate potential above 2000
volts.
In experimental or temporary wiring, great
care should be taken not to confuse the wires
leading to the plate and grid. A high positive
potential applied to the grid which is close to
the filament and of relatively small mass may
overheat or even melt it.
Occasionally when a receiving or low power
type of tube has been operated over a con-
siderable period of time, a slight conducting
deposit will form over the glass of the seal,
giving an electrical leakage path between the
grid and filament terminals. This will greatly
impair the operation of the tube as an ampli-
fier or detector. This condition can be re-
moved by connecting the grid terminal and
one filament terminal across a xe i'n- to 3^ in.
spark gap of an induction coil. The thin
conducting fihn will be disrupted in a few
seconds by the high tension discharge.
Plate or Anode
Tubes are usually designed tmd constructed
to give a safe continuous dissipation of a cer-
tain amount of energy. It is desirable that
tubes should be able to operate for a long
period of time with rated voltage on the fil-
ament, and with the plate and the grid at zero
voltage. In the small transmitting tubes this is
possible, but in the higher power high voltage
tubes this involves too expensive a construc-
tion and fuses or circuit breakers are used in
the plate circuit. However, any tube in order
to be conservatively rated should be capable
of dissipating continuously from the anode
an amount of electron boinbardment energy
equal to the output. In other words, unless
the conditions of operation are unusually
favorable as regards protective devices and
attendance, an efficiency of over .50 per cent
between plate voltage source and output
should not be relied upon.
The three metals used mostly as anode
material are nickel, molybdenum and tungs-
ten.
By good design and careful exhaust treat-
ment, nickel can be operated at a just visible
red heat, but operated in this way the tube
has a very small factor of safety against over-
load due to a variety of causes.
Molybdenum is the most common anode
material for pliotrons. With ordinary exhaust
methods it will safely dissipate 10 times more
energy per unit area than nickel. The melting
point of molybdenum is about 25o5 deg. C,
while that of nickel is only about 1450 deg. C.
Molybdenum as an anode can run continu-
ously at a good red heat. This high melting
point, together with the favorable mechanical
properties of molybdenum, makes it almost
an ideal anode metal.
Tung.sten with its higher melting point and
low rate of evaporation will dissipate safely
even more energy than molybdenum, but its
mechanical properties oft'er many difficulties.
It is altogether probable that, as vacuum
tube engineering develops, one basis of the
power rating of a tube will be a factor based
on the material, area and form of the anode:
that is, each metal will have a certain allow-
able energy dissipation in watts per square
inch of exposed area. At the present time
molybdenum anodes are usually designed for
a total heat dissipation (filament watts and
electron bombardment watts) of 30 to 50
watts per square inch of exposed area.
These are conservative figures.
Bulb and Glass
In power tubes the glass bulb rttns at a
temperature of the order of 100 deg. C. or
even more. In order to avoid strains in the
glass which are liable to cause cracks, care
should be taken to prevent liquids or cold
metallic bodies from coming into contact with
844 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 10
the hot bulb. Care should be taken also not
to scratch the glass as it is possible thus to
start a crack which may ruin the tube.
In the small types of power tubes all the
lead-in wires are usually carried in through a
common stem and seal. If this is the case, the
plate voltage which may be used is limited by
electrolysis in this seal. Hot glass is a con-
ductor of electricity, and the conduction is
accompanied by electrolysis, that is, the me-
tallic elements in the glass appear at the
negati\-e pole and the negati\-e ions (usually a
gas in this case) at the positive pole.
This electrolysis ruins the seal, making it
leak air and sometimes even cracking it. An
early indication of this electrolysis which
appears long before leakage occurs is a black-
ening of the grid leads in the glass of the seal
just beyond the point of entrance of the lead
into the glass from the vacuum side. This
action takes place at the grid terminals as
these are the most negative when the tube is
oscillating. The black layer is, in most cases,
lead deposited out from the glass. As stated,
this is not a danger signal but merely an in-
dication that electrolytic action has started.
Therefore if small tubes of this type are
operated considerably above their rated plate
voltage the life is liable to be terminated by
leakage of air due to electrolysis rather than
filament burnout.
Vacuum Conditions
Glass contains gases and vapors which are
liberated by heating in a vacuum. During the
exhaust treatment of tube manufacture this
process is carried on to as high a temperature
as the glass will stand.
During operation of the tube it may be
possible to so overload it. without jjroper ven-
tilation around the bulb, that gas is liberated
from the glass in sufficient quantities to affect
its operation. Pliotron transmitting tubes
are designed and rated to carry their normal
load without artificial cooling, natural ven-
tilation only being required. However, in the
case of overload, or operation in a small,
entirely enclosed space, or in vers- hot sur-
roundings, an artificial cooling of the bulb by
a current of air is beneficial in maintaining
the vacuum.
In some of the larger types of transmitting
tubes the ])late or grid lead is brought out
through a separate seal in the bulb, a small
lead wire connecting to the electrode. Under
ordinarv conditions of operation this lead
wire will be called upon to carr\- only a few
milliamperes in the case of a grid lead or a
fraction of one ampere in the case of a plate
lead. However, under certain circuit condi-
tions, one of which will be described later,
ven.- high frequency oscillations may occur
(10,000.000 cycles per second or greater),
in which case the capacity current to the grid
or plate may reach 10 amperes or even more.
One result of the abnormal current is to over-
heat portions of the tube that will not carry
this ver\' heaw current, such as the sm.ail
lead wires mentioned. This overheating may
cause liberation of gas.
If from some cause the gas pressure in a
tube during operation rises to a sufficient
value a glow will appear in the bulb as a result
of the positive ionization. If this gas pressure
is due to gases evolved from the metal or glass
the glow will appear blue. If, however, it is
due to leakage air it will appear purple or pink.
If air leakage into the tube has increased the
pressure sufficiently to prevent a pure electron
current, a dark blue or black oxide of tungsten
from the filament will appear on the grid or
plate. If a still greater air pressure is present
the combination of hot tungsten and oxygen
will cause the formation of a yellow oxide of
tungsten which floats in the tube like a heavy
vapor and which deposits on the bulb or
electrodes.
Vacuum Tube Circuits and Their Operation
A great deal has been i)ublishcd on this
phase of the subject, particularly receiving
circuits, and therefore comments in this article
will be confined almost entirely to oscillating
circuits for supplying high frequency energ^•.
There is one point, however, in connection
with the use of vacuum tubes as amplifiers of
ver\- small amounts of energy, such as radio
signals, that does not seem to be as widely
appreciated as it deserves. This fact is the
effect of energy loss in the grid circuit which
may arise from a \ariety of causes.
One of the characteristics of a vaciuim tube
which gives it such value as an amplifier is
that its control can be effected by almost a
imre potential effect. In other words, when
operating under i)roper conditions for the
amplification of audio frequency currents the
grid current and therefore the energy ab-
sorbed is exceedingly small: it may even be
considered zero for most cases.
Still another way of expressing this fact is
to sav that the input impedance of the tube
is \'er\' high, several megohms.
In the design of audio frequency am]>lifiers.
using transformers between stages. a<l\ antage
is taken of this fact and the available energA"
TUNGSTEN FILAMICNT ELECTRON TUBES
S45
to be supplied the grid of each tube is stepjjed
up by a high turn ratio to as high a voltage as
possible. Therefore in order to realize the
full amplification possible it is necessary to
adjust the filament and grid voltage on the
tube to give a sufficiently high input im])ed-
ance, and to avoid any leakage resistance path
between filament and grid or grid and plate
in the circuit or at the temiinals. A leakage
of the order of one megohm will often consid-
erably reduce amplification.
In the use of a tube as a power oscillator,
there are a number of jjoints which often
cause trouble.
A type of oscillating circuit commonly used
in experimental work is shown in Fig. 2 and
some of these points will be taken up in con-
nection with this circuit. If a direct-current
ammeter is used to measure the input energy
to the tube, it should be placed as shown at
Ai, that is, between the source of the direct-
current voltage and the by-pass condenser
Co. If placed in some part of the circuit
carrying a considerable component of high
frequency current it will indicate the average
value instead of the effective value, and cal-
culations of input energy based on its indica-
tion will be low. If a very high voltage cir-
cuit is being used so that a meter in the high
side is dangerous, it may be placed in the
negative high voltage lead, providing it is
])laced on the generator side of the by-pass
capacity C2.
It is advisable to use fuses or a circuit
breaker in the high voltage circuit. They
should be located in the circuit as near the
generator as possible. If obtainable, a fuse
rated at about double the normal plate current
should be used.
In a capacity coupled type of circuit (the
type shown in Fig 2) the lead wires from the
plate and grid terminals of the tube should
be short to the point of connection to C1L2 and
CLi res]3ectively. If these wires are a few
feet in length, the tube is very liable to oscil-
late at a very high frequency that is independ-
ent of the constants Ii, L2, and C, but deter-
mined by the inductance of the lead wires
mentioned and the capacity between the
electrodes inside the tube.
If for certain measurement work, it is quite
necessary to keep the generated frequency
very constant even though the voltage supply
varies somewhat, a great deal can be accom-
plished in that direction by using a very high
value of grid leak resistance Ri. This steady-
ing effect is due to the fact that an increasing
supply voltage decreases the impedance of the
tube and also increases the output which, if
Ri is high, increases the negative grid voltage,
which in turn tends to lower the impedance
again towards its initial value.
With a decreasing voltage supply the re-
verse effects occur.
+ A
i — 6 +
Fig. 2. Common Form of Oscillating Circuit
for Experimental Work with Pliotrons
In capacity' coujjled circuits in general the
coupling and oscillating circuit capacity C is
connected across the terminals of the plate
direct-current supply through the inductances
Li and Lo. In case of a breakdown of con-
denser C this direct-current supply would be
short circuited. It is therefore advisable to
include a second condenser Ci in series with
C. It may be of any capacity, large compared
withCi, and should have a dielectric strength
sufficient to insulate for the value of direct-
ctirrent plate voltage used.
In any type of oscillating circuit employing
a high power tube and particularly several
in parallel the ultra high frequency oscillations
mentioned are liable to occur. One good way
of preventing these oscillations is to insert a
very small inductance (a few microhenries)
in the grid lead of one or more tubes as close
to the grid terminal as possible.
In radio telephony the type of circuit most
used includes a modulator tube, the function
of which is to amplify the energy from the
telephone transmitter and thus to cause an
audio frequency variation of voltage in the
S4G October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 10
plate circuit of the oscillator tube. Therefore
the plate voltage.s and currents in the oscilla-
tor tube at certain parts of their cycle reach a
peak value about double the peak value they
reach in an oscillator circuit alone. Thus for
satisfactory operation, the oscillator tube in
such a circuit must have approximately double
the emission necessary for a simple oscillator
tube. Under these conditions the tube is also
delivering double the energy output.
Power Supply
In the use of direct-current generators to
supplv plate potential to pliotron tubes
operated as power oscillators there are several
factors which are of importance.
The plate current in a three-element
vacuum tube is controlled by the grid voltage.
The plate current can be instantly brought to
zero by a sufficiently negative grid voltage.
For this reason the plate current may be
brotight to zero far more quickly than is pos-
sible by opening a switch of some sort.
This very sudden cessation of current
causes high voltages to be built up across any
inductances in circuit including the generator
arm.ature windings.
On voltages above .500 and power outputs
above .50 watts, some sort of protective device
to safely limit and discharge this voltage
should be used. For this service aluminum
cell lightning arresters are very suitable.
They should be connected across the genera-
tor terminals.
Generators should not be over compounded
to any great extent, because a breakdown in
the tube or circuit will often cause a current
heavier than normal to pass, which will tend
to greatly overheat the anode. Over com-
pounding will aggravate this effect.
For radio telegraphy, using beat reception,
speed regulation of the driving engine or
motor from no load to full load is of more im-
portance than voltage regulation of the gen-
erator. This is so because in the generator
the voltage decrease with load is due to arma-
ture drop, and this occurs instantly and so
does not afTect the tone of the received note.
The drop in generator voltage, due to speed
drop, takes place more slowly, however, and
is appreciable in var\'ing the tone of the re-
ceived signal during a telegraphic dash. This
effect is increased in small machines and at
short wave lengths.
In radio telephony the commutator ripple
in the generator may introduce a voltage
variation which interferes with the clear re-
ception of speech. This difhculty can, of
course, be overcome by a "smoothing out"
combination of inductance and capacity
which, however, becomes large and expensive
when the ripple voltage is high.
Birdscyc View of the U. S Naval Radio Station. Sitka. Alaska
.S47
The Production and Measurement of High Vacua
PART V. MANOMETERS FOR LOW GAS PRESSURES ( Continued i
By Dr. Saul Dushman
Research Laboratory, General Electric Company
The present installment of this series is a continuation of Part IV and deals with the Knudsen, Pirani-
Hale, and ionization gauges. The next installment will treat of physical-chemical methods of producing
liigh-vacua and will appear in our January, 1921, issue. — Editor.
RADIOMETER GAUGES
Crookes' Radiometer
One of the first instruments to be used for
detecting low gas pressures was the radio-
meter devised by Sir William Crookes in
l.S7;5. The instrument, which is described
in all text-books, consists of a glass bulb in
which a small vane or fly is mounted on a
vertical axis. The vane has four anns of
aluminum wire on which are attached four
small plates of thin mica, coated on one side
with lamp black. These plates are set so
that their planes are parallel to the axis.
If a source of light or heat is brought near
the bulb, and the rarefaction is just right,
the fly rotates, but at very low pressures the
rotation practically ceases.
The theory of the device was apparently
not very well understood for a long time, and
attempts to use it as a gauge for low pres-
sures yielded very unsatisfactory results.
Dewar has stated the case for this instru-
ment as follows:
"The radiometer may be used as an ef-
ficient instrument of research for the detec-
tion of small gas pressures. For quanti-
tative measurements the torsion balance or
bifilar suspension must be employed."'
Some years ago W. E. Ruder, of this lab-
oratory, developed a method of using the
radiometer for the measurement of the re-
sidual gas pressure in incandescent lamps.
The following account was prepared by him
at the request of the writer:
"It was found that when exhausted to the
degree required in an incandescent lamp the
radiometer could not be made to revolve,
even in the brightest sunlight. In order to
get a measure of the vacuum, the radiometer
vanes were revolved rapidly by shaking the
lamp and the time required to come to a com-
plete stop was therefore a measure of the
resistance offered to the vanes by the gas,
together with the frictional resistance of the
bearings. The latter quantity was found
1 Proc. Royal Soc, A, 79. ,529 (1907).
- Two recent papers bv G. D. West (Phys. Soc. London. 3S.
166 and 222. 1920) deal rather fully with the theory of the
radiometer, especially at medium pressures, and also with the
forces acting on heated meta foil surfaces at low pressures.
' Ann. Phys.. 3J. 809 (1910).
to be SO small in most cases that a direct
comparison of the rates of decay of speed of
the vanes gave a satisfactory measure of the
degree of evacuation. In this manner a com-
plete set of curves was obtained which showed
the change in vacuum in an incandescent
lamp during its whole life and under a variety
of conditions of exhaust. The chief objec-
tions to this method of measuring vacua
were the difficulty in calibrating the radio-
meter and the difference in frictional resist-
ance offered by different radiometers. For
comparative results, however, the method was
entirely satisfactory."
As a result of his investigations of the laws
of heat transfer in gases at low pressures,
Knudsen arrived at a clear explanation of the
radiometer action and furthermore developed,
along the same lines, an accurate gauge for the
measurement of extremely low pressures.
According to Knudsen, there is a mechan-
ical force exerted between two surfaces main-
tained at different temperatures in a gas at
low pressure. This is due to the fact that
the molecules striking the hotter surface re-
bound with a higher average kinetic energy
than those that strike the colder surface.
In the case of the radiometer the blackened
surfaces absorb heat from the source of light
and the molecules rebounding from the vanes
are therefore at a higher temperature than
those striking the walls of the bulb. Conse-
quently a momentum is imparted to the vanes
which tends to make them rotate.-
Knudsen Gauge
The principle of the gauge constructed by
Knudsen^ may be explained by referring to
Fig. 42. Let us consider two parallel strips
.4 and B placed at a distance apart which is
less than the mean free path of the mole-
cules. Let A be at the same temperature
T as the residual gas, while B is maintained
at a higher temperature Ti. On the side away
from B, A will be bombarded by molecules
having a mean velocity G, corresponding to
the temperature T, as given by the equation
j:fRf
Si8 October, U)2(l
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 10
These molecules will of course rebound from
A with the same velocity. However, on the
side towards B, A will be bombarded by
molecules coming from B, and having a
higher velocity (7i corresponding to the tem-
perature Ti. Consequently .4 will receive
1
1 t
! /
1 1
-1
1 1
M
A
1
Fig. 42. Elementary Diagram of Knudsen Gauge
momentum at a greater rate on the side
towards J5, than on the opposite side, and will
therefore be repelled from B.
From theoretical considerations Knudsen
has shown that the force of repulsion K per
sq. cm. of the two parallel surfaces, when the
distance between them is less than the mean
free path, varies with the pressure and the
temperatures 7" and 7i, according to the
equation
. p It,
= 2\t-
(:j()a)
For small differences of temperature, and
for the purpose of pressure measurements,
this equation may be written in the form*;
7"
P = 4 K 7p ; -dynes per sq. cm. (.JOb)
In order to measure this force of repulsion,
Knudsen uses the arrangement shown dia-
grammatically in Fig. 42. The strip .4 is
rejjlaced by a rectangular vane, cut out in the
center and sus])ended by means of a fibre 5.
Two strips BB which can be heated are placed
symmetrically on op|josite sides of this vane,
and the force of repulsion is then balanced
* A simple derivation of this and the following equation has
been given by G. W. Todd. Phil. .Mag. SS. 381 (1919).
by the torsion of the fibre. By means of the
mirror M. the deflection can then be measured
in the same manner as in the case of galva-
nometers.
For this arrangement, equation (3()bj as-
sumes the form:
„ 4X-/Z) T ^ ._ .^
F = — 7-in ■ -^ ^jrdvnespersq.cm. (30c)
r At- a 1 1— 1 '
where
7 = moment of inertia of the moving vane,
r = mean radius of the moving vane,
2.4 = area of the vane .4 opposite each strip 5,
/ = period of vibration of the vane,
/? = scale deflection, and
(f = scale distance.
Since all these quantities can be measured
directly, it follows that the device can be
used as an absolute manometer, without the
necessity of calibrating against any other
gauge. It is also evident that the indica-
tions of this gauge must be independent of
the gas to be measured.
In his first paper on this subject. Knudsen
mentions several different forms of construc-
tion which may be used in making a gauge
on the foregoing principle, but gives ven.- few
constructional details. One form which looks
very simjjle is that shown in Fig. 43.
Fig. *3. One Construction of the Knudsen Gauge
.4.4 is a glass tube about 1.4 cm. diameter
in which is sealed a narrow tube BB. The
latter has a rectangular piece cut out at ( .
0.41 cm. wide by 2.!'.") cm. in length. A ])ieoe
of mica P is suspended in front of this open-
ing, bv means of a fibre which is fastened at
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
S49
E. The tuljc .4.4 can be heated by means of
an external water-jacket FF. As the tempera-
ture of the water in the latter is raised, the
mica |)latc is repelled by the "hot" molecules
traveling through the opening C, and the
amount of deflection can be obser\-ed by
means of a microscope.
Variations of this construction are de-
scribed by Knudsen in a later paper^, but very
few details are given. E. V. Angerer" has de-
scribed a Knudsen manometer which consists
of a silvered mica vane between two electri-
cally heated platinum strips, arranged as
shown in Fig. 42. He states that pressures
as low as SXKI^' mm. of mercury could be
measured with it.
The same type of design has also been used
by J . W. Woodrow" on the one hand, and by J .
E. Shrader and R. G. Sherwood** on the other.
Fig. 44. Woodrow's Modification of the Knudsen Gauge
Woodrow's Modification of Knudsen Gauge
The following description of Woodrow 's
form of Knudsen gauge is quoted from the
original publication:
"Several different gauges were constructed
varying in sensitivity so as to be used at
different pressures. A typical gauge is shown
in Figs. 44 and 45 and the electrical circuits
are given in Fig. 4(i. The glass rods (^G scr\'cd
as supports for the metallic parts of the gauge.
All the internal electrical connections and
adjustments, with the cxccjjtion of the final
Fig. 45.
Cross-sectional View Through the Middle
of the Gauge Shown in Fig. 44
^ Ann. Phvs.
* Ann. Phvs.
' Phvs. Rev.
' Phvs. Rev.
ii. 525 (1914).
41. 1 (1913).
4. 491 (1914).
12. 70 (1918).
leveling, were made before the outer glass
walls 00 were sealed on at 55. The suspen-
sion ir was a phosphor-bronze ribbon 50 mm.
in length which had been obtained from
W. G. Pye & Co. and was listed by them as
No. 0000. The movable vane \'\' consisted
of a rectangular frame of aluminum 0.07(1
mm. in thickness, the dimensions of the outer
rectangle being 30 by 30 mm. and the inner
2(i by 30 mm. The heating plates PP were
platinum strips 4 mm. wide, 40 mm. long
and 0.025 mm. thick. The deflections of
the movable vane were obtained in the usual
way by the reflection of a beam of light from
the mirror M. Fig. 45 is a cross-sectional
view through the middle of Fig. 44.
"All of the platinum connections were
made by electric welding, as that was found
much more satisfactory than the use of any
kind of solder, especially when heated. After
a little practice, it was possible to weld the thin
platinum heating vanes to the heavy platinum
wire so as to make a perfectly continuous
contact throughout its width. The phosphor-
bronze suspension was connected at both
ends by threading through three small holes
drilled into the flattened extremities of the
platinum and aluminum wires respectively.
The small loops DD were so placed that they
supported the movable vane V except when
the gauge was leveled for taking readings.
This made the gauge readily portable and,
by placing in the inverted position when con-
nected to the molecular pum]3, the danger of
the breaking of the suspension by vibration was
eliminated. One gauge of medium sensitivity
was constructed so as to be sufficiently steady
to be used when connected directly to themolec-
vilar pump. Large glass tubing was employed
in all the connecting portions of the apparatus.
"A small electromagnet, .shown at E in
Fig 44, was employed in bringing the moving
S50 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11)
vane to rest. This was found to be quite
necessary in working with the most sensitive
gauges, since in a ver;/ good vacuum the
damping is so small that the vane will not
settle down sufficiently for the taking of
readings for some time after an accidental
MM/n-(A>-i'
P B
Fig. 46. Electrical Connections of the Gauge
Shown in Fig. 44
disturbance has set it vibrating. It should
be noted that the electromagnet must have
cither an air core or one of good, soft Norway
iron, for otherwise the residual magnetism
will produce a false zero if the aluminum vane
is at all magnetic, as was the case with the
samples of metal investigated in this lab-
oratory. Under these conditions it is obvious
that the electromagnet should be used only
for damping and that the exciting current
should be shut off while making observations.
"Several methods were tried for determin-
ing the temperature of the healing strips and
that shown in diagram in Fig. 4() was finally
settled upon as giving the most satisfactory
results. The potentiometer leads TT were
connected by electric welding to the very
extremities of the platinum heating vanes
PP. The heating current was regulated by
the \ariablc resistance p and its value was
read on the ammeter .4. The resistance ^2
was kept constant at 1(1, ()()() ohms and ri
varied to obtain a balance of the sensitive
galvanometer G'. The potentiometer bat-
tery C consisted of a carefully calibrated
Weston Standard Cell. This arrangement
gave an accurate method of measuring the
resistance of the ])latinum strips PP, plus the
heavy platinum wire ab. the total cold re-
sistance being 0.17 ohm. This cold resist-
ance was dctennincd by i)lott;ng the cur\'e
connecting resistance and heating current
under a constant low pressure and extrapo-
lating backward to the intersection with the
axis of resistance. If the resistance is meas-
ured for small currents, the value at zero
current, that is the cold resistance, can be
determined ver\' accurately. The tempera-
ture coefficient of resistance of the platinum,
which contained a small amount of iridium,
was carefully determined and was found to
give a linear relation within the range of
temperatures employed. The value of the
coefficient was 2.35X10"^ ohms per deg. C.
With this system one can determine the
mean temperature of the heating strips with
sufficient accuracy, the error for temperature
difference of about '■>() deg. C. being less than
four per cent. "
Woodrow also observes that in order to
avoid electrical effects it was necessarv' to
silver-coat the outside of the glass walls which
were then grounded. Similarly the moving
system was connected through the suspen-
sion to that terminal of the heating strips
which was grounded.
"With the gauge whose dimensions are
given above, the period of a complete oscil-
lation was 10 sec, and the calculated moment
of inertia of the moving vane was 0.074
gm. cm.- This gives for the pressure.
P = 2.9X10-=
T,-T
d (bars)
T
= 2.2X 10"" .J- ..-d (mm. of mercurv)
where d is the deflection in mm. on a scale at
a distance of one meter from the mirror."
Thus with a temperature difference of 100
deg. C, the gauge could be used to read
pressures as low as .JX 10-* mm. of mercur>-.
Shrader and Sherwood's Modification of Knudscn
Gauge
The construction used by Shrader and
Sherwood differs in a few details from that
used by Woodrow. In view of the importance
of the Knudscn gauge for low pressure meas-
urements, the description of this modification
is worth quoting:
"The gauge is shown in Fig, 47. It is en-
closed in a hard glass tube two inches in
diameter and nine inches long. The heating
strip aa is of platinum. O.OIS mm. thick
and 7.."> mm. wide with a total length of IS
cm. It is folded at the top forming a cross
piece and two parallel sides. The ends arc
brazed to 20 mil tungsten leading-in wires
at the bottom. Fifteen mil tungsten wires
b sealed into the glass-rod support serve as
a s|)ring supi)ort for the platinum strip. This
allows accurate adjustment of the strip and
sufficient tension is secured to keep the striji
taut during heating. One of tliese wires
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
S.5I
is carried up the glass-rod support, sealed
into it at f, leaving a free end d to serve for
electrical connection of the moving vane to
the heating strip. Connection is made by
the wire pressing under tension against the
tungsten wire e to which the suspension of
Fig. 47. Shrader and Sherwood's Modification
of the Knudsen Gauge
the vane is fastened. Potential leads of fine
platinum wires^are welded to the strip about
one centimeter from the ends and are brazed
to tungsten sealing-in wires.
"The movable rectangular vane g is made
of aluminum (i.(il)7(i cm. thick. A standard size
adopted is o.2 cm. by 4 cm. outside dimensions,
the width of the vane being O.o cm. Because
of liability of warping during heat treatment
the vane is stiffened by an aluminum wire
passing through slits at the top and a hole
at the bottom into which the wire is hooked
and fastened firmly. For portability, two
copper wires /; are sealed into the glass-rod
support while the free end formed loops
around the rod, these forming guides for the
vane. The mirror is fastened at the bottom
of the vane by leaving a small projection of
' L. F. Richardson (Phys. Soc. London, SI, 270, 1919) has
described a form of Knudsen gauge in which he balances the
force of repulsion by means of a magnetic field. The instrument
seems, however, quite complicated in construction.
A commercial form of Knudsen gauge brought out recently
in Germany bv H. Rieger is described briefly in Engineering,
July .30. 1920.
<.he aluminum at the lower edge and cutting
out small tongues from the material of the
vane on either side. The mirror is laid in
])lace and the projection and the two tongues
are pressed closely over it, holding it securely.
"Silver mirrors were tried, but failed to
withstand the heat treatment to which the
gauge and system were subjected. Mirrors
made by coating microscope cover glass with
china decorators' platinuin solution, followed
by baking at 500 deg., solved this difficulty.
"The distance between the heating strip
and the vane is adjusted from outside the
case by magnetic control on a piece of soft
iron i sealed into a glass stem to which the
suspension is fastened.
"The suspension is O.OOO.Viiich tungsten
wire. This is fastened to small aluminum
hooks around which the wire is wrapped
several times after which the hooks are
pressed firmly together. This method is not
difficult and holds the wire securely. A hook
on the end of a tungsten wire sealed into a
glass stem, the free end passing through a
capillary rod / on the glass support, serves
to hold the suspension.
"A gauge such as has been described, using
a 0.0005-inch tungsten susj^cnsion from (i to 7
cm. long, has such a sensibility that a scale
deflection of 1 mm. at a meter's distance with
a temperature difference of 150 deg. C. be-
tween the heating strip and Ithe movable
vane indicates pressures of 1 X 10"' to 5X 10~*
mm. Hg. One gauge of other dimensions than
those given above would indicate a pressure of
5X 10"' mm. Hg. under the same conditions."
The temperature of the heated strips is
measured by substantiallv the same electri-
cal method as that vised by Woodrow.'
RESISTANCE MANOMETERS
At ordinary pressures the heat conduc-
tivity of gases, like the coefficient of vis-
cosity, is independent of the pressure. How-
ever, as the pressure is decreased, a point is
reached at which the heat conductivity be-
gins to decrease with the pressure in a manner
which is Quite analogous to the phenomena
obseni'cd in the case of viscosity ineasure-
ments. Kundt and Warburg pointed out,
as a "result of their experiments on the coef-
ficient of slip, that a similar phenomenon was
to be expected in the case of heat conductance
at low pressures. Subseoucnt ex]jeriments by
Sutherland, Smoluchowski and Knudsen have
shown that such is actually the ease.
These experiments have led to interesting
speculations upon the mechanism of the
852 October, 1920
GENERAL ELECTRIC REVIEW
Vol XXIII. No. 10
heat transfer in gases between surfaces which
are separated by a distance which is com-
parable with or less than the mean free path
of the molecules, and the more detailed dis-
cussion of this subject will be taken up in
a subsequent connection.
Fig. 48. Hale's Improved Form
of Pirani Gauge
The im]Jortant exi)erimcntal fact from
the present ]3oint of view is that at very low
pressure the heat conductivity of gases de-
pends upon the pressure. Warburg, Lcit-
hauser, and Johansen'" applied this fact to
the construction of a gauge by measuring
the change in resistance of a small bolometer
strip; while \'ocge" used a small thermocouple
attached to a wire healed by a constant
alternating current. The temperature of
the wire as observed by means of the ther-
mocouple was found to be a function of the
pressure. Quite recently, W. Rohn'- has
developed a gauge on the same ])rinciple.
Pirani-Hale Gauge
Piranf pointed out that in order to con-
struct a gavige based on the relation between
the heat conducted from a wire and the pres-
sure, three different schemes could be used.
1. The voltage on the wire is maintained
constatit, and the change in current is ob-
ser\'ed as a function of the pressure.
2. The resistance (and consequently the
temperature) of the wire is maintained con-
stant, and the energy input ret|uired for this
is observed as a function of the i)ressure.
;i. The current is maintained constant,
and the change in voltage drop obsen-cd as
a function of the pressure.
The first scheme was tried using an ordi-
nary 110-volt tantalum-lamp. Better results
were, however, obtained when the tantalum
wire was clamped lightly to the anchor wires
in order to keep constant the heat loss through
the supports. With the improved instrument
the two other methods were tried, using a
Wheatstone bridge arrangement to measure
the resistance of the wire, and the third one
finally recommended as the most sensitive
for use in pressure measurements.
While the principle of Pirani "s gauge is
thus extremely simple, the sensitiveness actu-
ally obtained by him was not very great, the
lower limit of accuracy being around 0.1 bar.
An improved form of this gauge was con-
structed by Hale.'-'' which is shown diagram-
matically in Fig. 4S. The following descrip-
tion is quoted from Hale's paper:
"A piece of pure platinum wire. 0.02.S mm.
in diameter and 4.)0 mm. long, is mounted
upon a glass stem carrying two radial glass
supports near the top and three at the bot-
tom. The wire is anchored to these radial
supports by means of short pieces of platinum
wire ().0.')2 mm. in diameter. The anchor is
fused into the radial supports at one end.
and the other end is made fast to the man-
ometer wire either by an arc weld or by a
tiny glass bead. The leading-in wires at /-.
to which the ends of the manometer wire are
'" Ann. d. Phvs. li, 2,j (1907).
" Phys. Zeits'. T, 498 (1908).
'= Zcit.s. f, Elcktrochem. iO. J39 (1914).
" Trans. Am. Electrochem. Soc. iO. 243 (1911).
Fig. 49. A Diagram of the Electrical Connections of the Gauce
Shown in Fig. 48
welded, are of platinum. 0..'51 mm. in diam-
eter. All of the ])laliinim wire employed
in making the manometer was drawn from
the same lot of larger wire and was assvimed
to be of uniform inirity. The tem|)eraturo
coefficient of the manometer wire was found
THE PRODUCTION AXD MEASUREMENT OF HIGH VACUA
.s:)3
to be (l.()(io7ii per cent per degree. The
platinum leading-in wires are joined to heavy
copper leads (1.1 mm. diameter) by welded
joints, and these joints are fused into the
stem as in electric lamps. The stem is sealed
into a tubular bulb 3.2 cm. in diameter and
1 1 .4 cm. long. This size of bulb is easily
obtained, since it is the size regularly used
for .ol)-watt tubular lamps, such as are com-
monly employed for galvanometer illumina-
tion. At S is a tube by which the manometer
is connected with the system whose pressure
is being studied. The upper end of the stem
T is considerably elongated to permit the
complete immersion of the manometer in a
constant temperature bath, whose tempera-
ture was approximately zero deg. C. This
stem tube is made of sufficient length to
resistance of 92."). (i ohms, and R-^ a decade
plug box containing 1(),()U() ohms. The
strength of the current, as indicated by the
milliammeter Am. was maintained constant
at U. 25X10"^ amp. by means of the battery
and resistance Ra- This current was suf-
ficient to raise the temperature of the wire
in the manometer and compensator to about
12.> deg. at the lowest pressures.
In calibrating the gauge against a McLeod
gauge care had to be taken to keep mercury-
vapor out by means of a liquid-air trap in-
serted between the manometer and the re-
mainder of the system. Fig. 50 shows cali-
bration cur\'es obtained with air and hydro-
gen at different pressures.
The difference is due to the higher con-
ductivitv of hvdrogen, so that the indications
Fig. 50 Calibration Curves of the Gauge Shown in Fig. 48
leave lo cm. of it above the level of the bath,
a provision which we found to be necessan.-
in order to avoid the condensation of atmos-
pheric moisture upon the top of the tube and
the leading-in wires during humid weather.
For electrical insulation this tube is packed
with purified dr\- asbestos wool."
A diagram of the electrical connections is
shown in Fig. 49. A Wheatstone bridge ar-
rangement was used for measuring the re-
sistance changes; and in order to increase
the sensitivity of the gauge, an exact dupli-
cate was exhausted as carefully as possible
to an extremely low pressure, sealed off. and
inserted in one arm of the bridge as a com-
pensator. Both the compensator and man-
ometer were kept immersed in the constant
temperature bath. Ri was a manganin wire
" Proc. Physico-Mathem. Soc. Japan. .3rd. Ser., /, 1.52 (19191.
of the manometer are dependent to a certain
extent upon the nature of the gas used. Hale's
measurements show that the lower limit of
sensitivity for a gauge of this construction
is about 0.00001 mm. (i.e. 0.0133 bar).
Recently some further measurements with
a Hale gauge have been carried out by Misa-
michi So.'''
The construction of gauge used by him
difters in a few slight details from that of
Hale. It was found that the -sensitivity of
the gauge is higher, the lower the tempera-
ture of the surrounding bath. At zero deg.
C and using a heating current of 0.03 amp.
for a platinum wire 0.076 mm. in diameter,
the sensitivitv as measured bv -j— . — was ob-
" dp R
sen-ed to be 1.3SX10"'' per 1 X Ur^ mm.
of mercun,'. Furthermore, varving the heat-
S.54 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. It)
ing current from 0.03 to 0.0.3 amp. was
found to produce no change in sensitivity.
A hot-wire manometer based on the same
principle has also been described by T.
Tschudv.'^
rj/Li/ /)rTn£r£fi
GAt vMf^o^srir^
Fig. 51. Ionization Gauge and Connections
IONIZATION GAUGES
An electron stream passing through a gas
will ionize the latter when the velocity of the
electrons exceeds a certain minimum value.
In this process, an electron is knocked out
of the neutral atom by the incident elec-
tron, with the result that the residual por-
tion of the atom is positively charged. The
relation between the velocity u of the elec-
trons and the voltage V required to produce
this velocity is given by the equation :
1^ niu"= IV
where
e = charge on electron
>H = mass of electron.
.So that corresponding to the ionizing velocity
there exists for every gas a minimum ionizing
potential. These range from one or two volts
for the alkali metals to "2.") volts for helium
The amount of ionization produced by a
given electron current increases with the pres-
sure and while this fact has been used as a
qualitative guide for the detection of low
gas pressures, it is only recently that attempts
" E ekt. Zeits. S9. 235 (1918); Electrical World. 73. 137
(1919).
■• Proc. Nat. Acad. Sciences. S. 683 (1916).
'' A brief account of these experiments was published in
Phys. Rev. The complete paper on the subject will appear very
shortly in the same journal.
" See Part II of this series of articles.
have been made to apply this principle to the
construction of an actual measuring device.
0. E. Buckley"" has published a short paper
on the results obtained with a manometer
of this type in which no details are given as
to the actual construction. The gauge con-
sists of three electrodes which are used as
cathode (source of electrons i, anode, and
collector of positive ions respectively. As
source of electrons a Wehnelt cathode or in-
candescent tungsten filament is used. The
collector electrode is placed between the anode
and cathode, and connected through a gal-
vanometer to the negative terminal of a bat-
tery whose positive terminal is connected to
the most negative end of the cathode. The
anode potentials used range from lUU to 250
volts, while the magnitude of the electron cur-
rent is varied from 0.2 to 2.0 milliampcres.
At a pressure of 1.0"' mm. the ionization cur-
rent was observed to be about one-thousandth
that of the electron current and proportion-
ately less at lower pressures; so that with an
electron current of 2.0 milHamperes. pressures
below 10~^ mm. could be measured qui teeasily.
According to Buckley "the exact forms of
the electrode are not of great importance."
However, subsequent experiments by Mr.
Found and the writer'^ showed that certain
designs are much better than others. A r^uge
consisting of three hair-pin filaments placed
in parallel planes shows erratic effects at low
pressures because of charges on the walls.
Of the many types of constniction tested.
that shown in Fig. .">1 was found to have the
best characteristics for measuring low pres-
sures. This illustration also shows the method
of connecting up the electrodes.
The gauge consists of two tungsten fila-
ments, each wound in the form of a double
spiral and mounted co-axially on a four-
lead stem which is sealed into the upper end
of a glass tube about 4 cm. in diameter and
12 cm. long. The inner si)iral i.> made of
.") turns of 0.12.')-mm. tungsten wire wound on
a 2.2.">-mm. mandrel. The outer spiral is
made of three turns of ().12."i-mm. tungsten
wire wound on a 3.().">-mm. mandrel. Sur-
rounding the spirals is a molybdenum cylin-
der about 1 2 mm. in diameter and 1 2 mm. long
which is supported on a two-lead stem at the
lower end of the tube.
Before using the gauge for any measure-
ments, it is of course absolutely essential that
gases occluded in all metal jiarts and water
vapor on the walls should he thoroughly elimi-
nated. This can be accomplished in the man-
ner alrcadv described.'*
THE PRODUCTION AND MEASUREMENT OF HIGH VACUA
855
The best conditions for the operation of
the gauge were found to be as follows:
(a) For very low pressures (belou.' 1 bar):
250 volts on the anode, — 20 volts on the col-
lector cylinder, and a maximum electron
current of 20 milliamperes. Under these con-
ditions, 1X10"" amp. positive ionization cur-
rent corresponds to 0.0132 bar argon.
(b) For higher pressures (1 to 30 bars):
125 volts on the anode, — 20 volts on the
collector cylinder, and an electron current
of 0.5 milliampcre. In this case, 1X10"''
amp. ionization current corresponds to 0.5
bar argon approximately.
Fig. 52 shows characteristic cun.'es at dif-
ferent electron currents. The greater the
electron current used, the lower the upper
limit of pressure at which the linear relation
is still valid.
It will be observed from these curves that
at constant pressure the ionization current
is not quite proportional to the electron cur-
rent. For measuring a considerable range
of pressures it is desirable to have this pro-
portionality, since it is then possible to in-
crease the electron current as the pressure is
lowered and thus increase the sensitivity of
the gauge. The following method of connec-
tion has been found to give a linear relation-
ship between ionization and electron current
and may therefore be used instead of the
arrangement just described. The inner fila-
ment is used as collector, the outer filament
as cathode and the cylinder as anode. With
this connection the ionization current is prac-
tically independent of the anode voltage be-
tween 125 and 250 volts. The sensitivity
is not quite as good as with the first method
of connection, 1 X 10"'' amp. positive ionization
corresponding to about 0.032 bar argon.
An interesting result which was found on
studying the behavior of the gauge with dif-
ferent gases is that at constant pressure and
with the same conditions as to anode voltage
and electron current, the ionization current
increases with the number of electrons in the
molecule. Thus the number of electrons in
an argon molecule (or atom) is 18, while in a
^^ A preliminary account of this investigation has been pub-
lished by Mr. Found and the writer in Phys. Rev. The more
complete discussion will appear shortly in the same journal.
■» Phys. Math. Soc. Jap. Proc. I. p. 76 (1919).
mercury molecule (which is also monatomic)
the number of electrons is 80. The ioniza-
tion currents at constant pressure are found
to be in approximately this ratio. Experi-
ments with h and //2O showed that the ioni-
zation currents in these cases, as compared
Fig.
e7.£ a3 o^ f^S '^6 07 cff 0.3 Ao / / ^B
52. Characteristic Curves of the Ionization Gauge
with that for argon at the same pressure (and
same electron current) correspond to elec-
tronic numbers of 10(i and 10 respectively,
if that for argon is taken at 18. This general-
ization is apparently not quite true for H^,
He and "Ne, and further investigation is
necessary in these cases. For all ordinary
cases, however, the calibration for nitro-
gen (14 electrons per molecule) may be used
as a general guide to the value of the pres-
sure."
The ionization gauge as just described,
has been found by the writer to be very use-
ful in investigating the pressure changes in
incandescent lamps and hot-cathode devices
after sealing off from the pump. The ease
of construction and simplicity of manipula-
tion ought to make it a very useful device
in high-vacuum technique.
Recently some results with a three-fila-
ment ionization gauge have been published
by Misamichi So.-" The ionization currents
were, however, measured at constant cathode
temperature, and the relation obtained be-
tween pressure and ionization current is not
linear.
So6 October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 10
A Special Form of Phosphoroscope
By W. S. Andrews
CoxsiLTixG Engineering Department, General Electric Company
In OUT Alarch, 1917, issue the author described a form of phosphoroscope which he had developed for the
visual observation' of the phosphorescent and fluorescent light emitted by various compounds when excited
bv ultra-violet light. As this device, however, is not well adapted for making photographic spectrographs.
he has since developed another form of phosphoroscope that fulfils this purpose admirably. This latter type
of instrument is described in the following article. — Editor.
In using the new form of phosphoroscope.
designed for spectrographic analysis, a paper
ribbon is first prepared by coating it with the
phosphorescent material reduced to a powder,
a chemically neutral adhesive being employed.
When perfectly dry, this coated ribbon is
fastened around the periphery of a small flat-
faced wheel that is attached to the shaft of an
electric motor. Fig. 1 . The e.xciting rays are
preferably generated by a high-tension dis-
ruptive discharge between two iron terminals
that are conveniently concealed and protected
Fig. 1. Phosphoroscope Adapted for Use in Making Spectro-
graphs. Shield Removed Showing Flat-faced Wheel
with Phosphorescent Ribbon Attached
within a jcircular hood. This hood is located
close to the periphen,- of the small flat-faced
wheel, as shown in Figs. 1 and 3, and
diametrically opposite to it is fixed a screen
of sheet metal having a small window cut
in it. This opening exposes a portion
of the coated ribbon, see Fig. 2. It is
evident that when the electric discharge
occurs, its rays will excite the adjacent
material on the rim of the wheel, and when
the motor revolves, the excited portion will be
carried around until it is opposite the window.
This action being continuous, the phosphores-
cent light of the material to be photographed
will shine continuously through the window,
and if the collimator of the spectroscope is
pointed into it. the spectrum of the phos-
phorescence will be seen through the eye
piece of the telescope. When the camera
attachment is used instead of the telescope, a
spectrograph of the phosphorescence may be
made in the usual way.
As the electric motor may be operated at a
high speed, say 3000 to '4000 r.p.m.. the
phosphorescent light seen at the window in
Fig. 2. Front View of Phosphoroscope with
Shield in Place showing the
Observation Window
the screen is practically instantaneous, the
time for decay being only about one hun-
dredth of a second. In this way, phosphor-
escence of vcr\- brief period may be seen
and photographed; also by varying the speed
of the motor, its actual duration may be
approximately determined.
Figs. 4. ."). and (> show the spectrographs of
various jihosphorcsccnt substances, these being
arranged in pairs, with the spectrum of helium
interpolated between them to indicate the
phosphorescent regions in the spectrum pro-
duced bv the different substances.
A SPECIAL FORM OF PHOSPHOROSCOPE
S57
Attention may be inA-ited to the phosphores-
cence of calcined chemically pure ( ?) cadmium
])hosphate, Fig. 4, which appears snow-white
to the unaided eye and which the spectro-
graph shows to cover the visible spectrum
fairly well, being strongest in the yellow,
blue, and violet, and weakest in the red and
green. Students who have investigated
phosphorescent spectra will recognize the
striking peculiarity of this one.
In all other cases known to the writer the
phosphorescent color of a given substance
covers only some specific part of the visible
spectrum or, in other words, its color to the
unaided eye may appear in some shade or
ctnnbination of red, yellow, green, or blue.
To present a snow-white to the eye, means
that all parts of the visible spectrum must be
represented as seen in the spectrograph.
It is true that the blue-violet phosphores-
cence of Balmain's luminous paint (phos-
phorescent calcium sulphide) usually fades
to a whitish color, but at this stage its
luminescence is so weak that it can be seen
only in perfect darkness after the eye has been
rested; whereas, the jjhospliorescence of the
pure cadmium phosphate shows a white
light of great purity and considerable inten-
sity. When the e.xciting light is cut off, the
intimate mechanical mixture of phos]3hores-
cent substances that show respectively com-
plementary colors; such for instance, as violet
calcium sulphide and yellow zinc sulphide, or
a similar effect may be produced by painting
sections of a disk with suitable phosphores-
cent materials and then putting the disk into
Fig. 4
A — Fused C. P. carhiiium phosphate; white phosphorescence.
B — Hehum. Comparison spectrum.
C — Zinc sulphide; green phosphorescence.
A
Fig. 5
Fused cadmium phosphate plu, man^^anese; orange phos-
phorescence.
B — Hehum. Comparison spectrum.
C — Calcium sulphide, or Balmain's luminous paint; blue-violet
phosphorescence.
Fig. 3. Rear View of Phosphoroscope
Fig- 6
.-1 — Fused cadmmm phosphate plus manganese; red phosphor-
escence.
B — Helium. Comparison spectrum.
C — Zinc silicate or willemite; green phosphorescence
white phos])horescence gradually fades away
to a reddish color before it disappears, point-
ing to the existence of an exceedingly minute
amount of manganese in the compound.
It is well known that a white or nearly
white phosphorescence can be obtained by the
rapid rotation. The particularly interesting
feature of the pure cadmium phosphate prepa-
ration, however, is that all the colors of the
spectrum are apparently produced in its
phosphorescent glow, so that by their com-
bination a pure white light is produced.
SoS October, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 10
The Cooper Hewitt Lamp
PART II. DEVELOPMENT AND APPLICATION
By L. J. BrxTOLPH
Engineering Department, Cooper Hewitt Electric Company
This is the second of a series of three articles by the author on the theory and uses of the Coopor Hewitt
lamp. Part I, "Theorv and Operation," appeared in our September issue. In the following installment the
author outlines the development and some of the applications of the lamp from 1901 to the present time. In
Part III, the Cooper Hewitt quartz lamp and its characteristics will be described. — Editor.
In ApriL 1901, at a conversazione of the
American Institute of Electrical Engineers.
Nicola Tesla and Peter Cooper Hewitt
divided the honors in the display of their
inventions. The Electrical World and Engi-
neer said editorially at that time: "But with
all these things in its favor it is still a far cry
from even the brilliant exhibits of the other
evening to a lamp that will meet the ever}"
day requirements of commercial work." The
development of a lamp meeting these require-
ments began with the two types exhibited in
1902 and shown in Fig. 1. Engineering of
London said: "A lady's lips look purple, so
at present no attempt is being made to
titilize the light for domestic purposes, as
feminine opposition would be too strong."
The other problems to be solved were a
simple starting device and operation on
alternating current.
In 1903 the first industrial installation of
two 200-watt lamps was placed in the com-
m m
4
^
Fig. 1. Original Form of Cooper Hewitt Lamp Which Was
First Displayed to the A.I.E.E. in 1901
Fig. 2. Progress in the Construction of Cooper Hewitt Lamps
from 1902 to 1907
THE COOPER I HEWITT LAMP
859
loosing room of the New York Evening Post.
At that time very few features of the lamp
had been standardized. Iron, mercury or
graphite positive electrodes were variously
used, while the condensing chamber was
placed at the positive end in some cases and at
the negative end in others. The lamps were
started either by tilting the tube manually
until a thin stream of mercury connected the
electrode and then allowing the lamp to
resume the normal position when the break-
ing started the arc, or by an oil immersed
switch operating an inductance coil to pro-
duce a high voltage kick.
The progress from 1902 to 1907 is illustrated
in Fig. 2. During that time the condensing
chamber was standardized and iron adopted
as a positive electrode material. The start-
ing was made automatic by the development
of a magnetic tilting device and of an auto-
matic mercury switch or "shifter" for the
high tension method. The latter method
consists in short circuiting a small current
The development of a commercial form of
alternating current lamp marked the period
from 1907 to 1910. As developed then and
operated now the alternating current lamp is a
single phase Cooper Hewitt rectifier of highly
Fig. 3. Orthochromatic Cooper Hewitt Lamp
Fig. 4. Direct-current Cooper Hewitt Lamp. 385 Watts, 850 Mean Horizontal Candle-power
Fig. 5. Alternating-current Cooper Hewitt Lamp, 430 Watts, 950 Mean Horizontal Candle-power
through inductance in scries with the lamp
tube. This current is broken by the mercury
switch or "shifter" which is magnetically
operated by the inductance coil itself. The
resulting induced high voltage is sufficient to
start a cathode discharge which breaks down
the initial high resistance by ionizing the
traces of mercury vapor in the tube and thus
fomiing the arc.
specialized fonn. In common with all arc
lamps it required a current regulating device
which consisted of impedance instead of the
resistance used in the direct current lamps.
The resulting power factor of about 50 per
cent was a problem for later solution. In 1910
increasing business forced the Cooper Hewitt
Company to seek larger quarters at their
present location in Hoboken, and from that
860 October, 1920
GEXERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 10
Fig. 6. Installation of Cooper Hewitt Lamps in a Textile Mill
Fig. 7. Cooper Hewitt Lamps Illuminating Machmery Used in the Preparation of Rubber for Shoes
THE COOPER HEWITT LAMP
8G1
time production was concentrated on stand-
ardized outfits for industrial purposes.
A fluorescent reflector for the Cooper
Hewitt lamp was put on the market at this
time. It altered the quality of the light from
the unit by giving off fluorescent light of an
orange red color at a sacrifice of a correspond-
ing amount of green and blue.
Among the few radical developments of the
succeeding period was a so-called orthochro-
matic lamp whose arrangement is obvious
from Fig. '.i. It represented one of the last
attempts to change the color of the Cooper
color of the light would mean a sacrifice of
what was proving to be one of its unique
and valuable properties. In certain textile
mills careful tests were made to determine the
relative ease with which fine textile threads
could be seen by direct north daylight and
by Cooper Hewitt light. The results were
strikingly in favor of the latter, and there
was no question as to the relative value of
color vision and of visual acuity for general
industrial illumination.
During this period the Cooper Hewitt
Electric Company also manufactured a quartz
Fig. 8. A Machine Shop Illuminated with Cooper Hewitt Lamps
Hewitt light. An installation of orthochro-
matic lamps was placed in the editorial rooms
of the New York World, with satisfactory
results from an Eesthetic standpoint. The
relatively small demand for the outfit and the
fact that it was more novel than efficient led
to its discontinuance. From the beginning
extensive research has been done here and
abroad to change the color of the arc itself by
the addition of various substances but the
attempts have availed little. At about the
time that a solution of the problem began to
seem impossible Cooper Hewitt engineers
realized that a change in the nature of the
mercury arc, similar to the European type but
incorporating the Cooper Hewitt idea of
controlling the electrical characteristics of the
lamp by a condensing chamber, as in the
ordinary low jjrcssure type of Cooper Hewitt
lamp.
The first radical change in quartz burner
design came with the development of a means
of connecting quartz through intermediate
steps of glasses of increasing coefficients of
expansion to a glass fused directly to a tung-
sten electrode and forming with it a per-
manent vacuum tight seal. The result was
the greatly simplified quartz bitrner now
862 October, 1<)2I1
GENERAL ELECTRIC REVIEW
Vol. XXIII, No.- 10
manufactured by the Cooper Hewitt Com-
pany.
A great many of^the old type quartz
burners are still in' ser\'ice for industrial
illumination. The increasing importance
of the mercury arc in quartz as a source of
Fig. 9. Triple-tube Cooper Hewitt Outfit
for Photographic Work
ultra-violet light is opening up a large field
for the new type of burners. Among the
present uses of these burners may be men-
tioned water sterilization, the treatment of
skin diseases, paint and dye testing, and the
acceleration of a great many general photo-
chemical reactions.
The shapes and sizes of Cooper Hewitt
lamps have only been limited by the imagina-
tions of the designers and by the glass blower's
art. Sizes have ranged from a few watts to
3000 watts, while the standard tubes range
at present from 200 watts to 1600 watts.
A few of the standard outfits arc illustrated.
The larger tubes, some of them six feet long by
three inches in diameter, are i)rincipally used
in blue printing machines.
The problem of operating the alternating
current lamp on a high power factor was
finally solved in 191S by the development
of a means of substituting resistance for some
of the reactance in the current-regulating
impedance of the outfit. This is done by
operating hot iron wire resistance units under
conditions of high tcmi)crature coefficient,
such that with change of ctirrent the falling
voltage characteristic of the tube is counter-
acted by the rising voltage characteristic of
the resistance. By this means the present
alternating current outfit is given the high
power factor of <S5 per cent. In 1910 the
dcveloijment of glass working machinery
was begun and with the co-operation of the
Edison Lamp Works machines for perform-
ing the major tube making operations were
placed on factory production in May, 1920.
Xow that most of the states are supple-
menting their sanitary codes with lighting
codes, a brief for good industrial illumination
is superfluous. The codification of good
industrial illumination has, however, been
attended by difficulties of the same sort met
with in the evaluation of the service of a
skilled workman in the terms of physical
strength, manual dexterity, mental traits.
ner\'ous temperament, etc. — things which do
not readily lend themselves to tangible defini-
tion and to simple summation. Bearing on the
problem is the fact that the early Cooper
Hewitt installations did pioneer work in
creating the present demand for a diffused
general illumination of relatively high inten-
sitv and minimum glare and that the system
is now established on its merits in spite of
much "feminine" opposition. As a result of
the slow but cumulative effects of this semi-
educational work and of a remarkable record
of ser\-icc in facilitating the intensive pro-
Fig. 10. Two tube Cooper Hewitt Outfit
for Photographic Work
duction of war materials, the demand for
Cooper Hewitt lamps has increased to such
an extent that radical steps are being taken
to treble production and to develop a unit
system of manufacture providing for in-
definite expansion.
THE COOPER HEWITT LAMP
863
The Cooper Hewitt idea in industrial
illumination is: first, to equal average day-
light illumination in point of luminosity,
diffusion and apparent color; and second, to
surpass daylight for special purposes in point
of high visual acuity.
By average daylight illumination is meant
that secured by the best practice in modem
factory construction. Only such moderate
intensities arc advocated as have been found
to give a maximum production under daylight
conditions. Greater intensities than these
are recognized as useless and in many cases
as a positive menace to the comfort and
candles of illumination on a working jjlane —
is the result of three things. Its visible radiant
energy is largely concentrated in light- of those
wave lengths to which the eye is most sensi-
tive. Simple reflectors involving little loss by
absorption can be used and because of the
low intrinsic brilliancy no light is lost through
the use of diffusing media, the Cooper Hewitt
light being the only high power modem
illuminant excepted by many of the state
codes from the use of diffusing media. Thus
it is, that, for the distribution, diffusion, low
intrinsic brilliancy and minimum glare re-
quired in the modern industrial installation,
Fig. 11. Installation of Cooper Hewitt Lamps in a Motion Picture Studio
efficiency of the workman. The good dif-
fusion and monochromatic quality of the
Cooper Hewitt light makes it unnecessary
to increase the intensity on the working plane
beyond a moderate normal to secure good
distribution and proper visual acuity. A
satisfacton,- uniform intensity of illumination
as measured by any standard illuminometer
or foot-candle meter represents quantity only
and may be taken as fundamental in all light-
ing systems but unique in none. The high
practical efficiency of the Cooper Hewitt
light as rated in terms of quantity only — in
terms of total lumens of light flux or in foot-
the Cooper Hewitt light gives an excellent
illumination in foot-candles on the working
plane per watt of electrical energy.
Qualities unique in the Cooper Hewitt lamp
and not to be directly expressed in foot-
candles are the diffusion of its light, a mini-
mum of glare, and high visual acuity. Because
the source is in the form of a tube of light
one inch in diameter and some 50-in. long, in
the standard lamps, the diffusion of light is
often literally equal to or better than that
secured by daylight illumination while there
is the added advantage of a simple control of
the diffusion in those cases where shadows are
S(U October, 192U
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 10
an aid to stereoscopic or perspective vision.
As is well known, too perfect diffusion is in
such cases a disadvantage. The ver\- slight
shadows under machines and tables is the
first and obvious result of this high diffusion.
The real advantages of it are that a workman
can do fine machine work without the glare
of local lighting, that his own head and body
does not cause confusing shadows, and that
all parts of his machine are so well lighted.
Although the light from the Cooper Hewitt
lamp is apparently blue white, it is as a matter
of fact vers' largely monochromatic and of a
yellow green color. A fundamental limitation
of the human eye is its inability to focus in
one and the same plane light of different
wave lengths or color The result is the
fonr_ation of multiple superimposed images
giving the effect of a single image with
a blurred outline; i.e.. chromatic aberration
Fig. 12. Cooper Hewitt Lamp
for Photographic Portraiture
Fig. 13. Bank of Cooper Hewitt Lamps
for Motion Picture Work
The intrinsic brilliancy of the Cooper
Hewitt light is only some l.> foot-candles per
square inch so that no diffusing media are
needed. To an eye accustomed to the usual
working intensity, there is no discomfort in
looking directly at and into the luminous
source. The construction of the standard
reflectors provides sufficient lighting above
the source to avoid glare from contrast with a
lilack background. The size and shape of the
source and its low brilliancy reduces the glare
by reflection from polished surfaces to a
degree equaled only by skylight illumination.
This lack of visual acuity in daylight or
ordinary white light may be improved
strikingly by the use of monochromatic
light which results in the formation of a
clearly defined image. The coincidence of
maximum visibility anil maximum acuity
in some (>0 per cent of the Cooper Hewitt
light partially accounts for its markoil suc-
cess in the textile industry, and in machine
shops and inspection deijartmcnts. As a
])ractical matter the fine threads of textiles
and the lines on machined metal surfaces can
be seen with an increase in detail equal to a
THE COOPER HEWITT LAMP
-Ci.")
ig. 14. Illustrating Use of Cooper Hewitt Portrait Outfit
Fig. 15. Cooper Hewitt Lamps Arranged
for Photographic Enlarging
■^'^i^l^
Fig. 16. Long and Short of the
Cooper Hewitt Lamp
.^e>. -^
■il 1 1 !«»■»■
-\l _;
■ ■.::.J:;>y^;;;^.":.-^-^<^::»-^v-:^jaii^^<iatf^
Fig. 17. Banks of the Cooper Hewitt Lamp on the Metro Motion Picture Stage
866 October, 1920
GENERAL ELECTRIC REVIEAV
VoL XXIII. Xo. lU
magnification of one and one half to two
times under Cooper Hewitt light.
Since the Cooper Hewitt light is entirely
lacking in red rays it is obvious that red
objects do not seem colored but black or
brown under the light, even as dark blue or
violet objects look black under incandescent
sources. This has proved no handicap in
industrial lighting, since colors are matched
in the factory by numbers and not by com-
parison.
Interesting contributions to the psychology
of color have come from users of the Cooper
Hewitt system. Traces of green hue in the
light combined with the well known hue
imparted to a face of good color in ordinary
light invariably produces a strong reaction
against the light. Workmen who are very
prejudiced for these reasons are however won
over to the light by a subtle but definite per-
sonal reaction variously characterized by
describing the light as restful, soft, cool, etc.
The change is probably because the first
reaction is largely mental, the result of a
subjective green hue rather than the yellow
green light of the so-called green mercury
line; while the second reaction is fundamental
and in response to the recognized psychologi-
cal effects of true blue and violet light. Con-
tributor}' to the latter reaction is also the
physiological fact of reduced eye strain under
relatively monochromatic light. More strik-
ing still is the oft repeated testimony to the
greater apparent coolness of a room lighted by
Cooper Hewitt lamps than by any other artifi-
cial illuminant. While for a given intensity of
illumination approximately the same quan-
tity of energy is dissipated in a room regard-
less of the system of illumination, a smaller
proportion is radiant heat in the Cooi)er
Hewitt unit and that heat is of as relatively
low inten.sity as the light. This jirobably
ser\'es to strengthen the subconscious impres-
sion of coolness produced by the blue white
quality of the light.
Of the numerous special applications of
Cooper Hewitt lighting the most interesting
is in the motion picture industrv". Here the
system is competing directly with Califoniia
sunshine and apparently winning out by
virtue of its remarkable actinic power and
reliability. Feminine opposition has changed
to feminine approbation. New York must
have its movie shows regardless of rainy da>s
in California. Other special applications of
lesser importance are: general photography,
photographic enlarging without condensers.
photographic reproduction of drawings, blue
printing, and certain photochemical processes
not requiring the more intense Cooper Hewitt
quartz arc light.
That a properly laid out Cooper Hewitt
lighting system can be made to give an
illumination equal to or better than daylight
is proved by large and successful installations
in nearly every basic industry-. Night shift
production rates equal to and even greater
than day rates are the rule. In certain
inspection departments and in certain highlv
specialized industries the light is used con-
tinuously in preference to daylight because
of its unvarying luminosity and color.
The Cooper Hewitt lamp is unique from a
scientific standpoint in combining high
efficiency with low intrinsic brilliancy, largo
total light flux with good diffusion, and
maximum monochromatic visual acuity with
apparent whiteness. It makes possible the
unique idea in industrial lighting of adding to
the required luminous intensities those quali-
ties of light which will enable the eyes of the
workman to function to best advantage in
their essentially artificial work of the con-
tinuous observation of details.
TWO DOLLARS PER YEAR
TWENTY CENTS PER COPY
GENERAL ELECTRIC
REVIEW
VOL. XXIII, No. 11
Published by
General Electric Company's Publication Bureau.
Schenectady. N. Y.
NOVEMBER, 1920
AUTOMATIC RAILWAY SUBSTATION, SACRAMENTO NORTHERN RAILROAD
(See article, page 894)
For
Fractional H. P. Motors
NEVER vet has anyone arranged a successful
compromise between price and quality. Man-
ufacturers with an established reputation to main-
tain, never attempt it. Today — as for years past —
"NORfflfl Bearings are the accepted standards in hun-
dreds of thousands of high-duty, high-speed, electri-
cal machines which dominate their fields by virtue
of sustained high performance. It is a question ot
quality — simply.
See that your Motors
are "NORfflfl" Equipped
THE M^WmM CPMIF/^MY
inlEiiaM© /3^®InlM(l
M^^RT lf®ii°k.
Ball, Rollci*. Thrust and Combination Bearings
General Electric Review
A MONTHLY MAGAZINE FOR ENGINEERS
Associate Editors. B. M. EOFP and E. C. SANDERS
Manager, M. P. RICE Editor, JOHN R. HEWETT j„ Charge of Advertising, B. M. EOFF
Subscription Rates: United States and Mciico, $2.00 per year; Canada, $2.25 per year; Foreign, $2.50 per year; payable in
advance. Library and Student Rates: United States and Mexico, $1.50 per year; Canada, $1.75 per year; Foreign, $2.00 per year;
payable in advance.
Remit by post-office or express money orders, bank checks, or drafts, made payable to the General Electric Review, Sche-
nectady, N. Y.
Entered as second-class matter, March 26, 1912, at the post office at Schenectady, N. Y., under the Act of March, 1879.
Vol. XXIII, No. 11 6y C».S°r£&.fcl^a«y Nc^VEMBER, 1920
CONTENTS Page
Frontispiece: The Largest Electrically Operated Blooming Mill in America .... 868
Editorial ; Photo-Elasticity for the Determination of Stresses 869
Photo Elasticity for Engineers. Part 1 870
By E. G. CoKER, D.Sc, F.R.S.
The Advantages of the Modern Electric Locomotive 878
By A. H. Armstrong
The Electric Reversing Mill Considered from the Standpoint of Tonnage .... 886
By K. A. Pauly
Effect of Ultra-violet Rays on the Eye .893
Automatic Substation, Sacramento Northern Railroad 894
By W. H. Evans
Automatic Substations for Alternating-current Railway Signal Power Supply. Part I . 902
By H. M. Jacobs
The Cooper Hewitt Quartz Lamp and Ultra-violet Light 909
By L. J. BuTTOLPH
Step-by-step Integration of Curve Areas of Phase Significance. Correct and Incorrect
Methods ,. 917
By Chas. L. Clarke
Performance and Life Tests on the Oxide Film Lightning Arrester 928
By N. A. LouGEE
QEnmmL
r-H'
|-H
=i
=A
—K
—\.
—A
PHOTO-ELASTICITY FOR THE DETERMINATION OF STRESSES
In recent years, when economy in every
possible way has been of such prime impoi'-
tance in engineering, the problem of accurate
detemiination of stresses in construction
members and parts of machinery has attained
unprecedented prominence. While for simple
cases the ordinary engineering methods of
stress determination by calculation are suffi-
cient, it is becoming more clearly recognized
that these methods have to be used with
great care, and that for members of unusual
shapes they may give seriously incorrect
results.
There are two ways by which trustworthy
results may be obtained. The first is the
theoretical method, depending on the exact
solution of the well-known and accepted
equations of elasticity for the particular case,
in question. The second is the experimental
method, using a means by which the required
stress distribution can be determined by
direct measurement.
Each of these methods introduces diflicul-
ties of its own. The first, or the theoretical
method, gives good resixlts for those cases
which can be solved; but unfortimately it is
only in the simplest cases that even an ap-
proximate solution is possible. For complex
shapes no kn;.wn mathematical method can
solve the equations. For simple members
and stress systems the solutions obtainable
are embodied in the ordinary methods of
stress calculation used in the engineering
schools, and as stated, if these methods are
extended and applied to more complex shapes
in approximate or common sense ways,
mistakes will surely be made. Stress dis-
tribution in an odd-shaped member is by no
means as simple a thing as distribution of
magnetic flux or electricity in odd-shaped
paths, and greater errors will be made in
estimating the former than the latter. We
can thus sum up for the theoretical method
by saying that it helps us but little except
in the simplest cases.
The experimental solution of the problem
has until recently been worse off than the
theoretical, it having been impractical in
even simple cases. However, a unique
method has been developed by Dr. E. G.
Coker, F. R. S., of University College of
London, which gives reliable measurements
in difficult cases and which promises to take
care of all ordinary cases of plane stress, or
stress in two dimensions.
This method offered promise of assistance
in certain stress problems, and it was there-
fore decided by the General Electric Com-
pany to institute similar work in their
Research Laboratory at Schenectady. A
special set of apparatus to fill the require-
ments of this work was designed with the
co-operation of Dr. Coker and manufactured
in London and shipped to Schenectady the
past summer. Dr. Coker also made a visit
to Schenectad}^ to assist in getting the work
started.
Some of the problems which have already
been studied by this method are stress dis-
tribution in ctirved and notched beams, in
tension members of different shapes, stress
about elliptical and circular holes, that pro-
duced by a rivet in a plate, etc. During the
time that the apparatus has been in use at
the Laboratory study has been given to the
analysis of stress distributions in various
types of steam turbine bucket dovetails and
tenons with different types of load. A fact
which is clear!}' shown by .the method is the
importance of avoiding sharp re-entrant
angles in design. Many engineers appreciate
this fact, but there are also others who do not,
as designs in use at the present time bear
witness. There are many cases where re-
entrant angles are necessary but there are
few where a sharp angle cannot be rounded
with a curve or fillet which may easily reduce
the local stress to one-half.
This method of stress determination and its
application are described in a series of five
articles which Dr. Coker has prepared for the
General Electric Review. The first arti-
cle appears in this issue.
A. L. Kimball.
870 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
Photo Elasticity for Engineers
Part I
By E. G. CoKER. D.Sc, F.R.S.
Professor of Engineering in the University of London, University College
Written specially for General Electric Review
This is the first of a series of five articles by Professor Coker on the investigation of stresses by means of
polarized bght transmitted through models of transparent material under stress. Professor Coker has been
able to solve with this method many problems of stress that have defied solution bv the usual methods of
mathematical analysis This installment mcludes a brief explanation of the principles involved in this method
and then proceeds to show how it is applied in practice. — Editor.
The strength and properties of materials
under load play so great a part in con-
structional work of ever>^ kind that it is not
too much to say that a study of those branches
of science concerned with the distribution
of stress in a material are fundamental for
engineers who are for the most part engaged
in the design and construction of appliances
capable of bearing relati\'ely great loads and
often subjected to stresses of much com-
plexity.
Moreover, modem needs and scientific
discoveries are ever leading to new and more
difficult problems in engineering, for which
solutions must be found and expressed in the
form of definite machines and structures.
In most of these problems stress distribution
occurs of a somewhat complicated kind since
the exigencies of construction necessarilv
lead to forms and dispositions of material
which are not amenable to exact calculation.
Although a large munber of solutions of stress
distribution in elastic bodies have been
obtained, and are continually being added
to by the labors of mathematicians, yet it
seems impossible to keep pace with actual
needs by philosophical reasoning alone; and,
as is well known, most of the difficulties met
with in engineering construction have to be
solved by bringing to bear upon them all the
theoretical and experimental knowledge avail-
able and utilizing this store of accirmulated
knowledge coupled with practical experience
of allied problems to obtain a safe and
economical structure or machine for the
purpose required. One of the most useful
experimental methods of attacking a new
engineering problem is to examine the
properties exhibited by a model of the
proposed form in the same or in a different
material, under loads bearing a proper scale
relation to those to be borne by the full sized
structure. Such methods need no words of
commendation here since they are employed
extensively in research and have recentlv,
I understand, been useful in demonstrating
some of the most interesting phenomena
relating to rapidly rotating disks used in
turbines.
This series of articles is intended to explain
another method of experimental investigation
by the use of models which has certain
advantages peculiar to itself, and although
it is not new in principle its applications have I
been much neglected in the past. I
The starting point of all photo-elastic
research is due to the discover\- of Sir David
Brewster, in 1816, that when a piece of glass
is loaded and viewed in polarized light under
suitable conditions it shows brilliant color
effects due to the internal stresses produced
in the material. This property is shared by
most transparent bodies in more or less
degree, and its application to engineering
problems was immediately ob^-ious to its
discoverer, who suggested' that the stress
distribution in masonr>- bridges might be
in\-estigated by constructing glass models,
subjecting them to suitable loads, and exam-
ining the optical effects produced thereby.
Little use appears to have been made of
this suggestion at the time, and, so far as T
have been able to find, no useful contribution
to the study of bridges was ever obtained
until recently, when Professor Mesnager, the
head of the State Laborator\- of the Depart-
ment for Roads and Bridges in Paris, investi-
gated the stresses in a large reinforced con-
crete bridge in Southern France by aid of a
model in glass.
Here and there one finds attempts to
utilize this temporan*- double refraction due
to stress to investigate engineering problems,
but the difficulties of fashioning glass to the
required shapes and the high stresses required
to produce optical eff^ects have always stood
in the way, and it is only in recent years that
new kinds of transparent products have
PHOTO-ELASTICITY FOR ENGINEERS
{A) Beam subjected to uniform Bending moment.
{B) Simple tension member.
(C) Equally stressed tension members arranged crosswise.
(D) Circular ring m pl^nt- polarized light with black band effect.
Fig. 1. These Color Plates Illustrate the Effects Obtained by the Use of Polarized
Light in the Study of Stresses in Material
Part I
PHOTO ELASTICITY FOR ENGINEERS
871
become available for various commercial
purposes which can be utilized for the
experimental investigation of engineering
problems of stress distribution. We are
fortunate in now having at our command
many transparent bodies possessing great
optical activity under load. For example, a
simple tension member of nitro-cellulose
exhibited in a plane polarized field, when un-
stressed, is hardly visible in the dark field
produced between crossed Nicol's prisms, but
immediately load is applied it shows vivid
colors. When a small load is applied, the
specimen in the field gradually becomes visi-
ble, and shows a white color, which gradually
changes to a uniform lemon yellow as the
load is increased, and becomes successively
orange, red, blue (as Fig. IB shows), and
later a somewhat changed white as the load
increases; while if further stress intensity
is produced these colors are repeated with
some slight modifications to a second, third
and even higher order, until the specimen
fractures, when it is usually found that there
are residual colors owing to the pemianent
internal stresses produced. If instead of
white light a homogeneous light had been
employed the phenomena obsen-ed would
have been different. We should then have
found as the load varied that the changes
consisted merely of alternations from dark-
ness to brightness, according to the load
employed.
From the practical point of view these
results are important from the fact that the
colors indicate a definite stress intensity
which can be obsen.'ed with ease in the
polariscope, and if the stresses are simple
tension or compression their intensities are
immediately obvious by aid of a color scale
like that shown in Table I or by comparison
TABLE I
Order
Color
Stress
f
1
1
I i
'■
" ■
I
Black
Grev
White
Straw
Orange
Brick Red
Purple
Blue
Yellow
Red
Purple
0
3.5
5.5
8
10
10.5
11
13
18
21
22
with those observed on a simple tension
member. Thus, for example, if a beam of
rectangular section, Fig. lA, is subjected to
pure bending it shows color bands parallel
to the contours, each of which marks a stress
which can be definitely stated by reference
to the comparison tension member, since
it is found (subject to small corrections for
change of thickness) that tension and com-
pression stresses produce exactly the same
effects in a polarized field, and it is therefore
possible to map out the distribution of stress
across the section of the beam and to show
that the intensity varies according to the
distance from the central longitudinal section,
at which place a persistent dark band indi-
cates that there is no stress for an applied
bending moment.
Having indicated the nature of the phe-
nomena obser\-ed in the simple cases, it will
be useful to form some idea of the physical
conditions which obtain, and it will probably
be sufficient if we content ourselves with an
elementar\' explanation based on the wave
theory of light and to ignore the fact that
light is an electro-magnetic phenomenon in
which there are electric and m.agnetic dis-
turbances, mutually at right angles to one
another, and also to the path of the ray. We
will, in fact, select one of these groups and
ignore the other.
An ordinary beam of light, under this simple
hypothesis, may be considered as consisting
of vibrations in the ether transverse to the
direction of the ray, and of all azimuths. If a
transparent specimen under load is viewed
in such a light, there are no visible effects of
stress in the m.aterial, and none can be
obser\-ed unless a more simple type of light
vibration is employed. A convenient method
of showing the existence of stress is to pick
out from this composite beam only those
constituents of it which have the same
transverse plane of vibration, and this may
be accomplished by reflecting the beam from
a plate of black glass at a suitable angle, or
by passing it through a series of transparent
glass plates arranged at a suitable angle; or
best of all, through a prism of the form
invented by Nicol and composed of two
wedges of Iceland spar cemented together
with Canada balsam. Any of these arrange-
ments exercise a selective effect, and the
emergent ray is found to be more or less'
uni-directional as regards its transverse vibra-
tions, or, as it is commonly termed, is polar-
ized. Such an arrangement is indicated in
Fig. 2, in which a beam of light after passing
through a polarizing prism A emerges as a
uni-directional ray B, and is afterwards
87
November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
passed through a transparent plate C under
load. The effect of the interposition of such a
plate depends on the stress imposed thereon,
but in general the stressed plate causes the
uni-directional beam to break up into two
systems of transverse waves, both of which
mf
Fig. 2
suffer retardation in passing through the
stressed plate and are, moreover, found to
execute their vibrations in planes at right
angles, so that when they emerge from the
stressed member they are uni-directional
wave systems D and E out of phase with
one another and executing the \-ibrations in
planes at right angles. So far as the eye is
capable of judging there seems to be no
difference in the emergent beam except a
diminution in the intensity of the light, and
there are no visible color effects; but if these
wave systems are passed through a second
polarizer F which exercises a selective effect
and allows only components of each system
to pass through which execute vibrations
parallel to its principal plane, there emerge
two trains of uni-directional waves in the
same plane which are out of phase with each
other and therefore gire interference effects
which, when white light is used, show brilliant
color effects on a screen, or if homogeneous
light is used show bands of color separated by
black fringes wherever the two wave systems
are in phase or opposed respectively. The
arrangements of two polarizing prisms or their
equivalents, termed respectively polarizer
and analyzer, have found many applications
besides those mentioned here anel form a
useful combination, for examjile, in micro-
scopes used by mineralogists.
The most usual arrangement of the polarizer
and analyzer is to place them in such a
position that their principal planes are
crossed at right angles so that the uni-
directional beam from the polarizer is stopped •
by the analyzer and little or no light passes
through. This arrangement of a dark field
is so convenient that it is almost always used
in practice, but in this arrangement a
peculiarity is observable which is ven,- useful
in one branch of photo-elastic work, but is
highly inconvenient in another.
When a stressed specimen like the circular
chain link shown in Fig. ID is turned round
in the field of view there is in addition to the
color effects a system of black bands which
continually change their form as the specimen
is rotated. They only occur when the speci-
men is loaded, they change in appearance
when the type of load is changed, but the>"
are practically independent of the- stress
intensity. Their fonr_s are, in fact, dependent
on the kind of stress distribution which the
loading imposes, and on the disposition of the
planes of the crossed polarizer and analyzer.
If it were possible to rotate both these latter
devices at a sufficiently high velocity and still
keep their principal planes crossed, the unaided
eye would no longer be able to follow the rapid
changes of these black bands, and the effect
would appear as a slight darkening of the
field of view, but the color effects of the stress
would remain stationary-. This mechanical
method of obliterating the dark fringes, or
iso-clinic bands as they are usually termed, is
obviously not ver\' convenient, although it
has been occasionally resorted to, and a
much more convenient arrangement has been
devised which permits all the apparatus to
remain stationarv-.
It is found that certain natural substances
like mica and selenite in the form of plates
have the property of di\-iding a plane polar-
ized ray into two constituent rays, and that
if their axis and thickness are properly
adjusted a definite amount of retardation
between the two rays is produced. Such
plates are in common use and are known as
wave plates; thus a half wave plate is one
which gives a relative retardation of half a
wave length of som.e definite light vibration.
If, therefore, such a plate Ri. Fig. 3. gi\-ing a
retardation of a quarter of a wave length, is
interposed after light has passed through
the polarizing prism P. we have a system
Fig. 3
which gives a circularly polarized beam.
The effect may be described as analogous so
far as a mechanical illustration will afford a
parallel, to the effect of two simple harmonic
disturbances with a quarter-phase difference,
applied to a particle moving in a plane and
PHOTO ELASTICITY FOR, ENGINEERS
873
thereby giving a circular motion to the
particle. The circularly polarized beam
produced by the quarter wave plate cork-
screws through the stressed material, and the
emergent rays 0 and E remain circularly,
polarized, although they have been retarded
differently by the stressed specimen; but
exactly the same amount of retardation is
produced as for plane polarized light. The
circular jjolarization is afterwards annulled
by the interposition of a second quarter wave
plate R«, and the analyzer .4 picks out those
constituents parallel to its own principal
plane as before and affords the opportunity
for interference of the two emerging plane
wav-es V , W as described before.
It is worthy of note that since the wave
plates can only act perfectly for one definite
wav-e length, they only fulfill their function
exactly for the corresponding homogeneous
light, and if used with white light which is
heterogeneous they give in general an elliptic
form of polarization which does not allow a
perfectly black field. This, however, is not
found to be of any serious inconvenience for
most of the applications in which engineers
are interested.
So far we have seen that stress can be
made visible in a transparent material by
aid of polarized light, and as may be surmised
the direction of the stress is picked out by a
plane polarized beam, and for convenience
of practical work it is convenient to use
plane polarized light for measuring stress
direction, and to measure its intensity in
circularly polarized light. It is therefore
necessary to address ourselves to the task of
determining the laws which govern the
phenomena observed in the polariscope.
It will be convenient here to consider some
elementary matters relating to stress distri-
bution in plates for which photo-elastic
methods are especially suitable, and for ease
of demonstration of some of these the usual
method of notation of stresses p will be
adopted following Rankine, in which the
normal to the plane of the stress considered
is denoted by a suffix, and the direction of the
stress by a second suffix. Thus a stress p,s
indicates that the stress considered acts in a
plane perpendicular to r lin the direction s. ,-
If we now consider the state of stress
produced in an element of any plate, say
for example a rectangular element ABCD,
Fig. 4, with sides ])arallel to the co-ordinates ;
The most general system which can be im-
posed upon it consists of inclined stresses at
angles a and /3, as shown, which for equi-
librium must have shear components of equal
intensity along the meeting edges.
The stress across any other plane AC in
the plate will in general be of intensity p
inclined at some angle y to the normal to the
plane AC. If for convenience the equilibrium
Fig. 4 "^ ;
of the wedge ABC be considered, the stress
on the plane AB may be replaced by stresses
pxx normal to AB and a shear pxy in the plane
of AB, while for the face BC the stress
system will be pyy and pyx in which for equi-
librium pxy = pyx- Similarly, the stress intensity
p on the face AC inclined at d can be resolved
into a normal stress p„ and a shear stress xd.
Resolving horizontally and vertically we
obtain as the conditions for equilibrium that
p cos {Q+y)=pxx-COS d+pxy.sind
p sin (9+7) =pxy.cos d-\-pyy.sin 6
giving
(i)
p^ = pxx" . cos'- e + Pyy- . sin- e + pxy- +
pxy ipxx + pyy) sUl 2 d
(ii)
ian (d+y) = (pxy.cosd + pyy.sind) /
ipxxcose + pxy.sind)
for the intensity and direction of the stress
on a plane AC inclined at an angle 6.
This stress is wholly normal when y = ()
giving the condition
^^ (iii)
tan 2d--
P XX- Pyy
for which there are two values
and d-.
connected by the relation 62 = ^1 + ^- The
stress at a point in a plate under the most
general system of plane stress is, therefore,
wholly normal on two planes at right angles
drawn through the point and defined by
equation (iii).
It is moreover easily shown by differentiat-
ing the first equation of (ii) with regard to 6
and equating to zero to obtain maximum and
minimum values of the stress that the
874 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
criterion is of the form (iii) precisely. There-
fore at a point in a plate under plane stress
the maximum and minimum stresses are
normal to the planes on which they act and
these latter planes intersect at right angles.
If therefore the magnitudes and directions
of the normal stresses at a point in a plate
are determined, the stresses on any other
plane through the same point can be cal-
culated. These results are important since
the principal feature of a photo-elastic
investigation is the ease with which the
maximiun and minimimi stresses and their
directions can be determined experimentally,
and a solution is thereby obtained of the
stress distribution in any plate, no matter
what its form may be, which is complete and
independent.
It is convenient to point out here that the
general relations between the stresses on the
element are immediately obtainable from
Fig. 5 by resolving perpendicularly to and
along the plane AC and that they are of the
form
Prr = Pxx . C05' 9 + pyy . Stir d + pxy. St ft 2 6 ]
pee = pxx-sin'^e + pyy.cos'e — pxy.sin2e\ (iv)
pre = {pyy — pxx) sin d . COS e + pxy.cos2d J
where pee is derived from pxr by substituting
for 0 in the first of these equations.
0-)
This system is equivalent to
pxx = prr .cos- e+pee .sin- e - pre. sin2 e 1
pyy = prr ■ Sttt^ 6-\-pe9. COS' d-\-p,9. SUl 2g\ ( v)
pxy = {prr — pee) sitt 6 . COS 0+ pre. COS 2 d \
as may be shown by direct substitution from
(iv) or by direct resolution.
These formulae also give the useful trans-
formations
prr + pee = px
prr - Pee -2i. pre = e-'" (^xi •
as mav be readilv verified.
-Pyy-2
t.p:
.) } (^^)
The Law of Optical Behavior of Transparent Ma-
terials Under Stress
The optical properties of glass have been
studied with much care and thoroughness
and the laws of the particular phenomena
presented by stress effect, discovered by
Brewster, has been subjected to careful
investigation, especially in later times by my
colleague. Professor Filon.
It may be taken to be established that a
plane polarized ray passing through a stressed
plate is dix-ided into two separate rays having
planes of vibration in the direction of the
principal stresses, and" moreover, that the
retardations suffered by each are proportional
to the principal stresses in these planes so
that the relative retardation R is proportional
to the differences of the principal stresses
P and 0 at a point. Further, this relative
retardation is dependent upon the thickness
of the plate T and to the optical character
of the material. Hence, the law governing
the stress-optical effect is a linear law and is
expressible in the form
R = C (P-O) T
where C is an optical constant.
Numerous investigations have shown that
this fundamental law is approximately ful-
filled by ver\' many kinds of glass and until
lately it had been assumed to hold for other
transparent substances used for investigating
stress distribution, although the evidence for
assuming its truth was somewhat scanty.
Recent investigations on some nitro-cellu-
loses have, however, justified this assumption,
and in a later instalment some account will
be given of the nature of the evidence on
which the law rests.
In this form it is evident that the pictures
of stress shown by investigations with the
polariscope are not of a simple type, since
it is quite possible for a material to be highly
stressed at a place and yet exhibit little or no
color effect.
Thus, for example, if a rectangular plate is
exposed to equal and normal stresses along its
boundaries it will show no color effects under
any intensity of stress within the range of
this optical law, but immediately a difference
of stress is established color effects are
observed proportional to (P — 0). but so far
no means have been described for separating
the constituents.
In many problems, however, this is not
important since only one stress is present, or
if both are present one is so small in value
that it may be neglected in all but the most
accurate investigation. A particularly interest-
PHOTO ELASTICITY FOR ENGINEERS
875
ing group of cases arises at the boundaries of
plate models at which there is no direct appli-
cation of external load. In all such cases a con-
sideration of the equilibriujm of an element of
the boundary shows that there can be no stress
normal to the boundary and that the total stress
is tangential to the boundary. The color effect
can therefore be utilized to obtain its magni-
tudes directly. Although it might appear to
be the easiest plan to use a color scale for such
cases, and compare it with the observed color
at a boundary, yet experience shows that this
is not the case.
The unassisted eye is not a very perfect
instnmient for comparing colors, especially
under the severe conditions usually imposed
by the presence of a number of brilliant bands
in the field of view, and moreover there is no
certain way of deciding whether the part
under inspection is in tension or compression.
It is therefore expedient to adopt a uniform
method which evades these difficulties, and
this is accomplished very perfectly by inter-
posing a member under simple stress in the
field of view which will reduce the color at
any required point to the condition of no
stress corresponding to the dark field between
crossed polarizer and analyzer. The principle
of the method is indicated in the accompany-
ing photograph, Fig. IC, in which two equally
stressed tension members are shown crossed
and the common field of view is then found
to be the same as that produced by the crossed
polarizer and analyzer alone.
In the majority of cases a simple tension
member loaded in any convenient way is used
for this purpose, and is applied along the
direction of one of the principal stresses to
neutralize the color effect. It is clear from
the optical law given above that only two
cases can arise along a boundary free from
external load, and if there is tension the stress
can only be neutralized by placing the tension
member across the boundary, while if the
stress is compressive the tension member
mu€t be set along the boundary. An illustra-
tion of this is afforded by the case of a circular
hole in a wide plate in tension in the direction
of the arrows YY, Fig. 6. It will be found
under these conditions that the boundary
of the hole is in compression for an angular
distance of about 30 degrees from the center
line as the calibration member must be
applied tangentially to secure color extinction,
but all the rest of the boundary is in tension
as is evidenced by the necessity of placing
the calibration rnembers athwart the bound-
ary to restore the dark field.
In a few minutes it is easy to verify in this
way that the distribution of stress has the
form shown in the polar diagram. Fig. 7, in
which the maximum stress is a tension, at
the point A, of approximately three times the
average stress applied and this gradually
Fig. 7
diminishes as the contour is transversed until
at the point B at an angular distance of about
30 degrees from the central line there is no
stress at all as a minute black patch indicates.
Thence there is compression which reaches a
maximum intensity at the points C on the
center line, and is approximately of the same
value as the average stress applied.
As a rule, circular holes in engineering
practice are subjected to much more compli-
cated stresses owing to loads applied at their
boundaries by bolts and the like and such
cases require more elaborate investigation
involving the separation of the principal
stresses, and in general in the body of a
plate and away from the unloaded contours
there are two principal stresses P and 0, and
these must be determined separately, point
876 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
by point, before the problem of stress dis-
tribution can be considered as solved.
One method of carrying this out is to obtain
the values of {P-Qh 6p and Bq at a number
of points along a line starting from the
contour where the stress can be determined
accurately. A process of graphical integration
then enables one to separate the stresses in
the following manner.
Consider the most general state of equi-
librium of an elemental rectangle ABCD,
Fig. S, in which the stress components vary
continuously from point to point. It is
required to express the relations between the
variations of stress when we pass from the
point A, with coordinates Xiy to a neighbor-
ing point C with coordinates x+dxi y-\-dy.
On the face AD there is a normal stress pxx
If the plate is not too thick the stresses in
these equations m.ay be taken to represent
the conditions of equilibrium; provided they
are average values across the thickness of the
plate. Moreover, along any line the stress pxy
at any point is obtained by the relation
Pxy = -
'-^. sin 2 6
(viii)
where (P — Q) is the optical efifect there and 6
corresponds to the direction of the m-aximvm
principal stress.
If, therefore, we know the values of {P — Q)
and e for a suff cient nimiber of points along
coordinate lines passing through a point
XiV and terminating at a boundary- where the
stress distribution can be easily determined,
we can infer the stresses pxx and pyy at the point
\l,x <-
da
-dx.
*-h^^-y . . . . ^ .
--*»"fr "<•
Fig. 8
and a tangential stress, pxy, and on the
corresponding face BC there are stresses
- ''^f-'.dx, and pxy+'^^'.dx, with similar
dx •"■^' """ ''^' ' dx
stresses on the faces AB and CD as indicated
in Fig. 8.
Resolving horizontally and vertically, and
neglecting quantities of the second order in
comparison with the first, we obtain in the
absence of stresses, dtie to the vokime, two
equations of the form
d.pxx, d . pxy ^ jj
dx
d.pxy
dy
d.pyy
dx
dy
= 0
(vii)
* Experimental determination of the distribution of stress and
strain m solids by Professors Filon and Coker. British Associa-
tion Report, 1914.
-/■
.<i.v-|-G )
.c/v+G
(ix>
by a process of graphical integration since
from equations (viil
rdpxy
dy
py"--,) -dV
where the constants d
respectively, the values of px
boundary.
It has also been shown by Professor Filc.n*
that it is possible to completely deterrrine
the stress distribution in a plate if I he
isoclinic bands are accurately trapped and
the stress at a few points is accurately
known.
The methods described have the great
advantage that the measurements are purely
and L\ represent
and pxy at the
PHOTO ELASTICITY FOR ENGINEERS
877
optical, but there may be some difficulty in
obtaining accurate values of the stresses at a
considerable distance from a contour, and
an independent measurement is in general
preferably based on the lateral strain which
a plate experiences when subjected to forces
in its own plane. As is well known, simple
tension member under a stress pxx in the
direction of its length experiences a strain
fix expressed by the relation pxx = E txx,
where E is the modulus of direct extension
and this is accompanied by lateral strains in
the directions of both width and thickness of
amounts — (Te. V.V where 0" = — is a constant
m
for the material. Similarly a simple shear
stress pxy is accompanied by an angular strain
exy expressed by the relation pxy = fJLepxy where
/i is a rigidity modulus which latter is not
an independent constant, but has a value
expressed in terms of M and E given by the
relation
fx = mE/2{m+\)E (x)
In an elementary rectangle, therefore, with
sides parallel to principal stresses the relations
between stress and strain can be written
down at once by the relation
niE. tp = m.P — Q 1
mE eQ = mO — P > (xi)
-mE tR = P+Q J
and the last equation of (xi) shows that the
sum of the principal stresses can be obtained
if the strain (r is known together with the
values of ni and E.
As the strain cm can only be measured
across the whole thickness of the plate this
method gives the average value of the
principal stresses and therefore corresponds
exactly with the optical determinations of the
difference of the principal stresses at the same
point.
An instrument for obtaining measurements
of the requisite accuracy should be capable
of measuring stress to within 5 pounds per
square inch in all cases, and if this is adopted
as a criterion of performance the lateral
change which an instrument should be able
to detect is easily calculated, since frequent
measurements have shown that a fair average
value of m is.2,5 and £ = .3,000,000 in lb. and
inch units. If, therefore, a plate 0.2 inches
thick is taken since 5 (P+Q) = o lbs. per
square inch we have 6^=1/1.50,000 and
7". eR = l 7.50,000 or rather less than one
millionth of an inch.
In the lateral extensometer, designed for
this purpose, measurements of changes of
thickness of the order of one millionth of an
inch are obtained by aid of a multiplying
lever system actuating a tilting mirror and
no difficulty is experienced in measuring these
small changes provided the temperature
conditions are satisfactory.
The methods described above, therefore,
afford a means of determining solely by
experimental means the distribution of stress
in an}^ plate subjected to loading in its own
plane whatever be its form and the type of
load applied provided the material obeys the
optical law and also that the stresses do not
exceed the elastic limit of the material . Under
certain conditions, however, these limits can
be extended as will be demonstrated later.
(^To be continued)
878 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 11
The Advantages of the Modern Electric Locomotive
By A. H. Armstrong
Chairman Electrification Committee, General Electric Company
This article presents the arguments of the electrical engineer for the electrification of main line railroads.
The extremely favorable showing made for the electric locomotive is not deduced from electrification on
paper; the facts are based on actual operating results on main line operation covering a period of several
years. It would seem that about the only favorable comparison that can be made for steam operation is the
lower initial cost of equipment; but this factor is very quickly offset by higher maintenance costs, standby
losses and lower efficiency generally, smaller hauling capacity, and the fact that almost 20 per cent of the
gross ton mileage consists of company coal. This article was one of several papers read at a joint convention
of the A.I.E.E. and A.S.M.E., New York City, on the relative advantages of steam and electric locomo-
tives.— Editor.
A comparison of the modem steam and tives, some of the fundamental characteristicb
electric locomotive leads immediately to a of each are but too briefly discussed herewith.
discussion of the relative fitness of the two
types of motive power to meet service con- Possibilities of Design
ditions. At present railway practice has A locomotive is primarily a hauling machine,
closely followed steam engine development, Its design is defined by recognized limits, such
but are we not justified in looking at the trans- as maximum degree of track curvature, co-
portation problem from the broader stand- efficient of adhesion between driving wheels
point of a more powerful and adaptable type and rail, gross weight and dead weight per
of motive power? axle, tracking qualities at high speed, etc.
Place at the disposal of an experienced Furthermore, the locomotive should be simple
train dispatcher a locomotive capable of in construction, reliable and adaptable in
hauling any train weight that modern or operation, and capable of being maintained
improved draft gear can stand, at any speed in condition for a reasonable percentage of its
permitted by track alignment regardless of cost (Table I).
ruling grade or climatic conditions, that can Owing to handicap of precedent and prej-
be run continuously for a thousand miles with udice, electricity must take up the railway
no attention but tliat of the several operating problem where steam leaves off. In other
crews, and witness what he can accomplish words, the proof is up to the electrical
in his all-important task of expediting freight engineer proposing any marked departure
movement. It is not merely a question of from commonly accepted standards as estab-
replacing a Mikado or Mallet by an electric lished by long years of steam engine railroad-
locomotive of equal capacity. The economies ing. Thus while a maximum standing load of
thus effected are in many instances not 60,000 lbs. per axle has been generally
sufficient in themselves to justify a material accepted for steam engines, it is well known
increase in capital account. The paramount that an impact of at least 30 per cent in excess
need of our railways today is improved of this figure is delivered to rail and bridges
service and this can be brought about by due to unbalanced forces at speed. Impact
introducing the more powerful, flexible and tests taken on electric locomotives of proper
efficient electric locomotive. Marked changes design disclose the feasibility of adopting a
in present railway practice will undoubtedly materially higher limiting weight per axle
follow the adoption of a type of motive power than 60,000 lbs. without exceeding the
that is free from many of the limitations of the destructive effect on track and roadbed now
steam engine. As this touches upon the inher- experienced with steam engines. However,
ent possibilities of steam and electric locomo- owing to the flexibility of electric locomo-
TABLE I
COMMONLY ACCEPTED CONSTANTS
Limiting gross weight per axle tiO.OOO lbs.
Limiting dead weight per axle 18,000 lbs.
Limiting coefficient adhesion, running 18 per cent
Limiting coefficient adhesion, starting 25 per cent
Ruling gradient 2 per cent
Maximum curvature 10 deg.
Maximum rigid wheel base 18 (t.
Maximum speed on level, passenger 65-70 m.p.h.
Maximum speed on level, freight 25-30 m.p.h.
Maximum draw bar pull 150,000 lbs.
THE ADVANTAGES OF THE MODERN ELECTRIC LOCOMOTIVE
879
tive design, there is no immediate need of
exceeding steam practice in this respect,
although this and other reserves may be called
upon in the future.
Accepting the Mikado and Mallet as the
highest developments of steam road and
helper engines for freight service, a general
comparison is drawn with an electric locomo-
tive that is entirely practicable to build with-
out in any respect going beyond the experi-
ence embodied in locomotives now operating
successfully (Table II).
This analysis brings out the fact that to
equal the hourly ton mile performance of
one electric locomotive it would require three
and four engine crews respectively for the
Mallet and Mikado types.
The electric locomotive has demonstrated
its very great advantages in relieving con-
gestion on single track mountain grade
Regenerative Braking
The hazard of mountain operation is great-
est on down grades although the perfection of
automatic air brakes has done much to modify
its dangers. It is left to electricity, however,
to add the completing touch to the safe control
of descending trains by supplying regenerative
electric braking. Not only are air brakes
entirely relieved and held in reserve by this
device, but the potential energy in the
descending train is actually converted into
electricity which is transmitted through the
trolley to the aid of the nearest train demand-
ing power. Aside from the power returned
from this source (14 per cent of the total on
the Chicago, Milwaukee & St. Paul Railway),
the chief advantage of electric braking lies in
its assurance of greater safety and higher
speeds permitted on down grades. The heat
now wasted in raising brake shoes and wheel
TABLE II
COMPARISON OF STEAM AND ELECTRIC LOCOMOTIVES
Type
Weight p.er driving axle
No. driving axles
Total weight on drivers ■. '. . .
Total weight locomotive and tender
Trac. efficiency at 18 per cent coefficient
Gross tons 2 per cent grade
Trailing tons 2 per cent grade
Speed on two per cent grade
Horse power at driver rims
Indicated horse power at 80 per cent efficiency . .
Trailing ton miles per hr. on 2 per cent gradient.
Mikado
2-8-2
60,000 lbs.
4
240,000 lbs.
480,000 lbs.
43,200 lbs.
940
693
14 m.p.h.
1,620
2,030
9,700
Mallet
2-8-8-2
60,000 lbs.
8
480,000 lbs.
800,000 lbs.
86,400 lbs.
1,880
1,495
9 m.p.h.
2,080
2,600
13,500
Electric
6-8-8-6
60,000 lbs.
12
720,000 lbs.
780,000 lbs.
129,600 lbs.
2,820
2,430
16 m.p.
5,570
38,800
divisions. The number of meeting points
on a single track line increases as the square
of the number of trains operating at one time,
and is proportional to the average speed, so
that it will be appreciated what an advance
in mountain railroading is opened up by the
adoption of the electric locomotive. Further-
more, the electric performance as tabulated
above can be obtained with each individual
locomotive practically regardless of climatic
conditions, efficiency of the crew or time that
has elapsed since shopping, and with a
demonstrated reliability that has set a new
standard in railroading. In view of the facts,
it is therefore a modest claim to make that the
daily tonnage capacity of single track moun-
tain grade divisions will be increased fully
50 per cent over possible steam engine per-
formance by the adoption of the electric loco-
motive.
rims often to a red heat is returned to the
trolley system and becomes an asset instead
of a likely cause of derailment.
Cost of Maintenance
Probably in no one respect does the electric
locomotive show greater advantage over the
steam engine than in cost of maintenance.
Special importance attaches to this item of
expense in these days of high labor and
material costs. In order to draw a fair com-
parison, however, there should be added to
back shop repairs, all expenses of round-
house, turntable, ash pit, coal and water
stations, in fact the many items contributing
to rendering necessary steam engine service
as most of these charges are eliminated by the
adoption of the electric locomotive. Spare
parts can be substituted so quickly that,
excepting wrecks, there is no need of the back
SSO November, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII, No. 11
TABLE m
ELECTRIC LOCOMOTIVE MAINTENANCE, YEAR 1919
Xo. locomotives owned. .
Locomotive weight, tons.
Annual mileage
Repairs per mile
N. Y. C.
73
118
1,946,879
6. .39 cents
C. M. & St. P.
45
290
2,-321,148
14.65 cents
B. A. & P.
28
84
566,977
6.48 cents
shop for electric locomotives, unless turning
tires and painting may be considered heavy
repairs. Electric locomotives are now being
operated 3000 miles between inspections on at
least two electrified railways and the figures
of Table III are available.
On the basis of pre-war prices, maintenance
costs weri approximately 613 per cent of these
figures for the year 1919. In contrast, it can
be stated that the present cost of maintaining
a type 2-S-8-2 Alallet may be taken at 60 cts.
per engine mile, without including many mis-
cellaneous charges not shared b}- the electric
locomotive. Possibly m.ore direct comparison
may be drawn by expressing maintenance in
tenns of driver weight. (Table IV.j
Including all engine service charges, the
facts available give foundation for the claim
that electric locomotives of the largest type
can be maintained for 2.5 to 30 per cent of the
upkeep cost of steam engines operating in
similar service.
Fuel Saving
Much has been written on the subject of
fuel saving effected by steam railway electri-
fication. The estimates of electric engineers
have been called extravagant by steam
engine advocates, who in turn have been
charged with an incomplete knowledge of all
the facts available. Fuel economy figured
prominently among the several reasons lead-
ing up to the replacement of the steam engine
on the Chicago, Milwaukee & St. Paul Rail-
way as brought out by a careful analysis of the
lierformance of the steam engines then in
serA'ice. The results of the many tests are
doubly interesting when compared with the
daily performance of the present electric loco-
motives now running over the same tracks
and operated in some instances by the same
engine crews. Although the steam engines
tested may now perhaps be considered
obsolete and not within the scope of this
discussion of the modem engine, nevertheless
TABLE IV
STEAM AND ELECTRIC REPAIRS
Steam Mallet
C. M. & St. P. Elec.
Cost repairs per mile
Weight on drivers
Cost repairs per 100 tons loco, weight on drivers. . .
60 cts.
240 tons
25 cts.
14.65 cts.
225 tons
1 6.52 cts.
]
TABLE V
LOCOMOTIVE DATA
C. M. & St. P. Tests
Type
Weight of engine . .
Weight of tender
\V light total engine and tender.
Weight on drivers
Ratio driver weight to total.
Rigid wheel base
Diameter drivers. .
Cylinders
Boiler pressure
Heating surface
Grate area ...
Water capacity
Coal :
Steam
Electaic
21
2-6-2
206,{KHt lbs.
1, ".4, 00(1 lbs.
.SliO,lKHI Ihs.
I."i2,0(t(l lbs.
42.2 per cent
13 ft.
63 in.
in. by 28 in.
20(1 lbs.
2346 sq. ft.
45 sc). ft.
8000 gal.
14 tons
4-4-4-4-4-4-4
568,0tH) lbs.
568,000 lbs.
450,000 lbs.
79.3 per cent
10 ft. 6 in.
52 in.
THE ADVANTAGES OF THE MODERN ELECTRIC LOCOMOTIVE
SSI
it is not without value to compare the results of
steam and electric locomotives operatins^ over
such long distances under identical conditions.
The following data are therefore submitted as
applying to a particular equipment only. No
claim is made that these figures are rei^resenta-
made in identical time on the basis of lOOU
total gross tons moved in each instance. The
fuel furnishing x'ower to the steam train was
coal having the analysis shown in Table VI.
Electric power was furnished by water and
hence no direct coal equivalent is provided by
TABLE VI
COAL ANALYSIS
Fixed Carbon
Volatile Carbon
Ash
Moisture
B.T.U's
47.99
38. 9S
8.35
4.68
11,793
tive of the best modern steam engine perform-
ance, although many thousands of steam en-
gines still in operation will show no greater
economies than those given in the table. The
general data applying to the steam and elec-
tric locomotives tests are give in Table V.
Other engines were also tested over other
sections of track, biit the following particular
runs are chosen for illustration as bringing out
most strikingly the inherent disadvantages of
operating a steam engine over a single track
mountain grade division and handicapped by
the usual delays attending freight train service
under such conditions. The run of 11 1.1
miles from Harlowton, elevation 41()2 ft., to
Three Forks, elevation 4060 ft., over the Belt
Mountain divnde at Loweth, elevation 5S7'.)
ft., was mace by steam with S71 tons trailing
in 20 cars and by electric locomotive haviling
64 cars weighing 2702 tons. In order to picture
a direct comparison of the results of the steam
and electric runs, all test data are reduced
to a common basis of 1000 gross tons moved,
this unit of measurement including the loco-
motive and tender weight. The running
speed of the electric train was but slightly
higher than the steam and the additional
correction in the power demand rate of the
former is made proportional to the lower
speed. Both runs are therefore shown as
the test result. To aflord a common basis of
comparison, however, a single assumption
seems permissible and a rate of 2]/2 lbs. of coal
per kilowatt hour is taken as representative of
fair electric power station practice. Coal
burned under the steam engine boiler was
determined by weighing at the end of the nm
and by detailed record of scoops en route.
Power input to the electric locomotive was
obtained by carefully calibrated recording
wattmeters as well as curve-drawing volt and
ampere meters. These values of locomotive
input were raised to the value of three phase
power purchased in the ratio of OS per cent
given by R. Beeuwkes in his A.I.E.E. paper
of July 21, 1920, and the kilowatt-hours so
obtained reduced to coal equivalent in the
ratio of 2\-> lbs. coal per kw.-hr.
The picture thus secured affords a most
striking illustration of one of the principles
upon which advocates for electrification base
their claim for fuel economy (Fig. 1). While the
electric locomotive demands power only when
in motion, the steam engine requires coal at all
times during the twenty-four hours, whether
doing useful work, standing idle or coasting
down grade. In fact so called "standby losses"
were such a large percentage of the total coal
consumed that a careful record was kept of
their several amounts. (See Table VII.)
TABLE VII
FUEL COMPARISON
Doing useful work . ,
Making up fire
Delay at Harlowton. . .
Held up at Lennep, . .
Held up at Loweth. . .
Held up at Dorsay. . .
Fire, banked 9 hrs.. . .
Coasting down grade -
Total standby losses. .
Regenerative braking.
Total net coal
Steam
1,535 lbs.
2,270 lbs.
394 lbs.
128 lbs.
230 lbs.
1 ,425 lbs.
3,060 lbs.
23,640 lbs.
9,042 lbs.
32,682 lbs.
Electric
8,100 lbs.
1,430 lbs.
6,670 lbs.
882 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
The run of a more modem steam engine
would have effected a material reduction in
the 23,640 lbs. of coal burned in doing useful
work, but the amount of coal wasted in standby
losses (9042 lbs.) might have been dupli-
cated or even possibly increased with larger
grate area. As standby losses constitute so
large a proportion of the total coal burned
(27 J/^ per cent in this instance) it is apparent
that enormous economies over the simple
engine tested must be realized in the modem
superheater and other improvements since
introduced to offset in part the high inherent
efficiency of the electric locomotive.
To assist in arriving at a truer comparison
of modern steam and electric locomotive
operation, a further analysis of the above test
is made. The cormnonly used unit, pounds
of coal per 1000 ton miles, is at best a ver\^
rough and unstable comparison of steam
engine runs over different profiles, with
variable quality of fuel and operating con-
ditions. For illustration, the data of Table
VIII may apply.
6000
5000
4000
3000
2000
1000
3000
2000
1000
0
1000
6000
5500
5000
4500
4000
Pounds coal per gross 1000 ton miles may
thus vary from 650 to 50.5 according to
gradient and with no standby losses whatever
included. The boiler must be kept hot at all
times, however, and fully 33 per cent can
safely be added to the figures to indicate the
inevitable standby losses inherent to steam
engine operation. Except over ven,' long
runs with terminals at the same elevation it
seems hardly possible therefore to accurately
compare engine performance over different
profiles by such a variable unit as pounds
coal per 1000 ton miles.
A truer understanding of what takes place
under the engine boiler may be shown by
continuous records of coal burned, tons moved.
profile, delays, etc., all reduced to pounds of
coal burned per useful horse power hour work
done at the driver rims with segregation of the
many standby losses. (Table IX.)
L'nder the same conditions a modem
engine would undoubtedlv have consimied
much less than 9.02 lbs. coal (11,793 B.t.u.)
while doing work measured at the driver rim.
TEA
LLY E
1
1
1 ■ s
COAL U5EFU
M RUN
3URNED 23640 LBS.
POU
NDS
STAND-BY LOSSES 9042 1 RS.
COAL TOTAL 32682 LBS.
COA
PEF
L
I
HOU
R
1
1
1
-^^■M
L.^^H
1
i 1 1
1 1 1
ELECTRIC RUN
COAL TOTAL 8100 LBS.
RETURN BY ELECTRIC BRAKING- 1430 LBS.
COAL- NET TOTAL 6670 LBS.
POUNDS
COAL
PER
HOUR
1
1
II
1
■
1
ll
1 1 1 1 1
—
-;
> - ;
5 - i
,- \
j - (
A.
M.
' - {
5 - <■
) - 1
0-1
1 12-
- 5
) - '■
J - i
I- !
5 - (
i- ■
p.
r- j
M.
5 -<
t -1
0- 1
1 -
/i
\
1 1
FEET
/
■
/ I
\
s
ELEVATION
hArlowton /■
loweth
\
three forks
V ll 1
y
\J
1
J
COALR
:cop
D.5TEA
M AND ELECTRIC
1
RUNS HA
^LOWTONfO]
'HRE
Era
RK5 M
3 GROSS TON
5M
DVE
0
Fig. 1
THE ADVANTAGES OF THE MODERN ELECTRIC LOCOMOTIVE
883
TABLE VIII
POUNDS COAL PER 1000 TON MILES
Horse power hours at driver rims
Indicated horse power hours at 80 per cent eff.
Lbs. water per i.h.p. hr
Lbs. water per lb. of coal
Lbs. coal per i.h.p. hr
Lbs. coal per 1000 ton miles
Lbs. coal per 1000 T. M. Trailing
2% Grade
Level Track
123
18.8
154
23.5
20
16
6
8
3.33
2.0
513
47
650
50.5
The addition of superheaters gives greater
output and economy, while mechanical stokers
add output only and, it is claimed, at some
expense in economy over good hand firing.
However efficient the power plant on wheels
may reasonably be developed without too
seriously interfering with the sole purpose of
the steam engine, namely, the hauling of
trains, it can never approach the fuel econo-
mies of modern turbine generating stations.
Whatever transmission and conversion losses
are interposed between power house and
electric locomotive are more than com-
pensated for by the improvement in the load
factor resulting from averaging the very
fluctuating demands of many individual
locomotives.
Every electrical engineer has learned the
lesson of the fuel economy resulting from
replacing several small and inefficient power
stations by one large power house of modern
construction. It therefore brings no surprise
to his mind that the comparison of steam and
electric railway operation discloses such
enormous fuel savings in favor of the latter,
for as a matter of fact, while our railways
carry on a wholesale transportation business
of the greatest magnitude, they are never-
theless engaged in burning coal and oil at
retail on some 65,000 individual engines.
The average output of each engine during
the time it is at the call of the transportation
department is but a small fraction of its
rating. The ftiel economy is further effected
by the condition of the boiler and climatic
conditions. Hence the average performance
of many thousands of steam engines must
reflect all the many handicaps of construction
and service under which they operate.
It would be a simple matter to carry
through a series of runs over the electrified
zone of the C. M. & St. P. with a inodern
Mikado equipped with all the up-to-the-
minute fuel saving devices and thus provide
the necessary data to draw direct comparisons
with the electric locomotive. Such tests
with modern steam equipment would un-
doubtedly discredit the above comparison,
which is based upon the economies of six
years ago and might lead to something
approximating the blend of fact and theory
given in Table X.
This table is based upon actual electric
locomotive performance, Harlowton, to Three
Forks, coal taken at 23^ lbs. per kw-hr. at
assmned steam power station. Steam engine
values are based upon the known working
efficiency of a Mikado equipped with super-
heaters but penalized with the same standby
losses actually determined with simple engine
tested Harlowton to Three Forks. A test
run from Harlowton to Three Forks with a
modem Mikado engine hauling 1420 tons
may possibly show a lower average fuel rate
than 3 lbs. per indicated horse power hour at
drivers, and lower standby waste than 9042
lbs. coal, but the average annual performance
of many such engines would be most excellent
TABLE IX
ANALYSIS OF STEAM AND ELECTRIC RUNS, HARLOWTON TO THREE FORKS
PER 1000 TONS MOVED
Steam
Electric
Kw-hrs. at driver rims
H.p. hrs. at driver rims
Coal per h.p. hr. driver rims
Credit regenerative braking. .
2038
2625
9.02 lbs.
2.47 lbs.
11.49 lbs.
2038
2625
*3.09 lbs.
.55
Standby losses, 27 J 2 per cent
Total coal per rim h.p.-hr
2.54 lbs.
'Measured at power house and includes locomotive losses and 32 per cent transmission and conversion loss.
884 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. II
if it reached the net figure arrived at, viz.,
5.9 lbs. coal per actual horse power hour
work performed at drivers. The electric
run, however, is being duplicated daily as to
relation between kilowatt-hours and ton miles,
and it is just this reliability of electric opera-
tion that may at times give rise to mis-
understanding in the comparison of steam
and electric data.
Each individual electric locomotive will
reproduce almost exactly the record of all
others in similar service, little influenced by
either extreme cold or skill of the engineer;
while the firemen so-called and still retained
has nothing to do with the matter at all.
There is no creeping paralysis gradualh'
impairing the efficiency of an electric loco-
general adoption of the electric locom.otive
would probabh' result in saving fully two-
thirds the fuel now burned on present steam
engines, and possibly one-half the amount
of fuel necessary for steam engines of the
most modem construction.
Comparative Cost
The superior operating advantages of the
electric locom-otive are admitted by many
who believe the first cost to be. prohibitive.
largely due to the trolley construction,
copper feeders, substations, transmission lines,
etc., which are necessary to complete the
electrification picture. It is true that such
auxiliaries add an amount that may equal the
electric locomotive expense and the task of
TABLE X
THEORETICAL COMPARISON, MODERN STEAM AND ELECTRIC LOCOMOTIVES
HARLOWTON TO THREE FORKS
Type
Weight on drivers
Weight engine and tender
Trac. efficiency 18 per cent coefficient.
Trailing tons 1 per cent grade
H.p. hrs. at driver rims
Coal per indicated horse power hour. .
Coal per driver horse power hr
Standby loss, test result
Standby loss per h.p.hr
Total coal per driver horse power hour
Coal at power house, k\v-hr
Coal at power house h.p. hr
Coal at locomotive driver, h.p. hr
Coal credit due regeneration
Net coal at driver h.p. hr
Total net coal
1000 trailing ton miles
Coal per 1000 ton miles
Ratio coal burned
Mikado
2-8-2
4-4-4-4-4-4
240,000 lbs.
450,000 lbs.
480,000 lbs.
568,000 lbs.
43,200 lbs.
81,000 lbs.
1,420
2,836
4,360
3
3.75
8,200
9,042 lbs.
2.15 lbs.
5.90 lbs.
5.90 lbs.
24,800 lbs.
157,500 lbs.
158 lbs.
2.37 lbs.
Electric
2.5 lbs.
1.86 lbs.
3.09 lbs.
.,55 lbs.
2.54 lbs.
20,900 lbs.
314,000 lbs.
66.7 lbs.
lib.
motive until temporary relief is obtained
through frequent washing of boiler and round-
house tinkering, inevital)ly ending up in the
major operations annually laerformed in the
back shop hospital on the steam engine to keep
it going. It is for such reasons that the
electrical engineer is slow to accept general
statements of average service operation based
on the results of tests usually made on steam
engines in excellent condition and skillfully
handled. Then, too, there is insufficient data
available as to standby losses, which must
finally largely account for the wide dis-
crepancy often noted between the amount of
fuel purchased and fuel presumably burned or
computed on the basis of test run records.
It is with some knowledge of all these facts
that the broad statement is made that the
proving the electric case is not made easier by
the fact that steam engine facilities arc already
installed and may have little or no salvage
value to offset new capital charge for electri-
fication.
Comparing the cost of equivalent steam and
electric motive power, it is apparent that on
the basis of the same unit prices for labor and
material, the first cost is approximately
the same. While electric locomotives cost
possibly .')() per cent more than steam for
equal driver weight, the smaller number
required to haul equal tonnage may quite
offset this handicap, especially with quantity
production of electric locomotives of standard
design.
The steam engine also demands a fonnida-
blc array of facilities ijeouliar to itself, as
THE ADVANTAGES OF THE MODERN ELECTRIC LOCOMOTIVE
885
shown in the following table of expenditures
made on 14 railways included in the North
Western group from 1907 to 1919. This
expense covers fuel and water stations, shops
and engine houses, shop machinery-, turn-
tables, ash pits, etc.
EXPENDITURES FOR ENGINES AND
FACILITIES, NORTHWESTERN GROUP
1907-1919
Engines Facilities
$68,000,000 $42,200,000
Proper facilities for rendering adequate
steam engine ser\-ice apparently add some
62 per cent to the cost of the latter and no
crv of extravagance has ever been raised in
this respect.
One of the advantages of electric loco-
motives rests in the longer engine divisions
which they make possible. Two of the four
steam engine divisions comprised in 440 miles
of the St. Paul were wiped out by electri-
fication and certain sidings and yard tracks
were dismantled. To these exclusively steam
engine facilities should be added therefore
the expense of engine division points not
necessarv to successful electric railroading.
Further credit is due to cover coal cars
released. Therefore, considered as a problem
of construction only, electrification of a new
road may in some instances compare quite
favorably with the complete first cost of steam
engines and all facilities incident thereto. As
the general problem, however, is one of
replacing steam engines now running, the
economic advantages of electrification are
rather individual to the particular railway
under consideration. The operating econo-
mies effected under favorable conditions have
been found sufficient to show an attractive
return upon the additional capital charge
incurred besides providing the improved
service which was the main objective in view
in replacing the steam engine.
No discussion of electric railway economies
would be complete without comment upon the
increased value of real estate brought about
by terminal electrification. Not only is
neighboring real' estate benefited thereby,
but the "air rights" over the electrified tracks
may become so valuable as to largely pay the
cost of the change from steam. With the
work but partly finished the Grand Central
Terminal District, New York City, is already
a remarkable example of the indirect benefits
derived from electrification.
Summary
Some of the principal advantages claimed
for the electric as compared to the steam loco-
motive are briefly :
1. No structural limits restricting tractive
effort and speed of electric locomotive
than can be handled by one operator.
2. Practical elimination of ruling grades by
reason of the enormously powert'ul
electric locomotives available.
3. Reduction of down grade dangers by
using regenerative electric braking.
4. Very large reduction in cost of loco-
motive maintenance.
5. Very large saving of fuel, estimated as
two-thirds the total now burned on
steam engines in operation.
G. Conservation of our natural resources
by utilizing water power where available.
7. Material reduction in engine and train
crew expense by reason of higher speeds
and greater hauling capacity.
8. Increased valuation of terminal real
estate following electrification.
9. Increased reliability of operation.
10. Material reduction in operating expense
due to elimination of steam engine
tenders and most of the Company- coal
movement, the two together expressed in
ton miles approximating nearly 20 per
cent of present gross revenue ton mile-
age.
1 1 . Large reduction in effect of climatic con-
ditions upon train operation.
12. Postponement of immediate necessity
for constructing additional tracks on
congested divisions.
13. Attractive return on cost of electri-
fication by reason of direct and indirect
savings in operation.
14. Far reaching improvements in operation
that may revolutionize present methods
of steam railroading.
886 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
The Electric Reversing Mill Considered from
the Standpoint of Tonnage
By K. A. Pauly
Power and Mining Engineering Department, General Electric Company
This discussion of the operation of the electric reversing steel mills is unique in that it is based on tonnage
production. Since the current-limit setting, which is necessary for the protection of the equipment, measures
the peak load capacity of the drive and places a limitation on production, a very careful analysis is made of
the speed-torque characteristics of the drive as affected by the current setting when rolling the several passes
required to break down an ingot. It is shown that an increase in production can be secured with less than
the same percentage increase in first cost of equipment. This fact should receive most careful consideration.
Originally, this article was presented as a paper before the 1920 Annual Convention of the Association of Iron
and Steel Electrical Engineers at New York City. — Editor.
Many articles have been written bearing on
the subject of the electrically driven reversing
mill but a perusal of them reveals the fact that
the matter of tonnage, so important to the
steel mill man, has been given almost no place
in them. Interesting installations and systems
of control, having as their object the protec-
tion of the electrical equipment from peak
loads, have been described, frequent references
have been made to the power consumed in
rolling and to its cost, but nowhere has the
writer found that we have squarely faced the
question of tonnage, the real issue between
the steam driven and the electrically driven
mill. Can the latter produce, at a reasonable
cost, a tonnage equal to or greater than the
steam mill? This question we can unquali-
fiedly answer in the affirmative, as is proved
by the records of some of the mills to which
reference is made later. From the stand-
point of production the electric mill has many
advantages over the steam mill, chief among
which is the extremely small loss due to delays
caused by the drive and because of this an
electric mill of the same hourly capacity will
e.Kceed the steam mill in monthly or yearly
capacity. However, the task of the electric
mill is not an easy one and the handicap
which the steam mill has over its competitor,
due to the lower moment of inertia of its
moving parts, must not be treated too lightly.
Obviously the principal factor which affects
the hourly capacity of a mill is the time re-
quired to roll an ingot without exceeding the
capacity of the drive to repeat this cycle
indefinitely. Consequently, this article is
devoted largely to a discussion of the influence
of the control on the rolling time, partiai-
larly to the effect of limiting the maximum
power delivered to the direct-current motor
driving the rolls. As we are concerned only
with the reversing mill proper, it is assumed
that no unnecessary' delays occur in the manip-
ulation of the steel on the live roll tables, in
the transfer of the ingots from the soaking pits
or in the removal of the finished steel from the
mill. If any improvement in these details is
possible, no time should be lost in taking the
necessar}' steps whatever the type of drive
used for the main rolls. We are therefore
assuming that the mill is properly tuned up
and that the inter\-als between passes are the
same for all mills, referring briefly later to some
of the characteristics of the different types
of drives which tend to affect this inter\-al.
The time required to roll an ingot, assum-
ing the inten-al between passes to be the
same, is increased directly by an increase in
the time the steel is under the rolls. This in
turn is affected by the speed at which the steel
enters the rolls, the time required to accelerate
to the maximum speed after the steel has
entered, the maximum speed attainable,
the time the rolls run at this speed during the
pass, and the time to retard the mill to a
proper delivery speed at the end of the pass.
The entering speed is limited largely by the
section of the steel, the design of the rolls, the
draft, etc., and is independent of the drive.
The deliver^' speed should be as high as is
permissible without increasing the inten'al
between passes, and as the time for retarda-
tion is extremely short with any drive it may
be assumed without appreciable error to be
the same for all. This leaves only the time
required to accelerate after the steel enters,
the maximum speed, and the time of rolling at
the maximum speed, as being subject to
variations due to the speed-torque character-
istic of the drive.
Before entering further into a discussion of
the subject, the writer wishes to caution
against a ven* common mistake made by
engineers when discussing reversing mill
problems. We frequently hear the expression :
"The time of reversal." This should not be
confounded with the time required to
accelerate. If the word "reversal" or the
expression "acceleration and retardation" is
used, it should be qualified by the limiting
ELECTRIC MILL CONSIDERED FROM THE STANDPOINT OF TONNAGE SS7
speeds between which the reversal or accelera-
tion and retardation takes place, e.g., from 90
r.p.m, forward to 90 r.p.m, reverse. By
simply rocking a machine, it may be made to
reverse many times a minute; and yet it may
be extremely slow in getting up to speed with
steel in the rolls.
The reversing mill was first driven electri-
cally abroad where the requirements of the
mills are very different from those in America.
In general the rolls are larger in diameter, their
speeds are higher, and the tonnages are very
much smaller than is our practice. Naturally,
therefore, we find the motors, generators, and
controls designed to meet these conditions
rather than those of producing the maximum
tonnage from the mills which they drive. A
complete description of one of the early
German electrically driven reversing mills is
given in Stahl und Eisen, January 23, 1907,
from which the following short description
of the system of control is taken:
"The main rolls are driven by three direct-
current motors direct connected to the mill
and connected electrically in series and
supplied from an Ilgner Ward-Leonard fly-
wheel set having two generators designed for
.500 volts each, connected in series, making a
total of 1000 volts applied to the mill motors.
The excitation for the motors and generators
is obtained from a small exciter motor-
generator, consisting of two direct-current
generators driven by an induction motor. One
of these generators serves to excite the shunt
windings of the generators and main roll
motors. The other generator of the exciter set
supplies special compound windings provided
for strengthening the fields of the roll motors
as their loads increase, thereby causing a
reduction in their speed and thus relieving the
generators of the overloads which would
otherwise be occasioned."
In America conditions are very different.
Tonnage is usually the all important considera-
tion and a fraction of a second per pass lost
because of insufficient capacity in the drive,
because of the slowing down of the roll motors
due to special windings as described above, or
because the motors are prevented by any
means from taking the peak loads required
to accelerate properly with the piece in the
rolls, will cost the operator (through the loss
in production) many times the few dollars he
will save in the first cost of the equipment.
That reversing mill requirements are severe
must be recognized, and the machines used to
drive them together with their controls must
be designed to meet these conditions with
sufficient momentary peak load capacity to
take care of the combined acceleration and
rolling load of each pass, and with sufficient
continuous capacity to roll constanth^ at the
rate necessary to produce the required ton-
nage. Because of the confusion which now
exists due to the different methods of rating
the equipments now in use, it is imperative
that standard specifications be prepared to
cover reversing mill main roll drives. That
the need for this is fully appreciated by both
the manufacturers and the operators is
evidenced by the discussion which followed
the reading of the paper, "Standardization of
Ratings of Large Rolling Mill Motors," pre-
sented by the writer at the 12th Annual Con-
vention of the Association of Iron and Steel
Electrical Engineers, held at Baltimore in
September, 1918.
There is, of course, a limit to the capacity
of any drive; and the characteristics of
electrical equipment are such that unless some
automatic means is provided for limiting the
power of the roll motors a careless operator
can abuse the equipment. Current-limit con-
trols have been developed for this purpose,
their function being to limit the current
taken, and, therefore, the maximum power
developed by the roll motors to a value which
they and their generator can safely carrv at
frequent inter\'als. The current-limit setting
for the main roll motors is therefore a real
indication of the relative peak load capacities
of two eauipments designed for the same
voltage. This limitation of current taken by
the roll motors should not be confused with
the control of the induction motor driving the
flywheel motor-generator through the slip
regulator. As time is required for the current-
limiting device to function, the current tends
to rise above the \'alue corresponding to the
relay setting, but the peak in excess of the
setting is of such short duration that it pro-
duces little effect on the acceleration of the
roll motor. The effect of the setting of the
current-limit relay on the tonnage produced
in the mill, and the importance of setting it
for as high a current as is possible can be very
well understood by a brief study of the effect
of different current-limit settings upon the
time to make the pass.
For the purpose of comparison, we have
assumed the mill to be driven by a motor with
the current-limit set for three different values
(9.500, 8500, and 7000 amp.), the potential of
the generators supplying the motors being
1200 volts when delivering the currents for
which the current-limit relaj-s are adjusted.
888 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
and with the motor-generators running at the
minimum speed occurring during the normal
rolling cj-cle. Such a generator or generators
will develop 1450 to 1500 volts when running
at full speed and carrying no load, and care
must be taken in computing the horse powers
ZO 46 60 BO 100
Rol' Motor ioeed IPPM
Fig. 1. Speed-torque Curves of a Reversing Mill Motor Corresponding to
Three Current-limit Settings, and Torques Required to Roll Three Passes
corresponding to any specified current-limit
setting that only the voltage delivered by the
generators when delivering these currents is
used and not the running light voltage, as is
frequently done.
^The curves in Fig. 1 are the approximate
speed-torque curves of a typical reversing mill
motor corresponding to these three current
settings, together with the torques required
to roll three of the passes in breaking down
an ingot 22 by 24 inches on the butt end and
weighing 8100 lb. to an 8 by S-inch bloom in
15 passes. These passes arc the first pass
after the ingot has been squared up, referred
to later as pass "a"; the middle pass "b";
and the last pass "c."
Now if the mill is tomeetitstonnage'require-
ments, it must be able to accelerate rapidly
after the steel has entered. Of the motor
torque available for acceleration after the steel
has entered, we have only that which is in ex-
cess of the amount required to roll and this
decreases at a much greater
rate than the reduction in
current-limit setting. This
is clearly shown by Fig. 2,
from which it will be seen
that although the 8500-
amp. setting is only 133^2
per cent less than the 9500-
amp. setting, the torque
with the (S500-amp. setting
available for accelerating
the mill, after the steel has
entered, is for pass "a"
only one-half of that with
the 9500-amp. setting, and
for pass "b" is only two-
thirds of that with the 9500-
amp. setting. This differ-
ence becomes still more marked as the current
limit is further reduced until wc reach the 7000-
amp. setting for which there is no torque avail-
able for acceleration after the pass "a" has
entered, although the 7000-amp. setting is
only approximately 27 per cent below the 9500-
amp. setting. Also, with the 70()0-amp. set-
ting we have available for acceleration after
pass "b" enters only one-third of the torque
available for the 9500-amp. setting. At this
setting the pass "a" cannot be rolled faster
than the entering speed, and the speed
during pass '"b" can increase only slightly
above the entering speed. As the current
limit is reduced below 7000 amp. it becomes
necessary to increase the number of passes
Ik SOOfiOO
\
SV
^.^.
^.
^?^;:
::;:^;^^;
r^s:^:^
0 20 40 60 80 la
^Qfi Motor 5pce</f.PM
%.i(y
^ SOO.OOO
\^
^^;
^.^.
s^,
:y;:;:n:
:^5:=;^:
40 60 60 HX)
Roll Motor SDft'SdFPM
Fig. 2. Speed-torque Curve's of Fig. 1 Analyzed with Respect to the Torques Required by the
Passes Shown in Fig. 1
ELECTRIC MILL CONSIDERED FROM THE STANDPOINT OF TONNAGE SSI)
and to rearrange the rolling cycle, reducing
the drafts of some of the passes and rolling
others at the speeds at which the steel enters,
the efTect of which, on the tonnage output of
the mill, is obvious.
On the other hand, it is the peak load as de-
termined by the current-limit setting and not
the magnitude of the torque available for accel-
eration ai, ao, as, bi, etc.. Fig. 2, which affects
the first cost of the equipment, driving the mill,
and any saving in first cost conditioned upon a
limitation of the current below the value re-
quired to roll, with a reasonable margin avail-
able for acceleration after the steel enters, is
false economy . Furthermore , the first cost is af-
fected by the continuous as well as the momen-
tary' peak load capacity. Therefore, the per-
centage increase in first cost will often be less
than the percentageincreaseinthecurrent-limit
setting which this increase in first cost buys.
The current-limit setting is, therefore, a matter
which should be given very careful considera-
tion in comparing equipments of this type.
The effect of the current-limit setting on
the length of time to roll can be determined
mathematically by the use of the following
formula?, by subdividing the roll motor torque
cur\-e into two parts; the portion correspond-
ing to full motor field and that corresponding
to weakened motor field.
The time t reauired to accelerate from a
speed So to 5i and the distance traveled D bj-
the steel during this time can be determined
by the following formula:
When 5o and Si fall within the full motor
field portion of the speed-torque cur\'e:
0.195 ir/?M5i- So)
t-
D =
Tr
s,+t
"-Li
Where
ll'i?^ = Weight in pounds times radius of
gyration in feet scuared, of the revolv-
ing parts of the mill and drive.
7^ = Motor torque minus load torque; ai,
ao, etc., Fig. 2.
L = Distance traveled in inches for one
revolution of rolls.
When So and Si fall within the weakened
motor field portion of the speed-torque curve:
'4/
D =
A'
A (S,- So) + HP, log.
HP, + ASo-\
A-
AS\-S\)
HP, + AS,\
HPi.4(Si-S„)-
HPi+Asol
A
^^'here
K = 0.00223 U'R-
'ST.o
//Pi = Maximum horse power of motor
as determined by the current-limit
setting.
Ti = Torque required to roll.
-:x:::tj:;::::HI: = H = J:::::::;::
> rvroffs J. ^a -f f, '3^ . C W//j r 7//1 -"^
_ -^s -_ -
=^~;;
'' rttl" L n 11 1 1
•0
• JO
K :;::::::::;::::^: ::::::::::
'■^ ~ ~~ per pas J tea >-r 0/3 ^ .
:::::::::::: ::^:::-^---^ ---=;---
laoo »oo xoo aooo sooo eooo tooo oooo 9oao foooo
Current Limit Settiifff irt Jlmper^s
Fig. 3. Curves Showing the Increase in Time Required
to Roll Resulting from Lowering the
Current Limit
The effect of lowering the current limit on
the time required to roll is shown by the
curves in Fig. 3, which are based on rolling
the 8 by 8-inch bloom from the 22 by
2-l:-inch ingot referred to. This cur\'e has been
extended back only to approximately 7000
amp. as it becomes necessar\' to re-arrange
the passes, increasing the number and reduc-
ing the drafts, as the current limit is reduced,
and thus the time for rolling is further in-
creased with corresponding reductions in ton-
nage output. The cur\'e shows that lowering
the current limit from 9.500 amp. to approxi-
mately 7000 amp. increases the average time
the steel is in the rolls by approximately XYi
seconds per pass, which for a 15-pass cycle
means a loss of 22J^ seconds per ingot, an
amount which will seriously reduce the output
HP VI log.
HP,+AsA
TABLE I
Pass
Approximate
Current-limit
Setting; Amp.
Maximum Speed
During Pass;
R.p.m.
a
a
a
9500
8500
7000
53
42
10
b
b
b
9500
8500
7000
67.5
62
52.5
c
c
c
9500
8500
7000
124
116.5
104
890 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 11
of any mill designed for large or moderately
large tonnage.
It is recognized that the rate of acceleration,
and therefore the time of rolling, is affected by
the moment of inertia of the revolving parts,
but the variations in modem standard
. Second intervats
Fig. 4. Speed-time Curves Taken at the Trumbull Mill
Before and After Changes Were Made in the Control
reversing mill drives from the values used in the
preparation of the curves in Fig. 3 will produce
no material changes in the results shown.
It is also interesting to note the effect of the
variations in current limit on the maximum
speed attained during the pass as given in
Table I on preceding page.
Appreciation of the im-
portance of tonnage to
American operators led us
to adopt the shunt-wound
motor for driving reversing
mills in preference to the
compound-wound motor so
generally used throughout
Europe, as well as to some
extent in America. The
machines are designed to
commutate large currents,
our current limits being set
at 9500 amp. and higher
for 1200-voIt units. The
changes in the field fiux of
the motors and generators
have been carefully stud-
ied, and the control so de-
signed as to take maximum
advantage of the charac-
rapid acceleration and retardation without i
the least interfering with the complete control
of the speed of the motors for rolling. Special
means are provided for forcing the generator
and motor fields to act quickly during
acceleration and retardation. The magnet
yoke as well as the pole pieces of the genera-
tor fields are laminated to reduce to a mini-
mum the eddy currents that tend to reduce
the rate of acceleration and retardation.
However, in spite of the complexity of the
problem involved in the control of these large
units, the system of control as finally worked
out is readily maintained by the plant
electrical department.
It is essential in entering any new field of
application to include a larger factor of safety
in the design of the equipment involved than
is customary in established fields; and after
our early reversing mill equipments had been
in operation a short time, we appreciated that
by certain minor changes in the control we
could make a material reduction in the time
required to accelerate and retard, and accord-
ingly a complete series of tests were made at
the works of the Trumbull Steel Company
with this in view. The Trumbull Mill is a
36-inch reversing blooming mill driven bv a
standard 6250-h.p. (A.I.E.E. rated; 5000-
h.p., 35 deg.) 50/120 r.p.m. direct-current
reversing mill motor, having a maximum
momcntar\^ capacity of 17,000 h.p. at 45
r.p.m. and supplied with power from a 5400-
kw. (A.I.E.E. rated; 4000-kw., 35 deg.) fly-
wheel motor-generator, consisting of two
tn.^J.ljlf . ._ •- .._
iuiiJU\ru\jmnnnm\ftArTi\ftiu\AnnAfuiMimiajumA.'y^- ■
.•vnTixvAnnr..
tss^— ,«l-■A-■.v^^^.,gk,
\rr\ /w\/SA/'
teristics of each to produce
Fig. S. Typical Speed-time Curvea of a Steam Mill and an Electric Mill
ELECTRIC MILL CONSIDERED FROM THE STANDPOINT OF TONNAGE 891
2700-kw. A.I.E.E. rated direct-current gener-
ators and a 50-ton flywheel, driven by a
3750-h.p. A.I.E.E. rated induction motor.
Tests of this nature of necessity require a
great deal of time, as one of the factors in-
volved is commutation and inten.-als of several
weeks and often months are required to deter-
mine with any degree of certainty the out-
come of any change.
Fig. 4 shows curves taken on the Trumbull
Mill before and after the changes were made
in the control. In order to avoid the intro-
duction of errors due to variations which
would unavoidably occur if steel were being
rolled because of the irregularities in draft,
temperature, etc., these curves were taken
without steel in the rolls.
From examination of the curves it will be
found that the time required to reverse from
90 revolutions forward to 90 revolutions
reverse was reduced from four to three
seconds, as a result of the changes which were
made in the control. Practically all of this
saving was made during the accelerating por-
tion of the cycle, approximately one second
being required to retard in either case. This
reduction in acceleration time either shortens
the time required for the mill to reach a given
speed or makes it possible for the rolls to reach
a higher speed during the pass, both of which
increase the average speed during the pass.
These results compare very favorably with
those obtainable with modern steam reversing
engines under similar conditions, and any
slight advantage which the steam engine may
have over the electric motor in the rate of
acceleration, as shown in Fig. 4, is more than
offset by delays due to slowing down and
frequent stalling as the steel enters steam
driven mills. This difference between the
steam engine and the shunt-wound mill motor
is clearly shown by the curves in Fig. 5,
which are the speed-time curv^es for the
Trumbull Mill and a modem steam mill.
Note that it is approximately one second after
the steel enters the steam mill before the en-
gine has regained the speed at which it was
running when the steel entered, while with the
electric mill there is no drop whatever, as is
shown by the Trumbull curve. It is here that
the shunt-wound differs from the compound-
wound mill motor, the latter dropping in
speed as the steel enters although not to as
great an extent as the steam engine. These
cur\^es are selected to illustrate the speed
characteristics of the two types of drives
rather than to indicate record speeds.
The developmental work which we carried
on at the Trumbull Mill, resulting in increas-
ing the rate of acceleration of the motor be-
yond the contract obligations, we feel has
fully repaid us for the time and expense
involved, because of the reduction in roll-
ing time and consequent increase in tonnage
made possible by it as evidenced by some
of the remarkable records made on this
mill. The Trumbull Steel Company has
rolled one 22 by 20 by 60-inch ingot
weighing 6700 lb. down to a 6% by
6^-inch billet in 11 passes in 57 seconds
and has rolled 57 of these ingots, 190 short
tons, to the same final section in one hour,
taking 13 passes per ingot to bring about
the reduction.
The 40-inch blooming mill at the Sparrows
Point Works of the Bethlehem Steel Com-
pany is driven by a duplicate of the equipment
driving the 36-inch Trumbull Mill, except that
the changes necessarv' to bring about the
results shown by curve "b," Fig. 4, have not
been made, in fact the control is the same as
that at the Trumbull Mill when curve "a,"
Fig. 4, was made. The Bethlehem Steel Com-
pany has rolled one 23 by 43-inch ingot weigh-
ing 16,500 lb. to a slab 9 by 38 inches in one
minute and twenty seconds and to a bloom 8
by 8 inches in two minutes and fifteen seconds,
and has rolled 330 tons of 10 by 40-inch slabs
in one hour, and 198 tons of 8 by 8-inch blooms
in one hour, and during the month of October
this mill has rolled 64,000 gross tons of ingots
to blooms and slabs.
These curves shown in Fig. 6 were taken on
the 40-inch blooming mill at Sparrows Point,
curve "a" being taken while rolling slabs 5 b}^
28-inch from a 28 by 39-inch ingot weighing
18,000 lb. and cur\'e "b" when rolling an 8
by 8-inch bloom from a 26 by 26-inch ingot
weighing 9100 lb.
These records made at the Trumbull
Plant and at the Sparrows Point Works of the
Bethlehem Steel Company exceed those made
by any other electrically driven reversing
blooming mill and put the electric mill in the
class with the steam mill from the standpoint
of tonnage. Coupling these records with the
advantages of the electric mill over the steam
mill from the standpoint of lower power costs,
lower maintenance cost, greater flexibility of
control, etc., leaves little room for argument
in favor of the steam reversing mill, and con-
vinces us that it is now as out of date as the
non-reversing steam mill has been for many
years.
S92 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
I'ig
m
m '
>l
x:;
^
11^
.^-
rH'
u
V
B
al'
» ' ,i'^
y--'^--'::^
■vf
ODD
•Ul-
■M-f
S93
Effect of Ultra-violet Rays on the Eye
Dr. C. R. Kindall, vSurgeon of the Bureau
of Mines, has issued a report in which it is
stated that 'M) men were recently viewing
the demonstration of a new portable electric
arc-welding outfit and a few hours later 17 of
the 30 men reported to the doctor for treat-
ment. They were suffering from traumatic
conjunctivitis. In two cases the pain was
very severe and the symptoms were similar
to those of iritis. Morphine had to be ad-
ministered to afford relief from pain. Only
two men of the 30 were not affected in some
way from this exposure. These two men wore
thick-lensed orange-colored glasses. Several
of the men wore orange-colored glasses with
thin lenses, but the latter were not heavy
enough to afford protection against an ex-
posure as long as took place. The distance
of the eye from the arc also influences the
possibility of injury.
Conjunctivitis is an inflammation of the
conjunctiva; the conjunctiva is the mucous
membrane covering the inside of the eyelids
and part of the eyeball. Traumatic con-
junctivitis is caused by foreign bodies in the
eye, exposure of the eyes to high winds, dust,
smoke, intense light from electric arc lamps,
and from electric welding apparatus. In the
instance mentioned above, the inflammation
was due to the ultra-violet rays. In some
cases the effect is so severe that, in addition
to conjunctivitis, an inflammation of the
skin similar to sunburn is produced.
The symptoms of conjunctivitis caused by
intense light or by the ultra-violet rays arc
abnormal intolerance to light, excessive
secretion of tears, intense smarting of the
lid, contraction of the pupil, sometimes
swelling of the lid, and small ulcers develop-
ing on the eyeball or cornea. Unless properly
treated by a physician immediately, chronic
inflammation of the conjunctiva, cornea, iris,
or retina, and possibly blindness, may result.
Under proper treatment most cases get
well in a few days. All treatments should be
under the direction of a physician. That
usually advised is to place ice packs on the
patient's eyes three or four times daily. The
pack should be left on from l.o minutes to an
hour. The eyes should be irrigated with
normal salt solution (a teaspoonful to a
quart of sterile water) or a saturated solution
of boric acid several times daily. If there
is a discharge of pus, a few drops of a 25 per
cent solution of argyrol or a 5 per cent solu-
tion of protargol should be placed in the eyes
three to six times daily. The patient should
be confined to a darkened room until his
condition improves in order to avoid com-
plications. These treatments will reduce the
swelling, give the patient comfort, and
prevent the development of chronic con-
junctivitis. In severe cases it may be neces-
sary to administer morphine to relieve the
pain.
All of the eye trouble recounted was caused
by neglecting to observe simple and well
known precautions. The glare from an
intensely bright point of light like the electric
arc, even at a distance of 20 or 30 ft., may
prove a source of injury to many eyes, al-
though at this distance all injurious ultra-
violet rays would be absorbed by the air.
The only safeguard against glare is a dark
glass, and a flashed dark ruby glass between
two pieces of emerald green glass forms a
very good combination. Blue glass should
generally be avoided, and orange-colored
glasses, unless very dark, will not sufficiently
subdue the glare of a strong arc.
At every plant where electric arc-welding
outfits are used, there should be an adequate
supply of these glasses. There should also
be on hand at the plant dispensary or hos-
pital a supply of boric acid, sterilized water,
ordinary table salt, argyrol and protargol for
immediate use. As previously mentioned,
all cases of traumatic conjunctivitis, caused
by exposure to bright light or ultra-violet
rays, should be treated under the direction
of a physician.
For more complete discussion of eye pro-
tection from injurious rays see article in
General Electric Review for December,
1918, entitled "Eye Protection in Iron Weld-
ing Operations," by W. S. Andrews.
Addenda to article," A Special Form of Phosphoroscope," by W. S. Andrews, October issue.
Through error the following paragraph was omitted from the article:
The general features of this phosphoroscope are described in a paper by Dr. Wallace Goold
Levison, published in "Annals" N. Y. Acad. Sci. XI, N. 17, pp. 401 to 403, October 13, 1898.
894 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
Automatic Substation, Sacramento Northern
Railroad*
By W. H. Evans
Electrical Engineer, Sacramento Northern Railro.\d
The electrical equipment of the portable automatic railroad substation described here is of interest not
only to railroad engineers but also to electrical engineers in all branches of the profession as the use of auto-
matic apparatus is becoming of widespread application. — Editor.
Although only the highest type of ser\'ice
can well be tolerated by any public service
compan^^ yet there must be certain definite
economical relations maintained between the
cost of ser\'ice and the results gained. If the
gain in service is not achieved economically
the ends will be defeated. If a gain in ser\-ice
may be made with economy, that advantage
will be taken. In the case in point, the Sacra-
mento Northern Railroad had been operating
its line with substations normally spaced ten
miles apart. However, between Sacramento
and the first substation north, there were
fourteen miles and as the traffic was particu-
larly heavy there the voltage conditions were
not the best. The results were slow speed
for both freight and passenger trains, and
undue heating of motors.
GENERAL FEATURES
To remedy this condition it was decided to
install at a point about 5.6 miles north of
Sacramento, a portable automatic substation
which our figures showed could be installed
for something less than $19,000, whereasfeeder
cable to produce the same voltage regulation
would have cost in excess of $40,000.
The portable substation consists brieflv of a
300-kw., 600-volt, 60-cyclc, 1200-r.p.rn., 6-
phase synchronous converter, a 240-kv-a.,
2344/445-volt, oil-insulated, self-cooled, 3-
phase transformer, together with the neces-
sary automatic control equipment. All of
this apparatus is installed in a box car con-
structed in our own shops from an SO,0()0-lb.
capacity, 40-ft. flat car, with the necessar\'
siding and roof added.
Energy is delivered to the railroad's port-
able substation at 2300 volts, 3-phase, from
the power company's OO.OOO-volt to 2300-volt
transformers, which are located on a concrete
platform and with the pole-top switches and
fuses are enclosed by a high wire fence for
protection against trespassers.
Since the substation is automatic, normally
the doors are always closed and locked, and in
order to provide ventilation, louvers were let
♦ Reprinted with changes from Journal of Electricity.
into both sides and ends of the car; in addition
screened openings are placed in the floor of
the car for further ventilation.
METHOD OF OPERATION
It may be of interest to give a short outline
of the sequence of operations which takes
place in automatically starting up and shutting
down.
Starting
With the station shut down, and a train
coming into the substation zone on either side,
the third rail voltage is gradually lowered until
it reaches the value at which the relays in the
automatic substation are set to govern
starting.
Relay 1 (Fig. 1) is a contact-making volt-
meter which is adjustable for any particular
trolley voltage desired, in this case 500 volts.
In connection with the underload relay 37,
which functions to shut down the station,
these two relays are the primary- control in
starting and stopping the station.
Relay 1 closes instantaneously when the
voltage drops to 500 volts, short circuiting the
coil of relay 2 which is a time-limit circuit-
opening relay whose function is to provide a
time delay in starting up the equipment as the
low voltage conditions appear, because obvi-
ously a momentary- swing below 500 volts
should not be permitted to start up the
station; the time setting of relay 2 can be
adjusted to suit the particular conditions at
any point.
The contacts of relay 2 arc normally closed
when the station is not rvmning. When relav
1 operates, closing its contacts, relay 2, after a
predetermined time, opens its contacts and
permits the coil of relay 3 to be energized.
Relay 3 then closes its contacts, and control
current from the 5-kw. auxiliary-control trans-
former is admitted to the control circuits of
the station. A circuit is then established from
the alternating-current control bus through
the contacts of relay 27-X, the contacts of
relays 3 and 2(5, operating coil of relay 4,
auxiliary switch on circuit breaker and hand-
AUTOMATIC SUBSTATION, SACRAMENTO NORTHERN RAILROAD 895
896 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 11
Fig. 2. Converter in Background is Equipped with Flash Bar-
riers and Motor Operated Brush Lifters. Automatic Control
Board to the Right and Series Resistances to the Left
Fig. 3.
Transfurmer End of Car Showing Arrangement of
Louvers to Assist Rapid Cooling
Fig. 4. Starting Grids, Running Contactors and Motor
Operated Drum Controller
Fig. 5. Near View of 240-kv-a. Starting Transformer. Motor
Operated Oil Breaker and 5-kw. Control Transformer
AUTOMATIC SUBSTATION, SACRAMENTO NORTHERN RAILROAD
SO 7
reset switch on oil-switch motor mechanism
back to control bus. The closing of con-
tactor 4 establishes a circuit from the control
bus throujjh one of its contacts to segment 13
on the controller, then to segment Hi upper
contact of auxiliary switch on brush-raising
device, and to the operating coil of contactor
() and back to the control bus.
Operation of Controller
Contactor (i closes and starts the motor
driving the controller which is very similar to
the ordinary type K street-car controller.
Through its various contact fingers and seg-
ments the controller establishes the necessary
sequence of operation of the various switches
for starting up and shutting down the station,
each succeeding step, however, being checked
electrically by means of various relays to
insure that the electrical and mechanical con-
ditions have been properly fulfilled.
ad\-ances beyond segment 15. Segment 2 on
the controller then makes contact, completing
a circuit through the contacts of the relay 32
and the operating coil of contactor 10.
Start of Converter
The starting contactor 10 now closes, plac-
ing reduced voltage from the transformer
upon the slip rings of the converter, which
starts. If the converter has come up to syn-
chronous speed by the time the first gap in
segment 16 is reached, a circuit is established
from segment 14 through the contacts of 13
to segment 20 and thence to segment IS and
the operating coil of contactor G. This holds
contactor 6 closed until the gap in segment 16
is passed. However, if the converter has
not come up to speed by the time the gap in
segment 16 is reached the circuit to the
operating coil of contactor 6 is broken and the
controller now comes to rest until svnchro-
M^^'^tmw
E
I
Fig. 6. The Automatic Railway Substation and Its Outdoor Transformer Installation
Segment 1.") on the controller closes the
operating coil of contactor 5 which establishes
a circuit through one of its contacts to seg-
ment 1 on the controller and simultaneously
completes a circuit from the same contact to
the closing circuit of the oil-switch motor
mechanism.
The oil switch now closes, energizing the
power transformer, and if the proper alternat-
ing-current voltage exists on all three phases of
the low tension side relays 32 close. These re-
lays are so connected that no further operation
can continue unless the proper phase voltage
exists. Segment 14 on the controller then
makes contact, completing a circuit through
the auxiliary switch on the oil circuit breaker,
one of the contacts and the operating coil of 5.
This operation thus establishes a holding
circuit for contactor 5 as soon as the controller
nous speed on the converter is reached, i.e.,
until the speed control switch 13 has closed its
contacts.
Segment 3 makes contact, closing the cir-
cuit to the operating coil of field contactor 31.
This closes and connects the fields of the con-
verter to the 2.50-volt exciter on the controller,
thus fixing the proper polarity on the con-
verter, and as the converter is brought to the
proper polarity, the polarized relay 36 closes
its contacts. Segment 3 then breaks contact,
opening contactor 31.
Segment 4 makes contact, energizing the
operating coil of full-field contactor 14, which
closes and places the field of the converter
across its own armature for self-excitation.
The field contactors 31 and 14 are mechani-
cally interlocked so that 31 must open before
14 can close.
898 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 11
Running Conditions
Segment 2 breaks contact, opening the
starting contactor 10 and segment 5 makes
contact, energizing the operating coil of run-
ning contactor 16 which closes and puts
full alternating-current voltage on the slip
Fig. 7. 60,000/2300-voIt Transformers Provided with External
Separate Pipes for Natural Cooling of Oi). They are
an Unusual Design
rings of the converter. At the same time relay
30 closes due to the establishment of full volt-
age across the armature of the converter.
Segment 2G makes contact, establishing a
circuit through the upper contacts of the limit
switch on the brush-raising device and starts
the motor of this device, thus lowering
the brushes upon the converter. If the
brushes reach their lowest position, and the
lower contact of the auxiliary switch on the
brush-raising device is closed before the
controller runs off the second gap in segment
16, a circuit is established from segment 17
through thelowerauxiliar>- switch of the brush-
raising device to the operating coil of con-
tactor 6, thus holding 6 closed and permitting
the controller to continue to revolve. If the
controller runs off segment 16 before the
brushes are in their lowest position the
operating coil circuit of 6 is opened and the
controller stops until the lower auxiliar>'
switch on the brush-raising device closes and
completes the circuit from segment 17 de-
scribed above. These steps insure that the
brushes have been properly lowered upon
the converter.
Segment 7 makes contact, giving direct-cur-
rent potential to segments S, 9, 10 and 11.
Segment S makes contact, establishing a cir-
cuit through the contacts of polarized relay
36, the contacts of relay 30, the electrical
interlock on contactor 16 and the operating
coil of contactor IS. Contactor IS now closes,
connecting the converter to the bus through
all three sections of the load limiting resist-
ance. The converter is then feeding the line
through the total series resistance.
Segment 9 makes contact, establishing a
circuit through the operating coil of con-
tactor 21 and the contacts of relay 25 and, if
the current demand is below the overload set-
ting of relay 25, contactor 21 closes, short
circuiting section R-3 to R-4 of the resistor.
Taking Load
Segment 10 makes contact, establishing a
circuit through the operating coil of contactor
20 and the contacts of relay 24 and, if the
current value is below the setting of relay 24,
relay 20 closes, short circuiting the section R-2
to R-3 of the resistor. In a similar manner,
segment 11, making contact, closes a circuit
through the operating coil of 19 and the con-
tacts of relay 23, thus cutting out the last
section of resistance. The machine is now
connected directly to the bus and delivering
load. During the last several operations men-
tioned above after contactor IS closed, the
contacts of relay 37 open as the converter
picks up load, inserting the section BC of the
resistancein series with thecontact-making volt-
meter 1 . Simultaneously thcvoltageonthe bus
has been broughtup tonormal but thecontacts
of the voltmeter still remain closed due to the
resistance BC which has iust been inserted.
Segment 17 breaks contact, opening the cir-
cuit previously established through the lower
contacts of the brush-raising device and the
operating coil of contactor 6, the latter
opens and the controller comes to rest at the
running position, being stopped immediately
by its solenoid brake.
Shutting Down
When the load demand decreases and
reaches the setting of relay 37, which in this
case is adjusted for 100 amperes, the contacts
of the latter close, short circuiting section BC
of the resistance in the coil of the contact mak-
ing voltmeter 1, causing the voltmeter to open
its contacts. This removes the short circuit
from coil of relay 2, closing its contacts instan-
taneously and short circuiting coil of relay 3
which starts to open its contacts. If the load
does not increase long enough for 2 to reset
at any time during the setting of the dash-pot
on relay 3, the lattcr's contacts open, inter-
rupting the circuit of contactor 4. Should the
AUTOMATIC SUBSTATION, SACRAMENTO NORTHERN RAILROAD 899
load increase before contacts 3 have opened,
37 would open, inserting resistance section BC
and causing the voltmeter to make contact.
The voltmeter contacts short circuit coil of
relay 2. Contacts of relay 2 open after time
delay and re-energize 3.
After 3 has opened, contactor 4 opens, inter-
rupting two circuits simultaneously; the first
being the alternating-current supply to con-
troller segment 13 and the other direct-current
circuit including the operating coil of contactor
18. The holding circuit for contactor 5 through
segment 14, the auxiliary switch on the oil
circuit breaker and the contacts of 29 are
broken and line contactor IS and control con-
tactor 5 now open.
The opening of contactor 5 interrupts the
supply to segment 1 on the controller and
establishes a circuit through its electrical inter-
lock to segment 19. Contactors Ki and 14
open, disconnecting the converter from the
transformer and discharging its field which in
turn drops relay 30 out.
The operating coil of contactor G is then
energized through the electrical interlock on
contactor 5 and segments 19 and 18. The
controller motor starts and contactors 19, 20
and 21 open. Segment 24 makes contact,
energizing the trip circuit of the oil switch
mechanism; also segment 25 makes contact
through the lower limit switch on the brush-
raising device. The high tension line is now
disconnected from the transformer de-energiz-
ing relays 32 which open, and the brushes are
raised from the commutator. Segments 18
and 19 break contact, and the controller comes
to rest at the off position. In the meantime
the motor of the brush-raising device con-
tinues to operate until reaching the end of its
travel when the lower limit switch is opened,
breaking the supply to the motor. As the
voltage on the converter armature dies down
after contactors 14 and Ki are open, relay 30
also opens, and the station is completely shut
down.
Protective Features
Direct-current Overload. — Relays 23, 24 and
25 are calibrated at alternating-current loads
corresponding to direct-current loads of 900,
1 200 and 1 500 amperes and upon reaching these
successive loads the series resistance of 0.15
ohms, 0.25 ohms and 0.35 ohms are inserted
in circuit with the converter, causing a reduc-
tion in the trolley voltage supplied to the third
rail and consequently reducing the ampere
output of the machine.
Alternating-current Overload. — ^Should trou-
ble develop on the direct current side of
the converter inside the connection of the'
load limiting resistance, relays 2(1 are ener-
gized from the current transformer on the low
tension side and will open after a set time
and shut down the equipment. Relays 26 are
set at a higher value than relays 23, 24 and 25
and are also time-limit opening. This time-
limit feature allows momentary swings to
occur without shutting down the machine.
In our case these relays, 25, 24 and 23, are
instantaneous circuit opening and time-limit
circuit closing, being adjusted to close at 3
seconds and at 10 seconds after the current has
fallen to a certain value for each relay. This
time delay permits of the acceleration at a low
voltage of heavy trains which when starting
up cause the resistance to come in ; and when
the trains have accelerated and the current
demand fallen off, the time setting permits of
their receiving full voltage at the end of their
accelerated period.
Additional alternating current protection is
provided by relay 28 which is energized from a
current transformer in the high tension wind-
ing and is set considerably higher than the
other overload devices. When this relay
operates, the coil circuit breaker is tripped
open and with it the hand-reset switch, thus
completely shutting down the station. The
opening of the hand-reset switch interrupts
the coil circuit of contactor 4 and simulta-
neously with it the opening of the auxiliary
switch on the oil circuit breaker interrupts
the holding circuit of contactor 5. The
operation of either of these devices shuts
down the equipment. After the oil circuit
breaker has been tripped in the above manner
and the hand-reset switch opened, the station
will not start up again until the hand-reset
switch is closed by the inspector. Con-
sequently relays 28 are set very high and are
expected to operate only in cases of severe
trouble where the attention of an inspector
would be necessary.
Low Voltage. — Relay 27 provides the alter-
nating-current low voltage protection. When
low voltage occurs, the left hand contacts
of 27 are closed, short circuiting the coil of
27-X, opening it and interrupting the supply
through the contacts of relay 3 to the coil of
contactor 4. Relay 29, in a certain sense, per-
fonns the functions of an alternating-current
low-voltage relay whenever the converter is
running, since, should the alternating-current
voltage fall too much, the converter would
900 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
invert and supply power from the trolley to
the alternating-current system. Reverse-
current relay 29 would then open, inter-
rupting the holding circuit of contactor 5,
shutting down the machine.
Over Speed. — Speed limit device 12 on over
speed closes the circuit of the shunt trip of the
direct-current circuit breaker. When this
open^, the auxiliary switch on the circuit
breaker interrupts the supply to the coil of
contactor 4 and the equipment shuts down.
When this happens it must be hand reset by
the inspector.
Under Speed. — -The speed control switch 13
is a centrifugal device, the contacts of which
remain open until approximate synchronism
is reached.
Sequence. — -The sequence of events is fixed
primarily by the controller but in addition to
this there are electrical interlocks on con-
tactors 10 and 16, as well as the holding circuit
of contactor 5, all of which are additional safe-
guards against incorrect sequence.
Polarity. — The 250-volt excitation gen-
rator, direct-connected to the motor of the
controller, fixes the polarity of the converter,
but as an additional precaution the polarized
relay 36 must be energized in the proper
direction before allowing line-contactor IS to
close.
Temperature. — -Should the load-limiting re-
sistance or bearings overheat the thermostats
will open, de-energizing relay 27 which, when
de-energized, closes the left hand contacts
of 27, thus shutting down the equipment.
The thermostats over the resistor are self-
resetting when the resistor cools off, while
those on the bearings of the converter are hand
reset and require the attention of the inspector
before the converter will again start.
A thermal relay is provided whose rise in
temperature is proportional to the heating in
the converter winding and in case of a long
continued overload which would injure the
insulation, this relay operates and shuts down
the station. This relay in one of the
illustrations is shown mounted on a small
panel sui)ported on the resistor grid iron
framework.
The thermal element is the fuse-like object
connected in one phase of the converter trans-
former secondary. A small tube containing
a volatile liquid connects from the thermal
element to the relay on the right — ^the expan-
sion of the liquid under heat actuating the
relay whose contacts, when opened, interrupt
the circuit to relay 27 and shut down the
station .
Balanced Polyphase Voltage. — This protec-
tion is provided on the low tension side of the
power transformer by means of the two relays
32 which are connected across different
phases. All three phases of the power trans-
former must be excited to approximately
normal voltage, otherwise one or both of
these relays will remain open and prevent the
starting contactor 10 from closing.
Position oj Converter Brushes. — ^Proper posi-
tion of these brushes is assured by means of the
auxiliary- switch on the brush-raising device.
The con\'erter is equipped with flash
barriers which completely surround the brush
holders and in case of an attempted flash-over
between brushes the hot metallic vapors are
scooped up from the commutator and dis-
sipated in two sections of wire mesh. The
system is subjected to frequent short circuits
between the third rail and traffic rail, owing to
section -men dropping tamping bars across the
conductor rail, and from various other causes,
so that flashovers on the commutator of the
motor-generator sets have been quite fre-
quent and severe. These flash-overs were
usually accompanied by a spill-over to the
pedestals of the machines and it has been
found necessar\- to remove the grounds from
the machine frames in order to reduce this
spilling-over. There has been, during five
months' operation of the automatic sub-
station, what was evidently a severe short
circuit on the third rail in the immediate
vicinity of the substation ; the flash barriers no
doubt took care of the resulting flash-over at
the commutator, some of the flash-screen
metal having been vaporized, but the sub-
station cleared itself and when the writer
visited the station that day the machine was
carrying 50 per cent overload without any
evidence of the flash-over having incon-
venienced the converter as far as normal
operation was concerned. It was evident,
however, that a spill-over had taken place to
the pedestal of the machine. These spill-overs
are believed to be due to the inductive kick
occasioned by the sudden extreme variation in
current in the third rail, the magnetic effect of
which accentuates the short circuit con-
ditions on the machine commutators. In
addition to removing the ground from the
machine frame an electrolytic lightning ar-
rester has been connected between the posi-
tive of the machine and the traftic rail, with
the belief that the arrester will take care of
any extreme inductive kick occurring across
the converter armature, the fields, or the
series resistors of the machine.
AUTOMATIC SUBSTATION, SACRAMENTO NORTHERN RAILROAD 901
GENERAL RESULTS OF OPERATION
Our experience so far with the automatic
control seems to show that this type of equip-
ment is particularly advantageous for inter-
urban service. The cushion of resistance
which is introduced in extremely heavy
demands results in much better operation
than the manually' operated stations, in that
improper handling by a motorman of his
train, or in case of two or more trains pulling
on a station, does not result in opening the
station breaker, with the resultant slowing
down of trains and probability of again pull-
ing the breaker when the station operator
closes his switches. With automatic control,
there is no breaker to open. The station
simply cuts in the proper resistance which
should have been cut in on the train by the
motorman if he had handled his train prop-
erly. The voltage to the train is thereby cut
down and a lower current demand follows;
but in the meantime the train continues to
accelerate under this reduced current and in a
few seconds the amperage falls to a value
which allows the resistance contactors to
again close, short circuiting the resistors and
delivering full voltage to the trains.
This method of operation naturally results
in better conditions as regards flashing at
commutators of car equipment due to poor
handling of trains, as the station resistance
automatically takes care of any such defective
train operation. For those interurban lines
which operate heavy freight trains the auto-
matic control, with the current-limiting resis-
tors and particularly in combination with a
200 per cent overload characteristic in the
converter or motor generator set, is partic-
ularly fitted for handling this class of service.
IMPROVED OPERATION
In addition to the large saving in operators'
wages which the automatic control gives, it
also provides a considerable saving in eliminat-
ing idle running of a substation with its
attendant running-light losses. Substation
operators are instructed to cut in or off the
line either at defined time intervals or upon
certain current and voltage indications upon
their station instruments, but we are aware
that even under these regulations there is a
very considerable am.ount of idle running.
Under automatic control, however, running-
light losses are cut to a minimum as the
station does not start except upon a predeter-
mined demand for power and then shuts
down when this demand no longer exists.
The greater the interval between trains, the
larger will be the saving of energy obtained
through the elimination of running-light
losses. The use of automatic control there-
fore reduces both of the predominant items
in the total cost of power, i.e., the energy
charge itself and the item of substation I
wages. Our equipment is adjusted so that
approximately three minutes after the demand
for power falls below 100 ami:)cres, the station
shuts down, this three-minute inter\'al in our
case being sufficient to take care of the time
consumed by a train in the substation zone,
coasting, braking, and stopping. The station
delivers current to the line thirty seconds after
relay 3 closes, or about thirty-five seconds
after the demand for current occurs, there
being about a five-seconds delay in the action
of the relay 2 to provide against momentary
swings bringing the station into action.
TROUBLE EXPERIENCED
This station will be regularly inspected at
intervals of about every four or five days, this
being done at present by an extra operator
who also spends part of his time in line work
and affording relief to other station operators.
To date the equipment has been remarkably
free from trouble, our main difficulty having
been loose contacts at terminals of relays
which had not been thoroughly tightened up
and were shaken loose by the vibration of the
car. These only resulted in shutting down
the station, and since going over all these
contacts thoroughly there has been no further
difficulty.
In addition to this imit the railroad has
recently ordered another similar equipment
to be installed at another point on the system
where present substation spacing is also too
great and voltage conditions poor. This addi-
tional equipment includes a converter with
high reluctance poles which is expected to be
practically free from all flash-overs incident to
shorts on third rail. Cur\'e drawing meters
will be provided to give us a record of what
is taking place in the station. In addition
the relays 2.3, 24 and 25 will be controlled
from direct-current shunts instead of from
alternating-current transformers, thus pro-
viding an easier means of adjusting the relay
settings.
The company has in operation nine manu-
ally operated stations, in four of which the
apparatus is located in a building, and in the
other five in a portable structure similar to the
automatic equipment. It is planned to pro-
vide all of these nine stations with automatic
control and probably take advantage of the
902 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
portable nature of five of them to shift their
relative locations so as to provide better
regulation over the system.
This unit was completely installed at a cost
of approximately $18,500 including the electri-
cal apparatus, the car in which it is installed,
the protection fence, concrete platform for
power transformers, and spur track on which
the car is mounted. This cost does not include
the cost of the 60,000 2300-voIt transformers
and open-air type switching equipment which
the power company provided.
The electrical apparatus used in this auto-
matic substation was designed and manu-
factured by the General Electric Companv of
Schenectady, New York.
Automatic Substations for Alternating-current
Railway Signal Power Supply
Part I
By H. M. Jacobs
Railway Department, General Electric Company
This article is the first of two which describe the application of automatically controlled equipment to
railway alternatmg-current signal substations for the purpose of maintaining an effectively continuous supply
ot power to the signals. The maximum interruption which can occur in case the regular power suoplv fails
vanes with the type of system employed but at most it is a matter of only a few seconds before the reserve
supply IS switched into operation. The present article deals with that type of substation in which the reserve
power supply is a secondary commercial source, either nearby or distant, and the concluding article will treat
ot the equipment used in the substation in which a storage battery and conyerting apparatus constitute the
reserve supply. — Editor. o ft-
So much publicity has been given to
automatic substations for electric railways
that one is apt incorrectly to consider all
automatic substations, especially for rail-
way service, of this class. An' automatic
railway substation functions with the direct-
current load demand; an automatic substation
for alternating-current railway signaHng pro-
vides against outages due to failure of power
supply.
The rapid and efficient handling of railroad
traffic depends in a great measure on the
dependability of the signaling system. The
first step in determining the responsibility for
an accident is an investigation of the condi-
tion of the signals. It is therefore of extreme
importance that the power for operating the
signals be as free from interruptions as pos-
sible.
Primary or storage batteries furnish the
power for signals operating on direct current.
The former operate over long periods with-
out renewals; storage batteries of the proper
ampere-hour capacity will operate several
days without recharging. As long as these
batteries are properly maintained there is
very little danger of signal failures due to
failure of power supply.
The continuity of service in alternating-
current signaling depends directly on the
continuity of the source of supply. Even a
temporary failure is liable to cause serious
trouble. When the engineer on the "fiver"
sees a signal suddenly go to "danger" with-
out having received a "caution" indication
from the preceding signal he does not know
the cause. There may be an obstruction,
a washout, a broken rail, or a stalled train in
the block ahead. It is his duty to stop his
train as quickly as possible; and in doing so he
causes discomfiture among the passengers and
may flatten every wheel on his train. These
and many other considerations make it im-
perative that ever\- precaution be taken to
maintain either continuous power or a means
for cutting in a reserve source in the shortest
possible time. The automatic substation was
developed to meet this latter condition and
eliminates serious time delays due to hand
switching, especially at night when certain
attendants are off duty.
There are two general classes of automatic
substations: one in which the reser\-e power is
a second commercial source, and the other in
which the reser\-e power is derived from a
storage battery through converting power ap-
paratus. The first class is again subdivided
according to whether the reserv-e source is at
the same location or at some distance from
the preferred source; the principle of opera-
AUTOMATIC SUBSTATIONS FOR A-C. RAILWAY SIGNAL POWER SUPPLY 903
tion is the same in either case. Where both
sources are at the same location, either one
may feed the bus to which the feeders are
connected. Where the sources are at dif-
ferent locations, either one may supply a
transmission line to which the feeders are
connected; in other words, the bus becomes
a transmission line.
The second general classification of auto-
matic substations, that in which the reserve
power is obtained through apparatus from
a storage battery, is likewise subdivided. In
one case the apparatus floats continuously
on the storage battery, and in the other case
the apparatus remains idle until a power
failure occurs. In the first case there is no
interruption in service when failure occurs,
but in the second case there is an interruption
for the period of time required for the appa-
ratus to start and switch in. Traffic condi-
tions determine which system is required.
Only the first general classification of auto-
matic substations will be treated in this
article.
The ordinary transmission line carries the
bulk of the transmitted power throughout
its whole length. A transmission line for sup-
plying power to railway signals is dift'erent.
The power is distributed in approximately
equal increments to every signal location, and
to every "interlocking" or signal tower con-
trolling power-operated, hand-controlled
switches, signals, and signal devices through-
out its length. To eliminate interruptions
due to prolonged failure of power supply,
provision is made for supplying the line from
either end. With power available at both
ends of a line, prolonged outages due to fail-
ure of the main source are eliminated because
the line may be connected to the source of
supply at the other end. Of course with
manually operated substations, the signaling
system is " tied up " and all trains are delayed
until this connection is made. The greatest
delays usually occur at night, and instances
are on record where it was necessary to get a
man out of bed to switch in the auxiliary
source, causing a delay of over half an hour.
If automatic switching equipment is added
to the regular equipment of a manually oper-
ated substation, the delays due to power
failure may be eliminated. This automatic
equipment consists of a magnetically oper-
ated switch, energized from a potential trans-
former connected to the local source; a low-
voltage relay to connect the switch to the
transformer, the relay being energized from a
potential transformer connected to the trans-
mission line side of the magnetic switch;
indicating lamps; and auxiliary switches
mechanically operated from the magnetic
switch to seal the switch closed as long as
power is available from the local source, and
to connect the indicating lamps.
t— I I Lightning
hokeCoilsfl i 1 I] Arrester
l3ion FysCB -J S ~^
Choke C
Expub
i\»rJS\ '•' i :*
Current Transfbrmer
a Oil Circuit Breaker
"" Hand Operated
arvfmpersClcswJ
When Contactor 19 Closed
Potentlol , ,
Transformer CwwC^D.
mportont.-
Piace no Fuaes in
Control Circuits
Choke Coils
Expulsron Fuses — '
,*- Power Transformer
Fig. 1.
wiring Diagram for a Typical Single-circuit,
Single- phase Automatic Substation
Fig. 1 is a simple diagram of a complete
single-phase automatic substation taking
power at 2200 volts and delivering it at a
higher voltage. The source and the delivery
may be any voltage ; it is best from the stand-
point of safety and economy to have the con-
trol equipment in the low-voltage side. If
it is desired to transmit at the same voltage
as the source of supply, no power transformer
is necessary; if it is desired to transmit at a
lower voltage than the source of supply, the
transformer is connected ahead of the hand-
operated oil circuit breaker.
The principle of operation is as follows:
There are two duplicate stations, ^4 and B,
connected to the opposite ends of a trans-
mission line. Assume the power is supplied
from station B. All the equipment connected
to the line side of the magnetically operated
contactor switch in station .4 is thus energized
from station B. The low-voltage relay is
energized and contacts are open so that the
closing coil of the contactor switch is de-
energized. The equipment connected to the
904 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
supply side of the contactor in station A is
energized from the local supply. The green
lamps are lighted indicating "local power
available. "
When the power supply at B fails, the
following action takes place in station A:
The low-voltage relay is de-energized and
the contacts close; this connects the closing
coil of the contactor across the 220-volt
winding of the operating transformer. The
contactor immediately closes and restores
energy to the transmission line. The con-
tactor has two auxiliary finger contacts
mechanically connected to the moving ele-
ment, so that when the contactor is closed
they make contact. One of these bridges the
contacts of the low-voltage relay so that it
will not open the circuit to the closing (now
holding) coil of the contactor when the relay
is energized from the local -source through
the contactor. The other finger contact
closes the circuit to the red lamps. With both
green and red lamps illuminated, the indi-
cation is "transmission line is being supplied
from local source."
SVtoy
Ainmct«r
Important.—
PloonoFuMSIn
ContnlClixulu
1 Arresters
Fig. 2. wiring Diagram for a Typical Single-circuit,
Three-phase Automatic Substation
If it is desired to transfer the load back to
station 5, it is only necessary to open the hand-
operated oil circuit breaker. This simulates
a failure of power in station A , and station B
will immediately cut in as above described.
Should a short circuit or overload occur
on the transmission line, the oil circuit
breaker will be tripped out by the action of
the inverse time-limit overload relay. This
of course will de-energize the line and the
other station will be cut in on the short cir-
cuit; the oil circuit breaker there will open.
The line will then remain dead until the
fault is located, the line sectionalizing switches
adjacent to the fault opened, and the hand-
operated oil circuit breakers closed in both
stations. The unaffected portions of the line
will then be energized from the adjacent
station (see Fig. 10).
A three-phase automatic substation is
more complicated. The reserve station must
not cut in when only one phase fails, because
the two unsynchronized sources would be
connected together single-phase. Further-
more, a station must not cut in unless power
is available on all three phases. If it should
cut in single-phase, only one third of the
feeders connected to the transmission line
would be energized. To take care of these
conditions, two low-voltage relays energized
from separate phases are connected to the
transmission line side of the contactor; the
closing coil circuit is connected to the relay
contacts in series, so that both relays must
be de-energized before the closing coil of the
contactor can be energized. The closing coil
circuit is also carried through the contacts of
another relay which is energized from a phase
other than that from which the closing coil is
energized. The contacts on this relay are
arranged differently from those on the other
two relays in that they are closed only when
the relay is energized. Hence, for the closing
coil to become energized the latter relay must
be energized and the two former de-energized.
The arrangement for shunting the contacts
of the two low-voltage relays, and for con-
necting the red indicating lamps is the same
as for a single-phase station. Provision is
made for reading current and voltage on all
three phases, and two inverse time-limit over-
load relays provide polyphase protection.
Fig. 2 is a complete wiring diagram of a three-
phase station.
For the proper functioning of all automatic
substations, it is imperative that no fuses be
placed in the control circuits, especially those
including the relays that cause a station or
supply source to cut in.
When a transmission line is supplied from
one end only, the secondan.- voltages of the
various feeders may be kept fairly uniform
by the use of taps on the transformers sup-
AUTOMATIC SUBSTATIONS FOR A-C. RAILWAY SIGNAL POWER SUPPLY 905
plying them. This permits of quite a heavy
line drop. However, a transmission line ar-
ranged for supply from either end must have
a very small line drop and the feeders must
not be connected to taps on their individual
transformers. To make this requirement
clear, assume a line with 10 per cent drop
supplied from a source at either end 10 per
cent above normal. The feeder adjacent to
the preferred source will be connected to a
minus 10 per cent tap on the transformer.
The last feeder on the line will be connected
across the full winding of its supply trans-
former. The voltage of these two feeders
will then be normal. Now suppose the trans-
mission line is supplied from the other end.
The voltage on the feeder adjacent thereto
is now 10 per cent above normal. The volt-
age at the other end of the line is then normal,
but the signal apparatus, being connected to
the minus 10 per cent tap on the transformer,
is operating at a voltage 10 per cent below
normal. The difference in voltage on the
signal apparatus at the ends of the line is
therefore 20 per cent, whereas, if the feeders
had not been connected to taps on the trans-
formers the dffierence would only have been
10 per cent.
As has already been pointed out, a short-
circuit on one section of the line will tie up
traffic the length of the line until the faulty
section is located and cut out. Naturally
the shorter the line the smaller is the zone
of disturbance resulting from line trouble.
Hence, from the standpoint of lessened lia-
bility to disturbance from line troubles, and
more uniform voltage on signal apparatus
">"LSM
r
■am
y
Fig. 3. Automatic Substation for Alternating-current Railway
Signaling, Illinois Central Railroad, Chicago, 111.
Fig. 4. General Interior View Automatic Substation
for Alternating-current Railway Signaling,
Illinois Central Railroad. Chicago, 111.
Fig. 5. Interior View Back of Switchboard Automatic
Substation for Alternating-current Railway Signaling,
Illinois Central Railroad, Chicago, 111.
906 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
due to reduced line drop, a short transmission
line between two available sources of power
is advisable.
Generally speaking, substations should be
approximately 30 miles apart. The distance
will depend principally upon the location
ChoKeCoMs -
Red Lamps
RLIgWtning H I
ArrestersLJ |
CnoKe Coils
Inverse Time Limit
Overload Relay
Trip Coi
Contactor shown
n open position
Additionoi Feeders
— OS desired
Oil Circuit Breoker
hand Operated
-4."C NoL« - Mo Po*er Transformers
required wnen Feeders.ond
A s, I' I Ligntning sources are some voltage
Choke Coil3.»i g LJ Arresters if Source 13 above 4400 wits.
ano Feeders are lower voiwje
Expulsion Fuses*/ / ^ place transfonrwrset Aano B
IF Feeders are Higher voltage
than eitHer source. place mna-
Feeder formers at c
IF Sources one not soma vol-
tage ploce trorsrormera st A
or B to mahe control at Some
voltage
Fig. 6. Wiring Diagram for Single-phase, Automatic Substation.
Two Sources and One or More Feeders in
One Station. Push Button Control
of reliable sources of commercial power sup-
ply and upon the load. Four-track signaling
requires more power per block than single
or double-track signaling. In some instances
it is necessary to transmit lOU miles or more
because the power sources at intermediate
points are not reliable. With automatic sub-
stations installed, power may be taken from
some of these more or less unreliable points
because, when they fail, another substation
picks up the load immediately.
This brings up the point of the "interme-
diate" substation, or what may more prop-
erly be termed a "two-circuit" substation.
Ever>' substation except those at the extreme
ends are two-circuit substations. They feed
power when called upon to either or both
adjacent sections of the line. The equip-
ment per circuit is essentially the same as
for the single-circuit station described. The
switchboard panel is wide enough to accom-
modate the api^aratus for both circuits. Only
one voltmeter is necessary. Of course the
protective equipment for the incoming line
is the same as for a single-circuit substation.
It is advisable, however, to provide a set of
disconnecting switches in each circuit be-
tween the switchboard and the incoming line.
The equipment of either single or two-cir-
cuit substations occupies comparatively small
space. The power transformers and much
of the protective and disconnecting equip-
ment may be placed on poles or on a platform
outside the building. The remainder of the
equipment may be easily placed in an 8 by
lO-ft. concrete house as shown in Fig. 3- The
interior arrangement of this station is shown
in Figs. 4 and 5. The automatic contactor
stands at the side of the switchboard. The
current and potential transformers are sup-
ported on the switchboard framework and
wall braces. The operating transformer is
on the rear wall near the ceiling. Some of
the protective equipment is placed inside the
building and is mounted on the rear wall. In
this station power is received and transmitted
at the same voltage (4400). Since this in-
stallation was placed in ser\-ice a much smaller
contactor has been developed for load^ up to
30 amperes at 4400 volts. This is approxi-
mately 200 kv-a. three-phase; or at 2200
volts three-phase it can carr>' a 100-kv-a.
load. For a three-phase transmission line
50 miles long supplying signal lighting and
Emergency Soorce
1 Source
^ Potentlot Transformer
Potential | [ || control Relay
Tronatformer
Raiiwoy Signal FeeOer
Fig. 7. Partial Wiring Diagram i'otherwiae same as Fig. 6^ for
Single-phase Automatic Substation. Two Sources and One
or More Feeders in One Station. Relay Control Gives
Preference to One Source
some small station lighting, "lO-kv-a. is con-
sidered a good load, so that this new device
should meet practically all conditions for
4400 volts and below. It is so small that it
can be mounted on a pipe framework back
of the switchboard.
AUTOMATIC SUBSTATIONS FOR A-C. RAILWAY SIGNAL POWER SUPPLY 907
It is advisable, though not absolutely nec-
essary, to have a set of disconnecting switches
on each incoming and outgoing circuit inside
the station, so that the station may be isolated
for inspection and repairs without the neces-
sity of going up on the pole or platform to pull
the fuses.
Up to this point the discussion has been
on the method of keeping a transmission line
energized from two sources of supply some
distance apart. In some cases energy can be
obtained at only one location, or there may
be certain territory having congested traffic
where it is important to have energy always
available. In such cases it is advisable to
have a second source of power, if it can be
obtained, and a means of quickly changing
from one to the other. This can be accom-
plished by automatic equipment essentially
the same as two automatic substations com-
bined into one equipment. The common bus
bears the same relation to this as the transmis-
sion line did in the arrangement just described.
Fig. G is the wiring diagram of a simple single-
phase equipment. There are no hand-oper-
ated oil circuit breakers between the contac-
tors and the expulsion fuses. Normally closed
push-button switches for opening the circuit
of the holding coils afford a means of opening
the contactors when it is desired to change the
operation from one source to the other. It is
recommended that disconnecting switches be
placed between each contactor and bus to
afford greater safety to anyone who wishes to
inspect or repair a contactor.
It may be that one source is preferable to
the other, and that the feeders should be con-
nected to the emergency source only when
there is no power available on the preferred.
This can be taken care of by one "preferen-
tial" relay in place of the two push-button
PrererenLiol Source
JJlfit
Emergency Source
jm
-AAAA/«-i
JL
PC'
_l/ /
Preferential
Relay
Feeder
Fig. 9. Wiring Diagram of Simple Automatic Substation Using
4400-volt Contactors, Low-voltage Relay and
Potential Transformer
switches. The relay magnet is connected to
the preferred source, and when energized the
relay serves to connect the contactor on that
source; when de-energized it cuts off this
contactor and connects the contactor on the
emergency source. Fig. 7 shows the wiring
for this feature.
A much simpler form of automatic sub-
station, but working on the same plan as that
last described, consists of two contactors
recently developed for use as an outdoor
Fig. 8. 4400-volt, 30-amp. Contactor. Holding Coil Takes About
1 to IJj Amperes at 110 Volts
90S November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 11
remote control switch, together with a control
relay and potential transformer. The devices
are so small and compact that if the relay and
transformer are placed in a waterproof hous-
ing the whole equipment may be mounted on
one pole out of doors. The external appear-
ance of the contactor and arrangement of the
contacts is shown in Fig. 8. This is rated 30
amperes maximum for any voltage up to
and including 4400, three-phase. The upper
part of the case contains the operating mag-
net and its potential transformer. Two low-
voltage leads are brought out of the case for
connection to a control switch. Fig. 9 is a
wiring diagram showing how this equipment
may be used as an automatic substation.
terrupt the ser\-ice at either sectionalizing
point because the local load is connected to
the line through the other switch, each portion
of the line being fed from a separate station.
This arrangement is illustrated by Fig. 10.
The sectionalizing switches are of either
the oil break or air break type. All live parts
of the former are enclosed, and are therefore
insulated against accidental contact by the
workmen. However, an oil break switch
requires considerably more attention than an
air break switch. For ordinary' ser\"ice where
they are opened infrequently and on moder-
ate loads, and the climate is generally dr\',
it is recommended that the' breakers be in-
spected and oil tested every six months. ; If
Substation
B
Si:::^
To Signal Service
Sz:€:=a;qff::::iQ;qS::;
Faulty Section
Substation A
Fig. 10. Simple Diagram Showing Faulty Section of Transmission Line Cut Out
and the System Supplied from Adjacent Substations
This is simply an automatic change-over
equipment ; no provision is made for separate
control of the sources of supply or feeders
and no line protective equipment is included.
To permit of inspection and repairs on a
transmission line with minimum interference
to traffic, it is the practice to sectionalize the
line at many points, usually at each signal
location. Two sets of sectionalizing switches
are installed at each point, and the trans-
formers which supply the local signal and
lighting load are connected to the line be-
tween the two sets. To cut out a faulty sec-
tion of line the one set of sectionalizing
switches at each location adjacent to the
faulty section is opened. This does not in-
the puncture voltage is below 22,000 volts
between 1 inch diameter flat disks 0.1 inch
apart, the oil should be replaced by oil that
meets the test. If the climate is moist, the
inspection and tests should be more fre-
quent.
Air break switches are not ordinarily
recommended for opening a circuit under load.
However, the normal load of a railway signal
transmission line is so small, comparatively
speaking, that air break switches having a
large breaking distance may be safely used.
They must be of rugged construction be-
cause they are operated by means of a long
rod. The rod should have a weather shield
and grounding device.
909
The Cooper Hewitt Quartz Lamp and
Ultra-violet Light
By L. J. BuTTOLPH
Engineering Department, Cooper Hewitt Electric Company
The principles and applications of the glass-enclosed Cooper Hewitt lamp were described in our two
preceding issues. This lamp is of low intrinsic brilliancy and finds extensive application in the lighting of
industrial plants and for photographic work. The Cooper Hewitt quartz burner, which is described in this
article, operates at a much higher temperature, and hence greater intrinsic brilliancy, and its peculiarity is
the richness and intensity of its violet and ultra-violet radiations. These ultra-violet rays are screened out
by ordinary crown glass, but are transmitted freely by quartz. The quartz lamp is of value in research work,
and commercially is of great importance in photo graphic and photo-chemical processes and in the treatment
of parasitic and tubercular skin affections. Its therapeutic effects are similar to those of X-rays but are
less severe. — Editor.
The increasing importance of the mer-
cury arc in quartz as a source of ultra-violet
light has justified a stunmary of its latest
developments and a compilation of some of
the related technical data.
Dr. J. C. Pole in "Die Quartz Lamp," 1914;
W. A. D. Evans in the Trans. I. E. S., and
Dr. E. Weintraub in the General Electric
Review, 1914, have detailed the early de-
velopment of the quartz mercury arc. Bas-
tian, in England, and Heraeus, in Germany,
mediate steps of glasses of increasing coef-
ficients of expansion to a glass fused directly
to a metal lead-in wire and forming with it
a permanent vacutim tight seal. The abil-
ity of this glass-metal seal to stand high tem-
perature permitted the use of an anode elec-
trode of infusible tungsten instead of mer-
cury. As a result of the use of the new method
of sealing-in a greatly simplified quartz burner
was developed and is now manufactured
by the Cooper Hewitt Electric Co.
Fig. 1. Cooper Hewitt Quartz Lamp, 110 and 220volt Direct-current Burners
contributed to the development of the first
commercial type of quartz burner which was
later manufactured in the United States. A
burner of this general type was also made by
the Cooper Hewitt Electric Co. some years
ago. These burners required an elaborate
temperature control of the mercury elec-
trodes which was secured in one case by
fins of metal and in the other case by a con-
densing chamber as in the ordinary low pres-
sure type of Cooper Hewitt lamp.
The first radical change in quartz burner
design came with the development of a
means of connecting quartz through inter-
The Cooper Hewitt quartz burner is essen-
tially a vacutun arc in a fused quartz chamber.
In contrast with the standard glass-enclosed
mercury arc the quartz burner operates at
temperatures, in certain parts of the arc,
which approach the softening temperature
of fused quartz, some 1400 C, and at a
mercury vapor pressure even above atmos-
pheric pressure. At this pressure and tem-
perature there is added to the discontinuous
spectnim of the hmiinescent mercury vapor
a continuous spectrum because of its incandes-
cence. It is this circtunstance which accounts
for a shift in the relative radiation intensity
910 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
towards the longer wave lengths with increased
energy input. This is in contrast with Wein's
displacement law for common incandescent
light sources.
As shown in Fig. 3, the quartz mercury
arc starts with high current and low voltage
when the burner is cold and the vapor pres-
sure low. As the temperature rises the vapor
pressure and the voltage follow while the
current drops. Normal operation is reached
when the heat radiated from the burner
equals the electrical energy input and the
vapor pressure no longer rises. This point
is largely determined by the ventilation of the
burner and by the room temperature. For ex-
ample, with a given maximiun burner voltage.
line supply voltage to a quartz arc outfit.
Any increase of line voltage makes little
change in the burner voltage because of the
temperature and vapor pressure lag. shown
graphically in the starting characteristic
cur\'e. This increase of voltage at first only
affects the current through the series resistance
and hence through the burner arc. As in start-
ing, with the temperature increase the burner
voltage increases, the series resistance voltage
decreases, the current decreases and normal
operation is resumed at a higher burner voltage
but with practically the original current.
The radiation of the Cooper Hewitt mer-
cury vapor arc extends from the extreme
infra-red to the region of 1S.50 Angstrom
Fig. 2. Sectional Drawing of Cooper Hewitt Quartz Lamp
determined by the line voltage, the current
is increased by cooling the burner. For this
reason the Cooper Hewitt laboratory outfit
with its open hood and excessive ventilation
requires about 4.5 amperes at ISO volts, while
the standard illuminating outfit with an en-
closing glass globe has a current consumjjtion
of approximately .3.j amperes. As the arc
changes from the low to the high pressure
condition the limiinous arc column becomes
concentrated in the center of the tube. There
is then some thirty volts drop per inch in the
arc as compared with one and one third volts
in the mercury arc in glass.
Fig. 4 shows the "stationary" volt ampere
characteristics of a quartz mercury vapor
arc. The significant feature is the steepness
of the curve for voltages above 100. There
is a range of some SO volts over which the
current is nearly constant. This condition
holds only for temperature equilibrium at the
various operating wattages along the curve.
The broken lines show approximately what
happens when there is a sudden change in the
units in the ultra-\-iolet. The relative spectral
distribution for normal operation is as shown
in Figs. .) and (> where the principal lines have
been shown of lengths proportional to their
relative radiant energ\-.
For convenience in discussion the quartz
mercury arc sjicctrum will be considerd as
of two parts, the violet and ultra-violet part
extending from \SM) to 4500 and the visible
part from 4500 to 7700.
The visible part of the spectrum is of
unique value as a source of high intensity
monochromatic light for polariscopic, spec-
troscopic, and interferometer work. The
radiation from 4500 to 14.000, one third of all
the radiation of wave length less than 14.000.
is largely concentrated in a close pair of yel-
low green lines at 5704 and 5791 and a green
line at .5401. In addition to the relatively
high radiant intensity of the.se lines is the
significant fact that they lie in a i^art of the
spectrum corresponding to nearly maximum
visibility or eye sensibility. These lines arc so
brilliant that for most purposes they may be
THE COOPER HEWITT QUARTZ LAMP AND ULTRA-VIOLET LIGHT 911
separated by refraction through a prism and
used directly. Formulfe for filters to isolate
any of these lines are readily found in the
standard handbooks. For example, a solu-
tion of eosin dye in ethyl alcohol will isolate
5764 and 5790 while a double cell filter of
neodymium ammonius nitrate and potassium
dichromate will isolate 54(5 1 which is one of
the finest monochromatic light sources known.
Wratten filters have been developed especially
for use with the mercury arc and the trans-
missions of three of these filters are shown on
Fig. 5, No. 22E2 isolating 57()4-9(), No. 77
isolating 5461, and No. IcS isolating 3650.
There is also a Wratten filter to isolate 4358,
although cobalt blue glass and a solution of
quinine sulphate in ethyl alcohol will serve
the same purpose.
For very accurate polarimetric readings
and measurements of rotary dispersion the
quartz mercury arc is unexcelled. The
best practice is illustrated in the Hilger polari-
meter with a three-field Lippich system to
which is added a slit on the polarizer and a
direct vision dispersing prism in the analyzer
eyepiece. Interchangeable dispersing prisms
transmitting green, blue or violet, as the case
may be, enable rotary dispersions to be
quickly detennined.
The Cooper Hewitt low pressure glass-
enclosed mercury arc and various forms of
Aron's lamp have been used for spectroscopic
work, for lens testing and for the study of
1'>A
ifiA-
\^
"^:
..^
/
Al^
/
fiA
A
r
411^
X
/
20
^^
^
BOOM t:mp!
EATUEE-
25C
^ 1 }
<i
t
!i
1 1
12
\
10
\
\
a
icn
\
,-00
SEi-
-Si
tNt.'
1ft
ti —
V
r
\
^^
■^
I'M
V
/
/ •
\
M
fin
/
"~~-i
.
>^M«
?rep
s
/
/
O
in
/
y
eo
)M
TEMf
^EEA
Tue
:-2
j'c
''l
»
I
t
I 6 ^ 10
MINUTES AFTE"e STAeTiMG
%X
^4
Fig. 3. Volt-ampere Starting Characteristics of a 220-'
Cooper Hewitt Quartz Lamp
Fig. 4. Volt-ampere Stationary Characteristic of a 220-volt
Cooper Hewitt Quartz Lamp
polarized light. These, however, have fallen
short in that they have either a low intrinsic
brilliancy or a very small source. The quartz
mercury arc provides an intrinsic brilliancy of
some fifteen hundred candles per square inch
as contrasted with fifteen for the low pres-
sure burner. Furthermore the light source is
equivalent to a slit source one-fourth
bythreeinchesinthe 1 10-voltburner
and six inches in t he 22()-volt burner.
The utility of a powerful source
of mono-chromatic light for inter-
ferometry is increasing with the
application of the interferometer
to gas analysis and to the testing
of solutions.
While designed to operate in a
nearly horizontal position, for in-
dustrial and routine laboratory work
the burner can be removed from
its holder, and with the cathode
chamber clamped in a laboratory
support, may be operated in a ver-
tical position with only a slight
change in the electrical character-
istics. In general when operated
out of the hood a cup-like shield
should be placed over and around
the ends of the burner to decrease
the rate of heat dissipation and to
maintain the high vapor pressure
characteristic of the burner.
912 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 11
THE COOPER HEWITT QUARTZ LAMP AND ULTRA-VIOLET LIGHT 913
914 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
It is vitally important when working with
the visible spectrum of a quartz mercun,-
arc that the operator's eyes be protected
from the invisible but injurious ultra-violet rays
by interposing a sheet of ordinary clear glass.
The intensity of its violet and ultra-vio-
let parts is the most unique quality of the
quartz mercury arc spectrum. The radia-
tion of a wave length less than 4500 repre-
sents two thirds of the total radiation of wave
length less than 14,000 and like the visible
radiation is concentrated in a few spectral
lines of high intensity. The region from 3000
to 4000 while transmitted by ordinary glass
lies largely outside the range of visibility. It
is easily detected by a photographic plate and
by the fluorescence of such substances as wil-
lemite. This part of the ultra-violet seems
to have relatively little bacteriological or
therapeutic effect. It is however of great
importance, photographically and photochem-
ically. Its effects on plant life are very similar
to those of intense sunlight and are being
studied for possibilities in plant growth con-
trol. The photochemical effects while not so
striking as in the more extreme spectrum, are
nevertheless of extreme importance in the
chemistry of plant life, the formation of vege-
table dyes, and the fading of synthetic ones.
A pair of especially strong lines occur at
oC)riO-4, one at 3984 and a pair at 404{)-7N.
There is a striking coincidence in the position
of these lines near the maximimi of photo-
graphic sensitivity as well as in the position of
line .5461 at the secondary maximum as shown
on Figs. 5 and 6. These lines, isolated by
cobalt glass filters similar to Wratten No. IS
but of higher transmission, have been used
with remarkable success for invisible signalling
by means of receivers whose fluorescence
transforms the invisible signal light to a
visible light in the blue. The light of these
lines, isolated from visible light in the same
manner as for signaling, is also used for paint
testing. Certain paint pigments instantly
show their instability by a visible fluores-
cence undoubtedly accompanying the photo-
chemical reaction, which in time changes the
composition and color of the paint.
The region from 2000 to 3000, trans-
mitted by quartz, but not by ordinar\- glass as
indicated in Fig. cS, is unique in the Cooper
Hewitt quartz mercury arc. The radiation
in this region is ver\- remarkable for its pho-
tochemical, therapeutic, and abiotic effects.
Its application as a catalyzer in gas reactions,
the halogenation of organic substances, the
deblooming of oils, the "ageing" of paint
materials, and the testing of dyes are a few
well known examples. In the latter case it
has been found that for practical working
conditions, the mercury arc light is eight to
ten times as effective as sunlight. The
application of the ultra-violet to the steriliza-
tion of liquids is limited only by their trans-
missions. Water sterilization has been a prac-
tical process for some time. The limitation of
milk sterilization is its opaqueness, but tur-
bulent flow in a thin film offers a solution
of the problem. There are immediate pos-
sibilities in the application of ultra-violet to
qualitative analysis and to factory- control
of chemical processes. For example, it is now
used to control the composition of the tail-
ings in the reduction of zinc ores. It has also
been used for the routine analysis of certain
unsaturated hydro-carbons by noting the
voliime of oxygen or of halogen absorbed.
The Finsen light was one of the first
attempts to apply ultra-violet to the treat-
ment of disease. It is of value in parasitic
and tubercular skin diseases. The phys-
iological effects of ultra-violet are more
superficial than those of Roentgen rays and
less destructive to the tissues. They are
ver\' similar to those of direct sunlight but
much more intense. The methods of using
a quartz burner for the therapeutic effects of
the ultra-violet are similar to those of X-ray
practice. They involve placing the burner
in a protective hood whose position is univer-
sally adjustable and providing the hood with
accessor^' adapters canying suitable dia-
phragms and screening devices.
Figs. 5 and 6 need little explanation.
Transmissions have been plotted against a
wave-length scale for crx'stalline quartz, water,
Uviol glass, crown glass, and certain eye pro-
tective glasses designated by their trade
names. Three representative filters are
shown from the so-called Wratten mercun,-
monochromats. The relative visibility and
photographic sensitivity curves have been
plotted and the relative luminosity and
photographic effect cur\-es for any source of
light arc then the continuous product of these
cur\-es by the relative radiant power of that
light source. For sources having continuous
spectra these curves are vcr>' similar to the
dominant fonn of the visibility and sensitivity
curves. For a line spectral source the relative
liuninosities and photographic effects would
then be represented by lines.
A change in the relative spectral distri-
bution of radiant power in the mercur>- arc
spectrum with change of- power input has
THE COOPER HEWITT QUARTZ LAMP AND ULTRA-VIOLET LIGHT &L5
been roughly indicated and shows a shift
towards the red end of the spectrum due to
a disproportionate increase in the tempera-
ture of the quartz tube itself and of the tung-
sten electrode. For line spectra no simple
relationship corresponding to Wein's dis-
placement law has been established.
Fig. () shows the reflectivities of a few
polished metal surfaces. Mach's magnalium
alloy with its peculiarly high reflectivity is
impractical because of its non-uniformity.
Sam])les of apparently the same composition
and polish differ very widely and, in general,
show very much lower values than those
indicated. High reflectivity in the region of
2200 and low values in the visible spectrum
is the unique property of polished silicon.
The quick drop below 2000 and the shape of
the curve suggests the possibility of sub-
stances having selective reflectivity and little
fluorescence in the shorter wave lengths. At
present, however, polished nickel remains the
most practical reflecting material for the
ultra-violet.
A vast amount of interesting research
remains to be done on the selective nature of
the various ultra-violet reactions. Definite
relationships have been noted between cer-
tain gas reactions and the selective absorption
of the reacting gases for certain portions of
the ultra-violet spectrum. It seems probable
that for every distinct type of photo-effect
whether chemical, abiotic, or therapeutic,
there is a curve analogous' to the relative
visibility and photographic sensitivity curves.
More exact data along these lines would be
invaluable.
Those planning to do research work
should note that to cover the spectrum com-
pletely the mercury arc must be supplemented
by a tungsten or titanium arc in a \-acuum
and in quartz which will give a fine-line ultra-
violet spectrum of low intensity. Experience
has shown that for practical work the high
intensity of the mercury arc lines compen-
sates for the gaps in the spectrum although
maximum efficiency for certain photo-
chemical reactions may make some qualifica-
tion necessary.
The quartz mercury arc as ordinarily
made is essentially a direct-current device
and must be supplemented by a rectifier or
motor-generator set to adapt it to use
on an alternating current. It is however
possible to make the quartz burner itself
function as a rectifier and an alternating-
current quartz arc of this type is under devel-
opment.
For most experimental and industrial
purposes the users devise their own holders
and accessory apparatus for use with the
burner. As a basis for these special opera-
tions the Cooper Hewitt Electric Company
manufactures a simple but effective labo-
^1^3:^
Fig. 7. Laboratory Outfit, Cooper Hewitt Quartz Lamp
ratory outfit, shown in Fig. 7. The burner
is of new transparent quartz which transmits
the ultra-violet light freely. Although quartz
glass will stand very high temperatures and
sudden changes of temperature it is, like or-
dinary glass, very fragile and must be han-
dled with care and kept chemically clean of
grease and dust. The aluminum reflector
and hood serves as a support for the burner
and holder as well as a protection to the
operator. The auxiliary consists of a react-
ance coil and adjustable resistance enclosed
in a ventilated metal case. The reactance,
resistance, and burner are connected in series
as shown in the wiring diagram. To operate
the burner outside of the outfit it is only
necessary to extend the lead wires from the
binding posts on the top of the hood to the
burner terminals. Directions for the instal-
lation and operation of the laboratory outfit
are contained in separate instruction books.
The one additional precaution to be observed
916 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
in operating a burner outside of the regular
outfit is to keep the negative or cathode end
lower than the positive end.
The following brief bibliography will, in
addition to the standard ph^'sico-chemical
tables, serve as a guide to the usual methods in
ultra-violet research, to some of the original
data, and to more extensive bibliography on
the subject:
J. C. Pole, "Die Quartz Lampe," 1914.
W. A. D. Evans, Trans. I.E.S. Vol. X, No. 9.
E. C. C. Balv, Spectroscopy.
R. W. Wood^ Physical Optics.
S. E. Sheppard, Photo-chemistry.
Theo. L>TTian, Spectroscopy of the Extreme
Ultra-violet.
M. Luckiesch, Color and Its Applications.
Kuch and Retchinskv, Ann. der Phys. (4)
20,563.
VerhoefE and Bell, "Pathological Effects of
Radiant Energy on the Eye." Proc. Amer.
Academv, Vol. .51.
Bureau of" Standards, Bulletin T119, "The
Ultra-violet and Visible Transm.ission of
Eye Protective Glasses."
Bureau of Standards, Bulletin S330, " Decrease
in Ultra-violet and Total Radiation with
Usage of Quartz Mercur\- Vapor Lamps."
Eastman Kodak Co. " Wratten Light Filters."
Rubens and Hagen, Ann. der Phvsik. 1352,
1900; IS, 1,1902.
E. O. Hurlburt,
42,205-03 (1915)
Bureau of Standards, Bulletin No. S34
"Spectrum Lines as Light Sources
Polariscopic Measurements."
Bureau of Standards, Bulletin No. T148,
"The Ultra-violet and Visible Transmission
of Various Colored Glasses."
Astrophysical Journal
m
i
A-SPECTBUM OF COOPEE HEWfTT QU*eTZ f«eCUBY ABC LIGHT
B- SPECTEUM 'A' APTEE TEAN5MISSIOM THEOUGH OEDTJAEY CEO^N GLASS
Fig. 8. Spectrum of Mercury Arc Transmitted Through Crown Glass and Through Quartz
917
Step-by-step Integration of Curve Areas of
Phase Significance
CORRECT AND INCORRECT METHODS
By Chas. L. Clarke
Consulting Engineering Department, General Electric Company
Several ordinate rules are in books for calculating more or less closely, as may be desired, the true area
of an irregular plane figure or curve, such as the mean-ordinate, mid-ordinate, trapezoidal, Simpson one third
and three eighths, Durand, and Weddle rules, all too well known to require explanation here. Of these the
first three mentioned are, in principle, applicable with practically equal accuracy, if rightly used, when con-
ditions of a problem require the area to be determined, or integrated, separately step-by-step between the
successive evenly-spaced ordinates. When, however, the step-by-step areas of the total area of a curve thus
integrated have a phase relation to another curve or curves, these elementary areas should be taken as con-
centrated, or located, at points midway between their respective bounding ordinates, that is, at the mid-ordinate
of the mid-ordinate rule, and not at one of the bounding ordinates, which are the only ordinates of the mean-
ordinate rule and of its substantial equivalent the trapezoidal rule. The necessity for observing this pre-
caution as to assumed position of the elementary areas, when of phase significance, is pointed out and
illustrated by examples in the following article. — Editor.
of a definitely related fundamental and triple
harmonic; and although in practice the wave
shape and value of the exciting current have to
be determined, we shall here pre-assume this cur-
rent, and that it also is composed of a definitely
related fundamental and triple harmonic.
Then since the maximum instantaneous
values of current and magnetic density must
be in phase, and the latter has a definitely
known mathematical relation to the voltage,
we are able correctly to calculate the phase
angle between the equivalent sine wa\'es of
voltage and current, and thus find the true
angle of hysteretic lead, and can also de-
termine the corresponding hysteresis loop
for the core iron, although in an actual case
the loop must be known beforehand.
Next, beginning with the same voltage wave
and hysteresis loop, we shall calculate from
these the exciting current and angle of hyster-
etic lead by the step-by-step method, intro-
ducing, however, the error before referred to in
handling the method, and thus obtaining an
incorrect result. The reason for the error and
the manner to avoid it will then be dealt with.
Finally, by way of demonstration, the
current and angle of hysteretic lead will be
recalculated by the step-by-step method
handle in the correct manner, whereby a
result for the angle is obtained practically
agreeing with the true angle derived in the
first instance by precise mathematical means.
MATHEMATICAL SOLUTION
For the sake of brevity in this mathematical
solution, only the results and the major steps
in obtaining them are given.
Voltage
Assume, for example, an impressed voltage
e having a triple harmonic of one fifth the
amplittide €,„ of the fundamental and in phase
therewith.
Introduction
The purpose of this article is to direct
attention to errors that are bound to occur
if the mean-ordinate rule, instead of the mid-
ordinate rule, is used in progressively integrat-
ing uniform step-by-step areas going to make
up the whole area of a cur\'e, when such
integration bears a phase significance to
other related curves.
For the purpose of demonstrating the
incorrectness of using the mean-ordinate rule
in such a case, and the practical accuracy of
results obtained by applying the mid-ordinate
rule, a convenient example is afforded in the
step-by-step approximation method of calcu-
lating the phase angle between equivalent
sine waves of impressed voltage and exciting
current,* in a transformer primary, and
therefrom determining the angle of hysteretic
lead of current phase (hysteretic angle of
advance), or phase position ahead of the
90-degree lag behind the voltage that the
current would have were there no con-
sumption of energy by hysteresis, and thus
no power required from the circuit to stipply
this energy. The phase relation between the
equivalent waves is determined from the
known hysteresis loop for the iron used in the
core, and the known effective value and wave
shape of the primary impressed voltage.
To demonstrate the nature of the error and
how to avoid it, we shall first consider a case
in which the impressed voltage is assumed to
be a distorted wave composed, for example,
* A sine wave of voltage or current that has the same effective
value and frequency as a wave of other form is the "equivalent
sine wave" of the latter, and is capable of the same effect.
Waves of voltage and current of different form in a circuit have
no definite phase relation, but their equivalent sine waves may
have a fixed phase relation which must be such that the waves
together represent the same average power as the actual waves.
In this article the "exciting current" does not include that
component of the primary exciting current that may be called
for by eddy currents set up in the transformer structure, but is
limited to that part of the primary current required to produce
the magnetic cycle in the core, composed of the magnetizing
current and hysteresis current.
91S November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
The equation for instantaneous values is
<
The relative
given by
sin a — T sin 3 a
instantaneous values er are
)
(1)
-(
stn a — — sin 3 a
o
)
(2)
in which g is a constant of any convenient
value greater than zero and less than infinity;
which means that the wave shape of relative
values may be plotted to any desired scale.
The vakies of Cr corresponding to the time
angle a for each five degrees of a half-cycle
between zero values of voltage are given in
column (2) of Table I.*
* All the calculations in this article have been carried out to
more decimal places than are given in numerical coefficients
in the equations, or written in the tables.
Squaring (1), multiplying through by da
and integrating between the limits zero and
IT, we obtain
average e- = E- = e-
from which
em = 1.3S6S£
and from (1) and (3)
<i)
e=1.3S6S£
(
1 . .,
sin a — — sin 3 a
)
(3)
(4)
in which E is the effective value of the actual
voltage wave, which must also be the effective
value of the equivalent sine ivave.
The values of c corresponding to a for each
five degrees of the half-cycle are given in
column (3) of Table I.
TABLE I
(1)
(2)
(3)
(4)
(5)
(6)
(7)
Degrees
a
er
e
B
i
/
P
0
0.0000 g
0.0000 E
-1.0000 Bmai.
-1.5865/
- 1.5865 /n I
0.0000 IE
5
0.0354 g
0.0419 E
10
0.0736 g
0.1021 £
-0.9933 Bmax.
-1.5231 /
-1.5231 In/ 1
-0.1556 /£
15
0.1174 g
0.1628 £
20
0.1688 g
0.2341 £
-0.9711 Bmax.
-1.3320/
-1..3320/n 7
-0.31 18 /£
25
0.2294 g
0.3182 £
30
0.3000 g
0.4160 £
-0.9279 Bmax.
-1.0312/
-1.0312 In, I
-0.4290 IE
35
0.3804 g
0.5275 £
40
0.4696 g
0.6512 £
-0.8565 Bmax.
-0.6608/
-0.6608 /n/
-0.4303 IE
45
50
aeii^l-
0.7845 £
, 0.9236 £
-0.7506 Bmax.
-0.2731 /
-0.2731 In/l
-0.2522 IE
55
0.7674 g
nil 1. 0642 £
60
0.8660 g
1.2010 £
-0.6071 Bmax.
+0.0798 /
+0.0798 In/l
+0.0959 IE
65
0.9581 g
1.3286 £
70
1 .0397 g
1.4418 £
-0.4283 Bmax.
+0.3584 /
+0.3584 In/l
+0.5167 IE
75
1.1073 g
1.53.56 £
80
1.1.580 g
1.6059 £
-0.2218 Bmax.
+0.5446 /
+0.5446 In I
+0.8746 IE
85
1.1894 g
1.6494 £
90
1 .2000 g
1.6641 £
0.0000 Bmax.
+0.6451 /
+0.6451 Inl
+ 1.0736 /£
95
1.1894 g
1.6494 £
100
1.1,580 g
1.60,59 £
+ 0.2218 Bmax.
+0.6878 /
+0.6878 In I
+ 1.1045 /£
105
1.1073 g
1 .5356 £
110
1.0397 g
1.4418 £
+0.4283 Bmax.
+0.7123/
+0.7123 /m 7
+ 1.0269 /£
115
0.9,581 g
1.3286 £
120
0.8660 g
1.2010 £
+ 0.6071 Bmax.
+0.7582 /
+0.7582 In I
+0.9106 /£
125
0.7674 g
1 .0642 E
130
0.6660 g
0.9236 £
+ 0.7506 Bmax.
+0.8531 /
+0.8531 In I
+0.7879 IE
1.^5
0.5657 g
0.7845 £
140
0.4696 g
0.6512 £
+0.8565 Bmax.
+ 1.0035/
+ 1.0035 In I
+0.6534 IE
145
0.3804 g
0.5275 £
150
0.3000 g
0.4160 £
+ 0.9279 Bmax.
+ 1.1925 /
+ 1.1925 111 I
+0.4961 IE
155
0.2294 g
0.3182 £
160
0.1688 g
0.2341 £
+0.9711 Bmax.
+ 1.3836/
+ 1.3836 /n 7
+0.3239 IE
165
0.1174 g
0.1628 £
170
0.0736 g
0.1021 £
+ 0.9933 Bmax.
+ 1.5299 /
+ 1.5299 /n 7
+0.1562 IE
175
0.0354 g
0.0419 £
180
0.0000 g
0.0000 E
+ 1.0000 Bmax.
+ 1.5865/
+ 1.5865/(1//
0.0000 /£
True av. e = 0.8240 £ True av. i =0.8971 / Truecff. i=/ True av. p = /> =0.3579 /£
STEP-BY-STEP INTEGRATION OF CURVE AREAS OF PHASE SIGNIFICANCE 919
Similarly, from equation (2),
average ^> = g"l ;5^ I
from which
effective gr = 0.72 11 g
(5)
Magnetism
The equation for the wave of magnetic
density, in terms of maximum density, may
readily be obtained by reference to the general
solution in the appendix to this article.
Comparing the specific equation (1), for
impressed voltage, with the general equation
(18), it is obvious for the purpose of the exam-
ple under consideration that in the general
equation Ci = £>„ ; G = — -\ and /3i, ft, . . . . /3„
r>
and Ci, Ch, ■ ■ ■ ■ C„ are each zero.
We may, therefore, pass to equation (23)
and bv substitution therein at once write
B
COS a — irz COS 3a
lo
)
15
as the equation for the wave of magnetic
density in the specific case, which reduces to
5= 1.0714 B„
{h
cos 3 a — cos a
)
(6)
where B is the instantaneous magnetic density
and Bmax. is the maximum density in the mag-
netic cycle.
The wave shape of magnetism is shown in
Fig. 1.
From the last equation the instantaneous
values of magnetic density are obtained
independently of any direct consideration of
the voltage, upon the basis of any assumed
value for the maximum density.
The values of B corresponding to a for
each ten degrees of a half-cycle are given in
column (4) of Table I.
Exciting Current
Preassume a current having, for example,
a triple harmonic of one-fourth the amplitude
im of the fundamental and lagging 90 degrees
behind the latter, and write the equation for
instantaneous values
i = i,„ \sin /3 - J 5m (3 /3 - 90°) 1
or
<
sin /3 + — C05 3 j3
)
(7)
Squaring (7), multiplying through by d/S
and integrating between the limits zero and
X, we obtain
average i-
=-=.©
Fig. 1
from which
j„,= 1.3720/
and from (7) and (S)
1.3720 /( sinfi+jCos3fi
0
)
(8)
(9)
in which / is the effective value of the current
wave.
We have now to establish the time relation
of (9) for current to (4) for voltage, and so
modify the former equation that the variable
time angle therein is also a. The first step is
to find the angle included between maximum
current and zero value of the fundamental
component of the wave.
Differentiating (9) and placing the differen-
tial coefficient of i with respect to /3 equal to
zero:
'(
'^'i = 1.3720 l(cos p-^sin 3 /3
)
= 0
* By closer approximation, ff =110° 39' 1.6".
dl3 ^.--r- ^
from which we have maximum current when
3 .
cos l3 = —sin 3 /3.
4
This relation is found by trial and approxi-
mation to be satisfied for the angle
/3=110° 39'*
that is, the maximum exciting current is
110° 39' behind zero value of the fundamental.
And since, from physical considerations,
maximum magnetism must occur at the same
instant as maximum current, the former is also
110° 39' behind zero value of the fundamental
of the current wave. But from (4) and (6) the
920 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 11
corresponding positive maximum magnetism is
ISO degrees behind zerovalueof the fundamen-
tal of the voltage wave ; hence zero value of the
current fundamental lags 180° -110° 39' = 69°
21' behind zero value of the voltage funda-
mental, from which follows the relation
/3=a-69°21'
(10)
Fig. 2
and we mav rewrite (9) in the form
i= 1.3720/
r5iw(a-69°21') + ^<:o53(Q;-69°21')l
as the equation for current referred to the
same time angle as is the impressed voltage in
equation (4). The wave shape of the current
is shown in Fig. 1.
The effective value I of the actual excit-
ing current wave, in (10), must also be the
effective value of the equivalent sine wave.
When i is zero, from (10) we have by trial
and approximation a = 57° 35'*, that is,
zero current is 57° 35' behind zero voltage.
The values of i corresponding to a for each
ten degrees of a half-cycle are given in column
(5) of Table I.
Hysteresis Loop
From (10) and the well-known equation
f=f (11)
in which / is the magnetizing force, ni the
ampere-turns of magnetomotive force, and
/ the length of the magnetic circuit, we may
write
1.3720 w J
^ I
\ sin (a- 69° 21')+^ cos 3 (a- G9° 21'"] (12)
The values of / corresponding to a for each
ten degrees of a half-cycle are given in column
(6) of Table I.
* More exactly, 57° 34' 42.4'.
t More exactly, $ =69° 1' 52.3'.
From the simultaneous values of B and / in
columns (4) and (6) of Table I, the hysteresis
loop is plotted in Fig. 2. In calculating the
angle of hA-steretic lead of phase in practice,
however, the hysteresis loop is, as previously
stated, known beforehand.
Power
Multiplying (4) and (10) together, we have
for the instantaneous power or rate of energy
delivered to the primar>' for supplying the
energy consumed by hj'steresis:
ie = p = 1 .9026 IE \sin a sin(a- 69° 21') +
1
1
sin a cos 3{a-69° 21')-
69° 21')-
sin3 aro5 3(a-69°21')l
(13)
— 5m 3 a sin(a-
o
J_
20'
The values of p corresponding to a for each
ten degrees of a half-cycle are given in column
(7) of Table I. The wave shape of power is
shown in Fig. 1.
Expanding (13), multiplj'ing through by
da, and omitting terms having sine and
cosine products, since they become zero when
integrated between the limits zero and r,
results in the equation
p da =1.9026 IE (cos 69° 21' sin- ada +
1_
20
sin 2S° 3' sin'- 3 a da
)
which, integrated between the limits of zero
and TT, finally gives
average /> = P = 0.3579 /£ (14)
Equivalent Sine Waves and Their Phase Angle
Equivalent sine waves of voltage and
exciting current, that is, sine waves of the
same effective values E and / as the foregoing
actual waves, will also represent the same
average power P as the actual waves, when
P = IE cose (15)
in which d is the phase difference between the
sine waves, or the equivalent phase angle, as
indicated in Fig. 3; hence from (14) and (15)
cose = 0.3579
and
e = 69°2't
Since the maximum \-alue of a sine wave is
\/2 times the effective value, we may write,
as the equation for instantaneous values of the
equivalent sine wave of voltage.
Csin = \^ E sin a
STEP-BY-STEP INTEGRATION OF CURVE AREAS OF PHASE SIGNIFICANCE 921
and for instantaneous values of the equivalent
sine wave of current
isin = V^ I -5"^ {a — d).
Multiph-ing the two last equations together
and remembering that 0 = 69° 2' we have, as
the equation for instantaneous values of
equivalent wave of power
isin esin = Pcquh. = 2 IE sin a sin{a — 69° 2')
the wave form of which is shown in Fig. 3.
The effective hysteresis (energy) com-
ponent of the equivalent sine wave of exciting
current, in phase with the equivalent sine
wave of impressed voltage, is
I cos 6 = 0.3579 I
and the effective magnetizing (reactive watt-
less) component, in quadrature with the
voltage, is
I sin 0 = O.933S I.
Angle of Hysteretic Lead of Phase
Since the lag of the equivalent sine wave of
exciting current behind the corresponding
sine wave of voltage would be 90 degrees, as
indicated by the dotted wave in Fig. 3, were
there no hysteresis and thus no power required
to supply the energy consumed thereby, the
effect of hysteresis has advanced the equiva-
lent sine wave of current by an amount
represented by the angle of hysteretic lead
7,° = 90° -69° 2'
= 20° 5S'
STEP-BY-STEP SOLUTION
Erroneous Method
Assume that a transformer is operating
under an impressed voltage of known wave
shape and effective value, and that the
hysteresis loop corresponding to the maximum
magnetic density in the core is also known,
the problem being to determine the angle of
hysteretic lead. Let these factors be the same
as heretofore.
First. On a diagram of the voltage wave,
nowadays based on an oscillogram record,
divide the time of one half-cycle between zero
values of voltage into a suitable nirmber of
equal parts, in the present example eighteen,
and write the corresponding time degrees a
in column (1) of Table II.*
* The number of divisions best to employ is largely determined
by the shape of the wave, and in general should be such that the
corresponding ordinates meet the wave as near the maximum,
minimum and inflection points as is feasible. Naturally, the
greater the number of divisions the more accurate the final
result, and the less the importance of the ordinates touching
the wave as near to the points mentioned.
Second. With any suitable scale measure
the lengths of the corresponding ordinates to
the wave, which will represent the relative
instantaneous values e, of the voltage —
although in the present case we find these
values directly by equation (2) — and write
them in column (2).
Fig. 3
Third. Enter the squares Cr^ of the relative
values in column (3), the square root of the
average value of which will be the relative
effective value, or
effective ^r = 0.7211 g
Fourth. As the relative effective voltage
eff. Ct bears the same ratio to the actual effec-
tive voltage E as relative instantaneous values
Cr bear to corresponding instantaneous values
e, we have
er E
e =
0.7211 g
from which the values of e in column (4) are
obtained.
Fifth. Remembering that under the condi-
tions of the problem the impressed voltage e and
counter induced voltage e,- are substantially
equal, although of opposite sign, it will be
seen from equation (16), Appendix, that the
magnetic density B for a given time angle a
is proportional to the integration or simi-
mation of e, that is, proportional to the area
of the voltage wave up to the same angle,
plus an integration constant. The pro-
portional and interrelative integration values
of e, that is 2(?, up to the respective angles in
column (1), are obtained by progressive
addition of the values in column (4) , as given
in column (5).
Sixth. The integration constant is de-
termined from the condition (see Fig. 1) that
B at zero degrees must be negative, and equal
and opposite to B at ISO degrees. Therefore,
922 Novembsr, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 11
ui « -.i -.i :i s; :s -.c ;i ■<; s: u; i; :i li IS -.s -.a
fc5
w
I '^
o c: ci * cr. X ci -2 o i-^ re ?: — -r i'- d iH ci
o ^ X Le c X t- -JM^ ci i-e -.c i-e 1- -r o
o — oq re re q ci -.0 cr. — — q q x i-- i-e re —
1
Si
II
w
C^^C«^J^^J^CJ^^^=,„..«,.
CTi
5
"^
2.5171
2.1282
1.4586
0.7465
0.2254
0.0091
0.0514
0.2134
0.3649
0.4.508
0.4897
0.()403
0,8503
1,2131
1,0,897
2.1835
2.5171
X
•--1
o
D
-1.5865
-1.4628
-1.2077
-0.8640
-0.4748
-0.09,53
1-0.2208
-0.4019
-0.6041
-0.6714
-0.6998
1-0.7310
-0.,S002
-0.9251
hi. 10 10
-1.2999
-1.4777
1-1.5865
IM
d
H
1
ssaa=aa=easa=a=aae
■5
Ss
From Myst
Cycle
5805
4028
2077
8040
4748
();),53
2268
4619
6041
0714
69'.I8
7310
8002
9254
1016
2999
4777
.5805
0
777°° t ++++++++++++
<
S
o^c -
<
re-r-piH»-eiHic — — cici»o
oooooooooooooooooo
o ^ --J- -r -t -^ -r -r -* ic o L*^ »-e Le u': i.'c LC o
—' oj ce Tt"-'; -.o t~ X c: o — ci re ^ ue 3 X
1
H
u
fcj
£££ = SSSSS = SSS5SSS = S
t-
5
«5
o
.0000
).9802
),9,54()
),89,84
),8105
).0859
)..5237
.3291
».1123
1.1123
.3291
),.5237
.0859
).8105
.89,84
.9.546
,9862
.0000
.0000
1 1 1 1 1 1 1 1 1 ++++++++++
t;
[i3K!fe;fej'.ii^[ijsjbife3i:tjiUfe:fe^:i;i;iisq
s
00
1
X :D i-e ic re t- i^ c. — — ~. w r^ re i-e i-e -.c X X
'~ £ -^ K; i — -■ 1; y JJ I; ~ - 3 "-J "" 'J3 t; '~
■* S q q q q X ^ X X ^ X q £ -i; •£ S -^ ^
t^ t^ t^ d -.s i.-: re ci d d ci H Le d -d t-^ f-^ t>^ t>^
1 1 1 1 1 1 1 1 1 ++++++++++
(«5
fe5fe5Km-i!::;i::-.a:aaqfe3te3[5jb5tsj[tjtj3fejij5
00
s
U
M
o — o re ue — — X 1- X t^ i.e le — ^: re -r o i-o
o ci -.s c) re I- X c: "-e — tr. i» x ci re c re o -c
o o re uo c ci ci -xi 1- re -t- X X — «; 1- — — —
o — re I- -f re i-e —. o c) X CI -t -r o I- 1^ X X
0
-t
II
0003 — cire-r-.2xc-. — cire-r-rTCTCTi,
(
'
bjtejti;'<'jCtqu;'<tc^ii:t:t:-.ai:;u;tjbjfej
w
^5
0000
1021
2.341
4160
6512
9236
2010
4418
60.59
664 1
6059
4418
2010
9236
6512
4160
234 1
1021
0000
"
OCCOOOr- — — — . ^ — oooooo
1
1
*W) *a« '&«'&<'&<%<'&<'&<''&< %e °a< %<■&< ^&< 'a* *!« '&f%4: %*
s
0.0000
0.00.54
0.0285
0.0900
0.2205
0.4436
0.7.500
1.0810
1.3410
1.4400
1..3410
1.0810
0.7.500
0.4436
0.2205
0.0900
0.0285
0.0054
0.0000
3
7i
w
b<tu:::urC4Q4&cbca4^&ca<c^&«Q<o£C4«<&£0C
1
M
k
«
0.0000
0.0736
0.1088
0.3000
0.4090
0.()660
0.8600
1.0397
1 . 1 580
1.2000
1.1.580
1.0397
0.8600
0.6000
0.4090
0.3000
0.1688
0.0730
0.0000
1
3
(
i«
oooooooooooooooooo©
X „ o
c:i~ci
II
II
>
<
00 (2o ox
n n II
to
o
00
c«
w
a
>
<
oc 5;
X
d
I u
-r> «|
Q4|— 04
o xi .>' =
•— — CI _• ci
ce I i.e — 1^
= 1 c>d
II II
II
the constant must have such
a value that if added to the
first and last values of 2^ in
column (o) the sums will be
equal, the former minus and
the latter plus, which obviously
obtains when the constant is
equal to the negative one half
total Ze, or - 7.4U7S E. Thus
by adding — 7.-107S E to the
interrelative values in column
(o) the interrelative instanta-
neous values Br of magnetic
density are found, as given in
column (6).
Srcenth. Since the max-
imimi relative density is 7. -1078
E for any given maximiun
actual density5„,ai.,theactual
instantaneous densities B cor-
responding to the relative
densities Br are expressed by
B =
BrBn
It:
a
7.4078 E
by which are obtained the
values entered in column (7).
Comparing the values of B
in column (7) of Table II,
obtained by this step-by-step
method, with the values in
column (4) of Table I derived
by exact mathematical means,
we notice a considerable dif-
ference between them. and the
suspicion should arise that
something is WTong with the
method up to this point. Since
there would arise no occasion,
however, for this error to
present itself in handling the
step-by-step method in prac-
tice, let us proceed.
Eighth. The next step is to
ascertain the magnetizing
forces f corresponding to the
densities in column (7) of
Table II. In practice the
values would be obtained di-
rectly from the hysteresis loop
for the maximum density, as
found by test on a sample of
the iron used in the trans-
former core, and at once
entered in column (S). Since
in the present example, how-
ever, f bears an assumed
mathematical relation toS,bv
STEP-BY-STEP INTEGRATION OF CURVE AREAS OF PHASE SIGNIFICANCE 923
substituting in equation (6) the values of
the latter from column (7), we obtain by
trial and approximation the corresponding
values of the time angle ««, as given, to the
nearest minute, in column (S). Here again is
evidence of error, since the angles in column
(8) do not correspond in value to those in
column (1), as they obviously should, but
widely differ therefrom.
Finally, by substituting for a in equation
(12) the values of as in column (S) the
magnetizing forces in column (9) are obtained,
which in the hysteresis loop of Fig. 2 corre-
spond to the values of B in column (7).
Ninth. From equation (11)
.JJ_
n
thus from the values of / in column (9) we
may at once write the corresponding values
of the exciting current in column (10).
Tenth. Enter the values of r in column
(11), and find the average value, the square
root of which is the effective current /', or
/' =1.0080 /
Correctly, /' should obviously equal /.
Eleventh. Multiplying the values of im-
pressed voltage in column (4) by the corre-
sponding values of exciting current in column
(10), we obtain the related instantaneous
values of power, ie = p, as entered in column
(12), from which the average power required
to supply the energy consumed by hysteresis is
P = 0.4220 IE
which by reference to (14) for the correct
average power is seen to be 17.9 per cent too
large.
Twelfth. But to determine the hysteretic
lead we are here concerned with the effective
value of current /' calculated by the step-by-
step method, and not with the true effective
value /. Therefore from the last two equa-
tions we have
P = 0.4187 I'E
= I'E cos d
from which
e = 63° 15'
that is, the equivalent sine wave of current
lags 65° 15' behind the equivalent sine wave
of voltage.
And we have for the angle of hysteretic lead
7y° = 90°-65° 15'
= 24° 45'
whereas the correct angle, we already know, is
20° 58'. Thus the step-by-step method, as
thus far incorrectly handled, has resulted in
an angle of hysteretic lead 18 per cent too
large. Had the half-cycle been divided into
less than eighteen parts the error would have
been larger, and vice versa, smaller.
The hysteresis current is
/' cos d = 0.-ll87 r
= 0.4187X1.0080 /
= 0.4220 /
Fig. 4
whereas the correct value is 0.3579 /, and is
thus 18 per cent too large.
The magnetizing current is
/'jm 5 = 0.9082 /'
= 0.9082X1.0080 /
= 0.9154 /
whereas the correct value is 0.9338 /, and is
thus 2 per cent too small.
Thirteenth. The underlying cause of the
errors before noted will be made apparent by
consideration of Fig. 4, in which e is the
voltage wave plotted from column (4) of
Table II; 2e is the step-by-step integration or
summation curve of e from column (5) ; and
B is the wave of magnetism from column (7).
2e for any angle a should, as before
observed, be proportional to the area of the
voltage wave from zero deg., thus zero volt-
age, up to the angle a. But when, for exam-
ple, a is 10 deg., we have taken the area from
zero to 10 deg. as proportional to the value of
e at 10 deg., indicated by the rectangle a, b;
and when a is 20 deg., the area from 10 to 20
deg. has been taken as proportional to the
value of e at 20 deg., indicated by the rec-
tangle c, d; and we have added together the
two rectangles for the total proportional area
Xe between zero and 20 deg., and so on.
This procedure is wrong. For the area of
the voltage wave from zero to 10 deg. is
proportional to the value of the dotted
mid-ordinate, or value of e at 5 deg. (subject
to such small error as may be due to curvi-
linear deviation of the wave from a straight
line joining the points on the wave corre-
sponding to the angular limits under con-
sideration); the area from 10 to 20 deg. is
924 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. No. 11
~ — o
1-1
>-1
>-, ►-1
►~1 ■--<
•"I
»-l
>-i
fcj Kl tj fej fc^
fc5
•-1
cJ
•^^
ooiocicor^coocioio^^^oc— i^co
010r-icOCOlOCT!.-lt^t^O(N— IQOOOIMO
O.-iC0'*'l<<MOin00O^O03t^OTi<c-;,-i
o
CO
oddodoodd.-i«r-idddodo 'if
1 1 1 1 1 + + + + + + + + + + + + "
c
■n*
(N(N^r-;dddddddddd-^'-i'-!c<i ||
>-m'~<>--i>~<'~<'-s>-^>~i>~|1-^>~(|-i>-,>^>~,>-,i~,i^
>~l
o
K 'ta
GOC<)CO-rt^Ot^t^»OTf*fJ*OO^lOlOOC:OOCO
lOlOCOOO<NOCO»OOOt^t^OOO^COlO
IN
00
•^
^^^^ddddddddddr-.— 1^« i;^
1 1 1 1 1 I+ + + + + + + + + + + + III
1 N
.S2
£
scsacseasssasaaaaas
"3
2
s
»Ol0MOc0(NOc0U3CDOt^t^00O»-«C0i0»O
B
— .rt^^dddddddddd'-'.-ir-i-Hrt
1 1 1 1 1 I+ + + + + + + + + + + + +
<
IS -
<
CO
oooooooooooooooo ooo
Ocs050S0505CiOiOiOOOOOOOOOO
EESSSSEEEESEEEEEESE
i
r^
2*^
oco^oor^'-'ooaJNOcioioO'^t^oo^cco
OCTiOl0500t^O-*(NO<Mrt<tOt^OOOOOO
H
<-iOOOOOOOOOOOOOOOOO-<
lllllllll + + + + + 4- + + +
fc]
bl(iJ^(HmKltiJ(ilbl(HKimbJb)b3bJli3(Um
§
2 >.
■*coNooro>nm^too5D.-iin>ocooo<N«>'«<
1
lllllllll + + + + + + + + +
K|lilkJtel^^u;iiJ(i|Klbjbjtlklb3fejk5tiIbJ
05
s
0'-'C30tOO(NOO'J<00(NOO^tDOlOt^"5«0
OO(MiDO00OClt~-Tl<OOC5C:t^«Ot~00
■9>
c:
c
c
o --
" (N
•t
10
t--
05 O — Ol c
Tj* -^
Tt
■*
BO
t»5
K|
Kl :«!
« :«
«
W
«
b] :ct] Id] :«] lb]
bj :«
w
« :
^
o
00
IN
(N -lO
00 t-
CO .m
CD
00
<N
n
CD
lO
n
lO
•*
CD
■<1< "CD "CD "IN "in
05 'in '00 '^ Itc
')' n IN CD 00
CD in CO o , t—
m • IN
t- :oo
CI — •
lo n
00
IN
CO
S :
o
o
o -o
o ■-•
•-I
^^
,-i
_, ._ .^ ._< .o
o -o
O
o •
b4
bo
M :"m
e4 . ct
"«
"«
M : M :"« :■«-) :i«i
"« :"«
"ex
« :
o
S
n
u
o
o
00
o
O • 05
o -00
IN -00
03
05
CD
IN
IN
CD
CD 'IN • O -05 -O
•<»• ;co ■ t- -00 -o
— (N ■ -> '00 -IN
^ M o ;io -n
1^ -CD
■* -IN
2 :§
00
n
o
o
CO
05
o
o
o -o
o -o
O
'~'
'"
^ .-- .© -o -o
d -d
o
o ■
II
«0
Ci«
M . ao
M . M
u
<X
tu
r\A . ^* . ^^ , ^^ , Q|£
tx . ex
o<
B« .
N
«^
.-4
(31 O
<N '00
N 'CO
in t-
to ;o
in t~
o>
n
O
00
05 ■ t— '00 ■ t^ "in
00 'o 'in 'CO 'CD
i-t T-i 05 t— in
O • 05
00 • <N
CO O
in
o
o
o -o
o -o
o
—
r-t
^ .-. -c -o -o
O -O
O
o •
3
i«
c
IOC
IOC
>od>oc
(NCCCO-*
in o in c
Tr inincc
inc
inc
inc
inoinoioomoindiooincsiooiod
0500-^ — MMCOm'T'J'iOiOCDCDI-t^X
cq m
00
N
CO
d
II II
a.
II
■ei.
V
>
<
o
co^
Oco
_;c>
>
<
OO 05
00
>
(4
00
d
II
at
>
<
"o*
"m
s
228
05
o
8**
1(3 -I
> O
I II I
a
>
<
la
STEP-BY-STEP INTEGIL\TION OF CURVE AREAS OF PHASE SIGNIFICANCE 925
likewise proportional to the dotted mid-
ordinate at 15 deg. and so on.
Inspection of Fig. 4 demonstrates that by
this step-by-step method the rectangles
expressing the proportional areas of the
voltage wave for successive steps have been
given a leading shift, in respect to their true
position in angular relation to the voltage
wave, amounting to half a step or 5 deg.
Moreover, instead of occupying the eighteen
steps into which a half-cycle is divided, they
are condensed into seventeen steps, thus
leaving the eighteenth step open, and pro-
ducing discontinuity in the summation curve
Xe and in the magnetism wave B; also the
zero point of the latter has erroneously been
advanced half a step.
All the errors noted in the preceding step-by-
step method have followed from the manner of
summing up the area of the voltage wave.
Correct Method
Let US put the correct method (by taking
the mid-ordinates of the voltage wave steps,
as proportional to the areas of the correspond-
ing steps of the wave) to practical test, using
the same data to start with, as before.
First. Referring to Table III, %vrite in
coliunn (1) the angles of the eighteen steps in
a half-cycle, 10 deg. apart, between which
interpolate the mid-angles, to correspond
with the mid-ordinates of the voltage wave
steps, indicated by the dotted lines in Fig. 5.
Second. In column (2) write the relative
values Br of the voltage for the mid-angles,
as found in practice b}' scaling the mid-
ordinates on a diagram of the wave — ■
determined in the present case, however, by
equation (2). Then find the values given in
colvmans (3), (4), and (5), in the same manner
as before for Table II, writing the values of
2e in column (5) of Table III, however,
opposite the corresponding angles for 10 deg.
steps up to which, on the voltage wave,
they represent a value proportional to the
area of the wave up to the same angle.
Third. Continue in the same way as
before for Table II, to complete the remaining
columns (6) to (12) of Table III, noting that
the values of the angle ccb (to the nearest
minute) in colimin (S) closely agree with the
corresponding angles in column (1).
Fourth. From column (11) we have
/'= 1.0016 J
and from column (12)
P = 0.3572 /£
hence
P = 0.3567 /'£
= I'E cos d
from which
0 = 69° 6' 12"
that is, the equivalent sine wave of current
lags 69° 6' 12" behind the equivalent sine wave
of voltage.
And we have for the angle of hysteresis lead
,j° = 90°-69° 6' 12"
= 20° 53' 48"
Fig. 5
which by this correct method of handling the
step-by-step method is only ]/i per cent too
small, whereas by the incorrect method the
result was IS per cent too large.
The hysteresis current is
J' C05 5 = 0.3567/'
= 0.3567X1.0016 I
= 0.3573 /
a result that is practically the same as the
correct value, 0.3579 /.
The magnetizing current is
r sin 0 = 0.9.342 /'
= 0.9342X1.0016 I
= 0.9357/
which is only Vo per cent larger than the
correct value, 0.9338 /.
A still closer determination of the angle of
hj-steretic lead may be assured, and the more
conveniently in practice the more irregular
the shape of the voltage wave, by measuring
the areas included between the successive
voltage steps with a planimeter, having any
scale reading, and entering such measure-
ments in column (2) of Table III, opposite the
corresponding mid-angles in column (1), and
proceeding with the calculations for the rest
of the table after the manner already shown.
For example, the area between zero and
10 deg. in the present case, assumed to be
accurately determined with a planimeter,
would be 0.0359 g, which enter in the table
opposite the mid-angle 5 deg., and so on.
We have determined the hysteresis angle for
the present example on this plan with only
an insignificant error in the result, but deem
926 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
it unimportant to give here a table of the
work.
FORM FACTOR
Mathematical Solution
The form factor of a wave of voltage, or
current, is ordinarily taken as the ratio of
effective value to the algebraic mean or aver-
age value for a half -cycle between zero values
of the wave.*
To find the form factor of the voltage wave,
multiply (4) through by da and obtain
eda = 1.3868 E I sin ada — ~ sin 3a da I
whence, since e is zero, when a is zero or tt,
Av.e = l:^868£r rsinada-\ fsm 3q! Ja 1
= 0.8240 E
Hence the form factor 7, of the voltage wave is
E
'^' 0.8240 E
= 1.2136
To find the form factor of the current wave,
multiply (10) through by da and obtain
Av.t da =1.3720/
r5«M(a-69° •2\')da-^-xos 3 (a-69° 21')dal
which, as in the integration of the above
differential equation of voltage, must also be
integrated between the limits, -k distance
apart, that represent the angles where the
current wave crosses the time axis, that is,
where i is zero, therefore between the limits
57° 35' and 237° 35'.
We then have
, . 1.3720 7
A V . 1 =
■K
r ^37° 35' , -237° 35' T
I I nn {a-m° 21') da-\-\ ( cos 3 (a-69° 2\')da I
l-/57''35' "**^37°33' J
= 0.8971 /.
And the form factor 7; of the current wave is
/
'^' 0.8971 /
= 1.1147.
Steinmetz, however, considers that the
definition of form factor as the ratio of effec-
tive value to average value is undesirable
because it results in 1.1 107 as the form factor
for a sine wave, which being always taken as
the standard wave of reference, should, in his
opinion, be assumed to have unity form factor.
And to this end he has in effect adopted the
* A.I.E.E. St.indardiz.ation Rule, No. 16.
rule that the form factor is equal to 2\/'2/ir =
0.9003 times the ratio of effective value to
average value or, stated in another way, the
form factor is the ratio of the average value
of a sine wave, having the same effective value
as the actual wave, to the average value of the
actual wave. It is at once obvious from the
last definition, that the form factor is unity,
when the actual wave is a sine wave.
According to the Steinmetz definition we
have, as the form factor of the voltage wave
in the present example,
^0.9003 £
'''' 0.8240 E
= 1.0926,
and of the current wave,
^0.9003 7
■^'^ 0.8971/
= 1.0035
Step-by-step Solution
Considering the form factor as represented
by the ratio of effective value to average
value, we have from column (4) of Table III,
average ^ = 0.82-14 E\ hence the form factor
of the voltage wave is
E
'*'' 0.8244 E
= 1.2130,
which is only ver\' slightly smaller than the
correct form factor.
From column (10) of the table, neglecting
signs, average t =0.8990 /; hence the form
factor of the current wave is
/
'^' 0.8990 /
= 1.1123.
which is only 0.2 per cent smaller than the
correct form factor.
The Steinmetz form factor of the voltage
wave is
0.9003 £
'''' ~ 0.8244 £
= 1.0921
and of the current wave is
0.9003/
"^'^ 0.8990/
= 1.0015.
CONCLUSION
This article has dealt with calculations by
the step-by-step method apjilied to a par-
ticular case involving determination of phase
relation and some other characteristics of
certain interrelated periodic cur\-es. and has
differentiated the correct from the incorrect
method for that case.
STEP-BY-STEP INTEGRATION OF CURVE AREAS OF PHASE SIGNIFICANCE <J27
The obvious and useful conclusion there-
from, applicable to all cases involving the
interrelation as well as other characteristics
of periodic curves, is that the step-by-step
mid-ordinate, and not the mean-ordinate, rule
should be used in progressively integrating
the area of a curve, when such integration is of
phase or angular significance in the problem.
APPENDIX
General Equation for the Magnetism Wave in Terms
of Maximum Magnetic Density
The equation for the wave of magnetic density
may be derived from that for the impressed voltage
wave, if the resistance of the primary circuit is
relatively so low that the voltage drop due to the
exciting current is a negligible percentage of the
total voltage (which will be the case in well-designed
transformers), and therefrom the latter is sub-
stantially equal to the counter induced voltage, but
of opposite sign.
Write the well-known equation for induced
voltage
— ^^"
^'~ lOM/
in which (j),, is the number of magnetic linkages
between the primary windings and the lines of
force, the minus sign indicating the opposition of
this voltage to the impressed voltage producing
the current and magnetism by which the induced
voltage is generated. From this equation and the
known relation d a—2iTfdt
_ 2Tvfd<t,„
*"'" lOMa
d<t>„ = —
thus
lO^e, da
27r/
10' r J
which, obviously, may at once be recast in the form
B=-bBmax.Jeida (16)
where B is the instantaneous magnetic density
corresponding to the time angle a; B,„„.,- is the
maximum density in the magnetic cycle, and 6 is a
constant, to be evaluated.
The general equation for a periodic impressed
voltage wave is
e = A\ sin a-l-.43 sin 3 q: + • ■ ■ • -|--Si cos a-t-Ba cos 3 a
-I-.... (17)
which may be written
e = Ci sin(a+^x) + Cz sin (3a-|-^3)-|-
+ C„sin(na-\-p„) (18)
in which the sine and cosine terms of equal fre-
quency, in (17), have been combined.
Since the counter induced voltage e/is equal to the
impressed voltage, but of contrary sign, from (18),
we may write
e,= — Ci sin ( a +/3i) — fs sin (3 a -\-^i) — . . . .
-C„ iin (Ka-t-(3„) (19)
From (16) and (19)
B=bBmax.f[Cisin {a+0,)da + C3Sin (^a+0z)da +
. . . . +C„ sin (na+P„)da],
['
= -b Bmax. I Ci COS ( a -l-^i) -1-yCOS (3 a +0,) + .
+ —cos{noi + 0„)\+C
n J
(20)
To find the angle, or angles, for which B is of
maximum or minimum value, place the first differen-
tial coefficient of B with respect to a equal to zero, or
dB
bBn
Ci sin ( o -1-/3,) -I- Cs sin (3 a -|-03 -f •
-\-C„ sin {na+^.i
= 0,
that is, B is either maximum or minimum when
Ci sin {a+Pd + C, sin (3 a -f /Ss) -|-
-\-C,.sin(na-\-0„)=O (21)
thus, when
o-|-/3i, 3a-f ^3, etc. =0, or t, or 2ir, or3]r, etc.
From (18) and (21), and (19) and (21), it follows,
that e and ei are zero, when B is maximum or
minimum, as we otherwise know from physical,
independently of mathematical, considerations must
be the case.
The second differential coefficient of B with
respect to a is
^ = i Bmax. r C, f OS ( a -f /3,) -i- 3 C, f OS (3 a -t-^3) -I-
which reduces to
J'B
da-"
-\-nC„ cos {na+fi„) I,
-bBmax. (Ci-l-3G-h. . . . +nC„),
when a+/3i, 3a4-ft, etc. = tt, or 3ir, or 5ir, etc., and
is negative, and therefore B is of maximum value,
and (20) then takes the form
B„
- b Bmax. \CvCOs{a+0i) + y (-0S(3 O -1-^3) -I-
['
c„
']
= ftB„
. + — COS {na+0„) +C,
n J
which obviously holds, when the integration con-
stant C = 0, and
,[c, + f + ....+§l
thus when
6= ,^ TT. (22)
=]
= 1,
G+f+.
+
C.,'
Hence, from (20) and (22), and remembering that
C = 0, we may finally write the equation for the
wave of magnetic density in terms of the maximum
density,
rj _ Bmax. |-
cT. rc:,\c^cos{<x+0i)
G+f+.
■ +-
G
+ ^cos(3a+03) + .
d-'B
C„
+ ^cos{nct + 0„)\ (23)
Similarly, -3—, is positive, and therefore B is of mini-
o a-
mum value, when a+0u 3a-f-^3, etc. =0, or 2?r, or
47r, etc., and we then have B = Bmin.= —Bmax., vec-
torially. The minimum scalar value of B is 0, when
a +01, 3a+0z, etc. = 1 or '"l^, or y, etc.
928 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
Performance and Life Tests on the Oxide Film
Lightning Arrester'
By N. A. LouGEE
General Engineering Laboratory, General Electric Company
The development of the oxide film lightning arrester was first announced in a series of papers before the
A.I.E.E., in June, 1918. These papers were prepared only after extensive laboratory tests and a few installa-
tions on commercial circuits had demonstrated the merits of the device. Further laboratory tests and the
performance of an increased number of oxide film arresters on transmission lines during the thirty-month
interval have proved conclusively that this form of lightning arrester will satisfactorily fulfill the requirements
of service. The behavior of the oxide film arrester during these laboratory tests is recorded in this article.
The condition of one cell after four years actual service is also illustrated. — Editor.
design, which permits of an indoor setting.
Since the first papers on the oxide film
(O F) lightning arrester were presented a
little over two years ago,^ the arrester has
proved to be a worthy piece of apparatus by
performance in regular ser\'ice. Several
hundred arresters are now installed on both
indoor and outdoor circuits up to 73-kv.
rating, and higher voltage units will soon
be in service. Figs. 1, 2, and 3 show the
typical designs used. In Fig. 3, the three-
phase legs and the ground leg are all arranged
in one stack, the bottom section being the
ground leg. In Fig. 1, the three-phase legs
are the upper sections and the ground
leg is the lower section. In Fig. 2, the
three-phase legs and ground leg are set up
parallel to one another. Fig. 4 shows the
covered sphere gap used with the outdoor
1 Read before A.I.E.E.. Chicago. Nov. 12. 1920.
!"The O F Lightning Arrester," General Eectric Review,
1918; Vol. XXXVII, Transactions A. I.E. E,
..„, ._ _._ 1918.
'"The Effect of Transient Voltages on Dielectrics II
F. W. Peek. Jr., Transactions A.I.E.E., Vol. XXXVIII.
by
^rv
Due to the small leakage current of these
arresters (about 0.010 amperes), it is not
necessary to use horn gaps to aid in breaking
the arc, and it is therefore possible to use
the covered sphere gap which has previously
been described.^ Fig. 5 shows the testing
device used and its method of operation,
about which more will be said later.
The efficiency of a lightning arrester is gov-
erned by four factors; namely, sensitiveness,
discharge capacity, reseal. and life.
Sensitiveness
As most electrical apparatus is tested at
twice normal voltage, an arrester should be
able to begin discharging at about this volt-
age. This means a horn or sphere gap should
not be set for over double voltage for best
results. To care for steep wave impulses the
time lag of the arrester should be a minimum.
Fig. 1.
Oxide Film Lightning Arrester for Indoor Service on
Three-phase Circuits, 15.000-25.000 Volts
Fig. 2.
Oxide Film Lightning Arrester for Indoor Service on
Three-phMc Circuits. 37,000-50.000 Volts
PERFORMANCE AND LIFE TESTS ON OXIDE FILM LIGHTNING ARRESTER 929
Current Discharge Capacity
To discharge the energy from a surge, the
discharge path must be of sufficiently low re-
sistance to prevent the voltage drop being
above the insulation strength of the appara-
plished the better it will be, for if an arrester
has sufficient discharge capacity, the dynamic
or line frequency current that follows will
not only be apt to destroy the arrester but
may also cause' bad disturbances on the line.
Fig, 3. Oxide Film Lightning Ar-
rester for Indoor Service on
Three-phase Circuits,
5,000-7,500 Volts
tus connected to the line. Again, since a
double voltage test is given to apparatus, the
discharge capacity of an arrester is usually
given at double rated voltage.
Fig. 5. Oxide Film Cell Testing Device in Position for Testing
Reseal should also permit an arrester to be
ready immediately for another discharge, for
with a lightning storm over a large area of
transmission lines it is fair to assume that
impulses and surges can occur e.xtremely close
together; that is, at least a second apart, and
sometimes several per second.
Life
It is difficult to exactly define what the
life of a satisfactory arrester should be, but
a good arrester should easily withstand the
average surge or impulse. Arcing grounds
are the most dangerous type of discharges
and as they vary greatly in severity, depend-
ing upon the system and just where they
occur, it is difficult to state how long an
arrester should care for one.
Fig. 4. Covered Hemisphere Gap as Used on Outdoor Type Oxide Film
Arrester, 50,000-73,000 Volts. Section of Cover Omitted to Show Gap
Reseal
Reseal is the act of cutting off the dis-
charge path through the arrester when the
voltage across the arrester has returned to
normal. The quicker this can be accom-
TESTS
The following results of tests show how
the O F arrester fulfils the requirements
outlined above. A single cell was used
in all these tests in order to obtain as
powerful discharges through the cell as
possible with the power available.
The first set of tests was made with
a circuit as shown in Fig. 6. The usual
surge circuit was used, which superim-
poses the 25,000-volt, 2.300-cycle surge
on the dynamic 300-volt, 47-cycles cir-
cuit. Fig. 11 shows an oscillogram of
the discharges of an O F cell on this cir-
cuit. Vibrator 2 shows the dynamic
47-cycle voltage across the arrester with
the 25,000-volt, 2300-cycle surge super-
imposed. The voltage peaks are kept at
about double voltage and the cell reseals with-
out permitting any dynamic current to follow ;
that is, this test shows that reseal and sensi-
tiveness are satisfactory. Although the dis-
charge through the cell is about 50 amperes,
930 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 11 J
Fig. 6. Circuit Connection Used for Surge Tests
,4
^
AUTO TA^AfSf^O/fMf/P
\t=--
'*'O^C£-£L
Fig. 7. Circuit Connection Used for Double- voltage Surge Tests
Fig. 8. Oxide Film on Circuit in Fig. 7
Vibrator 1. Current through arrester at 600 volts, 1 mm. =100
amperes
Vibrator 2. Voltage across arrester, 1 mm. =22 vo!ts
Vibrators. Current through arrester at 300 volts, 1 mm. =5
amperei
Fig. 9. Inside of Electrodes of Oxide Film Lightning Arrester
Cell Returned from l3.000-voU Installation After
Four Years of Service
Fig. 10. Oxide Film Cell on a 600-volt, 40-cycle Circuit
Vibrator I.
Vibrator 2.
Current through arrester. 1 mm. =100 amperes
Voltage across arrester. 1 mm. =22 volts
Fig. 11 Oxide Film Cell on Circuit in Fig. 6
Vibrator 1. Current through arrester. I mm. =5 amperes iix.i
value)
Vibrator 2. Voltage across arrester. I mm =8'> volts iihm
va'uc)
Fig. 12. Arrester X on Circuit in Fig. 6
Vibrator 1. Current through arrester. 1 mm. =5 amperes (peak
value)
Vibrator 2. Voltage across arrester, 1 mm. =85 volts (peak
value)
Fig. 13. Arrester Y on Circuit in Fig. 6
Vibrator 1. Current through arrc«itcr. 1 mm. -.1 amperes (peak
value)
Vibrator 2. VohaKC across arrester. 1 mm. -8o vqU« (peak
value)
PERFORMANCE AND LIFE TESTS ON OXIDE FILM LIGHTNING ARRESTER 931
since the surge is supplied by a 15-kv-a. trans-
former, this test is not enough in itself to
demonstrate that the discharge capacity is
satisfactory- .
Fig. lU shows an oscillogram taken with
600 volts, 40 cycles (double standard voltage)
impressed across an 0 F cell, to show current
discharge capacity. The current peaks are
3500, 4200, and 3300 amperes, and the voltage
peaks 110, 89 and 154 volts respectively, giv-
ing an internal resistance of 0.031, 0.021 and
0.047 ohms respectively. Due to the low
resistance of the O F cell and its relative
value to the impedance of the circuit, the
impressed voltage of 600 was not sustained
across the coil when the high current flowed.
This discharge capacity is extremely high and
should be ample under all conditions. The in-
ternal resistance of a cell will vary between 0.01
and 0.1 ohm, depending upon the particular
path followed by the discharge through the cell.
Fig. 7 gives the connection used for a
double voltage surge test with normal volt-
age immediately following. This is accom-
plished as shown by bringing out a tap from
the transformer at 300 volts (standard volt-
age) and connecting it through a low resist-
ance to the arrester cell. The resistance is
necessary- to prevent the lower section of the
transformer from becoming short circuited.
With this connection, 600 volts is supplied to
the arrester until the fuse opens, and the
lower half of the transformer then being cut
off, 300 volts is continued across the arrester
cell. This is about the most severe test that
can be given a lightning arrester and only an
arrester which has a low breakdown, good
current discharge capacity, and good sealing
characteristics will act satisfactorily.
Referring to the oscillogram taken on this
circuit, shown in Fig. 8, the switch impress-
ing 600 volts across the cell closed at the
extreme left. The cell immediately broke
down and discharged 2700 amperes. This
current after one-half cycle blew the 10-
ampere fuse, thereby cutting off one-half of
the transformer, and causing the voltage
across the cell to drop to 300 or normal.
There was then a sealing current of about
2 amperes for several cycles shown by vibra-
tor 3, which caused the small breaks in the
voltage wa\-e. After a few seconds the cur-
rent through the cell had dropped to normal
or a few milliamperes.
To show the relation of protection and
current discharge capacity, oscillograms were
taken of single O F cells with external re-
sistance in series, on the circuit of Fig. 6.
A' represents an arrester with a medium in-
ternal resistance and a discharge capacity at
double voltage of 60 amperes. \ ' represents an
arrester with a higher internal resistance and
a discharge capacity of 20 amperes at double
voltage. Fig. 12 shows an oscillogram taken
Fig. 14. Circuit Connection Used in Intensive Life Run Tests
with arrester A', and Fig. 13 an oscillogram
taken with arrester Y on this circuit. It will be
noted that the voltage peaks with A' arc 1600
and with Y 3650, as against 900 with the
standard cell, which was shown in Fig. 11.
Moreover, if the frequency were nearer what
is obtained in actual ser^-ice, that is, from
10,000 to 100,000 cycles instead of 2300 cycles
which had to be used for oscillographic work,
this difference would have been much greater,
due to the higher impedance of the trans-
former at the higher frequencies. Therefore,
to give satisfactory' protection an arrester
must have a good current discharge capacity
on double voltage and more than these A' and
y arresters show ; A' and Y also show the bad
effect of a poor ground connection.
Sensitiveness in service is limited by the
gap setting, but since no dynamic current
follows a surge discharge and the leakage
current is only a few milliamperes, this gap
setting can be small. The gap settings used
in ser\'ice correspond to line voltage so that
breakdown between phases is double voltage
and the breakdown to ground is 1.7 times the
voltage to ground. Since the covered gap is
used for outdoor installations a dr>' or indoor
setting can be used.
The life of a lightning arrester is a very
important factor, and one that has to be
estimated from both operating and labora-
ton- data. Operating data obtained during
the past five years show that little deteriora-
tion has occurred in the O F cells. Cells have
been returned and tested from typical installa-
tions, and little if any change has been found.
Fig. 9 is a view of an opened returned cell,
and shows the film side of the electrodes and
932 November, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 11
the porcelain spacer. The lead peroxide
(PbOa) filler has been removed. This cell
was returned recently from a 13,000-volt
arrester installed early in 1916, which has
been subjected to much more than average
service due to its location and surroundings.
The few white spots in the photograph are
discharge areas covered with yellow litharge
(PbO). This PbO area or plug is what has
caused the cell to reseal after the surge has
passed through and is reduced from the Pb02
filler by the heat of the current through the
small discharge spot in the film. The larger
dark areas are where some of the PbO-j filler
is still adhering to other discharge areas, and
the light background is the varnish film. The
lead peroxide filler showed no change.
To obtain information on the life of O F
arresters several years ahead of outside re-
ports, however, an intensive test has been
running during the past few years. Fig. 14
gives the general scheme of circuit used.
In Fig. 14, the surge circuit is shown to the
right and consists of the usual inductance,
capacitance, and air gap, used to obtain oscil-
lations. The 50,000-volt transformer charges
the condensers, which, upon breaking down
the air gap set for a little under 50,000 volts
cause the surge through the arresters. The
transformer to the left supplies the dynamic
60-cycle voltage to all the arresters running
on this particular voltage. Ordinarily all the
lever switches are down. Fig. 14 shows only
one particular voltage, and only one arrester
and one set of switches. With this arrange-
ment the arresters are separate from the
surge circuit. When it is desired to surge
any one particular arrester, the upper lever
switch corresponding to this arrester is
thrown, thus paralleling the two transformers
supplying dynamic voltage. The lower lever
switch is then opened and this particular
arrester is still on dynamic voltage, but also
on the surge circuit, which can now be thrown
on. After surging, this arrester is thrown
back on the regular dynamic transformer and
the next arrester put through a similar opera-
tion. This arrangement of transformers and
switches pennits the regular dynamic voltage
to all the arresters to be uninterrupted during
surging operations.
O F arresters were placed on 330, 2300 and
12,700 volts respectively at 60 cycles with
no series gap, and all arranged as shown in
Fig. 14. These arresters have been surged
daily, the surge current through the arresters
having a maximum peak value of about 50
amperes and dying down to about 20 amperes
at the end. This surge, having a surge im-
pedance of 370 ohms, is representative of an
actual surge on a line, except that the average
actual surge has a higher frequency and may
at times be more powerful. It has been
found, however, that the lower the frequency
of a surge, the more difficult it is for an
arrester to seal.
330-voU Circuit, Single O F Cells
These cells take from 5 to 75 milliamperes
leakage current and run at a temperature of
about 50 deg. C. It took about four years
to record a failure with these cells. A failure
then occurred by reduction of a sufficient
amount of the PbO.> to cause high internal
resistance and hence loss of protection. As
the voltage across the cells is. of course, al-
ways the same, this group of cells is more
permanent than the 2300 and 12, 700- volt
arresters. With these latter arresters the
voltage distribution across the various cells
may change. This adds one more variable to
the action of arresters consisting of more than
one cell.
2S00-volt Circuit, Eight 0 F Cells in Series
These arresters so far have acted about
like the single cells; that is, voltage distribu-
tion has remained normal. Voltage distribu-
tion is obtained by means of shunting vacuum
tubes, which break down or glow at various
voltages, across each cell in turn. It is the
same idea that is used to t^st the cells in
sen-ice as shown in Fig. 5. For sen-ice con-
ditions a vacuum tube which will glow at
about 1000 volts alternating current is used.
As the internal condition of a cell changes, and
more particularly the film, the voltage drop
when in series with a number of cells may
change. Although this is not an infallible
method of picking out poor cells in scn'ice,
it docs give a reliable indication in most
cases. For voltage distribution tests tubes
breaking down between 100 and 2300 volts
alternating current arc used. For conven-
ience in interpolating these data, a cell hav-
ing a voltage drop of less than 200 is desig-
nated low; from 200 to 400 nonnal. from 400
to 600 high, and above CiOO very high. The
results can then be plotted against the re-
spective cells by using a ditTcrent color for
each of these four groups. The units on 2300
\-olts ha^•e shown with one or two exceptions
only normal cells on voltage distribution, and
the few low or high cells, which have appeared
from time to time, have returned again to
normal. The leakage current of this group
of arresters varies between 1 and 10 milli-
PERFORMAXCE AXD LIFE TESTS OX OXIDE FILM LIGHTXIXG ARRESTER 933
amperes and the cells run at a temperature
of about 40 deg. C. A few units have failed
or lost their protection after four years of
continuous sen.'ice.
12,700-volt Circuit, Forty-seven 0 F Cells in
Series
These arresters have been running almost
two years with no appreciable deterioration.
To obtain the relative effect of dynamic
and surge, similar arresters were run with
different sendee characteristics, as follows:
(a) dynamic only, (b) dynamic and surge,
(c) surge only, and (d) idle. The leakage cur-
rent is from 5 to 10 milliampercs and about the
same through all the arresters. The tempera-
ture is about 35 deg. C. at the top of the stack,
45 deg. C. in the middle and 30 deg. C. at the
bottom of the stack. Results to date show
that (a) and (b) types of arresters give about
the same characteristics; that is, the daily
surge has no ill effect on the arresters. Both
(a) and (b) show a gradual tendency for low
voltage cells to appear at the bottom of the
stack and higher voltage cells at the top.
Here again no change has been found to be
absolute; that is, unless a cell is extremely
high, it may go from low to high and back
again. The low cells at the bottom of a stack
may be due to either capacity or tempera-
ture, but probably the latter, as all the cells
are about normal when first put on circuit.
The (c) and (d) types of cells show a general
scattering of high and low cells throughout
the stack.
This sort of intensive test has been found
extremely valuable in determining ahead of
time what might occur in servdce and also
for determining the effect of changes. So
far as applying to standard arresters in serv-
ice, it seems fair to assume that if an arrester
will stand up say four years under such an
intensive test, it will stand up several times
four years in actual ser\'ice. Of course, it is
always possible that a more or less direct
lightning stroke or a long arcing ground will
destroy an arrester, so that this conclusion
should apply to normal average service. As
yet the factor between test and actual serv'ice
is not known, but it should be determined
when longer ser\'ice results are available.
The author wishes to express his appre-
ciation to E. E. Burger for his valuable assist-
ance in obtaining the data used in this paper.
Oxide Film Lightning Aircstrr for Outdoor Service on Three-phase Circuits. 50.000-73.000 Volts
Ground stack shields removed for cell inspection and test
18
GENERAL ELECTRIC REVIEW
NOVEMBER. 1920
Where to Get G-E Service —
Quick service is best obtained from the nearest G-E
sales office, distributing jobber, or foreign representative
List of G-E Factories
Conn.c[ii
Til
Bndic«pon
WinJior
iDdiUU
Fl Wayne
B«lQmor.
Muucbiu
etU
Ev,t.n
GroncMier
HDdion
Lynn
Piiiifitid
MUmof
Sl Loai
Ne» Jer.<7
Ampff.
Btll.vill*
RomU* Park
Ne- York
N>« Vo.k Ciiy
Rocbrtlcr
Clf vfUnd
Sanduiky
PduuyUaaia
En*
PhiUdrlphta
\. Rhod. LUnd
V P'^--^i'r
^
For Basiness in tLe United States
G-E Sales Office G-E Dixtributins Jobber
Alabama. Blrmlni:riiini MattJiews Eire. Supply Co.
Alabama. Mobllel , , .Matthews Elec. Supply Co.
Arkansas. Little Rock
California, Los Anectest Parlftc Stai« Elcriric Co.
California. Oakland: Paflrtc States Klwt/lc Co.
California. Pan Franclacojt Pacific States Elecirlc Co
Colorado. Denver t Tbe Hendrle & Bolthon Mfg. ft
Sup. Co.
Connecticut. Hartford
Connecticut, New Haven
Cinnectlciit. Wat*rl)ur7l New Enitland Eng. Co.
DIstrlctorColiimlfla.WashlnEtoQ National Elec'l Supply Co.
Florida, Jacksonville . Florida Elec. .Supply Co.
Florida. Tampa: Florida Elec Supply Co.
Georsia, AUantaSt Carter ElecUlr Company
Georcla. Savannahl Carter Elecirlc Company
Illinois. ChlcagoSf ^ Central Electric Company
Commonwealth Edison Co.
Indiana, Fort Wayne
Indiana. Indianapolis. .Indianapolis Eter. Siinnly Co.
Intilana. South Dendt South Bend Electric Co.
Iowa. Dea Moines .Mld-Wesl Electric Co.
Kentucky, Louisville Belknap Hardware &. Muiatee*
turin? Co.. Inc
Louisiana, New Orleans Gulf states F:iectrle Co.. Inc.
Maryland. Baltimore Southern Electric Co.
Massachusetts. Boston f Pet Uocell- Andrews Co.
Massachusetts, Sprlngneld
Massachusetts. Worcester
Michigan. Ucuoit Frank C. Teal Company
Michigan. Grand R^iplda
Minnesota. Duluth Northwesui-n Electric Equip-
ment Company
Minnesota. Mlnneapollst Peerless Electrical Co.
Minnesota. St. Paul: NorUiwcstcm Elec. Equip. Co.
Missouri. Joplint
Missouri. Kansas City t The B-R Electric Co.
MLssnurl, St. Ixiulsjt Wesco Supply Company
Montana. Uuttet Butte Electric Supply Co
Nebraska. Oma&a MId-weti Electric Co.
New Jersey. Newark TrhClly Electric Co.. Inc.
New Jersey. Trenion , . .
New VorK, Albany: Havens Electric Co,, Inc.
INoG-n Office
WoritMat Erie. Pa. Growiwl arta tS
acres. Floor jpocc JJOOOfiOO «a. ft.
C-E Sales Office G-E Distribatins Jobber
New Vork. Buflato Rot>eri«>n-Catarac( Elec. Co.
New York , Elmlra
New York Clty|t .E. B Latham ft Comnaay
Royal Eastern Elec'l Sup. Co.
(Also Borooffh of Br«oUTB
and Jaataiea. L. I.)
Blbley- Pitman £iec Corp.
New York. Niagara Falls
New Ynrk. tlochestef . Wbeekr-Greeo Elec'l Sup. Co.
New York , .■^cheneclady
New Vori:, s>Tacuse Mohawk Elec'l Sup. Co.
North Carolina. Cbarlotle Eler Sup[dy ft Fqulpmettt CO.
Ohio. ClnclnnaUt The F I) Lawrence Elec. C©.
Ohio. Cleveland Republic Electric Co
Ohio. Columbus The Emer A Hopkins Co.
Ohio. Dayton Tbe Wm Hall ElectrleCo.
Ohio. Toledo W. C Nagel Eleeir*c Co.
Ohio. Youniatown
Oklahoma, oklahome Cliyt Southwest C-E Co.
Oregon. Portland t PaclOe SUtee Eleetrte Co.
Pennsylvania. Erie
Pennsylvania. PbltadHptilalt PhlladHphla EleetrV Conpear
Supply Deparureot
PennsyU-anla. Plttuburght L'nion Electne Company
Rhode Island. Providence
South CaroUna, Columbia; Perry-Mann Elec. Co.. iDO
Teoneasce. Chattanooga Jamei 8upply Company
Tenoeawe. Knoivllle
Tcancaace, Memphis Electric Supply Compear
Tenoenee, NaahrlUe
Tetas. l>allajt Southwest G-E Co.
Texaj. El Paaot Southwest C^E Co.
Texas. llouJtont Southwest C-E Co.
l"ian '*a1t 1.ake Cliyt Capital Elec trie Compear
Vir.' fii^ ir vr, .Tinri'l ...SouthrCT) Electric Compeaj
''^ ttiet PadOc Siatei El«rirte Co.
.ine
"ina .
\\ . I tcfleld
\^ 'Til V ir,-inia «. harlntoaT
Wianinsln. .Milwaukee ...
For tiawalun bu*lne« aditreM Cstton. N«1U ft Coaipear. LM.
Honolulu
tWarehoiJse. fSerrke Sbop.
Diatnbulora for the General Electric Company Outside of ihe United States
INTERNATIONAL GENERAL ELECTRIC COMPANY. INC.
120 Broadway, New York, N. Y. Scbcoeclady. N. Y. 83 Cannon Street, Loodoo
Foreign Offices and Repeesentalivee
Greece aod^ Colonic*: Compagnle FraacelM
Argentina; General Electric. S. A.. Buenos Aires.
Australia: Australian General Electric Co., Ltd.. Sydney and
MelOournc.
Belgium and Colonies: Socieie d'Electrlclte et de Mecanlque
Proredes Thomson-Ilousion A. C&rels Soclete Anonyme.
Brussels.
Bolivia: International Machinery Co.. La Pai and Oruro.
BrasU: General Electric, S. A.. Hlo de Janeiro and Sao Paulo.
Canada: Canadian General Electric Co.. Ltd., Toronto.
Chile: International Machinery Company, Santiago, Anto-
fagaata and Valparaiso.
China: Aniier*en. Meyr A Company. Ltd , Shansbal;
International General Elecirlc Co. Inc (General repre«enu-
tiveeof the For East, excluding China and J aiiaiO Shanghai.
Colombia: Wrssclhueft ft Poor, Medetlln, liarranquilla and
Bogota.
Cuba: General Electric Company of Cuba. Havana.
Dutch East Indies: International General Electric Co.. Inc. ,
Socrabala, Java,
Ecuador: Carlos Cord<>ves, Guayaoull and Quito.
Egypt: British Thomson-Houston Co.. Ltd., Cairo
France and Colonies: Compagnle Francalse Thomson-Houttoa.
Paris.
Great Britain and Ireland: BrltUhThomson-Hou9tonCb..Lt(].
Rugby; International Ueocral ElectrleCo.. Inc.. London.
and Colonic*:
uouJiou, Paris. France.
India: Hrliuh Thornton-Houston Co.. Ltd.. Calcutta and
Itumbay. International General Electric Company, lac
Co .
Ltd .
■aki.
I General Etecuie Co., City of Mexico and
Guadalajara.
New Zealand: National Electrical A Engineering Co., Lid.
Auckland. Uunedln. Cbrblctiurch. and Weillngtoo.
Paraguay: General Electric. S A., Buenos Aire*., Argentina.
Peru: \V . It Grace A Co . Lima.
Phlilnplne lalands: Pattflc I ommerclal Company. Manila.
Porio HIco: loiernatlunal lleneral Electric Co . Inc.. San Juan.
Portugal and Colonin: Compagnle Fraoc&lH TlMMasoo-llotaioa,
Parts. Irance.
Russia: \\ •mbshtchala ElecUlchcskala Kompaola. PHrograd
and Vladivostok.
South Africa: South African General Electric Co..Lld..J(^
hannesOurg and <.'a[>elo«n
Spain and colontrs: Compag&le Francalse Tbomeoo • UooMoa
Madrid. Spain.
t'Mi:ua> : General Eteeirlc, S. A., Moatevldco.
Vrii'iurla: Wesselhoett A Poor. Caracea.
(Dtrfl©
General Office
Schene<Jtad>^N.Y.
C©inni
.nj
Sales Ot'lices in
all laige cities
Say you saw it advertised in the General Electric Review
TWO DOLLARS PER YEAR
TWENTY CENTS PER COPY
GENERAL ELECTRIC
REVIEW
VOL. XXIII, No. 12
Published by
General Electric Company's Publication Bureau,
Schenectady, N. Y.
DECEMBER, 1920
300-TON ELECTRIC SHOVEL IN OPEN-CUT MINING
(Sec article, page 938)
For
Fractional H. P. Motors
NEVER yet has anyone arranged a successful
compromise between price and quality. Man-
ufacturers with an established reputation to main-
tain, never attempt it. Today — as for years past —
"NORfflfl Bearings are the accepted standards in hun-
dreds of thousands of high-duty, high-speed, electri-
cal machines which dominate their fields by virtue
of sustained high performance. It is a question of
quality — simply.
See that your Motors
are "NORm/q" Equipped
THE M^MM/M €^MF/^MY
ILs'InKol IglairiKo!
Ball, Rollci', Thrust and Combinalion Bearings
General Electric Review
A MONTHLY MAGAZINE FOR ENGINEERS
, ^„.^^^ Associate Editors, B. M. EOFF and E. C. SANDERS
Manager. M. P. RICE Ed.tor. JOHN R. HEWETT j_^ ^^^^^^ „f Advertising. B. M. EOFF
Subscription Rates: United States and Mexico. $2.00 per year; Canada. $2.25 per year; Foreign. $2.50 per year; payable in
advance. Library and Student Rates: United States and Mexico, $1.50 per year; Canada. $1.75 per year; Foreign. $2.00 per year;
payable in advance.
Remit by post-office or express money orders, bank checks, or drafts, made payable to the General Electric Review, Sche-
nectady. N. Y.
Entered as second-class matter. March 26. 1912. at the post office at Schenectady. N. Y.. under the Act of March. 1879.
Vol. XXIII. No. 1--' ,yGeferTlilfJrilTon,p.,.y DECEMBER. 1920
CONTENTS Page
Frontispiece: The Integrating Sphere Photometer ■ . . ■ • 936
Editorial: The 300-ton Electric Shovel in Open-cut Quarrying 937
The Electric Shovel in Open-cut Mining 938
By C. R. Fisher and H. G. Head
Automatic Substation for Alternating-current Railway Signal Power Supply. Part II . 949
By H. M. J.\coBS
Commercial Photometry. Part I 954
By A. L. Powell and J. A. Summers
Photo-elasticity for Engineers. Part II 966
By E. G. CoKER
One Hundred Years Since Oersted, Ampere and Arago 975
By Elihu Thomson
Dr. Elihu Thomson 983
By D. C. Jackson
Studies in Lightning Protection on 4000-volt Circuits 985
By D. W. Roper
Discussion by C. P. Steinmetz 998
J. L. R. Hayden 1001
V. E. Goodwin 1003
The Bowl-enameled Mazda C Lamp 1005
By Ward Harrison
a =■
0 •-
H
GENERAL ELECTRIC
REVIEW
THE 300-TON ELECTRIC SHOVEL IN OPEN-CUT QUARRYING
Carl D. Bradley
President and General Manager, Michigan Limestone & Chemical Company
About ten years ago the Michigan Lime-
stone & Chemical Company purchased a
tract of several thousand acres extending for
several miles along the shore of Lake Huron
and containing limestone which the company
intended to develop commercially for blast
furnaces, chemical plants, etc. The lime-
stone lay close to the water front and delivery
to steamers was comparatively easy and
economical with proper facilities.
In order to establish a market for its
product the company made sales contracts at
very low prices which required ver}- careful
consideration of all details relating to con-
struction and operation of the plant in order
that costs might be kept within the limits
prescribed by the selling prices obtainable.
Large scale operations were involved and
many engineering problems had to be solved.
The general problem was to drill, blast,
quarry and transport the stone to the mill
and there crush, size, wash and convey it to
storage, and thence load it into steamers.
Changes were necessarily made in the plant
from year to year, and the difficulty of
handling the great tonnage was finally over-
come by the installation of large crushers,
large screens and similar equipment. The
loading facilities have developed to a point
where steamers of 13,000 gross tons are
loaded in six hours, and the management is
convinced that theoretically the problem of
quarrying the limestone is no different from
that of handling and loading it. However, no
adequate means of getting large output from
open cut quarry operations at low cost had been
developed 1 and therefore attention has lately
been forced upon production at the quarry.
Quarrying operations are being conducted
against the natural bluff of limestone which is
now in excess of one and one-half miles long
and more than one hundred feet high, requir-
ing two benches. This bank is too high for
the economic and safe operation of the 100-
ton steam shovel, and because of this fact and
the high costs of labor and material the
management has become deeply interested
in the application of large digging and trans-
portation units which will permit the quarry-
ing operation to keep pace with the mill and
loading system. If a digging machine can
be had which will take care of 5000 tons of
material , in ten hours and operate satis-
factorily under this punishment day in and
day out, the problem is approaching solution,
with a resultant economy in all operations.
For the future the quarry will approximate
two miles in length in one face with five large
electric shovel units working against it,
served by locomotives and cars of comparable
capacity. One man will operate the shovel
and another the train, and the tonnage per
man hour will be multiplied by five over that
of present day equipment. With the intro-
duction of the 300-ton electric shovels, quarry-
ing on a large property such as that under
consideration is reduced to a scientific basis.
The modern trend in industrial develop-
ment has been toward increasing the effi-
ciency of the individual, or in other words,
the rate of commodity handling per man
hour; only by such a test have we the right
to measure accomplishment. While the
management has been able to satisfactorily
increase the rate in crushing limestone, in
screening, conveying and loading it, until
recently efficiency at the digging end has not
kept pace with that of other operations, and it
was specifically for the purpose of improving
this performance that the 300-ton electric
shovel was installed. The restdts that have
been accomplished by the new equipment
have been most gratifying, and it is firmly
believed that through the proper application
of these large electric shovels the quarry
operations will be made entirely satisfac-
tory.
The economical reasons that have dictated
the employment of these largest electric
shovels by the Michigan Limestone &
Chemical Company will be better understood
by a review of their operations as outlined in
an article by Messrs. Fisher and Head, pub-
lished elsewhere in this issue.
938 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 12
The Electric Shovel in Open-cut Mining
By C. R. Fisher
Electrical Engineer, Michigan Limestone and Chemical Company
and
H. G. Head
General Electric Company, Detroit
Coal deposits with over-burden too shallow to permit shaft mining now constitute the most profitable form
of coal mining when large electrically driven stripper shovels are used. The principal contributing factors are
the greater depth of cut which the larger shovel can make, the increased bulk of material which its greater radius
of action permits it to remove before taking out time to make a forward movement, and the lower labor and
power costs. The performance of the large electric shovel in stripping operations has naturallv led to its
adoption in other branches of mining, and this article describes the construction and operation of the latest
type 300-ton electric shovel as applied to open-cut limestone quarrying, where it has every promise of dupli-
cating its success in coal mining. — Editor.
The electric shovel is a new develop-
ment, the possibilities of which are being
recognized more and more by quarr\dng and
mining companies. Heretofore the use of
steam as the motive force for mechanically
operated shovels has seldom been questioned,
but the present high prices of coal, shortage
of labor, and general need of economical and
increased production are causing progressive
companies to scrutinize their operating con-
ditions ver\f closeh' and to seek ever\' method
which will increase the efficiencies of their
operating systems.
The development of the large electric
shovel has been made possible largely through
application in the middle west coal fields and
in the iron ranges of Minnesota. Up to a
few years ago the 100-ton steam shovel was
the largest excavator in general use, but
material handling has been greatly improved
by the adoption of large electrics.
In coal mining, the usual procedure was
for the sur\^eyors to go over all prospective
coal mines and buy up mining rights from
the owners. Wherever the coal had an over-
burden of 25 to 30 feet or more the option
was bought, but when the over-burden was
10 to 15 feet, the prospect was considered
uncommercial because of insufficient material
above the coal vein for safe shaft mining.
This meant that there was a large amount of
very good coal too close to the surface to be
mined. However, with the advent of the
large coal strippers, all of this coal land was
reclaimed and it is now the most remunera-
tive form of coal mining. The art has devel-
oped so rapidly that burdens up to 100 feet
in thickness have been removed from coal
seams 4 to 5 feet thick. In such cases, the
large 300-ton shovel is commonly used, fol-
lowed by a small friction electric shovel
which loads the coal into small dump cars.
The success of these large shovels in strip-
ping has led to their development in other
fields, such as excavating and loading ore
directly into the dump cars. The success of
large shovels for this work is due largely to
the great difference in the amount of material
available in front of the large shovel without
moving ahead. A smaller shovel, say the
100-ton size, would have to move ahead
about twenty times to handle the same
amount of material. This would necessitate
lengthening the loading track eight times,
whereas the large shovel would load the same
amount of material from one position.
Figuring in these delays of the 100-ton shovel,
the 300-ton shovel should dig approximately
twice as much yardage as the 100-ton shovel
per shift.
This article was written primarily to dis-
cuss the construction and performance of
the model 300 E Marion electric shovel with
General Electric automatic control, recently
installed in the Michigan Limestone and Chem-
ical Company quarry. In general the shovel
is of the large capacity type fitted with an
SO-foot boom to give a digging radius of
approximately 54 feet at the rail and 99 feet
at 40 feet above the rail. The present dipper
has a capacity of six cubic yards. Ultimately
an eight-yard dipper will probably be sub-
stituted. All of the electric equipment
except the crowd motor installed on the
boom is located in the cab, which is appro.x-
imately 50 feet long, 22 feet wide, and 15
feet high. The cab and boom revolves as a
turntable on a large square platform made
up of steel beams and plates and mounted on
four trucks for locomotion on eight 130-pound
rails.
The electrical machiner\- portion of the
equipment consists of one four-unit motor-
generator set with direct-connected exciter,
THE ELECTRIC SHOVEL IN OPEN-CUT MINING
939
a
O
>
940 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 12
two hoist motors geared to a common shaft,
one swing motor, one crowd motor and one
trip motor. The ratings of the units of the
generator set are :
Synchronous motor ATI, 6 poles, 435 kv-a ,
1200 r.p.m., 2300 volts, 3 phase, 60 cycle.
Hoist generator MPC, 6 poles, 250 kw.,
330 volts no load, 250 volts full load, differen-
tial compound wound.
Direct connected exciter EC-5, 4 poles,
20 kw., 125 volts, flat compound wound.
Swing generator and crowd generator each
MPC, 4 poles, 50 kw., 330 volts no load,
250 volts full load, direct current, differen-
tial compound wound.
The ratings of the direct-current motors
are:
Two hoist motors each ]MDS 109, 6 poles,
175/140 h.p., 400/450 r.p.m., 230 volts.
Swing motor MDS 106, 4 poles, 75/55 h.p.,
485/550 r.p.m., 230 volts.
Crowd motor MD-106, 4 poles, 70/25 h.p.,
500/900 r.p.m., 230 volts.
The hoist, swing and crowd motors are
series wound. They are doubly rated, the
larger horse power value being the inter-
mittent rating at the lower speed, the smaller
horse power value the continuous rating at
the higher speed. The trip motor is rated
KTE, 8 pole, 50 lb. torque, 900 r.p.m., 220
volt, 3 phase, 60 cycle. This motor is of the
high resistance squirrel cage rotor type.
The two hoist motors are connected in
parallel and supplied with power by the
hoist generator. By a system of gears the
hoist motors are used for hoisting the dipper
by means of cable and drum, for propelling
the entire shovel outfit on its track to new-
digging positions, or for raising or lowering
the boom. The swing motor is supplied with
power by the swing generator and is used
for revolving the cab and boom on the turn-
table. The crowd motor is supplied with
power by the crowd generator. This motor
operates the dipper stick through a system
of reduction gears, pinions and racks. The
trip motor operates the trip mechanism on
the bottom of the dipper by means of cable
and drum.
It should be noted that the control of this
new electric shovel is radically different from
that of the old type which consisted of series
motors all operated from one constant poten-
tial generator, speed variation being taken care
of by a resistance in series with the armature
of these scries motors, which resistance was
short circuited in steps by contactors to
produce acceleration. These motors were
protected by stalling relays and resistances
which automatically kept the current down
to a proper value whenever the motor was
stalled. It is evident that this type of control
meant a heavy energy loss and lack of smooth
operation. On the new equipment, the
crowd, swing, and hoist motors are each
supplied with power from generators of their
own and the speed control is entirely through
variation of a resistance in the shunt field of
each generator similar to the Ward Leonard
control on battleship turrets and high duty
mine hoists. Due to the inherent protection
features of the control, no protection of any
kind is necessar\' as the differential series
field of each generator bucks the generator
voltage to nearly zero when the motors
are stalled at maximum torque. The
operation with the new equipment is much
smoother than with the old type and does
away with the sudden stresses in the hoist
cable. These stresses are ver\- objectionable
as they shorten the life of the hoist cable
^"erv materiallv.
p£2!!2IHR}— I
Fig. 5. Diagram of Main Connections Between the Four-unit
Motor-generator Set and the Hoist, Swing, and Crowd Motor*
The diagram. Fig. 1, shows the main con-
necting circuits of motor-generator set and
motors The direct connected exciter is
used to excite the fields of all the generators
and the synchronous motor. The resistances
in the fields of the hoist generator, swing
THE ELECTRIC SHOVEL IN OPEN-CUT MINING
941
generator, and crowd generator are varied
by means of the three master drum control-
lers located at the operating station, thus
permitting the operator a great range of
control over the generator voltages, con-
siderable variation of which must be had
during process of operation to give torque as
required on the motors. The reversing
switches shown with the hoist motors are
operated manually by the operator by means
of one of two levers located at the operating
station. These reversing switches reverse
the armatures of the motors with respect to
the fields, thereby reversing the rotation of
the motors. The rotation of the motors is
reversed in this way only when their power
tor for this motor and also the time limit
relay are energized by means of a push
button. When the push button is pressed
the contactor closes, short circuiting the
permanent resistance in the stator circuit,
thereby giving full line voltage and full
tripping torque to the motor. The time
limit relay prevents the motor being left on
full line voltage any longer than is necessary
to operate the trip cable. If the push button
is held in too long, the time limit relay trips
the contactor and releases the trip motor
from the full line voltage. When the con-
tactor opens, the permanent resistance is
connected in the stator circuit of the trip
motor. This resistance allows enough cur-
Fig. 6. Diagram of Connections for Crowd Motor Control
is used to move the shovel as a locomotive
backward or for^vard on its track. The
swing generator is connected to the swing
motor through either of two pairs of con-
tactors depending upon the direction of
rotation desired for the swing motor. The
crowd generator is connected in a manner
similar to that of the swing generator. The
trip motor shown at the bottom of the dia-
gram, the air compressor motor, and the
lighting circuits are supplied with power
from three T^-kw., 2200/220-volt trans-
formers connected to the incoming line.
The trip motor is of high resistance squirrel
cage rotor type with permanent resistance
connected in the stator circuit. The contac-
rent to flow from the line into'^the motor to
give just sufficient torque to the motor to
take up the slack of the tripping cable and
maintain a small torque on the drum of the
tripping cable at all times except when the
motor is operating on full torqiie to trip the
dipper load.
The common countershaft of the two
hoist motors is furnished with a friction band
pneumatic brake normally spring set. The
crowd and swing motors are furnished with
a similar brake. The hoist drum is furnished
with two brakes, one similar to the type used
with the hoist, crowd and swing motors and
another operated directly through a foot
lever. The pneumatic brake on the hoist
942 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 12
drum is a revolving brake operated by air
introduced through the center of the hoist
drum shaft. This brake locks the hoist drum
to the gear drive of the hoist motor counter-
shaft and is disengaged when the empty
dipper is falling and when the hoist motors
1
150
■
'
/30
1
\
5
1
120
j^
,
\^
1
j
■K""
V
1
1
Is
,
,
1
1^0
»
1
(
1
s
1
■c ao
»
s
V
\
i
-c >^
"v
>•>
h
7,,
1 60
s
r
'
S
\
f
1^0
-
-
-
-
f^
^
J
'h/J Dipper
riTiTr
Y
1-
^40
'^
<
V,
MM:
1
-
-
L
-
-
-
-
:s
s
l1
t»
S
-
mn
r
r'^
er
■^
f*
X
,
1
30
'
V
s
'
■^
1
1
"■
^.
■^
^
; ■
■^
•k
1
'S
0
5
0
B
0
/
^/
/J
3C
0
ec
1-
f
ee
■)0
je
r
"0
n
jt
e
JC
10
J.
S"0
«.
X)
Fig. 7. Curves Comparing the Bail-speed Bail-puJI Characteristic
of Series Motors with that of a Steam Engine at Two Pressures
are used for propelling. This
brake is controlled by a sole-
noid valve from one reverse
point on the hoist motor
drum controller and from a
double pole knife switch.
The foot brake is mounted
on the hoist drum opposite to
the side on which the pneu-
matic brake is mounted. This
brake enables the operator to
catch the bucket before it hits
the pit on the down drop.
The period of maximum
regenerative braking is that
period of time during which
the hoist motors are driven
by the hoist cable as series
generators when the loaded
dipper is allowed to descend
while it is swinging from the
digging position to the dump
car. During the period of regenerative
braking, the foot brake of the hoist drum is
not used and its pneumatic friction brake is
in a locked jjosition.
The pneumatic brakes are all of the same
general type consisting of a friction band of
the action of a powerful spring. The friction
band is loosened from the clutch surface by
means of an air ram which acts against the
tension of the holding spring when air is
admitted under pressure by means of a
solenoid operated vahe energized on the
first point of the master controllers.
The regenerative braking feature of the
equipment is a very interesting one. This is
accomplished mainly when the dipper is
being dropped with load to the dumping
position. Thus some of the power con-
sumed in hoisting the dipper is returned to
the line, thereby increasing the operating
economy. When regenerating, the dipper
in its descent exerts torque on the hoist drum
through the hoist cable and drives the hoist
motors as series generators, their direction of
rotation being opposite to that when operat-
ing as motors for hoisting. As soon as the
hoist motor shaft reverses direction of rota-
tion a switch is closed by means of a small
belt from the shaft. This switch closes the
small equalizer contactor mounted at the
bottom of the hoist generator panel, thus
connecting the equalizer circuit and per-
mitting parallel operation as series generators.
When these motors arc operating as genera-
tors the shunt field of the hoist generator has
ioo -too 600 too
--" - "X "^ — IIX
t ;- - -- - - ___"±"'"±±
■■rtrt ""^ -^
! i — "ti i " it "
^ ■
'~>_ -^-i~-4.-^-+~^-f--^- -4--i-
^DO I
"■^ . i<-'' :: r^-c i.'U^
^^T^i-' -'-"'"> ■ ,1 "1" 1 "
^X^III^T ---t-^
sxitii^ ~ ^i .:
isit It i
-- -- X- i - --^^ -^-Z-^-^-Z
^^0 ^-- ""-":-"::
.''>_
j^^OO -- -- - 5 5 _- -
-" mn : ' :'','>rr>'>i'
ltd > yttor dtntr o :cr s , 1
0 - -fc^^ 2i_
Choractc-i.a.l ' " "'STi"^ '
t ,60 _ _- !p^^ii_ _
^ 1
t "" :::: — x s; ::
V , „ «,■*». ^
s — "" "T::t:x\ i"'
'^ , pn - - - - ^r-.
Pi* 1 , . , . \, . ,,.
r *».± T XT^ , i ^
O .
^'^° i'L^or^^arccicr
2-::;- = ;^"^ i T±T "-^--T '
QQ :_ ^i'lj^:
* si. 1 t ■ ' -J ■ — -
f* ^t^ \
wocorir^i£^ii^^-r-t-'c;
2Q
-■ TTTTTTrTtT: = s;T
_
M 1 1 1 t t t t ~> J
HOC
1500
400
1300
leoo
1100
'000
>00
too
700
too
soo
400
300
zoo
I l^co /40J
/K'O ISCC ^AO
wood blocks which are normally set through
Fig. 8- Characteristic Curves of the Hoist Generator. Motors and Their Combination
all its resistance cut in, causing a shunt field
so weak that the differential series winding
is predominant and the hoist generator
operates as a scries motor, exerting its torque
on the shaft of the synchronous motor-
generator set. This regeneration occurs only
on the first point for\vard on the hoist con-
THE ELECTRIC vSHOVEL IN OPEN-CUT MINING
943
trollcr. As the lever of this controller is
moved over to the second, third and fourth
positions, the shunt field of the hoist gen-
erator is strengthened, the hoist motors
stop regenerating, their shafts come to stand-
still, reverse, and then start up in the reverse
direction and the motors begin hoisting.
This action occurs onee every cycle and for
the time of its duration returns to the system
approximately the losses of the motor-
generator set and the power required to
operate the swing motor. This action,
together with the energy stored up in the
revolving elements in the motor-generator set,
keeps the power peaks down to a low figure.
The perfonnance of the swing, hoist and
crowd generators and motors is controlled
by the three master drum controllers men-
tioned above. The hoist controller and crowd
controller are operated by means of hand
levers. The swing controller is operated by
two foot pedals, one pedal for one direction
motor gears through a clutch for hoisting the
boom. The crowd controller and swing con-
troller give reversible operation, but the
hoist controller gives one way operation only,
although one reverse point on the hoist con-
troller releases the revolving pneumatic
brake on the hoist drum.
On the crowd motor, protection is supplied
by interlocks which are operated by a limit
switch on each end of the dipper stick motion
which opens the magnetizing circuit of the
contactors whenever the dipper handle is
driven too far at either end. This is the only
electrical protective feature which is installed,
all the other protection being inherent in the
design of the equipment.
The shovel proper is supported by four
trucks, one at each comer. Each tnick is
driven by a shaft with bevel gear from the
hoisting mechanism. This gives very flex-
ible control of the shovel proper and allows
a very quick move up. Each truck is set into
Fig. 9. Four-unit Motor-generator Set, Consisting of a Synchronous Motor, Exciter, and Three Direct current Generators.
The Latter Individually Supply the Operating Motors with Power
of swing and the other pedal for the reverse
direction. The controllers operate mainly
on the field rheostats of the swing, hoist and
crowd generators, but the first point of these
controllers operate the shunt contactors of
the contactor panels which close and open
the main circuits between the swing, hoist
and crowd generators, and their respective
motors. This point of the controller also
energizes the solenoids of the pneumatic
brake valves, thereby releasing the particular
motors controlled for work. On the lever
of the hoist controller is located the push
button for operating the control of the trip
motor. This tripping feature does away with
the serv'ices of one man always needed on
steam equipments to trip the dipper. Two
levers are located at the right hand of the
operator. One is used for connecting gears
to the hoist motor gears through a clutch to
accomplish locomotion of the shovel. The
other is used for connecting gears to the hoist
a hydraulic cylinder of large dimensions
similar to a step bearing. While in operation
this cylinder is pumped full of oil and the
shovel practically floats, giving automatic
alignment and cushioning against shock from
digging. Each truck runs on two 130-lb.
rail sections put together in short lengths to
allow for a small turning radius and carries
a load of approximately 85 tons, which is well
within the factor of safety for reliable opera-
tion.
The hoist controller is used for propelling
after the rotating drum brake is disengaged
by opening a double pole knife switch.
Reverse motion is secured by knife switches
in the field circuits of the hoist motors.
The shovel is served with 2300-volt, 3-
phase power from the plant power house at
a present distance of approximately 4000 feet.
The shovel is grounded by means of a ground
wire in the filler of the supply cable. This
ground wire is carried back to the generating
944 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 12
COMPARATIVE COST OF OPERATION, TEN-HOUR BASIS
(August 6, 1920)
First Cost 300 Electric
100 Steam
J13o.000.00
$45,000.00
Interest 87c— 440 10- Hr. Shifts
24.55
8.40
Depreciation 7%— 440 10-Hr. Shifts
21.87
io<r^
10.22
Maintenance 5%— 440 10-Hr. Shifts
15.61
Actu
al 16.85
Oil Waste Supplies
2.00
5.00
Coal 70 Bags 80 Lbs. 5600 Lbs. @ S9.20 Ton
0.00
25.80
Electric Power
15.00
Water Lines and Equipments
0.00
5.00
Labor Engineer
10.00
10.00
Crane Man
0.00
7.00
Oiler
5.00
0.00
Fireman
0.00
5.00
Pitmen (2)
10.00
(4)
20.00
Lighting
1.00
5.00
Operating cost for 10 hours
$105.03
$118.27
Cost per carload 100 ton steam 100 per cent
10-hour tonnage 100 per cent
Cost per carload 300 ton Electric 51.5 per cent
10-hour tonnage 172 per cent
These figures are digging cost — do not include large sa\'ings such as track labor, elimination of second cut, etc.
(August 4, 1920)
No.
of
Train
Cars
Began
Finished
Total
Time
Train
Loading
Loading
Loading
Min.
0
7:00
1
6
7:08
7:32
24
2
6
7-AQV2
8:0414
24
3
6
8:11
8:331^
2214
4
6
8:42
9:07^
2o}4
5
6
9:16
9:41
25
6
6
9:47
10:14Ji
27 J^
7
6
10:20
10:42 M
22 J^
8
6
10:54
11:20
26
9
6
11:26}^
U-AQH
23
10
1^2
11:57
12:02
0
Actual
Time
Loading
Min. /Sec.
Total
Delays
Min./Sec.
NATURE OF DELAYS
Waiting
for Trains
Min./Sec.
25:10
8:00
23:20
8:30
21:53
6:15
24:40
7:59
25:04
7:58
27:00
6:05
22:49
5:38
23:30
12:33
22:51
6:55
4:45
8:15
00:00
8:00
5:45
6:30
5:38
6:05
4:58
5:33
5:55
6:15
Cleaning
Tracks
Min./Sec.
:30
1:30
:40
3:00
1:00
2:00
Shovel
Min./Sec.
•8:00
:30
(Move up Ties) — 2:20
(Move up) — 4:00
Miscel.
Min./Sec.
Average
Angle
Shovel
Swing
120
120
130
120
120
150
145
90
110
Lunch Hour
10
iV?
1:04 H
1:21^
17
18:29
4:30
•4:30
110
11
6
1:30
1:55
25
24:23
8:53
8:53
120
12
6
2.00}^
2:27
2614
26:30
5:32
5:32
140
13
6
2:34
3:05
31
31:05
6:29
5:14
(Move up Ties) — 1:15
160
14
6
3:12
3.36
24
24:32
6:32
6:32
1
150
15
6
3:47H
4:11
23 H
23:30
10:01
6:30
(Move up)— 3:31
110
16
6
4:18
4:40
22
22:26
6:35
6:35
100
17
6
5:01
5:23
22
21:42
22:00
22:00
130
18
6
5:29
5:53
24
24:30
5:30
5.00
:30
130
* Inspection, etc.
RECAPITULATION
Stop watch time loading .6 hrs. 57 min.
Delay on inspection 12 'i "
Delay on trains (practically all shuttling) 1 hr. 55
Delay on mo\^ng up 7 HH "
Delay on ending day at 5:53 7 "
Delay on cleaning tracks 9 "
TOTAL DAY 9 hrs. 32 min.
Unaccounted for (personal errors, false starts, etc.) 28 "
TOTAL 10 hrs
Total cars loaded 108
Power consumption 1050 Kw-hrs.
Total digging time .... 70 per cent
THE ELECTRIC SHOVEL IN OPEN-CUT MINING
945
station, where it is effectively grounded. It
is evident that the shovel cannot be well
grounded locally on account of the nature
of the limestone formation. The 3-wire
transmission line is carried across the quarry
on poles up to a point near the shovel, where
it enters into a 3-conductor cable which is
wound on a drum attached to the shovel
turntable support.
Reference to the diagram, Fig. 2, will
show that the series motors give a bail-speed
bail-pull characteristic very similar to that
of the steam engine. The 120-lb. steam
The cycle of operation of the 300-ton
shovel with an SO-foot boom and G-yard
bucket will average about 55 seconds. This
is on the basis of a ISO-deg. swing, and load-
ing cars on the same level as the shovel.
Loading cars on top of the bank would
facilitate this operation, probably cutting it
down to 40 to 45 seconds. The working
cycle of the small 100-ton shovel with Sj^-
yard bucket is considerably faster, being
from 22 to 27 seconds. This cycle, however,
is based on a 100-deg. swing only, as this is
the usual arc of operation. With a SJ^-yard
Figs. 10, 11, and 12.
Dipper-trip Panel, Hoist Panel, and Crowd or Swing Panel Used in the Control of the 300-ton
Electrically-operated Shovel
pressure engine curve is the one which should
be used in the comparison, as this is the
pressure most usually maintained in practice.
It will be noted that the full dipper pull is
at the same speed for either engine or motors.
At lighter loads the dipper speed is slightly
lower with motors than with the engine,
while at heavy load the dipper speed is much
faster with the motors than with the engine.
It will also be noted that the motors have
much greater stalling pull at very heavy
loads, which is a very desirable characteristic,
provided it does not exceed the mechanical
strength of the shovel.
dipper load swinging an^angle of 100 deg. the
loading operation of the 100-ton shovel would
be very fast if the material were always
available in front of the shovel, but due to
the small radius of digging a great portion of
the time is used in moving ahead and a
considerable time in moving the loading
track. Also, the loading track is so close to
the bank that the small shovels seem to be
best served by shuttling trains to them. On
the larger shovel the operation can be con-
tinuous, as the loading track is entirely
clear of the bank. The larger shovel actually
operates on an average swing of about 120
94G December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 12
deg., which shortens the duty cycle to ap-
proximately 45 seconds or better so that it
will load one train slightly faster than the
smaller shovel, but at a swing of 120 deg.,
instead of 100 deg. The great gain by the
larger shovel is made in the time saved
through the necessity of the smaller shovel
moving ahead about twenty times and
changing the loading track about eight times
to handle the same amount of material.
This gain is so marked that at the end
of 10 hours' operation the larger shovel
show better results than this, as considerable
delay is experienced with the smaller shovels
due to slides from the bank occasioned by
blasting and by storms. The smaller shovels
must necessarily operate very close to the
bank on account of their small digging
radius and they are therefore very much
exposed to these bank slides. Besides the
loss of time occasioned by the slides, in-
creased repairs must be charged against the
smaller shovels due to them. The larger
shovel with its .54-foot digging radius at the
Fig. 13. Crowd Motor Input Current. Full Scale Equals 1000 Amperes. Interval Between Curved Lines IS Seconds
Fig. 14. Hoist Motor Input Current. Full Scale Equals 3000 Amperes. Interval Between Curved Lines 15 Seconds
delivers approximately 50 to 70 per cent ton-
nage excess over the smaller shovel under
unfavorable loading conditions. It is con-
fidently expected that later when loading
conditions become more favorable and an
8-yard dipper is substituted for the present
6-yard dipper the larger shovel will double
the output of the smaller one. During a
season's operation the larger shovel may even
rail stands well clear of the bank and is in
no danger of damage from it.
The operating cost of the electric shovel
shows marked advantage over the steam
shovel both as regards labor and fuel con-
sumption. While the comparison of costs
between a 300-ton electric and a 100-ton
steam shovel is manifestly unfair to the
electric, nevertheless such comparison shows
THE ELECTRIC SHOVEL IN OPEN-CUT MINING
947
interesting results. For example, operating
crews are as follows:
300-TON ELECTRIC I
1 Shovel Runner
No Craneman
No Fireman
2 Pittmen
1 Oiler
4 Men Total
100-TON STEAM
1 Shovel Runner
1 Craneman
1 Fireman
4 Pittmen
No Oiler
7 Men Total
less than 5 per cent transmission loss from a
highly efficient steam turbine plant. The
steam turbines operate condensing with
consequent low water rate and the steam
they use is derived from boilers designed
for high coal economy. The engines on the
steam shovels operate non-condensing with
full cut-off and their boilers cannot be de-
signed for high coal economy. A pound of
coal burned on the power house grates,
therefore, is much more effective in power
production than when burned under the
Fig. 15. Swing Motor Input Current. Full Scale Equals 1000 Amperes. Interval Between Curved Lines 15 Seconds
l>o U S. «■ NO. 2310.C
Fig. 16. Kilowatt Input to Shovel. Full Scale Equals 600 Kilowatts. Interval Between Curved Lines 15 Seconds
In addition to these there are men as-
signed to water line attendance and to coal
supply service. This attendance and service
is divided among eight steam shovels but
would easily add one more man chargeable
to each steam shovel.
The electric power for the 300-ton shovel
is obtained economically with low fuel
consumption, as this power is obtained with
steam shovel boiler and a great saving in
fuel for the same amount of power is the
result. Anothei> great source of gain in fuel
economy is due to the fact that the electric
shovel consumes ^-ery little power when not
in use, while it is necessar\' to maintain steam
at all times on the steam shovels.
The operating costs of the 300-ton electric
and the 100-ton steam shovel were com-
9481 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
puted by starting with the original invest-
ment in each case and charging against each
shovel, interest, depreciation, upkeep, labor
and fuel cost. Electrical power cost was
figured at IJ/^ cents per kilowatt hour and
coal at $9.20 delivered at the shovel. In
Fig. 17. Foot-operated Master Switch for Controlling the
Swing Motor
spite of the heavy handicap of much higher
first cost of the large electric as compared to
the small steam shovel, the detail figures
show the operating cost per shift to be ])rac-
tically the same. As stated above, for fair
comparison the same size electric and steam
shovels should be compared, so that this
result is a striking one. Considered another
way, the result shows that every ton dug in
10 hours by the .SOO-ton electric over the
100-steam shovel is dug at practically no
cost, and the cost per ton of all material
dug in the 10 hours is correspondingly de-
creased. For example, if the large shovel
digs 50 per cent more material in ten hours
than the small shovel, the cost per ton dug
is 663^ per cent.
The elimination of the second cut in the
bank was one of the most important con-
siderations in choosing the large shovel in
preference to the small one. The consequent
saving in track expense will prove a great
asset. This is made possible due to the fact
that the large shovel can work against banks
as high as 100 feet or more, which is inadvisa-
ble with a small shovel, as explained in a
previous paragraph, on account of danger
from bank slides.
Even now more progress is being made in
the perfection of the electrical control for
shovels, such that future installations will
undoubtedly show even better results than
Fig. 18. Hand-operated Master Switch for Controllinc the
Crowd Motor
at present. The probabilities are that they
will also be found advantageous in other ap-
plications as operators become more familiar
with their characteristics and excellent show-
ing in operation.
949
Automatic Substation for Alternating-current
Railway Signal Power Supply
Part II
By H. M. Jacobs
Railway Department, General Electric Company
The preceding installment of this article described alternating-current railway signal substations in which
the source of reserve power is a second commercial service. The present installment describes substations
that rely on storage batteries for emergency service and these are of two classes; namely, those substations
that demand uninterrupted service, and those that will permit of a short interruption of one or two minutes.
With the former class the motor-generator set must float continuously on the storage battery and the control
equipment must effect undisturbed operation when commercial power fails. Two motor-generator sets are
necessary. In the latter class of substation the converting apparatus stands idle until power goes off, when
the control equipment operates to start the motor-generator set in the prescribed time. — Editor.
In the preceding article, railway signal
automatic substations for providing against
prolonged failure of the signal system due to
failure of power supply were divided into two
general classifications, and substations in
which the reser\'e power is a second com-
mercial source either at the same station or
at some remote point were discussed in detail.
This article deals with substations relying on
storage batteries as the reserve source.
Although there is no reason why this class of
automatic substation is not applicable to the
supply of power to automatic block signals,
Fig. 1. Power Equipment for Alternating-current Signaling Installation on New York
New Haven fit Hartford Railroad, Worcester. Mass.
every installation that has come to our
attention has been at interlocking plants
where the movements of trains are under the
control of the tower man.
If traffic conditions are such that even a
momentary loss of power would be serious,
the converting apparatus must float con-
tinuously on the storage battery and the
control equipment must be so arranged that
the power feeding the signal system is
undisturbed when the commercial power fails.
If a slight interruption is permissible, say
one minute, the converting apparatus may
stand idle until the failure occurs ; the control
equipment must be arranged to start the
converting apparatus and connect it to the
signal power feeders within the prescribed
time.
To meet the first condition requires a motor-
generator set consisting of an
induction motor, an alter-
nating-current generator to
stipply power to the signal
system, and a direct- current
machine which acts as a gener-
ator for charging the storage
battery or as a motor for driv-
ing the set from the battery
when the power supply to
the alternating-current motor
fails. The generator may be
either self-excited or excited
from a direct-connected unit.
A duplicate motor-generator
set is necessary. In order to
take care of the change in
field current of the d-c.
machine when changing from
the generating to the motor-
ing condition and vice versa,
two field rheostats are pro-
vided, one for each con-
dition. These rheostats are connected in
series, and the one not required is short
circuited automatically by the control
equipment. A speed regulator is connected
permanently to the field rheostat which
governs the motoring condition in order to
9.50 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 12
maintain constant frequency on the a-c.
machine.
Fig. 1 shows such an equipment installed
at Worcester, Mass., on the New York, New
Haven & Hartford Railroad. Only one of
the motor-generator sets appears in the illus-
tration. As these sets are started by hand-
controlled starting compensators, it is neces-
sary to reconnect the motors to the power
supply on resumption of power after a failure.
This can be taken care of automatically by
substituting auto-starters for the hand-con-
trol type. With this arrangement the sets
will automatically shift from battery opera-
tion to induction motor operation. The
constant speed as before the failure of power,
so that the frequency and voltage of the
alternating-current generator is not affected.
At many places a slight interruption in
sendee is permissible; at these locations the
converting power apparatus may stand idle
until the failure occurs. The control equip-
ment must be arranged so as to start the set
in the shortest possible time. The equip-
ment required will depend largely on the
character of the signaling equipment with
which it will be used. Some interlockings
use alternating current for all functions,
whereas others have alternating -current track
circuits and lights but require a storage
rt^
g
@fe:
rOl
a
<^
r^
Mm
Tf
J©j^ ' — Ca- -r o L r««>i i*e< Of 6o*«»-
=n
-^:
Fig. 2. Wiring Diagram of Automatic Substation Equipment. Boston Switch-Central Falls Signal Tower.
New York, New Haven & Hartford Railroad. Pawtucket, R. I.
O
~M
automatic control employs a low voltage
relay energized from the a-c. supply con-
nected to some magnetically operated switches .
When power is available, the relay is ener-
gized and causes a magnetic switch to short
circuit the d-c. motor field rheostat; the d-c.
generator field rheostat is adjusted for the
desired current for the battery. When power
fails, the relay is de-energizcd ; this opens the
short circuit of the aforementioned field
rheostat and short circuits the other. A speed
regulator is connected across the motor field
rheostat now cut in circuit, making the direct-
current machine motor the set at the same
battery for certain other functions. For an
all a-c. plant the battery will be used only for
furnishing emergency power to the motor-
generator set. The battery may be charged
either from a separate motor-generator set.
a mercury arc rectifier, or the emergency set
running "reversed." that is. the alternating-
current generator operating as a s\ nchronous
motor and the direct -current motor operating
as a generator. Although the control equip-
ment is complicated by using one set for both
the emergency power supply and for charging
the battery, there is a considerable saving in
cost of cqui]imcnt and floor space.
AUTOMATIC SUBSTATION FOR A-C. RAILWAY SIGNAL POWER SUPPLY !)31
Two such equipments have been installed
on the New A^ork, New Haven & Hartford
Railroad, one at Pawtucket, R. I., in 1915,
and the other at Stamford, Conn., in 1917.
The motor-generator set at the former loca-
tion is designed to deliver single phase,
60-cycle alternating current at 120 volts, .')
kv-a., 0.7 p-f., with a 3-kv-a. intermittent
overload capacity of short duration. The
direct-current machine operates as a motor at
constant speed with the voltage of the storage
battery from 110 down to SO volts; as a
generator it will charge the battery at 50
amperes between 110 and 155 volts.
The control equipment is arranged to
fulfill four conditions;
(1) On failure of the commercial power
supply when the set is at rest, to dis-
connect the power supply, start the set
from the battery, and connect the a-c.
generator to the signal buses at normal
voltage and frequency.
(2) On return of power supply, to dis-
connect the set from the signal bus
and the storage battery, and reconnect
the power supply to the signal bus.
(3.) On failure of commercial power supply
when set is charging the battery, to dis-
connect the power supply and change
the field current of the two units of the
set so that the d-c. generator becomes
a motor and the synchronous motor
becomes an a-c. generator, and connect
the latter to the signal bus.
(4.) On return of power supply, to notify
the operator so that he can synchronize
the a-c. generator with the supply, re-
Fig
4. Motor-generator Set. Boston Switch-Central Falls Signal
New York, New Haven & Hartford Railroad, Pawtucket, R. 1
Fig. 3. Switchboard, Boston Switch-Central Falls Signal
Tower, New York, New Haven fit Hartford Railroad,
Pawtucket, R. I.
connect the latter, and either continue
charging the battery or shut down the
set by hand. Automatic synchronizing
does not seem of sufficient import-
ance to warrant the added complicated
control that would be involved.
From actual test, the elapsed
time to fulfill condition (1) was
13 to 15 seconds. The signals
" cleared " in 5 seconds. Hence
the total interruption from a
traffic standpoint is 20 seconds,
or less.
The interruption for fulfilling
condition (2) was so short that
the semaphore arms on the
signals did not drop to the full
"stop" position, but merely
"bobbed." Conditions (3) and
(4) are fulfilled with no interrup-
tion since the set is running and
energy is not cut off the line at
any time.
Fig. 2 is a wiring diagram of
this equipment and Figs. 3
and 4 show respectively the
952 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII. Xo. 12
switchboard and motor-generator set. A com-
plete description of this equipment and its
operation appeared in the General Electric
Review, January, 1916, under the title
"Power Equipment for Alternating Current
Signaling at Interlocking Plants."
The equipment at the Stamford, Conn.,
plant is similar but is of larger capacity
and provision is made for charging the bat-
tery from a 600-volt d-c. source in case
of extreme necessity. When the set is operat-
ing from the batterv, the a-c. machine will
simple. We will describe one of several plants
of this nature recently installed on the
Philadelphia & Reading Railroad, which is
illustrated in Figs. 5 and 6. Fig. 7 is the
wiring diagram of the switchboard. Power is
supplied to these plants from a three-phase,
60-cycle, 4400-volt aerial transmission line
and is stepped down by transformers mounted
on the pole structure, to 110 volts. The
storage battery is charged in the customary-
manner from a motor-generator set having
a three-phase induction motor started by a
^^^^^^HiT'^^'^'^^^HHhfV'^^ r ^^9^P^ <^
1
^^^^H^~ '.-'V-
To
(J
1
^^^^^Ht/
E
« ;I^^H
E»«
.'L'!ik 1
E
'{nA I,-;,
■^ ' "-"•■ ^ ' f a
Fig.
5. Switchboard in a Signaling Substation, Philadelphia
& Reading Railroad
deliver 12kv-a., .53 power-factor continuously
and 20 kv-a., .67 power-factor for 30 seconds;
when charging the battery the d-c. machine
will deliver S5 amperes or less between 115
and 175 volts. By actual test the set started
from rest and was connected to the signal bus
in 5 seconds — less than half the time required
for the smaller outfit at Pawtucket, R. I.
For interlocking plants using alternating
current for track circuits and lights, and a
110-volt storage battery for the operating
functions, the automatic equipment is very
IM
Fig. 6. Battery Charging Motor- generator Set and Alternating-
current Emergency Motor-generator Set in a Signaling
Substation. Philadelphia 8k Reading Railroad
self-contained compensator fitted with a
low voltage release. Should power fail while
charging the battery, an underload circuit
breaker will disconnect the generator from the
battery, and the compensator will disconnect
the motor from the line.
The signal bus, from which all signal
circuits are supplied, is normally connected to
one phase of the 110-volt power supply by a
double pole magnetically controlled contactor
switch energized from the source through the
"front" or upper contacts of a control relay
AUTOMATIC SUBSTATION FOR A-C. RAILWAY SIGNAL POWER SUPPLY 953
also energized from the power source. The
control relay has two sets of "back" or lower
contacts, one set being connected to a d-c.
auto starter between storage battery and the
motor of the emergency motor-generator set,
and the other connected to another double-pole
contactor switch energized
from the a-c. generator of the
emergency motor-generator
set to connect the latter to
the signal bus. When power
fails, the control relay is de-
energized and connects the
motor to the battery through
the auto starter, and the coil
of the other contactor switch
to the generator. When the
set has attained speed and the
voltage of the a-c. machine is
sufficient to energize the con-
tactor switch to the pick-up
value, the latter closes and
connects the generator to the
signal bus. The generators of
all the sets are rated 2 kv-a.,
0.6 power-factor, 110 volts,
ISOO r.p.m., 60 cycle, single-
phase and the motors are o
h.p., 110 volts, shunt wound
direct current. They restore
energy to the signal bus in
after power failure.
The switchboard is simple to operate and
compactly arranged. The automatic equip-
ment and starter are mounted on the sub-
panels and the instruments and hand switches
on the upper sections. All the field rheostats
are mounted on the back of the switchboard.
The alternating-current load on the emer-
gency generator is so constant that after
once adjusting the field rheostat for proper
voltage, when the set is running at normal
speed, no other adjustments are necessary.
For this reason this rheostat is not arranged
for front of board control. The frequency
of the a-c. generator may be varied by ad-
justing the field rheostat of the d-c. motor.
A frequency indicator and voltmeter are
mounted on one of the panels.
At some interlockings the switches and
signals are operated by compressed air. The
air valves are controlled bv direct current.
The track circuits and lights are sometimes
operated by alternating current. Such a
plant requires air compressors and battery
charging equipment. The control battery is
usualh' 12 to 16 volts. A power failure would
tie up .such a plant even though the switches
Battery
Twm?-
DCMotor^ ^
V '/' uifhts
IIOVA.C Signa/
Feeders
Am. Ammeter
CB. Circuit Breaf(er
F. Fuse
F.I Fraqut-hcy Indicator
PR Potzntial Plug 5h.Field-i\ ■
Re PeactancK
Rh Rheostat O.COanerator
He. Peceptacli
Rs Resistor
Sh Shunt
Sw. Switch
V. Voltmatcr
•nikY - 1 l!0V3PhaSi
" ' -2. toO Cycle
"J- Supply
LF
\Co'npertsator
Fig. 7.
Typical Wiring Diagram of Switchboard and Automatic Substation Equipment
for Railway Signal Power Supply, Philadelphia & Reading Railroad
2 to 3 seconds
and hand-controlled signals could be operated,
because the track circuits and lights would
be dead. If the air supply is taken from a
source not dependent on the commercial
power supply, delays due to temporary power
failure may be eliminated by installing a larger
ampere hour capacity low voltage battery
than ordinarily required, and emergency equip-
ment to operate from the battery. Two
such plants were installed several years ago
at Jamaica, Long Island, N. Y., on the Long
Island Railroad. The air is taken from the
railroad shops. Only one motor-generator in
each plant is used, the motor being a-c.
synchronous type, which acts as a generator
when running reversed.
If an inexhaustible air supply cannot be
obtained, the reservoir capacity may be made
large enough to supply the demand over
a long period, and the system can operate
without a-c. power until either the air supply
or the storage battery becomes exhausted.
954 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 12
Commercial Photometry
Part I
B}^ A. L. Powell and J A. Summers
Edison Lamp Works. General Electric Company
Satisfactory illumination for any given condition is largely dependent on physiological, and to a lesser
extent on psychological considerations. Neither of these factors is measurable and the only means
we have of gauging the degree of illumination is by the scientific process of photometry. The value
of proper illumination is now generally appreciated, and to a corresponding extent have the methods of meas-
uring illumination attained importance. This article is a complete review of the subject of photometry,
including descriptions of photometers, their principles, calibration, and use. The correct interruption of a
set of readings is necessary in practice and ample explanation is included on this point. — Editor.
SYNOPSIS
Part I
Introductory^
Elements of a Photometer
Means of Var>'ing the Light
Means of Comparing the Light
The Standard Lamp and Calibration
of the Comparison Lamps
Measuring Horizontal Candle-power of
Light Sources
Measuring Spherical Candle-power or
Total Light of lUuminants
Determination of the Distribution of
Light
Selection of Equipment for Test
Calculating Results from Test Readings
Portable Photometers and Their Cali-
bration.
Part II
Illumination Tests of Interiors
Brightness Measurements
Measurement of Reflection Factor
Illumination Tests on Single Units
Rough Distribution Determination with
Portable Photometer
Street Illumination Tests
Projector Tests
Introductory
Distribution cur\'es of lighting devices,
or the results of photometric tests, are often
misinterpreted. In view of their apparent
complexity, they do not receive the attention
they justly desen-e. Other items, such as
appearance and cost, are often given undue
weight in making a decision.
If the art of lighting is to advance on firm
ground a study of the qualities of equipment
to be used is necessary'. The fixture man-
ufacturer can well afford to spend some time
and money analyzing the properties of the
glassware he employs. A purchaser of any
large amount of equipment should insist on
knowing its performance. This is particu-
larly true where similarunits are in competition.
A knowledge of the fundamental principles
of photometry is of great assistance in inter-
preting results of the tests. It is the purpose
of this article to point out some of the
features often overlooked. The section on
Selection of Equipment for Test indicates
some of the diversified factors which must be
given consideration, and well warrants care-
ful study.
The subject of photometers is treated in
greater detail than any other phase of illumi-
nating engineering in text books on lighting.
It is therefore unnecessary to discuss minutely
the theory of photometry or photometric
instruments. On the other hand, there are
many phases of light measurements which
do not come under the category of laboratory
methods, and which only come to the atten-
tion of the investigator through actual ex-
perience. It is worth while, therefore, to
briefly discuss some of these features and to
describe the actual procedure necessan,' to
satisfactorily operate the photometric device.
The fundamental quantity which we de-
termine in photometn." is the strength or
power of luminous flux. Intensity expressed
in candle-power is the flux per unit solid
angle, while illumination expressed in foot-
candles is the flux per unit area.
The eye is very sensitive to light and is
the basic instrument, but the unaided eye
cannot determine, with any degree of ac-
curacy, the absolute intensity of light or
illumination. It can, however, determine the
equality of brightness of two illuminated
areas provided they are contiguous and not
too dissimilar in color.
Elements of a Photometer
The elements of a photometer are : A means
of obtaining adjacent fields, a means of vary-
ing the intensity of illumination on one or
both of the fields, and a standard light source.
COiM.MERCIAL PHOTOMETRY
955
Means of Varying the Light
A number of means are used to vary the
intensity of illumination on the photometer
screen, namely;
The distance of the standard light source
of the test lamp from the screen mav be
varied, the law of inverse
squares holding true. Such
a method as this is most
common.
A revolving sector disc may
be interposed between the
light and the photometerhead.
the proportional size of
opening determining the
amount of light transmitted.
A diaphragm or absorbing
media may be interposed.
The angle of incident light
may be changed by the use
of an inclined plate, the inten-
sity being proportional to the
cosine of the angle.
The candle-power of the
standard may be changed by
varying the voltage applied
to it.
These methods may be used
in conjunction with each other
and the tj'pical commercial
photometer uses at least two
of the possible schemes for
obtaining the desired varia-
tion. For example, a stand-
ard bar photometer is so
arranged as to vary the dis-
tance and also permit the
insertion of a rotating sector
disc. A typical portable
photometer varies the dis-
tance and supplements this
with neutral absorbing screens
of known transmission, permitting a wide range
of measurements with a given standard lamp.
Means of Comparing the Light
The development of the photometric head
is interesting from historical standpoints.
One of the first schemes was employed by
both Lambert and Rumford. An upright rod
was so placed in reference to the lamps under
test that it cast two shadows on a white back-
ground. (Fig. lA.) When the position of
these lamps was so adjusted that the shadows
appeared to the eye of equal density, a
photometric balance was obtained, and the
familiar law of inverse squares applied to
calculate the ratio of intensitv.
Ritchie employed a triangular shaped prism
with the apex toward the obser\-er, one face
being illuminated by the standard lamp and
the other by the lamp under test. When the
balance occurred, the two faces appeared
equally bright.
(A)
<B)
(C)
(D)
(E)
Fig, 1. Historic and Modem Photometer Heads
The Lambert photometer
The Bunsen screen in plan and elevation
Plan of Lummer-Brodhun head
Section of Lummer-Brodhun prisms
Type of field produced by the Lummer-Brodhun prisms
In the Joly-Block, two rectangular prisms
of translucent substance, such as paraffin
or milk glass, are placed side by side with
a very thin opaque diaphragm between
them. If one block is then lighted by each
of the lamps under test, the front of each
block is seen illuminated by the internally
diffused light from its respective lamp,
and the equality of brightness can be readily
observed.
The Bunsen sight box has probably had
the widest use of any type of photometer
head. Mirrors are arranged so that both
sides of a screen can be observed at the
same time. (Fig. IB.) The screen is made
of white opaque material, usually paper, with
9.:6 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
a sharply defined translucent spot, usually
made with paraffin, in the center.
The Leeson disc is a modification of the
Bunsen screen and is usually made by in-
serting an opaque sheet, such as tinfoil, with
a central opening, sometimes star shaped,
between two translucent sheets, such as
paraffined paper. The sensitiveness depends
upon the density of the translucent sheet.
The Bunsen and Leeson screens do not
require the use of a telescopic lens and can be
read with both eyes. They are particularly
advantageous for quick readings, such asare re-
quired for sources of rapidly varying intensity.
Two types of Lummer-Brodhun screens are
used, namely , the comparison type in which the
contrast disappears when a balance is obtained
with lights of the same color, and the contrast
type in w'hich graded contrast always appears,
the balance being judged by the eye.
The Lummer-Brodhun contrast screen is
the most satisfactory' form for precise work.
It is somewhat intricate as will be seen from
Fig. IC, which shows the sight box in plan.
The box is mounted on the photometer bar
with its axis of rotation U Z perpendicular
thereto. The screen proper c, c', d, d' is a
"disc of compressed magnesia which gives a
brilliant matt surface upon which the rays
from the sources of light to be compared fall
normally. This screen is simultaneously
viewed from both sides by the help of the
mirrors Fl, F2, and the right angle prisms,
A, B, shown in the plan in Fig. ID. Prior to
cementing together the hypothenuse faces of
these prisms, the surface of A is recessed by
sand blasting in vertical strips as shown.
When the prisms are cemented, the spaces
between the strips are transparent, but at
the strips there is a total reflection for light
entering normal to the free prism faces.
Therefore the odd numbered rays (Fig. ID)
received from c, c' via Fl, enter the sight
field only through the cemented faces, and
the even rays from d, d' via F2, only by
total reflection at the strips. The arrows in
the figures show plainly the course of the
rays. The result is a field resembling Fie. IE,
each half circle receiving light from one side
of the screen and having superposed upon it
a trapezoidal area received from the other
side of the screen. These areas are slightly
darkened by absorption from the glass strips
mc and gb, so that when everything is in a
balance there are two equally shaded areas
in a uniform field. The operator can work
either by uniformity of field or by equality
of contrast of the trapezoids.
When lights of two different colors are to
be compared, as for instance red and blue,
it is extremely difficult to judge when the
intensities are equal. For this work the
flicker photometer is used. In this type a
screen is illuminated b\' the two sources of
light in rapid alternation. When the speed
is adjusted between 10 and 20 alternations
per second the illumination appears to flicker
until the intensities of the two become equal,
or the flash from one bridges over the gap to
the flash from the other. It is essential that
the speed be regulated to correspond to the
degree of accuracy desired. With a low
speed the flicker cannot be eliminated, and
w-ith too high a speed the photometer loses
in sensitiveness.
The Standard Lamp and Calibration of the Compari-
son Lamp
There arc a number of priman,- flame stand-
ards of luminous intensity, or candle-power,
based on definite specifications, carefully
drawn. The satisfactory operation of any of
these is complex and difficult, and they are
adapted only to the standardizing laboratory-.
Incandescent lamps, carefully standard-
ized by comparison with the primar}- stand-
ards, are now- employed universally for all
photometry of electric sources. These elimi-
nate variations due to barometric pressure,
humidity, air temperature, etc.. which render
it difficult to reproduce the same light with
flame standards.
The incandescent lamp, as a standard of
candle-power, was established about 18S2 by
the Edison Lamp Works and transferred to
the Electrical Testing Laboratories. Later
the United States Bureau of Standards took
charge of the maintenance of the incandescent
lamp standards, verifying them by extensive
measurements and comparisons with similar
determinations in other countries.
Even- commercial laboratory should have
at least one certified standard incandescent
lamp. In the larger laboratories it is custo-
mary to have secondar>- standards which are
checked about once a month with the certified
standard, and with which the comparison
lamps arc calibrated each day.
In order that the standard lamp may not
change in value, it should not be used more
than necessary for this purpose. Care should
always be taken not to subject the lamp to
abnormally high voltage.
Comparison lamjis should be "aged" or
burned several hours before they are cali-
brated. After the initial variation in can-
COMMERCIAL PHOTOMETRY
957
die-power of an incandescent lamp takes
place, it remains quite constant for a con-
siderable period. The comparison lamp
should be set up in a definite position on the
photometer, and this position relative to the
photometric screen should be marked on the
bulb of the lamp with an arrow. For all
future work, the same position should be
maintained. The candle-power, voltage and
current should next be determined, using
the standard lamp and the method of
procedure discussed under Measuring Hori-
zontal Candle-power of Light Sources. It
described, a socket for holding the lamp
under test and a means of rotating it, a
comparison lamp and a means of varying the
distance from either the test lamp or the
comparison lamp or both to the photometer
head. This type of instrument is still neces-
sary for precision work and the checking of
standard lamps. Its commercial field, how-
ever, is rapidly diminishing.
A few years back all ordinary types of
incandescent lamps had the same shape of
filament and distributed the light in the
same general manner. A comparison of them
Fig. 2. A Precision Bar Photometer as Used in the Laboratory. (Note the Screens, Means of
Rotating the Lamp, Photometer for Maintaining Constant Voltage
Movable Head and Graduated Bar or Scale)
is obvious that constant candle-power will
be obtained by operating the lamp at a
constant wattage if there is no blackening of
the bulb, since at a constant rate of energy
supply there will result a constant light flux.
On the other hand, as blackening is likely to
take place slowly, it is desirable to check the
comparison lamp at reasonably frequent inter-
vals. When the point is reached where appre-
ciable variation is noted between checks, the
lamp should be discarded and a new compari-
son lamp employed.
Measuring Horizontal Candle-power of Light Sources
The standard bar photometer (Fig. 2) is
used for this purpose. It consists in brief of
a photometric head of one of the types already
on the basis of the average horizontal candle-
power was acceptable. Now, however, with
the Mazda B lamps having an extended fila-
ment and the Mazda C or gas filled lamps
having concentrated filaments of various
shapes, the horizontal candle-power of two
lamps giving identical total outputs may be
quite different. It is therefore wrong to ex-
press the efficiency in terms of watts per
horizontal candle-power and far more logical
to express this factor in lumens per watt.
Some means must therefore be provided for
measuring the total light emitted by a
source in a rapid and convenient manner.
This is accomplished by means of the
spherical photometer described in the next
section.
95<S December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
In the days of the carbon lamp it was
necessan- to test each lamp for candle-power
and efficiency and the bar photometer was
used for this purpose. The method of opera-
tion was as follows: A bo.xed-in photometer
of a type shown in Fig. 3 was ordinarily
Fig. 3. A Boxed-in Type of Factory Bar Photometer. Two
operators are able to photometer a large number of
lamps per hour. The working standards will be
noted on a rack at the left
employed. The "reader" so-called, because
she observ^es the "spots, " was furnished with
a resistance connected in series with the lamp
under test.
The photometer was set up so that it read
directly a given candle-power, 16, 20 or 32.
The lamp to be tested was put up in the
rotating socket, and the voltage applied to it
was varied until a balance was obtained.
The second operator or "marker" read the
voltage and amperage. A glance at a table
indicated the efficiency of the lamp. The
voltage at which it gave its rated candle-power
was marked on the bottom of the bulb and
afterwards the lamp was so labeled.
With the diminishing demand for carbon
lamps and the fact that Mazda lamps are
made of wire so accurately drawn to dimen-
sions and cut to the proper length that it
is necessary to photometer only a ver\- small
proportion of the total product, this type
of photometer is rapidly going out of use.
The bar photometer ser\-cs a good purpose
in the laboratory or classroom in demonstrat-
ing the principles of photometry and making
such determinations as the effect of voltage
on candle-power and the like. In installing
such an instrument, precaution should be
taken that surrounding walls do not cast
reflections on the photometric screen. Any
stray light can be cut off by the use of a
suitable number of shields or screens between
the lamps and the photometric head.
In calculating the results from the test with
this instrument, the standard fundamental
and simple formula of
applies, where CPi is the candle-power of
standard source, Di the distance of the
standard source to the sight box, D? the
distance from the test lamp to the screen,
CP2 the candle-power of the test lamp.
Measuring Spherical Candle-power or Total Light
of lUuminants
The total light emitted by a source rather
than the candle-power in a definite direction
is a measure of the energ>' available for
illuminating purposes. It is possible, of
course, to analyze the distribution of light
as described in the following section, and
calculate from this the total flux. Such a
method is tedious and unnecessary- unless tht
characteristic of distribution is desired.
The globe photometer or integrating sphere
(Fig. 4) gives us a means of determin-
ing the total light with one reading and is
now widely used. Its theor>- in brief is as
follows: When a source of light is placed in-
side of a spherical shell having a matt or
Fig. 4. A Small IJ-in. Sphere Photometer for
Testing Miniature and Low Wattage Lamps
depolishcd surface, the light received by any
part of the interior surface may be considered
in two parts, (a) that coming directly from
the lamp and (b) the light received from the
remainder of the interior surface of the sphere
after one or more reflections. The quantity
COM M ERCI AL PHOTOM ETRY
959
(a) is that which is measured in the ordinary
photometer, whieh determines the intensity
of Hght emitted in any one direction and is
not considered at all in the integrating sphere,
for an opacjuc screen is placed between the
lamp to be measured and the opening in
which the photometer head is in
serted. The quantity (b) is constant
all over the surface of the shell and
is proportional to the total amount
of light emitted by the lamp inde-
pendent of its position in the shell.
To calibrate the sphere, a lamp
of known mean spherical candle-
power of the same type and size
as being tested is placed within and
a reading made in the usual man-
ner with the photometer; in other
words, the substitution method is
applied.
Where approximate or compara-
tive readings are required and it is
not expedient to go to the expense
of having a carefully constrticted
sphere, a box photometer or mod-
ified sphere (Fig. 5) is often used.
This consists of a large cubical box with the
corners cut off approximating a sphere in shape,
painted with lithopone (barium sulphate),
and operated in the same manner as a
standard sphere. Its accuracy is not of as
high order, especially for lamps having dis-
plest to operate and to understand is shown
in Fig. 6. This is a twin mirror photometer
of constant radius. The direct rays from
the lamp are intercepted by a black screen
placed in the photometric axis. The light
which is measured is reflected bv the mirrors
Fig. 5.
Front and Rear Views of a Modified Sphere or Boxed Photometer
of Suitable Accuracy for Commercial Work
and strikes the photometer screen at an
acute angle. This necessitates calibrating
the apparatus by the substitution method.
A source of known candle-power is put [in
the same position as the test lamp and
suitable readings taken.
-E^^di
PhotWMtnc Aiiv^
Fig. 6. Sketch of Twin Mirror Constant Radius Distribution Photometer.
The component parts are clearly indicated
similar characteristics, but it serves a useful
purpose under commercial conditions.
Determination of Vertical Distribution of Light
There arc a number of tyjjcs of photome-
ters for this purpose which are described
in all good text books. One of the sim-
In the illustration the mirrors are shown
in a position not encountered in practice, but
this arrangement illustrates the construc-
tion. In actual operation each is similarly
placed on the opposite side of a vertical line
passing through the center of the light source.
Thus, if it is desired to determine the candle-
960 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
power at 10 degrees below the horizontal,
the mirrors are placed on the opposite sides
of the lamp 10 degrees below the horizontal
and so on. In taking the reading at zero
degrees or directly beneath the unit, one
mirror alone is used, the other being covered
with a piece of non-reflecting black felt.
Fig. 7. The Appearance of the Twin Mirrors as Viewed from
the Photometer Head When 45 Degrees Above and Below
the Horizontal
The mirrors serve to gather the light at
any given angle, and reflect it to the photo-
metric screen as shown in Fig. 7. Obviously
in theory one mirror could be used instead of
two, but two mirrors offer an advantage in
arc lamp photometry by reducing fluctua-
tions in intensities due "to the unsteadiness
and travel of the arc.
In photomctering Mazda B (vacuum)
lamps, the lamp and its equipment can be
rotated and only one mirror used. In photo-
metering Mazda C lamps, rotation is out of
the question if accurate results arc to be
obtained, for the whirling action of the cool-
ing gas affects the candle-power. Two mirrors
offer an advantage in overcoming any varia-
tion at difterent horizontal angles.
A rotating sector disc placed in the photo-
metric axis will be noted. This can be used
on either side of the sight box and increases
the range of the apparatus. The comparison
lamp is suspended from a track and manipu-
lated by a steel tape driven by a hand wheel
below the sight box. Varying the distance
of the comparison lamp from the ])hotometer
head enables one to obtain a balance. Any
reflected light from the comparison lamp is
cut oft" by the opaque screens with small
openings. The sight box remains in the same
position or at a constant radius so that the
incident angle of the light rays will be the
same as when the photometer was calibrated.
To calibrate the photometer by the sub-
stitution method, the sight box is set at a
definite point, say 10 feet, from the unit being
tested. The mirrors are set at a 90 degree
angle from the vertical and a secondary
standard lamp whose horizontal candle-
power is accurately known is inserted in the
test socket with its center on the photometric
axis. The comparison lamp is now placed
at the distance from the photometer screen
corresponding to the candle-power of the
secondary standard (assuming the scale is
graduated according to the inverse square
law) and the voltage applied to the com.pari-
son lamp adjusted until a balance is obtained.
This method takes care of the absorption
of the mirrors and other constants of the
device.
Having determined the voltage at which
to operate the comj aris-n la.T!p, the lamp
and reflector or globe equipment is placed in
position so that the center is on the photo-
metric axis. Care is taken to insure that the
light center of the lamp is in the proper rel-
ative position to the reflector. This is ac-
complished by raising or lowering the ad-
justable socket which is part of the equipment.
The voltage on both the test and comparison
lamp should be carefully watched and held
constant. In the case of lamps designed for
series burning, the amperage must be held
constant. It is generally desirable to see
that both amperage and voltage on any
lamp remains constant, for a chinge in
either indicates a change in light output.
Having the set-up properly adjusted, the
mirrors are revolved around the photometric
axis and readings taken at 10 or 15 degree
intervals. Three or more settings should be
taken at each angle. The number of settings
will depend on the constancy of the source.
In the case of a fluctuating source, such as an
arc lam|3, more readings are necessary than
with a steady source, such as the incandescent
lamp. With a fluctuating source a scries of
snap readings is preferable to a lesser number
of careful settings, otherwise the operator is
likely to follow the fluctuations up and down
the scale and not obtain a true average.
Selection of Equipment for Test
The purpose of the test has an impotant
bearing on the selection of samples. Indi-
vidual lamps, reflectors and globes vary, and
COMMERCIAL PHOTOMETRY
961
it is therefore necessary to use as much dis-
cretion as when testing samples of coal for
total heating value. Every engineer knows
how necessary it is in this case to obtain a
representative sample.
If incandescent lamps are to be tested for
efficiency or life performance, it is obviously
impossible to test one or two specimens and
obtain fair results. Incandescent lamps
resemble human beings — some burn out at
an early day, others last over the normal life.
The average life, however, can be determined
by a test on a suitable percentage. The
efficiency of individual lamps also varies
somewhat. For example, the standard speci-
fications for incandescent electric lamps, pre-
pared by the Bureau of Standards, state
that the test quantity shall consist of 5 per
cent each lot of lamps inspected of any one
type, size and voltage range, and in no case
shall be less than 10 lamps.
Individual opalescent enclosing globes and
similar accessories differ considerably in
density. It is impossible to blow glassware
with perfect uniformity. Thickness, and
hence density and absorption, will vary. A
visual inspection will indicate the general
characteristics of the glassware and if onh-
one is to be tested, a quantity should be
inspected and the globe 'selected for test
should be of average density. If it is possible
to test more than one in addition to the
average specimen, the globe which appears
particularly dense and one which is very light
should be selected to determine the maximum
and minim, um absorption values.
Porcelain enamel and similar reflectors
vary a'i to quality of reflecting surface, and
a similar procedure should be followed in
selecting the test unit. A thin coating of
enamel, through which the base metal is
visible, will have low reflecting power.
In selecting a lamp with which to make a
distribution test, particular attention should
be paid to its light center length and phys-
ical dimensions. If it is impractical to secure
a lamp of exactly standard light center length,
the position of the socket, with reference to
the reflector, should be adjusted so that the
standard filament position is attained.
Lamps as ordinarily manufactured vary
somewhat in total light output. Tests should
always be conducted with the test lamp emit-
ting the proper rated lumens of the clear bare
lamp. This is accomplished by placing the
test lamp in a sphere without any aux-
iliary equipment; setting the photometer
attached to the sphere at such a value that
it reads the mean spherical candle-power cor-
responding to the total rated lumens of the
lamp, adjusting the voltage applied to the
lamp under test until a photometric balance
is obtained, then operating this lamp during
subsequent tests at the voltage thus deter-
mined.
In the case of using bowl frosted or bowl
enameled lamps, the absorption of the frost-
ing or enameling should first be determined
by testing a group of lamps for total output
at a given voltage, clear, then frosting or
enameling the same lamps and testing them
for total output at the same voltage. Having
determined the absorption of the frosting or
enameling the lamps should be operated at
the proper percentage of the clear rated total
lumens throughout the test.
The purpose of the test determines, in
general, the procedure which should be fol-
lowed in selecting a sample. If one is en-
deavoring to find out what a certain equip-
ment will do as put out by the manufacturer,
as is the case with the purchaser, the standard
arrangement as to socket or length of fixture
and regular run of glass, as equipped, should
be tested without adjustments.
If it is desired to show from the manufac-
turer's standpoint what equipment will do
under proper conditions, then care must be
taken in selecting average glassware and ad-
justing for standard positions.
If in connection with development work,
it is desired to discover what is the best com-
bination of parts, readings can be taken with
various lamp positions and with various
glassware combinations, eventually deter-
mining the best possible distribution with
minimum absorption and maximum diftu-
sion, or other desirable qualities.
Calculating Results from Test Readings
Having measured the candle-power or in-
tensity of light from the unit at the various
angles in a vertical plane, the readings are
plotted on polar co-ordinate paper to any de-
sired scale. It remains to calculate the mean
spherical candle-power, the mean hemispheri-
cal candle-power, the zonal lumens, the down-
ward lumens and the total lumens.
There are a number of graphical methods
devised by Rousseau, Kennelly, Macbeth,
Wohlauer and others for determining the
mean spherical candlepower and total flux.
These are all based on the same fundamental
equations.
If we consider a lighting unit as suspended
at the center of an imaginan,' sphere of radius
962 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 12
R (Fig. 8), it is obvious that the light in an
angle from 0 to 15 deg. directh^ beneath the
unit will be spread over a comparatively
small area while the light striking the sphere
at, sa}-, 75 to 90 deg. will be spread over a
zone of much greater area. The light flux or
Fig. 8. Diagram Showing Variation in Zonal Areas from Zero
to 90 Degrees
lumens embraced by a zone will be equal to
the product of the average intensity and the
area of the zone. A summation of these
products for each zone gives us the total
value of the light flux or lumens emitted by
the source.
It can be shown by spherical trigonometn.'
that the area of the zones of the sphere are
to each other as their altitudes; thus the
area of any zone of this imaginars^ sphere of
unit radius is equal to
2 7r (cos Oi— cos 02)
where
(01 — 02) is the angle subtended by the
zone in reference, Oi and a« being measured
from the vertical.
Substituting in this formula for 10 degree
zones, we arrive at the following values:
Angle on Curve
Zone Represented
Deg.
Constant
0
0
5
0-10
.095
15
10-20
.284
25
20-30
.463
35
30-40
.628
45
40-50
.774
00
50-60
.897
65
60-70
.993
to
70-80
1.058
85
80-90
1.091
The candle-power readings at' the various
degrees multiplied by these constants will
give respectively the lumens in the zone under
consideration. It will be noted that the la?
constant is but half of the value that wou'. .
seem logical. This is because it takes intc
consideration only the flux from 85 to 90 deg.
The sum of these individual zonal lumens will
give the total downward lumens. The down-
ward (0-90 deg.) divided by 2 ir (6.283) wiU give
the mean lower hemispherical candle-power.
A similar computation can be applied to
the upper. hemisphere, obtaining the upward
lumens and the mean upper hemispherical
candle-power. The downward lumens plus
the upward gives the total Itmiens, which,
divided by 4 tt (12.57), gives mean spherical
candle-power. If it is desired to determine the
lumens in any particular zone, say from
0 to 60 deg., the sum of the individual
zonal limiens should be taken, bearing in
mind that the value for the mid-zone angle
of 60 deg. covers from 55 to 65 deg., and
therefore should be divided by 2 to obtain
the flux only from 55 to 60 deg.
The following data should be included in
a report of a test to determine the light dis-
tribution of a given unit:
Type of fixture.
Trade name and number of manufacturer.
Type of lamp.
Size and kind of bulb, whether clear, bowl frosted,
bowl enameled, or all frosted.
Rated volts.
Rated amperes.
Rated watts.
Mean lower or upper hemispherical candle-power.
Watts per mean lower or upper hemispherical can-
dle-power.
Downward or upward lumens per watt.
Mean spherical candle-power.
Watts per mean spherical candle-power.
Total lumens.
Total lumens per watt.
Per cent total lumens of clear bare lamp.
Lumens in various zones.
Dimensions of reflector, diameter and depth.
Light center length of lamp.
Distance from edge of reflector to top of base.
Vertical distribution curve with note as to whether
initial, or with depreciation.
Description of equipment.
Total lumens at which lamp is operated during test.
With a constant radius photometer, the distance at
which readings were taken.
Test number, date, number of curve and initials of
checker or inspector.
A typical distribution c\iT\-e containing
this information is reproduced in Fig. 0.
Portable Photometers and Their Calibration
The three portable photometers in most
common use in this country- are:
Sharp-Millar photometer
Macbeth illuminator
Foot-candle meter
COMMERCIAL PHOTOMETRY
963
The photometric principle of the first two
instruments is practically the same. Both
have a Lummer-Brodhun cube mounted in the
head, and use a low voltage Mazda lamp as
a comparison lamp. In operation the com-
parison lamp is moved back and forth on the
opticalaxisuntilabalanccissecured in thecube
as seen through a telescope. Both have a scale
calculated according to the inverse square law
and calibrated to read foot-candles direct.
In mechanical construction the instruments
are entirely different. The Sharp-Millar pho-
tometer (.Fig. 10) is a box 5 in. square by
28 in. long. An elbow opposite the photometer
head has a 45 deg. mirror at the elbow and a
translucent test plate at the end. The lamp
is mounted in a carriage on the inside of the
box and moves back and forth with a pulley
and cord. A slot in the lamp carriage il-
luminates the scale, so that when a balance
is reached the foot-candles may be read
directly opposite the slot in the lamp carriage.
A resistance is mounted on the box to keep the
Jyahhoe Metal ffef/eccor 8€DD-Z00
/?L M SCanaara OomG
It^nhoe-ffegent k/orfis (re^c-c/ CtectfC Co
ZOO ilALlt raiscn Mazaa CMuJt'p'e. lamp,P3-30 Bult>,B.€
lamp Botvl Enamef
falls /J5
Amperes /.7d
n'attsfffatea} 200.00
Mean Hemispnencaf C.P ^IfS.OO
niatta Per Meanhsmi5pnencal C.P ,65
Down^ara Lurr^ns / 995,00
Oo^nf^ra lumens Per Watt 9.97
Mean 5pfierical CP I5^.00
ifiztl Per Mean Spherical CP /.Z6
Total Lumens 1995,00
Total Lumens PeriVcU 9.97
Percent Total Lumens of Lamp 70.00
Photometric Test
Initial Distribution of Cardle-Poiver in a Vertical Plane
Peactings TaKen at IQ Ft Radiums
Bowl Enamel Lamp Operatea at 2^50 Total Lumens
Equipment
2i m holder '
Steel Reflector
iVhite Porcelain Enamel Reflectins Surface
Fig. 9. Typical Vertical Distribution Curve Containing
Data Essential for Coniparison or Repetition of Tests
voltage across the lamp constant. A battery
meter set consisting of milliammeter, 6-volt
storage battery and resistance, is provided
with the photometer. The elbow may be
turned in any direction so as to read normal
or horizontal illumination. The instrument
is quite sensitive and is capable of as great
accuracy as is desirable for portable pho-
tometric work.
The range of the foot-candle scale is from
0.4 of a foot-candle to 20 foot-candles, but by
means of neutral tinted absorbing screens
_«3_
Fig. 10. Front and Top View of the Sharp-Millar Portable
Photometer
Fig. 11. Sectional View of the Macbeth Photometer
the range may be increased from 4/1000 of a
foot-candle to 2000 foot-candles.
The Macbeth photometer (Fig. 11) consists
of a tube 9 in. long by X'^/'i in. in diameter.
Inside of this tube is the carriage which holds
the incandescent working standard. The
carriage is mounted on a brass rod extending
through the end of the tube. A rack and
pinion operates the rod and draws the car-
riage back and forth. On one side of the rod
to which the lamp carriage is attached is
engraved a direct reading scale calibrated
from 1 to 25 foot-candles. An index point is
attached to the bottom of the tube. This
index point may be changed so as to allow
for adjustment if variation in filament posi-
tion occurs when renewing standard working
lamps. At the other end of the tube is the
Lummer-Brodhun cube in a rectangular box.
The photometric field is observed through a
telescope. The opening opposite the telescope
is aimed or pointed toward the detached test
plate which is placed at the point where it is
desired to know the illumination. The test
plate made of glass is finished by a special
process so as to get the minimum error when
viewed at various angles. With a given
illumination a perfect plate would be of
964 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
equal brightness when viewed from all
directions. Such a surface has never been
secured. The plate used with the Macbeth
ilkxminator shows practically no error up to an
angle of 25 deg., and from that point the
error is much less than with most other
materials that have been tried.
Fig. 12. Exterior of the Foot candle Meter
In order to increase the range of measure-
ments absorbing screens are provided. These
screens are made of neutral tinted glass which
may be placed on one side or the other of the
Lummer-Brodhun cube, thus widely extend-
ing the normal range of the instrument.
There is no limit to the number of screens
which may be used, either neutral or colored,
for selective absorption. These screens are
easily inserted or removed. With the absorj)-
tion screens usually supplied the range of the
instrument is from 2/100 to 1200 foot-candles.
This may be increased by additional screens
3000 times maximum or minimum if desired.
The auxiliary apparatus supplied for this
instrument is the controller and the reference
standard.
The controller is a self-contained unit con-
sisting of a milliamm.etcr, two adjustable resist-
ances, double-])ole double-throw switch, the
necessary connectors, and two dry batteries.
A detachable shoulder strap makes it possible
to hang the instrument over the shoulder and
conveniently carry it about. Flexible leads
connect the controller with the photometer.
The reference standard is in a housing
which fits over the tube at the end of the
photometer head, and is used for calibrating
the photometer. In using this reference
standard the photometer may be calibrated
at any time, anj^ place, without a dark room,
and is a decided convenience because the
operator may calibrate the instrument him-
self and thus eliminate the personal factor
which is always present when using the illu-
minometer standardized by others.
The sensibility and accuracy of this instru-
ment are about the same as the Sharp-Millar.
The foot-candle meter (Figs. 12 and 13j is
based on the grease spot or Bunsen photometer
principle, which is modified to fit the design
and character of this meter. In the case of
the Sharp-Millar and the Macbeth photom-
eter the working standard lamp is moved
back and forth until the spot in the
photometer head is balanced by the external
source. In the case of the foot-candle meter
a series of spots are lighted from one lamp
in a fixed position, which illuminates a box
which forms the background of the series
of spots. In reading the meter the spot is
selected which Vjlends with the background of
the scale, this background being illuminated
b>' the external source or room illumination
which it is desired to measure.
The screen consists of a piece of clear glass
on which are two thicknesses of paper, one
of which is punched with a series of round
Fig. 13. Intciior of the Foot-candic Meter
holes and is fairly opaque, and the other is
highly translucent. This screen forms one
side of the light box, which is so constructed
that the screen is illuminated from within
to a much higher intensity at the right than
at the left. The exposed side of this screen
COMMERCIAL PHOTOMETRY
965
is ver\- nearly uniformly lighted, and con-
sequently the round spots appear brighter
than the screen surface at the right end and
darker at the left. It is evident that the
point where the spots change from lighter
tfian the screen surface to darker the illumina-
tion on both sides of the screen is approx-
imately the same. When the instrument has
once been calibrated the illumination inten-
sity indicated by a foot-candle scale on the
screen may be read at a glance.
A three-cell flashlight battery supplies
current for the lamp through an adjustable
rheostat. A voltmeter across the lamp in-
dicates when the proper voltage is supplied.
The entire equipment is built into a small
case, 6 by 8 in., weighing only 3 lb. It is not
as accurate nor as sensitive as the large
photometer, but it has thoroughly demon-
strated its value for a light weight non-com-
plicated instrument for general sur\-ey work
and for measuring the illumination in fields
where a large photometer could not be con-
veniently or practically used.
Calibration
Frequent checking or calibration of photom-
eters is necessarv^ if any degree of accuracy
is to be expected. The equipment necessary
to do this work is a voltmeter to keep the
voltage across the standard lamp constant,
a calibrated standard lamp and adjustable
resistance, and constant voltage supply.
Calibrated standard lamps for this purpose
can be secured from the Electrical Testing
Laboratories in New York or the Bureau of
Standards in Washington.
With a given voltage marked on the stand-
ard lamp a certain candle-power is given in
a specified direction. The lamp is set up
and the exact voltage of the lamp is impressed
on it, and maintained constant during the
calibration. Place the photometer a measured
distance from the lamp and calculate the
foot-candle intensity at that point. Set the
comparison lamp of your photometer at the
calculated foot-candle intensity' and varj'
the resistance on your photometer until a
balance is secured. For instance if an S c-p.
lamp is used set the horn of the photometer
two feet from the light source, or filament
of the lamp. At two feet from an S c-p. lamp
according to the inverse square law, we get
two foot-candles. Set the comparison lamp
of the photometer at two foot-candles, var\'
the resistance on the photometer until a
balance is secured with the calibration lamp.
Note the current or voltage on the photo-
meter instrument, and that is the point at
which to hold the comparison lamp in the
photometer when making the test. Care
should be used not to burn the reference
standard lamp longer than absolutely neces-
sary, as continued burning will destroy its
accuracy.
The only difference in calibrating a pho-
tometer with an external test plate, like the
Macbeth, is to set the plate in place of the
photometer, keeping its face normal to the
lamp. In pointing the photometer at the
test plate keep within 30 deg. of normal and
do not allow am* extraneous light to enter
the tube.
The Sharp-Millar photometer may also
be used with an external test plate by simply
removing the cap holding the translucent
plate at the end of the horn and calibrating
as described above. White blotting paper
may conveniently be used as a temporary
test plate, although a specially prepared
glass plate is more constant and more per-
manent.
If the calibrating devices that are furnished
with the instrument are used they should be
checked occasionally against the standard
lamp to see that they have not deteriorated.
When calibrating the foot-candle meter
it is necessarj' that the meter lie in a horizon-
tal position while being calibrated. This is
true because the needle of the voltmeter con-
tained in the instrument is balanced to read
correctly in the horizontal position. When
the voltage indication has once been deter-
mined, however, the meter may be used in
any position.
{To be Continued)
9 06 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 12
Photo-elasticity for Engineers
Part II
By E. G. CoKER, D.Sc. F.R.S.
Professor of Engineering in the University of London, University College
Written specially for General Electric Review
In this article the author first gives the results of some experimental work on the determination of stress
in the neighborhood of a circular hole in a tension mamber, using celluloid models. He also gives the results
of some independent determinations of stress in similar steel pieces, and compares the results with those
required by theory. A theoretical proof that stress distributions are independent of the elastic constants of
the material in many cases is outlined. Further experimental results are given covering the case of elliptical
holes in tension members and extending the study to the important matter of cracks and discontinuities.
Knowledge of these results is important in many practical cases of design such as boiler plating, steam tur-
bine wheels, and any case where a stressed member contains holes. — Editor.
Holes and Cracks
The effect of a hole, or a group of holes, on
the distribution of stress in any member or
element of a machine under load is of great
interest and importance, since in most
engineering operations holes are drilled or
othenvise shaped for connections like bolts
and rivets, or possibly as means of communi-
cation between neighboring chambers, for
assembling purposes and the like. Whatever
their use may be it can be shown that they
alter very^ greath" the stress distribution in
their neighborhood, and the combinations
in which such discontinuities can occur in
practical work are so immense in number
that it is only possible to deal with a limited
number of simple cases; in fact very few
have been solved, as yet. experimentally or
by calculation.
The importance of this group of cases,
however, warrants us in considering them
somewhat early as practical examples of the
use of photo-elastic investigation.
In the simplest case of a very wide tension
member, having a hole of moderate size
drilled centrally, we have already seen that
when load is applied the color effects are
marked around the boundary- of the hole,
even when the rest of the plate is under
little stress. The efTects are symmetrical
about the line of pull, and most intense at
points distant from the center line, and they
gradually decrease in intensity as we approach
the axis, until at an angular distance of about
30 degrees from this there is no stress at all.
Along that portion of the boundary nearest
the center line there is a compression stress
which attains its maximum value at the axis,
TABLE I
as may be readily verified by aid of the
exploration tension member.
Very many holes have been examined
optically and as an example of such measure-
ments we may take a hole '4 in. in diameter
in a plate on which the load applied gives a
uniform stress of 570 pounds per square inch
of cross section well rcmo\-ed from the dis-
continuity. An exploration of the stresses at
difTerent angular points of this boundary-
shows that the maximum stress reached is
1720 pounds per square inch in tension, and
Table I shows that it varies greatly with the
angular distance from the axis. Fortunately,
we arc able to compare these results with
calculation, since this example is one of the
cases for which an exact solution has been
found and it can be shown that the boundar\'
stress follows the law
s = p {1-2 cos 2 0)
Fis. 1
Angular Distance 6
from Axis
0°
- .-)4n
15°
-400
30°
-20
45°
+580
(50=
+ 121X)
75"
9cr
Stress in pounds per
square inch
-1-1. t30
+ 1T20
PHOTO-ELASTICITY FOR ENGINEERS
<)G7
This is substantially what is found in this and
all like cases. A cylindrical hole, in fact, raises
the stress to three times its normal value at
the sides of the hole as the linear diagram
(Fig. 1, Curve A) shows, and very much
increases the stress at places near to this;
but this is not, as a rule, realized by practical
men, and often a hole is considered as a mere
loss in cross section. It is much more than
this, however, as an inspection in the polari-
scope shows. It is, moreover, easy to prove
that a new load of the same type may actually
decrease the stress at the hole if it is applied
at right angles to the former direction. If
it has the same general intensity we then get
a uniform stress all round the boundary, of
twice the mean intensity (Curve C), due to
the combined effects of the boundarv' stresses
(Curves A and B), since the effect of the
extra load is to eliminate all variations due
to angular change. If, on the other hand, a
similar load of the opposite sign is imposed
(Curve D), in this new direction, the stress is
increased to four times the intensity at the
edge of the hole at four places (Curve E), two
along the ax'al line in compression and two
at the ends of the transverse diameter. Com-
bined stresses of this kind are of frequent
occurrence in practice and some of special
interest occur in the rotating disl<s of turbine
wheels where they are pierced by holes and
it is often found to be necessary to increase
Fig. 2. Stresses Around Circular Hole as
Shown by Polarized Light
the thickness of the metal around these
discontinuities in order to avoid fracture.
If now we proceed further to examine the
condition of stress away from the boundary
of the hole we need to know the directions of
the principal stresses since we have no bound-
ary to help us, and it becomes necessary,
therefore, to map out the region by the aid
of plane polarized light using the crossed
polarizer and analyzer. For such a case as
the present it is comparatively easy to
determine these directions and the process
Fig. 3
is indicated by the bands observed around a
somewhat larger hole and these are found to
have the forms of the type shown in the
illustration. Fig. 2, thereby fixing the direc-
tions of stress in the area around the hole.
A number of these bands with the directions
of the stress marked on them are shown in
Fig. 3, while the stress distribution is indi-
cated in the color photograph of Fig. 4A, and
although this diagram gives all the informa-
tion required it is not in a very convenient
form and it is generally preferaljle to re-cast
the information it gives by drawing curves
which show more directly the directions of
the stress at any point.
The simplest way of carrying this out
is, in the present instance, to take a cross
section at some distance from the discon-
tinuity where the lines of principal stress are
parallel and perpendicular to the direction of
the pull and then produce these lines in the
direction of the hole and guide their directions
by the isoclinic lines so that at any point
these directions correspond. Proceeding in
this way a new map is obtained as shown in
968 December, 1920
GENER.\L ELECTRIC REVIEW
Vol. XXIII, No. 12
Fig. 5 on which it is usually convenient to
project the stress picture forrned in circularly
polarized light and then proceed to determine
the stress difference opticaUv at the points
required. If further it is desired to know the
magnitudes of each principal stress one of the
methods described in the preceding lecture must
be applied in order to effect the separation.
In a case where a 14-in. hole pierced a plate
1 m. in width this separation was effected by
aid of lateral measurements of the strain
across the minimum section and it was then
found that the values of (P^Q) at various
pomts reckoned from the center of the hole
had the following values :
Hole }4 in. in diameter in a plate 1 in. wide
TABLE I
Distance from I
Center of Hole
STRESSES IN POUNDS PER SQUARE INCH
in Inches
P+Q ip-d) 1 p
9
—0.50
—0.40
—0.30
-0.20
-0.14
—0.125
0
+0.125
+0.14
+0.20
+0.30
+0.40
+0.50
530
550
650
930
1440
1350
910
660
560
500
540
580
630
750
1460
1720
1730
1400
760
660
580
560
535
565
640
840
1450
1720
1730
1375
835
660
570
530
—5
—15
+ 10
+90
—10
-25
+75
-16
—30
which at once gives the separation of the
stresses required and also shows that the
effect of the puU produces a cross stress at this
section which although small is perfectly
definite with a ma.Kimimi value near the
boundan,- of the hole.
It is of interest to compare these resiilts
with those obtained by the elastic theory- of
the effect of a hole in a ver\' wide plate. Fig. 6.
For such a case we obtain the stresses in
polar co-ordinates in the form
^-ii(-7:)+('-:-:+#-'«[
'"■M('+'^)-('+#'''-"[
poe —
P
I
sin 2 e
which become for the cross section measured
Prr-
pM-
prB = 0
in which the uniform stress p corresponds to
an infinite plate, so that for a finite plate some
correction is required. As the experimental
cur\-c for p^ is found to closelv follow the
theoretical value, it is probablv sufficiently
accurate to assume that it has the same law.
If then p„ is the mean average stress at the
section through the hole of radius a, where
2 ca is the width of the member, we obtain
fma{c — l)=J pM.dr
PHOTO-ELASTICITY FOR ENGINEERS
(.4) Circular hole.
(B) Elliptical hole.
(C) Slit.
(D) Silt with ends bored out to
reduce stress.
(£) Slit with ends bored with elliptical
holes to further reduce the stress.
Fig. 4. Tension Members with Various Shaped Holes in Them Showing Maximum Stress Intensities at Top
and Bottom of Holes. Tension is Applied in a Horizontal Direction
Part II
PHOTO-ELASTICITY FOR ENGINEERS
969
giving
P = P'
'li
c 2c- 2c^
)
an equation for the required value appropriate
to this case of 570 pounds per square inch.
Comparison of the experimental and theoret-
ical results then show almost perfect agree-
ment for the stress across the section, but not
so good for the cross stress.
This discrepancy arises from the manner
in which the latter value is determined as a
difference between the sum and difference of
the principal stresses, and as both are in
general large values any small errors in their
determination become still larger percentages
of the cross stress values. Although errors
tend to accumulate in this way, there is
sufficient agreement to show that the be-
havior of the cross stress follows the law
indicated by theory and that the maximum
values of this stress occur at a short distance
away from the hole at a radius r = a\/2
corresponding to — t^ = 0 for d = ^-
size of hole but the maximum stress is still
approximately 3 p even when the diameter of
the hole is half the width of the plate. Fig. 7.
It is noticeable, however, that the zero stress
at this boundary tends away from the central
line while the stress at the ends of the cross
Fig. 8
Fig. 7
It is interesting to examine what effect is
produced by successive enlargement of the
hole in such a plate and it is found that
around the boundars' of the hole the stress
distribution changes gradually with increased
section becomes less than that indicated by
the theoretical expression for pee-
This latter phenomenon is in fact just
perceptible with a 14-in. hole in a 1-in. plate
and becomes more marked with a hole of
greater size. Another peculiarity which
seems connected with this latter phenomenon
is that the stress along the straight edges of
the tension member appears to always be
less at the intersection with the central
cross section than at any other place.
It is quite perceptibly so with a .^i-in. hole
and the appearance of the specimen in the
polariscope shows that this must be so since
the bands of constant stress difference inter-
sect these boundaries at two places away from
the central cross section, showing that a
minimum stress value lies between. With a
J/^-in. hole (Fig. 7) in a 1-in. plate the stress
at places along the parallel sides about an
inch away from the minimum section is
actually about 40 per cent greater than that
at the ends of the central cross section and
as the hole becomes still greater the maximum
and minimum values become still more
pronounced. In a plate 1.1 in. wide, with a
central hole O.SS in. diameter. Fig. S, it was
found that the maximum stress along the
edges rose to three times the value of the
minimum and it seemed possible from
observations of the color bands shown in
Fig. 4A of the stress distribution across this
970 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 12
minimum section that the stress at the outer
edge might actually change to compression
if the hole became sufficiently large.
On tnang this with a hole about 5 in. in
diameter in a plate about 5.2 in. wide, it was
found that the minimum stress did not even
in this case become negative but was approxi-
mately zero, while the variation across the
section was approximately linear. For ver\'
large holes, therefore, the maximum stress is
approximately double the mean stress across
the section.
It seems natural to inquire whether the
stress distributions produced by holes in
nitro-cellulose are obtained in materials used
for constructive purposes, and for one of the
cases described here it happened that some
data for comparison existed which afforded
a completely independent test since the
measurements on a steel bar with a central
hole had been carried out by one of my
Japanese students, Y. Satake, who was
not at the time aware of the peculiarities
which were afterwards obser\-ed in trans-
parent specimens.
He had, in fact, undertaken a sun-ey of the
stress condition of a steel bar 1.5 inches
broad and 0.4S8 inches thick, by measuring
the lateral contractions observed when a load
of 2.S tons was applied to the ends of this
tension member. The measurements afforded
values of the sum of the principal stresses at
various points when the lateral strains were
multipled by the value of £, which latter
were obtained by direct measurement and
were found to have the mean value 98.3 X 10"'.
His measurements across the central cross
section gave the distribution shown in Table
II.
TABLB II
1
Distance from ,
Lateral Strain
(P+O) ="■£,...-
Center of Hole
^io-«
=98.3
in Inches
in Inches
Lateral Strain
0.256
232
22.820
0.266
227
22.310
0.320
205
20.150
0.376 1
179
17.600
0.476
147
14.450
0.576
114
11.210
0.676
105
10.320
0.710
92
9.040
There were no experimental values on a
plate of optical material which corresponded
exactly to this case, the nearest available
being for a ?^-in. hole in a 1-in. plate. For
comparison purposes, therefore, the linear
dimensions of this latter were increased in
scale to make the holes agree in size, while
the stresses were adjusted in the ratio of
their equivalent tensions. These values
were then plotted for comparison and they
are sho^Ti on Fig. 9, in which the upper cur\-es
show the value of {P+Q) obtained from the
elastic theory- with the actual stress distri-
bution cur\-es for the steel members immedi-
ately below. This latter cur\'e corresponds
ver\' closely therewith except at the ends,
while the nitro-cellulose specimen gives
slightly lower values and in both cases the
stress distribution near the parallel contours
have a steeper gradient than theor>- indicates,
although they agree fairly well with each
other.
An additional check on the accuracy of the
measurements was obtained by integrating
the normal stress across the minimum section
and comparing it with the pull. For this
purpose the cross stress was calculated and
deducted from the measured (P+0) cur\-e
for steel to give the stress distribution cur\-e
marked P on the diagram, and it was then
found that the average error of the measure-
ments was 3.5 per cent in excess, a satis-
factory- agreement having regard to the
diflicultv of accuratelv measuring lateral
I
I
Fig. 9
strains in the steel bar. the largest of which
was only two or three ten-thousandths of an
inch.
Additional evidence of the applicability of
experiments on nitro-cellulose to steel was
afforded by a comparison of the stress
PHOTO-ELASTICITY FOR ENGINEERS
971
distribution along tlie sides, also shown in
Fig. 9, where the agreement was still closer,
and would probably have been still more so
if the two specimens had been exactly
similar. Taken as a whole, this comparison
of the stress distribution in two specimens
of different materials warrants the conclusion
that photo-elastic investigations may be
safely used to infer stress distribution in
metals within the elastic limits of the material.
This, however, does not complete the
evidence, for there are theoretical grounds
supporting this conclusion which may be
briefly described here.
It has already been shown that the general
equations of equilibrium for a plate are
expressible in the form
dpxx . dpxy^^^
dx dy
dpxy , dpyy^Q
dx dy
where the stresses may be taken as mean
values throughout the thickness.
If u and V are the average values of the
displacements corresponding to a point x,y
of the plate the corresponding strains are
_du _dv _du dv , .
''"'d^' '"'^dy' ^"'"d^^Tx ^^'
Differentiating the first of the equations (1)
with respect to .v and the second with respect
to y we obtain the stress relation in the form
d^p^x_d^pyy_ d'p:
dy^'
dx^ dy'^ dxdy
An identity which is satisfied by a function
X (Airy's Function) if
d^X . d^x . d\
(6)
In a similar way the strains of equations (2)
may be shown to be connected by the relation
d-txx . d-iyy d-fxy_„ ,..
dy^ dx- dxdy
while the relations between stress and strain
are
mEtxx = mpxx — p yy
mEtyy = mpyy — pxx
mEezz= —{pxx + Pyy)
mEexy = 2 {m + l)pxy
If now the stresses in these latter equations
be expressed in terms of the function x by
aid of equations (3), (4) and (5), and the
values of the strains obtained are substituted
in the strain relation ((3) we obtain after some
reduction the relation
d'x , d'x , .^_d'x^
dx-dy
— -l--^-l-2
dx'^^dy''^
= 0
(7)
the fundamental equation of plane stress
which as will be observed involves no elastic
constants. Now since the stresses can all be
expressed in terms of x by equations (4)
these latter must also be independent of the
elastic constants for any plane stress.
The tacit assumptions underlying this
conclusion are the generalized elastic law
and a single boundary condition, but Michell
has shown that it is still true for bodies with
any number of separate boundaries provided
the applied forces over each have no resultant.
If, however, there is a resultant unbalanced
force on a boundary, then elastic constants
have to be taken into account and a cor-
rection made when applying the results from
one material to obtain the stress distribution
in another.
This is a subject which offers a very attrac-
tive field for further investigation, and until
further work is accomplished it is somewhat
difficult to form an opinion on the magnitude
of the correction to be applied in such cases.
Elliptical Holes and Cracks
The use of elliptical holes in practice is so
limited that the stress distribution around
discontinuities of this kind would have little
interest for engineers if it were not for the
information they give on the stress due to
cracks. These latter, if straight, may be
regarded as cases of an ellipse in which one
axis is very small, and it is probable that
even if a crack is of irregular shape its ends
may still be regarded as limiting cases of an
ellipse with a major axis in the directions of
these ends.
The stress around elliptical holes in a wide
tension member has been examined by Mr.
Kimball of the General Electric Company's
Research Laboratory, and the writer;
Fig. 4B shows the general characteristics of
the stress distribution at the boundary of
such a hole having a major axis 1.20 inches
long perpendicular to the line of pull, and a
minor axis 0.8 inches in length in the line of
pull. As will be observed, the concentration
of stress is very great at the ends of the
major axis, and diminishes as the boundary
is traversed, and finally becomes a com-
pression with a maximum value at the minor
axis. The variation in stress along this
boundary in terms of the uniform stress R
in the plate away from the hole is shown in
Fig. 10, from which it will be observed that
the maximum tensional stress is AR and the
maximum compression stress is R. These
results agree with the theory developed by
972 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 12
Professor Inglis, who has shown for such cases
that the maximum stress is i?(l + 2 a/b) where
a and b are the principal axes perpendicular
to and along the line of pull respectively,
and also that this diminishes as the boundary
is traversed in the manner shown by the
dotted curve of Fig. 10. The agreement
Stress at Boundary
of elliptical hole
Fig. 10
between experiment and calculation is in'fact
very close. In a further case examined with a
major axis of 1.25 inches and a minor axis of
0.375 inches, a similar agreement was ob-
tained. The experiments and calculations,
therefore, appear to justify us in assuming
that when there is an elliptical hole with its
minor axis in the line of pull and the minor
axis is very small compared with the major
« 2-
j
i
i
■^
';-.p.4
IJ
-1
o
X
u
o
o
NO-.,.
H^
/■
"~~-4-~£_
>-oT
L
)-— *
_
i
1.0 1,5 2.0 2p
DISTANCE FROM CENTRAL LINE
INCHES
Fig. II
axis the tensional stress developed approaches
an infinite value according to the law 7?
(1+2 a/b). This agrees with observ^ations
of the effect of loading a tension member
with a very fine slit cut in it across the line
of pull as even the smallest load stresses the
material beyond its yield point, while the
characteristic features of overstress appear
and remain after the load is removed. The
experiments also show that this intense stress
falls off very rapidly as we proceed outwards
along the minimum cross section as appears
from the measurements, Fig. 11.
Lines of Equal Lines of Principal.
Inclinatioh I Stress.
•^•:^)'<'&
Elliptical Hole S^a
INCLINA TION 49 TO TH E
LIKE OF Pull.
Fig. 12
When an elliptical hole is inclined to the
direction of pull the most intense stress is
no longer developed at the extremities of the
major axis, but at or near a point where a
tangent line parallel to the line of pull
touches the boundary of the discontinuity.
This appears probable from an inspection of
the lines of princijial stress drawn for a case
in which the inclination of the hole is at
49 degrees, Fig. 12, since in this neighborhood
the lines of principal stress arc most crowded
together and when this takes place we find,
in general, the most intense stress.
Practical men have long been aware of the
extreme concentration of stress produced by
cracks and have provided the well known
remedy of drilling holes at the ends to stop
their extension. This has the effect of reduc-
ing the local stress at the end of the crack
from a very large but indefinite value to
only a few times the average stress in the
material; but it seemed possible that this
method might sometimes be varied with
PHOTO-ELASTICITY FOR ENGINEERS
973
advantage having regard to the lower stress
which is usually found at the minor axis of an
elliptical hole. This view is^confirmed by
some experiments on slits in a wide plate with
variously shaped ends. The slits examined
are shown grouped together in Fig. 13 and
1670 lbs. at points on the rounded contours.
A slight increase to 1700 lbs. per square inch,
however, took place when this contour was
shaped to an approximately elliptical form
(Form D). There seems little to choose
between these last two forms, since the last
Fig. 13
the primary' form is a slit half an inch long
cut by a drill of -^ inch diameter, the ends
being left untouched. When an average stress
of 790 lbs. per square inch was applied the
stress at the extreme edge of this slit rose to
2900 lbs. per square inch, but fell to 2258 lbs.
when the ends were drilled out by a i^-inch
drill (Case B). A further reduction took place
when this hole was enlarged by drilling two
holes tangentially to the center line of the slit
(Case C) and the maximum stress again fell to
Fig. 14
stress was only 30 lbs. more per square inch
than for Form C. The distribution of stress
for the first case is shown in the colored
photograph, Fig. 4C, and as will be observ^ed,
it is extremely concentrated around the semi-
circular end, but when this latter was enlarged
by a drill, Fig. 4D, the distribution shows
the characteristic features we have already
obsen^ed with the cases already considered and
on further enlargement to an elliptical form
as the stress concentration became still less, as
Fig. 4E indicates. The experiments, therefore,
show that in drilling out the ends of cracks
it is an advantage to form the ends with
elongated holes of an approximately ellip-
tical form, with the part of greatest radius
parallel to the line of greatest stress.
(To be Continued)
974 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
Waste and What It Means in Industry
A far Eastern missionary, on returning to
his native land, once remarked that "a
Chinaman could live a day on what the
average American wastes at a single meal."
To the foreigner the American is always an
extremist. He admires our ingenuity, our
resourcefulness, our productiveness, and our
"pep," which is but a synonym for a New
World brand of enthusiasm, but he is aston-
ished at our inexcusable extravagance, our
shameful waste of foodstuffs, raw materials,
and the various by-products of certain com-
modities.
With all our so-called efficiency experts,
we are still a very wasteful nation, and though
the lesson of the war, with all its vicissitudes,
may have had some beneficial results, the
net gain in our favor is negligible. Through
the favor of the fates, America has yet to
know the actual meaning of the words pov-
erty, hunger and misery, but we must correct
and curb this great national shortcoming or
pay the penalty at some future date. A
policy of national extravagance is not con-
sistent with national prosperity, and the can-
dle burning at both ends soon scars the
fingers — sooner sometimes than we might
expect.
Let us consider the average employee in
the average American industrj'.
It must be admitted that the success of an
industry is just as dependent upon the ef-
ficiency and ability of employees as it is upon
the directing reins of the management.
Brilliant minds might design wonderful ships,
but they cannot float off the blue j^rints into
the water; it takes cflicicnt heads and muscles
to put together the plates and steel beams
that make a seaworthy vessel.
So much for the ability of the American
artisans; their calibre is conceded, but how
about their honesty to their employer from a
standpoint of extravagance in the use of
loaned tools, materials, and overhead main-
tenance? Economy is the first step toward
efficiency, and industrial efficiency is the long-
est and most important step toward national
prosperity.
Is the American workman as considerate
of the property of the man or company as
he would be of the articles were they his own ?
Hardly. He may not be maliciously destruct-
ful or wasteful, but he is often careless and
thoughtless. To be a little more explicit,
we will consider the average shop. How
many times a day is an electric light left
burning needlessly? How many hundreds
of bulbs are broken in a week by rough
handling. How many times is the power
turned on without producing anything, with
belt or gears speeding in vain? Nature
squanders immense quantities of power in
each electric storm, but man cannot help that.
But he can turn a switch or press a button when
light or power is wasting in the shop.
How many thousands of dollars worth of
new material is wasted yearly in the average
shop or mill through careless handling? It
is safe to say that this item alone mounts to
the equal of a Liberty Bond issue. It is
natural that there should be a certain amount
of shrinkage. No concern can hope to con-
sistently operate with a perfect percentage.
It is quite proper that production inspectors
should be needed to keep the quality of the
product at the highest possible standard. But
there is plenty of room for improvement in
the actuating motive of the employee toward
the property and product of the employer. A
better spirit of fairness is needed; a feeling
of " do unto others as you would have them do
unto you " is needed imperatively.
In many respects America is the most pro-
gressive country on earth, hut we must curb
our extravagant methods, or look to our in-
dustrial and commercial laurels. Envious eyes
from across the seas are watching every oppor-
tunity, and you can rest assured that it is only
the negligible minimum that goes to waste in
those lands where necessity has always taught
the lesson of thrift and economy. Play fair
with your employer as you expect him to play
fair with you, and take the same reasonable
care of his property as you would with your
own. Thrift is the right way to save, but the
"elimination of waste" is his left-handed
brother. — Specd-L 'p.
975
One Hundred Years Since Oersted, Ampere
and Arago*
By Dr. Elihu Thomson
General Electric Company, Lynn, Mass.
This account of the epoch-making discoveries in electricity and magnetism by the pioneers. Oersted,
Ampere and Arago holds much additional interest from the fact that it is presented by a man who himself
has won world renown by his discoveries and inventions in the same branch of science. The work of these
pioneers, including Faraday, forms the real foundation of the huge, diversified electrical industry of today. —
Editor.
Preamble
It is fitting that in Philadelphia we should
celebrate the centenary of the great discov-
eries in elcctromagnetism. It was here that
Franklin's investigations in electricity were
made, culminating in the kite experiment.
It was here that he and a few confreres founded
the American Philosophical Society, which
became a national institution for the spread
of that spirit of science and philosophy char-
acteristic of Franklin. It was here not many
years ago that under its atxspices a very
notable commemoration of the centenary^ of
Franklin's work was held. Not far from here
in Princeton the pioneer work of Henry in
elcctromagnetism, induction of currents, and
oscillations was done nearly a century ago.
Not far to the south from here the first
Morse telegraph line was established in 1844.
In Philadelphia, Robert Hare in the early
years of last century did his work with voltaic
batteries. Here Bell first exhibited his speak-
ing telephone at the Centennial Exhibition
of 1876, calling such witnesses as Sir William
Thomson (Lord Kelvin) to hear it speak.
Not far from here, in the Laboratory.' of
Edison in Menlo Park, the incandescent lamp
was bom in 1879. Here again in commem-
oration of Franklin was established the Frank-
lin Institute, the influence of which has been
so marked a factor in science and the mechanic
arts ever>'where. Under its auspices the
first investigation of the electrical properties
of the dynamo was made in 1877, and the first
Electrical Exhibition held in America in 1884,
the Paris Exposition of 1881 being the only
forerunner. It is a pleasure to note at this
time the possibility of great and increasing
lustre to its future in the electrical field has
come by a large bequest from one whom the
present writer knew well in his old Phil-
adelphia days, Mr. Henry Bartol. I am re-
minded of the fact that the first meeting of
the American Institute of Electrical Engineers
* An address before a meeting of the A. I. E. E.. Philadelphia,
October 8, 1920.
was held in Philadelphia. And now, to relate
very briefly more intimate but infinitely
less important matters, may the writer
modestly add that here over fifty-five years
ago he built his first electrical machine,
voltaic piles, batteries, electromagnets and
telegraph, acquiring through them in his
early years an insight of the science of elec-
tricity as it then existed. It was here that he
taught science for ten years in the old Central
High School at Broad and Green streets and
that during this period in 1875 there was
made, incidentally, the first wireless trans-
mission, using induction coil, spark gap,
ground and radiating conductor, briefly de-
scribed in the Franklin Institute Journal of
the time, and recently related more in detail
by Professor M. B. Snyder, of the school.
It was here in Philadelphia that the writer
did his first electrical engineering, and defi-
nitely chose that professional career which
has kept him alive and busy ever since.
There are times when an epoch-making
discoverv' gives rise to a new science or art,
or opens up new fields for experimental re-
search. When this has occurred before our
time we can at best visualize the antecedent
conditions imperfecth". The background of
such a discovery as was made in 1819, by
Hans Christian Oersted, of Copenhagen, and
announced in July, 1820, is scarcely repro-
ducible now. We shall not attempt it.
Simple as was the experiment of Oersted, the
fundamental character of his results was in-
stantly recognized by his contemporary
leaders in science, such as Ampere, Arago
and Davy, and served to stimulate them to an
intensity of research work which at once
brought wonderful additions to human knowl-
edge.
Oersted
Oersted, a Dane, born in 1777, was edu-
cated at the University of Copenhagen, and
in 1806 occupied the chair of physics there.
Though he had already done important work.
976 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 12
he was immortalized by his being the first
to discover and investigate the effects of a
current in a conductor upon a magnetized
needle. It was, or at least it may now seem
to us, a most simple discovery, the outcome
of experiment of equal simplicity. Neverthe-
less, subsequent events soon proved it to be
the foundation stone upon which now rests
the great science of electromagnetism.
Oersted found that a wire conveying the
currents of a voltaic battery, and called by
him a "conjunctive wire," affected a freely
pivoted magnetic needle in such a manner
that the needle tended to set itself at right
angles to the wire. The deflection was shown
to be definite as to direction, depending on
the direction of the current in the wire, the
position of the poles of the needle and the
relation of positions of the wire and needle.
It was recognized, also, that if the magnet
was fixed in position and the conjunctive wire
free to move, corresponding movements of
the wire would take place. This discovery
was given out in a brief work in Latin, the
title of which was in substance, "Experi-
ments Concerning the Effect of Electric Con-
flict on the Magnetic Needle."
Oersted had apparently convinced him-
self long before of there being a necessarj' con-
nection between electricity and magnetism
and had held perhaps more pertinaciously
than others to this view. In recognition of
the great scientific value of his discovery the
prize of the French Institute was awarded to
him. This had already been given to Davy for
his electro-chemical discoveries, such as that
of the separation of the alkali metals, sodium
and potassium, from their compounds.
Oersted also received the Copley medal of the
Royal Society of London and was honored
by the distinction of Knighthood. Dying
in 1S51 at seventy-four, he had lived to see
great progress in electromagnetism and to
witness some of its early applications, such
as the telegraph, to the needs of mankind.
The following translation of Oersted's
description appears in Barlow's "Magnetic
Attractions," a book published in 1824.
After assuming current passing in the con-
junctive wire — •
" Let the straight part of this wire be placed
horizontally above the magnetic needle prop-
erly suspended and parallel to it. If nec-
essary, the uniting wire is bent so as to assume
a proper position for the experiment. Things
being in this state the needle will be moved,
and the end of it next the negative side of
the battery will go westward.
"If the distance of the uniting wire does not
exceed three quarters of an inch from the
needle the declination of the needle makes
an angle of 45 deg. If the distance is increased
the angle diminishes proportionally. The
declination likewise varies with the power
of the batten,-.
"The uniting wire may change its place,
either towards the east or west, provided it
continues parallel to the needle, without any
other change of the effect than in respect to
its quantity. Hence the effect cannot be
ascribed to attraction; for the same pole of
the magnetic needle which approaches the
uniting wire, while placed on its east side,
ought to recede from it when placed on the
west side, if these declinations depended on
attractions and repulsions. The uniting con-
ductor ma}' consist of several wires or metallic
ribbons connected together. The nature of
the metal does not alter the effect, but merely
the quantity. Wires of platinum, gold, silver,
brass, iron, ribbons of lead, and tin, and a mass
of mercun.- were employed with equal success.
The conductor does not lose its effect, though
interrupted by water, unless the interruption
amounts to several inches in length.
"The effect of the uniting wire passes to the
needle through glass, metals, wood, resin,
stoneware, stones, for it is not taken away by
interposing plates of glass, metal or wood.
Even glass, metal and wood interposed at
once do not destroy, and indeed scarcely
diminish, the effect. The disc of the elec-
trophorus, plates of porphyry, a stoneware
vessel even filled with water were interposed
with the same result. We found the effects
unchanged when the needle was included in
a brass box filled with water. It is needless
to obser\-e that the transmission of effects
through all these matters has never before
been observed in electricity and galvanism.
If the uniting wire be placed under the mag-
netic needle, all the effects are the same as
when it is above the needle, only they are in
opposite directions; for the pole of the mag-
netic needle next to the battery declines to
the east.
"That these facts may be more easily re-
tained, we may use this formula: the pole
above which the negative electricity enters
is turned to the west; under which to the
cast.
"If the uniting wire be so turned in a hor-
izontal plane as to fonn a gradually increasing
angle with the magnetic meridian, the declina-
tion of the needle increases, if the motion of
the wire be toward the place of the disturbed
ONE HUNDRED YEARS SINCE OERSTED, AMPERE AND ARAGO
97
needle; but it diminishes if the wire moves
further from that place.
"When the uniting wire is situated in the
same horizontal plane in which the needle
moves, and parallel to it, no declination is
produced either to the east or to the west;
but an inclination takes place, so that the
pole next which the negative electricity enters
the wire is depressed when the wire is sit-
uated on the west side, and elevated when
situated on the east side.
"If the uniting wire be placed perpendicu-
larly to the plane of the magnetic meridian,
whether above or below it, the needle remains
at rest, unless it be very near the pole; in
that case the pole is elevated when the en-
trance is from the west side of the wire and
depressed when from the east side.
"When the uniting wire is placed perpen-
dicularly opposite to the pole of the magnetic
needle and the upper extremity of the wire
receives the negative electricity, the pole is
moved toward the east; but when the wire
is opposite to a point between the pole and
the middle of the needle the pole is moved
towards the west. When the upper end of
the wire receives positive electricity the
phenomena are reversed.
"If the uniting wire be bent so as to form
two legs parallel to each other, it repels or
attracts the magnetic poles according to the
different conditions of the case. Suppose the
wire placed opposite to either pole of the
needle, so that the plane of the parallel legs
is perpendicular to the magnetic meridian,
and let the eastern leg be united with the
negative end, the western leg with the pos-
itive end of the battery, and in that case the
nearest pole will be repelled either to the
east or west, according to the position of the
plane of the leg. The eastmost leg being
united with the positive and westwards with
the negative side of the batten,-, the nearest
pole will be attracted. When the plane of
the legs is placed perpendicular to the place
between the pole and the middle of the needle,
the same effects occur, but reversed.
"A brass needle suspended, like a magnetic
needle, is not moved by the effect of the
uniting wire. Needles of glass and of gumlac
remain likewise quiescent. "
On first thought it may seem singular that
as manj' as twenty years elapsed after the
Galvani and Volta discoveries before such a
simple experiment as that of Oersted was
tried. The only hint or suggestion of prior
observation appears in the statement that
about 1S02 Romagnasi of Trent (a town in
the Austrian Tyrol) had noticed an effect
on a compass needle in the neighborhood of
a voltaic pile. Evidently, however, the ob-
servation made was very imperfect, as it
led to no consistent recorded result. In this
connection we must consider that the early
years of the last century were disturbed b}'
wars stirring the whole of Europe and further
that the available voltaic currents must have
been relatively weak owing to the small area
of batter\- plates used, with a high resist-
ance electrolyte. Strong acid could not be
availed of as the zinc elements were not
amalgamated, a procedure which was later
almost universal. Again, the negative ele-
ment was usually copper, giving against zinc
a low voltage and subject to rapid polariza-
tion. There was, therefore, in the years be-
fore Oersted, little probability of large cur-
rents being available, such as would be needed
when a single wire was used for the deflecting
agency. This was, of course, before the
principle of coiling the conductor to increase
its effect was known. Dr. Robert Hare, the
inventor of the oxyhydrogen or compound
blowpipe, apparently, in 1816 first appre-
ciated the need of increasing the surface of
the zinc and copper to obtain, as it was after-
ward called, "large quantity. " The blowpipe
in his hands had become the source of heat
of highest temperature known to man, and
the known heating eft'ects of electric currents
naturally led Hare to investigate means of
intensifying them. He produced two forms
of apparatus which were known as the Hare
calorimotor and the Hare deflagrator. In the
prior "trough" batters' the plates were small,
rarely more than four inches square, with only
one side active. Hare rolled his zinc and
copper sheets into interlaced spirals, spaced
apart by wooden separators, so that not only
large plates could be used, but both sides of
the plates were active. Another form giving
a similar result was embodied in the "def-
lagrator" which was used to deflagrate strips
of thin metal in the same manner as the blow-
ing of a modern safety fuse.
In early youth it was the privilege of the
writer to see examples of the apparatus of
Hare which were preser\*ed at the University
of Pennsylvania, then located in Philadelphia,
on the west side of Ninth St., between Chest-
nut and Market streets. Hare had been
Professor of Chemistry there during the early
years of the past century. The Hare appa-
ratus is mentioned here because of a passage
occurring in a work on Heat and Electricity,
printed in 1830. Its author, Thomas Thom-
97S December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
son, M.D., was very eminent in the science
of his time. The passage reads as follows:
"The apparatus employed by Oersted, and of
the efficacy of which he speaks in high terms,
approached very nearly to this last one of
Hare. " This passage occurs just following a
description of the " deflagrator. " The state-
ment seems to imply that Oersted early ap-
preciated the value in a voltaic batterj' of
large active surface, or as we should now say
"low internal resistance." Maj^ it not be
that this condition was the secret of his
experimental success? Even in Franklin's
time it had been obser\'ed that electric dis-
charges had some obscure action on the mag-
netic needle, for sometimes compasses were
demagnetized wholly or in part, or even
reversed when in proximitj' to a lightning
stroke. Beccaria, the Italian contemporary^
of Franklin, tried manj^ experiments with
magnetized needles and heavy Leyden jar
condenser discharges sent through them, but
did not succeed in establishing any magnetic
effect as due to the discharge. Such effects as
he did obtain arc readily interpretable at
this time as the natural result of a vigorous
shaking up, mechanical, or thermal, while
the needle was in a magnetic field, such for
example as that of the earth. Subsequently
to Oersted's discovery, however, condenser
discharges sent through a helix or spiral sur-
rounding the needle were found to produce
decided magnetic effects thereon.
Ampere
The news of Oersted's discovery reached
Paris, as it appears, through Arago, on Sept.
11, 1820, who had witnessed the experiments
in Geneva, and on September IS, Andre
Marie Ampere presented a paper to the Paris
Academy of Sciences, remarkable for its
originality and for the variety and accuracy
of the experimental results recorded. In it he
dealt with the interactions and repulsions of
wires conveying currents and the magnetic
effects of helices of wire, showing these latter
to be possessed of magnetic poles.
The fact that Ampere's paper appeared
only a week after Oersted's discovery had
become known to him gives to it the unique
place in the history of science. It was the
production of a mind of the first order work-
ing at high pressure. Ampere was bom at
Lyons in 1775. As a child his precocity was
most unusual. His tendency was towardmathe-
matics, though his reading during youth
brought to him a wide range of information
on many branches of knowledge. The death
of his father on the scaffold as a victim of
political conditions almost wrecked his young
life. Owing later to fortunate environment
at a critical time, he gradually recovered, and
in 1809, at thirty-four, was made Professor
of Analysis at the Paris Polytechnic. The
Oersted announcement evidently stirred him
deeply and he went immediately to work with
wonderful zeal and sagacity, to unravel the
mysterious relationship between magnetism
and currents, suggested only in part by
Oersted's experiment. In his hands the fun-
damental character of these relationships be-
came plain, and their future possibilities clear.
It is recorded that he even suggested at the
time the plan of a telegraph of simple form.
As a side light, it appears that Ampere's
discover}^ of the fact that parallel currents
in the same direction in wires cause attrac-
tion, and when in the opposite direction re-
pulsion, seems to have puzzled some of the
philosophers of the time, for it was known of
course that similar electricities repel and dis-
similar attract. Why should not, therefore,
similar currents repel and unlike attract,
when in fact the exact opposite was the case ?
Ingenious, though altogether fallacious, was
an explanation first put forward and credited
to Oersted. He was apparently driven to
imagining that the current in wires did not
go straight along the axis, but was conducted
in a helical course in them always the same
for the same direction of current. This was
a pure invention without facts to support it.
But upon it he founded a theor\' of attraction
and repulsion of parallel wires involving at-
traction of unlike and repulsion of like elec-
tricities. In reality this theon.* bore with it its
own refutation, as a few test experiments
might easily have shown.
In relation to this fanciful theor\- I find in
the old book, before referred to as published
in 1S30, the following : "This way of account-
ing for the phenomena of electro-magnetism
was first employed by Oersted. It was after-
wards used by others, particularly by Dr.
Wollaston and M. Ampere, with much felic-
ity." The writer does not vouch for the
correctness of these statements. As soon,
however, as the effects of currents in estab-
lishing magnetic circuits around them was
worked out, the true cause of the attractions
and repulsions of parallel wires became clear,
and the fanciful notion of spiral courses for
currents inside a conductor was abandoned.
Such a notion has of course no relation to the
later theor\' of Ampdrc for accounting for
magnetism in iron or permanent magnets
ONE HUNDRED YEARS SINCE OERSTED, AMPERE AND ARAGO
979
in which he assumes each magnetic element
to consist of a closed circuit with a current
always circulating therein, a theory which to
this day has not been displaced, but rather
refined and strengthened by its further ex-
tension by Ewing and by the electron
theory. It was in fact Ampere who referred
all magnetism to electricity or electric cur-
rents, now interpreted as movement of elec-
trons. The need of Ampere's clarification is
perhaps made more evident from the fol-
lowing quotation, if it be a fact: "Oersted
originally believed that the negative elec-
tricity propelled the north pole of the magnet,
but had no effect on the south ; while positive
electricity propelled the south and had no
effect on the north pole. " The writer has not
verified this statement as expressing the
original ideas of Oersted, but if they at any
time represented his view they must soon
have been dispelled. They would be per-
haps a sort of survival of older notions, at
least in part, since before the "conjunctive
wire" was used fruitless efforts had been
made to connect magnetism and electricity
while using batteries on open circuit.
Ampere formulated a simple rule known as
Ampere's rule, for determining the direction
of deflection of a magnet or the direction of
development of magnetism by a wire con-
veying current. It may be stated (bearing
in mind that the direction, we assume, posi-
tive to negative, is merely a convention)
about as follows: Conceive one's self lying
or swimming in the current in such a way that
the current enters by the feet and leaves by
the head as we face the needle. Then the
action will be that the north pole of the needle
will turn to one's left. The writer must con-
fess that when he first learned this rule it
seemed rather clumsy to him, and he was
sometimes treated to the ludicrous spectacle
of an obese professor trying to twist himself
with respect to an immovable wire circuit
into curious attitudes so as to lie or swim in
the current, and so note the direction of mag-
netism produced.
Other ways of remembering the relation
given by Ampere's rule have been devised,
but perhaps none excel in ease and simplicity
of application a simple gesture of the hand
which has been used by the writer for about
fifty years. The hand is held out with the
index finger pointed away. If the hand be
now given a swing or turn in righthanded
direction, still keeping the forefinger directed
as at first, such swing, turn, or rotation repre-
senting direction of current in a circuit, the
north magnetic pole will be directed away as
the forefinger points. Reversing the gesture,
turning or swinging the hand counter-clock-
wise makes the north pole take direction
toward the wrist or forearm, or what is the
same thing the extended forefinger represents
the direction of south polarity. As the swing
or slight rotation given the hand from the
wrist and elbow represents all directions of
current, above, below, and to the right or
left of the magnetic axle considered, it is
easy to select any element of current course
matching actual conditions. Moreover, the
same gesture (for it is a gesture, not a rule)
applies equally to the relations of magnetic
field developed around the course of a current,
for if current passes in a wire in the direction
of the point of the index finger, the magnetic
circuit around it will have north polarity di-
rected righthandedly, and lefthandedly or
counter-clockwise if the current has opposite
direction, as from the tip of the index finger
inward toward the wrist. In any case it is
only necessary to make the proper gesture,
which requires no especial mental effort. This
soon becomes a matter of habit, a mistake
being practically impossible.
Arago
According to De la Rive, "Traite D'Elec-
tricite," Vol.1, it was Arago who was first to
show that a wire of copper or other metal
acquired, when traversed by a strong current,
the property of attracting and retaining
around it, under the form of a cylindrical
envelope, a quantity of iron filings, the filings
falling off immediately when the current
ceased to flow, and being reattracted on the
restoration of the current. This experiment,
prior to all those of Ampere, is the first
which established in a striking manner that
electric current impresses on conductors when
it is transmitted by them properties fully
analogous to those of magnets, and not alone
to magnetic bodies; in other terms, that it
magnetizes them and does not simply render
them susceptible of being magnetized. In
fact, the iron filings are magnetized by the
current as they would be by a magnet, and
are in consequence attracted by the wire
which transmits the current. This statement
is substantially that of the account given by
De la Rive, translated. Ampere and Davy
are credited with having made the same ob-
servation, but if De la Rive is right, it was
first made by Arago.
Here then was the first exemplification of
the phenomenon of temporary magnetism in
980 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
iron, so fundamental to the unlimited variety
of electromagnets and mechanisms founded
thereon; the basic principle of the Morse
telegraph and most other signalling or electric
recording systems and essential to the greater
machinery of electrical engineering, developed
for the most part in the latter half of the past
century.
Arago went farther and made his conduct-
ing wire into a spiral, and then succeeded in
magnetizing steel needles placed in the axis
thereof. In his pioneer experiments the spiral
was wound open around a glass tube as a sup-
port, the wire itself being presumably bare.
In Silliman's "Principles of Physics," a well
known and much used textbook from its
first edition in 1858, and for perhaps thirty
years thereafter, in describing the above
experiment there is the following statement ;
"If the helix is wound on a tube of glass,
paper or wood, these substances offer no
resistance to the passage of the power; but
if a tube of copper or other metal were em-
ployed, the magnetizing power of the current
on the enclosed bar would be destroyed."
Such a statement means either that the metal
tube short circuited the bare wire of the helix
or that currents of extremely brief duration,
such as condenser discharges, were concerned.
It was indeed soon found that even "if
common electricity be made to pass along the
spiral conducting wire, the needle is equally
converted into a magnet." Common elec-
tricity was evidently the frictional or static
electricity thus distinguished from the newer
or less common voltaic current, later called
dynamic electricity. The principle of coiling
or increasing the turns in the original rec-
tangular circuits of Oersted and Ampere was
soon appreciated, even for the deflection of
needles, a hank of insulated wire wound
around the hand, or upon a rectangular block
of wood and tied by string to preserve its form
was used to surround the needle; the pro-
totype of the taped coil of today. It seems
to have taken some time to develop the coil
consisting of a spool or bobbin wound closely
with insulated wire. Schweiggcr used the
rectangular coil of many turns on his "Elec-
tric multiplier" the term "multiijlicr" being
extant for at least fifty years as applied to
galvanometers with such coils.
The floating battery of De la Rive with its
conjunctive wire, or spiral connecting the
poles, was an exceedingly neat arrangement
for showing the neutral action of currents
and magnets, or the effects of wires convey-
ing currents on others more or less parallel
thereto. It seems to have been a very earlv
modification of Ampere's apparatus. It
avoided the problem of pivots conveying
current, mercur\- cups being usually em-
ployed.
Arago's famous disc experiment, involving
the discovery that a moving conducting disk
of non-magnetic metal, such as copper, pos-
sessed an effect in the nature of a drag on a
poised magnetic needle near it, was made
about 1824. It was found also that if the
needle was spun around over a conducting
disc or plate, it was rapidly slowed or damped.
These e.xperiments were carried on using a
great variety of materials, and wide varia-
tions in the magnitude of the effects were ob-
served. Precautions were taken to eliminate
any effects due to air currents. It was found
that discs of the best conducting metal, such
as copper, were the most effective. Here we
have, then, the prototype of dampers in
magnetic fields. Curious hypotheses were
advanced to account for the effects, such as
the assumption of special forms of magnetism
generated by revolution. In reality, had the
secret of the action of the Arago disc been
found, the generation of currents in a con-
ductor moving in a magnetic field would have
been discovered, and Faraday's discovery of
that great principle in 1831 would have been
anticipated by a number of years.
The full name of Arago was Francois Jean
Dominique Arago, and he was a scientist of
varied activities. There is no space here to
refer to his career, except in the briefest
possible way. Bom in 1786 at Estagcl, near
Perpignan, France, he displayed in his early
years great aptitude for learning, and at
eighteen became secretary to the Obscr\-a-
tory of Paris, which brought him into contact
with the famous La Place, and he was "colla-
boratcur" with Biot. He served in the de-
termination of the meter as the unit of length,
in measuring the ten-millionth of a quadrant
of the earth's meridian. This task, involving
travel into lands in turmoil, brought great
dangers, imprisonment, escape in disguise,
capture by the Spanish and prison again.
Even after release he remained a long time
in quarantine, on account of disease condi-
tions. Given the post of astronomer in the
Royal Observ-atory in Paris by Napoleon,
in 181G, he started the famous journal
"Annales de Physique ct de Chimic. " As
secretary of the Paris Academy of Science, and
Chief of Deputies, his life was a very full one,
and its responsibilities not light. He was
associated with Frcsncl in giving fonn to the
ONE HUNDRED YEARS SINCE OERSTED, AMPERE AND ARAGO
981
undulaton^ theory of light, proposing to
test the theory bj^ studying retardation in
refractive media. In fact his work on polar-
ization of light, invention of the polariscope,
and other researches rank scarcely less highly
than his work in electro-magnetism, with
which we are here chiefly concerned.
Subsequent Discoveries and Application
The later discoveries by Faraday and his
brilliant work on electromagnetic rotations,
especially his discovery- in 1831 of induction
of currents by magnetism, refined the early
theories and added greatly to the develop-
ment of electromagnetic science. Some-
what crude as the earlier ideas were, the
clarification given them by Faraday, Max-
well, Kelvin and many others had the most
profound eff'ect on its future. As a direct
outcome of Oersted's observations, mention
may be made of the discovery in 1823 by
Seebeck of thermoelectric currents in a
closed circuit. He used a rectangle in a
vertical plane surrounding a pivoted magnetic
needle. The base of this closed circuit so
arranged was a bar of antimony, while the
ends and upper side were of copper. By heat-
ing one of the junctions of the antimony bar
with the copper, deflection of the needle showed
the presence of magnetism in the closed loop
around the needle. This was followed by
examination of the effects of junctions of
different metals and conductors heated to
various temperatures, and led to the well
known table of thermo electric powers. The
Melloni Thermo-pile, so delicate as a heat
detector, was the outcome, used by him in
his beautiful researches on Diathermany.
Having in the foregoing traced briefl}'
the work of the pioneers in laying the founda-
tion of the science a century ago, it is perhaps
unnecessary,' to remind electricians and en-
gineers of the great scientific advances and
the important applications which soon fol-
lowed. Some of them became familiar
studies of the electrical student fifty years
ago. This progress has continued and ap-
parently at an increasing rate ever since.
To the consciousness of the writer the
period of a hundred years seems continually
to dwindle. He is reminded of the fact that
his own life's span has covered more than
two thirds of a century. Looked at in this
way, the Oersted, Ampere and Arago ex-
periments do not seem to have been made,
after all, so ver>' long ago. Outside of the
forms of electromagnetic telegraph, the years
following 1820 saw but few other applications
of importance, but there were many examples
of electromagnetic apparatus used for in-
struction in schools. The little book now
rare, entitled "Davis' Manual of Magnetism"
was and is interesting as a catalogue, with
brief descriptions of such apparatus, some of
which is doubtless still extant in the older
collections. The first edition was published
in Boston, in 1842, and the author, Daniel
Davis, Jr., called himself " ^Nlagnetical In-
strument Maker."
How many, or rather how few of us are
left of those who as boys experimented with
the sulphate of copper battery as their source
of current, with flat spirals such as Henry
used, or with such apparatus as Oersted,
Ampere and Arago used. We find there
Henr\''s electromagnet, De la Rive's floating
batten,', Faraday's revolving circuits and
magnets, Barlow's spur wheel, Page's re-
volving ring, his revolving magnet and re-
volving multiplier, and other examples of
the simplest types of electric motors with
commutator and brushes called "pole chang-
ers" and even apparatus with both com-
mutator, revolving brushes, and slip rings, so
that both elements of the motor might re-
volve oppositely. There were bell engines,
and reciprocating engines, elementar\' motors
driven by thermoelectric currents, or by bat-
teries revolving, all involving the simple
principles of interaction of circuits and mag-
nets permanent or temporary. These and
other simple forms of apparatus, besides the
so-called devices for static electricity, were
the things electrical with which the youngster
with an electrical bent became familiar either
in his reading, or better, by the fascination
of experiment with them. Such equipment
characterized the infant years of the science
now grown to a giant, with no limit to future
growth.
It was natural that the first great practical
application of electromagnetic principles
should be found in the telegraph.
Attempts had been made as early as
1774 to telegraph by the electricity of fric-
tional machines, which even as late as 1850
was called ' ' machine electricity " or " common
electricity from machines." Even in 1816,
Ronalds in England was attempting to signal
through long circuits by Leyden Jar dis-
charges. After the discovery of the voltaic
pile in 1800 there was a better prospect of
success, and Sommering in 1808 proposed
a system of 35 wires at the ends of which
were gold strips in water, upon which strips
gas appeared on the passage of current, which
982 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIIl, No. 12
appearance constituted the signal received.
There was a wire so arranged for each letter
or character transmitted. It was Ampere
who, just after Oersted's discovery, proposed
to substitute in Sommering's system deflected
needles for the voltameter receivers. Then
followed Schilling in Russia, in 1S32; Gauss
and Weber at Gottingen, in 1833; and
finally Cooke and Wheatstone in England,
and Steinheil in Munich, in 1837, to whom
perhaps more than to any others the develop-
ment of the needle telegraph for practical
work was really due.
The Morse type of telegraph was early dis-
tinguished as the electromagnetic telegraph,
or one based on the use of electromagnets.
Barlow in England seems to have made an
early suggestion of the kind, but it was not
until 1830, upon the construction of the first
powerful electromagnets by Joseph Henry,
of Princeton, New Jersey, that such a form
became possible. In his first paper on the
results of his experiments, Henry proposes
to apply them to the telegraph. Samuel
F. B. Morse conceived of such a telegraph in
1832, and with the assistance of Vail worked
it out practically and publicly exhibited the
Morse system working over a circuit of a
third of a mile in 1837, but after that it was
nearly seven long years before Congress
for consideration, when at last a modest
grant was made to establish the famous
Baltimore-Washington line first put into
operation in 1844.
In the subsequent numerous developments
of systems of signalling, from the simple
call bell to the fire alarm and printing tele-
graph, the electromagnet holds undisputed
sway. In annunciators of many types it is
found, as often in relays, in telephone re-
ceivers and the like, with polarized cores.
It is as indispensable in wireless transmission
as in transmission by wires. The extreme
sensibility of the telephone receiver, coupled
with the wonderful delicacy of the ear, make
it most effective for the detection of minute
electrical disturbances. With modern therm-
ionic amplifiers the possible extension in
range of telephonic transmission by wireless
waves seems to be without limit.
It is not necessary here to allude to the
great developments in the field of electricity
and electromagnetism as exemplified in
generation and transmission of electrical
energy. They have covered the past half
century, but the foundation principles belong
to those early years of upward of a century
ago. Do we cause movement of iron masses
by a current coil? It is the experiment of
Oersted. Do we cause movement of coils,
one with relation to another, as in our motors?
It is the experiment of Ampere. Do we gen-
erate currents in a conducting mass in a mag-
netic field? It is the experiment of the Arago
disc. When we measure current or energy
by galvanometer, voltmeter, electro-dyna-
mometer, or wattmeter we have the work of
Oersted, Ampere, Arago, illustrated. But
these early discoveries had a depeer signifi-
cance still. They showed that electric cur-
rents and magnetism are inseparable — in-
separable in practice, inseparable in theory.
Moving charges are, as shown by Rowland,
the equivalent of currents; and now we are
assured that moving charges and currents
are moving electrons. Hence moving elec-
trons are magnetic. Like charges or electrons
repel each other, but like charges moving in
the same general direction attract, for they
are the equivalent of parallel currents in the
same direction. They repel one another
electrostatically and attract one another mag-
netically. Conversely, oppositely moving
charges or electrons should repel, but still not
attract, but continue to repel electrostatically.
We are now certainly down to the fundamen-
tals. No vacuum, however perfect, lessens
or stops the development of the electric field
in it, nor prevents the existence of the mag-
netic field. Space itself is electromagnetic,
using the term in its broadest sense. Like
electrons moving in the same direction in
such space must attract one another and at
some speed the static repulsion of the like
charges will be balanced by this attraction.
The higher the speed, the closer is the ap-
proach, until the repulsion balances the at-
traction. And just here is the key phenom-
enon of nature. Space, whether empty or full
of ether, is fundamentally electromagnetic
and perhaps only that. Energy and mass,
interchangeable terms and due to relative
movements of electrons, is electromagnetic
and only that. Matter in all its forms,
systems of electrons in motion, is electro-
magnetic. Alive or dead it is electro-
magnetic and nothing else. All properties of
matter, all forms of energy arc electromag-
netic and electrostatic. If ether exists it is
purely and solely electric and magnetic, with-
out mechanical properitcs, for such proper-
tics depend on motion of electrons. Ether as
a medium, then, not being mechanical, can
neither be in rest nor in motion; it can
only be the theater of electrostatic and
magnetic conditions, whatever they may be.
DR. ELIHU THOMSON
983
These statements may be very sweeping, but
do not the recent notions of the relativity
of Einstein carry us even farther? Space is
empty, but has a warp in it; it is curved, of
four dimensions, (one being time), in which
the gravitational field of the sun, or for that
matter even the smallest speck of matter or
energy, is a local warping. Are electric and
magnetic fields but other kinds of warping in
this space which though empty is full of elec-
tric, magnetic and gravitational fields ? What-
ever all this may mean, we must remember
that scientific theories can never change the
facts; they are not creeds, they are means of
pointing the way to further additions to our
knowledge — ^to be modified, changed or
abandoned according to their usefulness in \
leading to further knowledge or discovery.
The facts of science are its bed rock founda-
tion, unchanged and unchanging. The dis-
covery, then, of the relation between elec-
tricity and magnetism was in reality the dis-
covery of a fundamental fact or principle
lying at the foundation of the Universe itself,
the soul of energy, as of matter; of electric
waves from zero periodicity up to the most
penetrating rays of the raduim emanations.
It is eminently fitting, then, that we cele-
brate the hundredth anniversary of dis-
coveries, the fruits of which have been of
stupendous influence and value, and at the
same time carry us to the very foundations
of existence.
The coming century will doubtless have its
wonders to unfold, but it is fairly safe to pre-
dict that they can hardly exceed in funda-
mental bearing those revealed to us in the
past hundred years.
Dr. Elihu Thomson
SCIENTIST, INVENTOR, AND EDUCATOR
By DuGALD C. Jackson
Dr. Thomson has been acting president of Massachusetts Institute of Technology since early spring of
this year, and the following sketch of his career by a member of the faculty will be read with interest by the
manv admirers of our eminent scientist. — Editor.
Professor Elihu
Thomson : If you
speak that name
in electrical engi-
neering circles, in
any part of the
world, it at once
brings a word of
recognition, for
Thomson is one of
the fathers of elec-
trical engineering
as well as a nota-
ble leader and mas-
ter. Though born
in England, Professor Thomson came to the
United States with his parents when he was
five years old and grew up and was educated
in Philadelphia. There he graduated from the
Central High School with the degree of A. B.
at the age of seventeen, and was immediately
appointed Assistant Professor of Chemistry
in the same school. Six years later he was
made Professor of Chemistry and Mechanics.
Here he began his researches in electricity and
started on his career as an inventor. Fame
has been added to the United States and hie
name spread in technical and scientific circles
for the great range and importance of his
achievements in electrical engineering.
Interest in electrical phenomena held him
for years before his graduation from the Cen-
tral High School, but it was during his pro-
fessorship that he started on his revolutionary
series of researches and inventions. Here he
invented and constructed the famous arc light
dynamo with spherical three-coil armature,
which went into commercial use in 18S0 and
continued a central figure in arc lighting serv-
ice as long as the arc lainp with carbon elec-
trodes held the field of illumination by large
lightingunits. Here also he made the researches
and discoveries which gave the foundation of
his later primary inventions in electrical weld-
ing. In ISSO Professor Thomson resigned his
post in the Central High School and moved to
New Britain, Conn., to become the technical
head of a company called the American Elec-
tric Company which was established on the
foundation of his inventions. Three years
later this companv was reorganized on a larger
scale, moved to Lynn, Mass., and was renamed
the Thomson-Houston Electric Company.
For many years the company occupied what
984 December, 1920
GENERAL ELECTRIC REVIEW
VoL XXIII. Xo. 12
are no^v called the West Lynn Works, -n'hich
were gradually increased in size and impor-
tance as the variety and extent of the business
grew. After the Thomson-Houston Company
became associated with the Edison General
Electric Company to make the great company
known as the General Electric Company, the
River Works were built at Lynn, but the West
Lynn Works continue to produce many of the
important products which the genius of Pro-
fessor Thomson has conferred on electrical
engineering.
To list all of his inventions would be beyond
the scope of this article, for more than six
hundred patents have been issued to him by
the United States. His inventions in dynamo
electric machinery-, electric welding, electric
watt-hour meters, lightning arresters and
magnetic arc extinguishers are fundamental.
The arc lamp, the incandescent lamp, the elec-
tric motor, the alternator, the alternating cur-
rent transformer, railway motors, high fre-
quency apparatus, and innumerable other
devices of electrical engineering have found
improvement at his hands. His touch has
ever been of originality and sound scientific
conception, so that ever>- part of the art
which has passed before him for review has
profited from the activity' of his illuminating
mind.
Broad recognition has come to him at home
and abroad. Yale University conferred upon
him the honorary- degree of Alaster of Arts in
1890, Tufts College the degree of Doctor of
Philosophy in 1894, and Har\-ard University
the degree of Doctor of Science in 1909. He
was awarded the Grand Prix at the Paris Ex-
position in 1889 for his electrical discoveries
and inventions, and was decorated by the
French Republic as Chevalier et Officier de la
Legion D'Honneur. Again at the Paris Expo-
sition of 1900 he was awarded the Grand Prix.
At the Saint Louis Exposition of 1904 he was
again awarded the Grand Prize for his electri-
cal achievements. He was president of the
American Institute of Electrical Engineers in
1889 and in 1910 was made the first recipient
of its famous Edison Medal for meritorious
achievement in electricity. In 1916 he re-
ceived the John Fritz iledal of the four
national engineering societies, awarded to
him for achievements in electrical invention,
electrical engineering, industrial developments
and scientific research. In 1916 also he was
awarded the Hughes Medal of the Royal So-
ciety of London. This award carried a
money prize which Dr. Thomson donated to
a war charity in England. Many lesser
medals have also been conferred upon him.
His recognition by scientific and professional
societies has been world wide. Having
been President of the American Institute of
Electrical Engineers in 1889-90, he was L'nited
States delegate to the International Electrical
Congresses at Chicago in 1893 and St. Louis
in 1904, and was President of the Chamber of
Delegates and of the -Congress at St. Louis.
For the three years 1908-11 he was President
of the International Electrotechnical Com-
mission which has for its function the impor-
tant duty of arranging international standard-
ization in electrical engineering, also of provid-
ing for the International Electrical Congresses
at which units and standards are adopted.
Of his numerous scientific and professional
societies only his membership in the National
Academy of Sciences and honorary member-
ship in the Institution of Electrical Engineers
of Great Britain can be mentioned here.
Professor Thomson has been noted for his
interest in the careers of younger men, and his
assistance and counsel are remembered with
affectionate gratefulness by many men who
themselves have come to distinguished places
in the electrical engineering profession. Men
who have had the fortune to ser\-e as his assist-
ants, rejoice in telling of his "many-sided-
ness" and fertility, for he himself has served
them as a L'niversity.
— The Tech Engineering News.
985
Studies in Lightning Protection on 4000-volt Circuits*
By D. W. Roper
Commonwealth Edison Company, Chicago
The author presents the results of an investigation of lightning arrester performance in practice extending
over a period of five years. The investigation was originally undertaken for the object of reconciling the dif-
ferences between results obtained in laboratory experiments and actual service, and some of the conclusions
previously arrived at were presented in a paper before the A.I.E.E. in June, 1916. The scope of the investiga-
tion broadened to a determination of the relative merits of the several types of lightning arresters which were
installed on the system under consideration, and the data thus compiled constitute the most valuable contribu-
tion ever made to the study of lightning disturbances in primary distribution networks.
Discussions by Dr. C. P. Steinmetz, W. L. R. Hayden and V. E. Goodwin accompany our abstract of the
paper. — Editor.
Introduction
The investigations forming the basis of
this paper as well as the previous paper^ on
the same subject had as their primary object
the determination of the relative merits of
the several types of lightning arresters which
were installed on the 60-cycle distribution
system of the Commonwealth Edison Com-
pany in Chicago. The previous investigations
had indicated in a general way the several
factors which affected lightning arrester
performance and also the extreme variability
of the distribution and intensity of the
lightning storms, from which it appeared
that in order to get reasonably accurate
results, it would be necessary to accumulate
the experience with a large number of arrest-
ers over a period of several years.
Description of the System
The system of distribution on which these
investigations were made is a four-wire three-
phase system, with the neutral grounded only
at the substations. The normal potential
on the distributing mains is 20S() volts
between phase and neutral wires. The dis-
tribution pole lines are in the alleys, or along
the rear lot lines in the center of the block
where alleys are missing. Single-phase trans-
formers are used exclusively and are con-
nected between the phase and neutral wires
except in the case of three-transformer three-
phase installations in which case the common
point of the transformer primaries is not
connected to the neutral wire. Secondaries
of power transformers are connected in delta.
Power and lighting customers are supplied
from the same primary mains, but the very
large customers are connected to a 12,000-
volt system. The feeders are all No. 0 wire and
the mains No. 6 A. W. G. About 85 per cent
of the feeders and 15 per cent of the mains are
*A comprehensive abstract of paper presented before A.I.E.E.
at Chicago. November. 1920.
1 Trans. A.I.E.E.. 1916. Vol. XXXV, p. 655.
underground. About 99 per cent of the
transformers are on poles and the rest in
manholes or in vaults on customers' premises.
At single transforiner installations a 2400-
volt arrester is connected to the same phase
wire as the transformer and a 300-volt
arrester to the neutral wire. Where three
transformers are installed for a power service
there are three 2-iOO-volt arresters, one con-
nected to each of the phase wires; and one
300-volt arrester is connected to the neutral
wire. Arresters are installed in this manner
on the same pole with all transformers. The
lightning arrester ground consists of one-half
inch galvanized iron pipe ten feet long, driven
into the ground at the base of the transformer
pole. Secondary circuits are usually less than
one block long and the secondary ground is
similar to the lightning arrester ground, but
3000
S
\
\
1
\
1 2000
\
\
o
\
\
at
\
3
z
\
/
/
\
1
\
\
1
\
\
\
1
\
\
\
\
/
\,
1 1.5 2 2.5 3 4 5 7,5 10 15 20 25 30 37.5 40 50 75 100 150
SIZE OF TRANSFORMERS KVA
Fig. 1. Diagram Showing the Number of Each Size of Trans-
former in Service on August I, 1918
is installed on the next pole. On long second-
aries there are at least two such ground con-
nections and in addition the neutral wire on the
customer's premises in many recent installa-
tions is grounded to the water pipes insideof the
building. The distribution system at this time
986 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
Arresters E and F have iden-
tical diagrams and differ princi-
pally in mechanical details, the
amount of resistance and the
length of the resistance rod.
^
E&F
Diagram G shows a type
which was installed in 1920.
Diagram H represents the
neutral 300-volt arrester in-
stalled on the neutral.
O
Q
H
Fig. 2. Electrical Diagrams of the Lightning Arresters
Used in These Investigations
The gaps are conventional and do not show the actual
shape of the gaps on all of the arresters.
0.7
06
05
0.4
SO-3
CJ
£0.2
Q.
^0.1
z
£ 0
£2
-0.5
CD
^0.4
o
o0.3
io2
a"'
1 0
as
0.1
(jUa — t^.
191.^- —'
1917 ...
-^
1919.
1920
-?^
y
T-
/
^
1
1
1 1
\,
1 u-^
\
A
\,
-/-
X"
y^
y
\
\
\
\
\
FUSM
42.„
y
\
^
\
B^„;:S
, '
y
\
1 ! i
20.000
Fig,
1912 1913 19M I4)5 1916 1917 1918 1919 1920
3. Graphical Record Showing the Number of Transformers
in the Distribution System and the Per Cent of Transformer
Troubles Each Year Over a Period of Years
LIGHTNING
I I Jjw Af?R£:5TER AREAi
-±.
0.4
OS
02
0.1
0
0.5
a4
0.3
0.2
01
MAR APRIL MAY JUNE JULY
Fig. 5. Diagram Showing the Percentage
out in Each Storm for the Years 19
AUG. SEPT. OCT.
of Transformers Burned
15-1919 Inclusive
Fig. 5. Outline Map of the City, Showing Section Lines
and Lightning Arrester Area Number*
The lightly shaded areas show the sections in which
arrester G shown in Fig. 2 was installed m 1920, The
heavily shaded portions show the sections m which addi-
tional arresters of type .-1 were installed m 1921) for the
purpose of getting more conclusive information regarding
this type.
STUDIES IN LIGHTNING PROTECTION ON 4000-VOLT CIRCUITS
987
includes about 100,000 poles, 20,000 trans-
formers with a total capacity of about 270,000
kv-a., 6500 conductor miles of overhead primar}'
line wire, 2200 con ductormilesof underground
cable and 2500 cable poles. The system
ser\'es about 400,000 customers.
In Fig. 1 is shown the number of each size
of transformers on the line as of August 1,
1918. This date was selected for the purpose
of the calculations, as the number of trans-
formers in service on that date was about
the average of the number in the five-year
period under investigation. Fig. 2 shows
electrical diagrams of all of the lightning
arresters used in the investigations. The
letters shown on this diagram are consistently
used throughout the several tables, diagrams
and curves. Fig. 3 shows graphically the
number of transformers on the distribution
system over a period of years, as well as the
percentage of transformer primary fuses
blown and transformers burned out by light-
ning each year. The increase in the percent-
age of fuses blown during the years 1918
and 1919 was due to causes defmitelj^ known
to be entirely distinct from lightning, but as
some of these fuses were blown during the
same day as lightning storms, they were
included with fuses blown by lightning
because of the impossibility of accurately
determining just which fuses were blown by
lightning and which by other causes.
In Fig. 4 is shown the percentage of burn-
outs of transfonners for each storm during
the five-year period and also for the year 1920,
the percentages being plotted cumulatively.
From these records it will be noted that it is
MARCH APRIL MAY JUNE JULY AUGUST SEPTEMBER
Fig. 6. A Composite Diagram of the Transformer Burn-outs for
1915-1919 Inclusive
not unusual to have over one-third of the total
trouble in any one year due to lightning
occur in one or two daj^s. A composite of these
curves for the five-year period is shown in
Fig. 6, from which it will be noted that on the
average, the lightning is quite uniformly
distributed throughout the 4}^ months from
May 1st to September 15th, and that there
is comparatively little trouble outside of this
period.
Fig. ^ is an outline map of the portion of the
city covered by the distribution system on
August 1, 1918, showing the section lines
and the lightning arrester area numbers.
These areas will be found to differ from those
shown in the previous paper as some changes
were made in 1917 for the purpose of trying
another type of arrester, a new scheme of
protection and incidentally securing a little
better distribution of the various types of
arresters over the different portions of the
city. The shaded areas on this diagram will
be referred to later in the paper.
Preliminary Investigations
From the previous paper and subsequent
studies it appears that the factors which
might affect lightning arrester performance
are as follows ;
1. The system of distribution and the
grounding of the neutral.
2. Primary terminal boards.
3. The shielding effect of trees, buildings
or wires of other companies.
4. The resistance of the lightning arrester
ground connection.
5. The maker of the transformer.
6. The size of the transformer.
7. The age and previous service record
of the transformer.
8. Variation in the distribution and inten-
sity of the lightning.
9. The density of lightning arresters, that
is, the number per square mile.
10. The design of the arrester.
In laying out the lightning arrester
areas which w-ere given in the previous
paper, and which are also shown in
Fig. 5 in this paper, it was the inten-
tion to arrange the boundaries of the
areas and to distribute the several
types of lightning arresters over the
city so as to eliminate variables 3 to 8
inclusive as given in the above list.
An investigation of the records dem-
onstrated beyond question that the
shielding effect of trees or buildings
immediately adjacent to the lines con-
siderably reduced the amount of damage on
our lines from lightning. This was shown by
the following facts :
(a) The percentage of poles in the dis-
tribution system shattered by direct strokes
is extremely small, being of the order of
OCTOBER
Years
988 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIIl, No. 12
1/400 of 1 per cent. This is very much
smaller than the corresponding percentage
for transmission line poles belonging to the
same company in the flat open country in the
OOHVCHTUH 1
=a^°°-'°° -(nV
I I
Fig. 7. Outline Map of Chicago, Shomng Section Lines and Area Actually
Covered by Distribution System. Density of shading indicates
the density of lightning arresters
southeastern portion of the city and is also
smaller than experienced in general bv
companies having transmission lines crossing
open country-. That there are many direct
strokes in ever\- severe lightning
storm is shown by the newspaper
reports on the day following light-
ning storms, which record the most
severe or unusual cases of damage
to trees, church steeples, chimneys,
or other portions of buildings and
structures.
(b) An investigation of the con-
ditions surrounding the installation
of 97 out of 529 cases covered by
these investigations where transform-
ers were burned out by lightning
failed to reveal a single case in which
the primarv- wires adjacent to the
transformer were over-shadowed by
high trees or buildings immediately
adjacent.
By "spot checking" selected por-
tions of each of the lightning arres-
ter areas in co-operation with the
representatives of the manufac-
turers, it appeared, although the
shielding effect of trees and buildings
was considerable, that as far as could
be determined without making a
detailed sur\xy and record of the
conditions in each block throughout
the city, no type of arrester was at
any serious advantage or disadvan-
tage on this account.
The records and the conditions
surrounding the transformer instal-
lations were carefully and thoroughly
examined to determine the effect of
the other points 5, 6 and 7. These
investigations included the assem-
bling of the complete histor>- of each
transformer that had burned out
during the five-year period and the
compiling and assembling of all data
which might scr\-e to add to the
information on the several points.
On the completion of the investi-
gation, the representatives of the
manufacturers concurred in the
decision that none of the arresters
appeared to be at any material
advantage or disadvantage on
account of the first seven vari-
able factors in the above list,
and these factors were, therefore,
ignored in the further investiga-
tion.
STUDIES IN LIGHTNING PROTECTION ON 4000-VOLT CIRCUITS
9S9
There still remains two variables, namely
the variability of the lightning, and the
density of lightning arresters. In determining
the relation between the density of lightning
arresters and their performance a method
was discovered of eliminating the effect of the
lightning as a variable as described at some
length later in the paper.
The Effect of Density of Lightning Arresters on
Their Performance
A preliminary investigation of the effect
of density was made by plotting the density
of arresters in each original lightning arrester
area against the percentage of burn-outs in
Fig. 8. Outline Map of Chicago Giving the Numbers
Assigned to Each of the Sections as Shown in the
Third Column of Table I
that area. The points plotted in this manner
were so irregular that they did not permit the
drawing of any curve which might be con-
sidered as representing the results, but the
method appeared to indicate that there was
a very marked decrease in the percentage
of burn-outs with increase in density which
would warrant further investigation along this
line. The results also indicated that some
further subdivisions of the original lightning
arrester areas would be necessary in order
to eliminate the lightning as a variable.
The manner in which the records were kept
enabled this change to be made very readily
by using the section (that is, the square mile)
as the unit, resulting in an increase in the
number of areas from 19 to 192. For each
one of these sections there was determined
from the records the number of transformers
in the section as of August 1, 1918. As there
is an arrester on the same pole with each
transformer, and comparatively few cases
where there were two transformers connected
to the same phase wire on the same pole, the
number of transformers in each section was
taken as the number of arresters. There was
also determined for each section the number
of transfonner bvirn-outs and primary fuses
blown by lightning during the five-year
period and the actual area covered by the line.
This latter quantity was determined by going
over the large scale maps of the distribution
system and assuming that a line through
the center of the block covered the width
of the block. (This width varies in the
different portions of the city from about 250
ft. to over 600 ft. and averages approximately
400 ft.) From these figures can be calculated
the percentage of burn-outs in any section or
group of sections. The data with the sec-
tions arranged in the order of density of
arresters are shown in Table I.
The data in this table and other data
regarding the system are shown graphically
in several drawings which give a better idea
of the conditions than can be obtained from
tables of statistics. In Fig. 7 is shown
an outline map of the city on which are
shaded the areas actually covered by the lines,
the number of arresters per square mile being
indicated by the density of the shading. The
distribution system extends into 192 sec-
tions covering 163.25 square miles within the
city, while the area actually covered by the
lines, determined in the manner above
described is 93.49 square miles. As there
were 17,529 transformers on the lines on
August 1, 1918, the average density of
arresters is thus 187 per square mile.
Fig. 9 shows these data in another manner,
from which figure it will be noted that in
the larger portion of the area covered by the
distribution system, the density of arresters
ranges between 100 and 300 per square mile.
990 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
0) w
ta2
u -^
1,R
OJ
•— - CJ
t,c
a
j2 S
u
z
"v5 cti
<
<i^ ^
S
fS-S
o
<x-2
K
£5
(ii
1>§
2
o
-^
o a
CO
■■S-*^
«
H
"1 S
H
2 lu
CO
•a >
Be:
5fM
<
(L.
c S
o
ta
>'
1?^
« H
O -M
3 i
D3 S
< u
aJ (i>
•^ Q
w 3
O
c^K
U
^.S'a
s
<^'^-?.
H
.you
0
2
l-«
■- ^ o
;2;
13 fe-
° " S
s
rf cu >
o C u
1^
■-= != 3
u
^ rt u
H
U
Q
K
O
M-C o
b
s::a
<
s-^.s
H
— J=i <u
<
° &-S
Q
w _, rt
S.-SE
>.2 M
■i-> G cs
0,-2 c:
"> c
oj ca
XiM
*J o
JB3A asd sino
-ujnq au3D
J3d 9S1313AY
CO -^^ c^ r-00
00 -^ -- i^ -r
C-. o - o o
ir^ u"^ ^- ^- lO
d d d -I d
0.20
0.82
,0.09
0.41
4.18
0.48
0.
0.37
1.18
0.85
0.
1.23
2.86
0.21
0.18
xin gx
ddddd
3|itu aj-enbs
J3Cl SJ3;S3iIB
}0 XIISU3Q
\ ooo oo
(M <M 03 O) CN
c: O 0(N-#
S o) c^ o) a
lOCOX -^ '-I
!N C^l !M O) C>1
— "H-H'*CO
(N C^l M IM O)
(M Ol C^) OJ O)
t^ X — 1 in m
Ol (M CO CO CO
O) C) Ol CI Ol
X X X c^ re
rerere-* -*•
CQ OJCqC^IN
sauri
Xq pajaAOD
d d d d d
r- -,2 O 3-. 1.0
05 c; iO «3 lO
ddddd
O C: CJ: c; en
05 1^ Ci X 3;
ddddd
X »0 t^ X CO
M in 3: CO CO
c dddd
3-. CO CO O •#
I^ ^ O X o:
ddddd
t^ incq t^M
ddddd
uoujas ui
UAio'iq sasrij
JO jaquinj^
CO C: t- O O
ci t~ cc — r;
= 52::'-
— -1- — wm
m 35 - Lo 2
m 3-. 3: 00 O
uoijDas ui
sin'o-ujnq
JO jaqumj^
iC iC ^ :d ^
tom^xro
<Nt>-H.*<M
IMO'* 05C0
O CO — CI Ol
— ' <r o — re
aonoas UI
sjauijajsu-BJi
JO jaqmn^
(M lO t^ lO O
CO " »o o o;
o O O^ -H
OJO) — -^ — ■
lO O CO 9 X
Ol — C^ = C-l
-r CI 3: M N
X cj o m ^
— Oi — —
o: X t~ X -<
in a: X c)
— —CI
ci 3; re in X
— t^-^-OCD
jaqoinu
uoiioag
t^ ^ lO C:l^-
'^ (M — ' r-^ O
(M lO O CO -,0
t^ Ol 'I' t^ lO
■o r- u'l CO X
m C^ CO X 1-
lo " r^ Oi "
-Tf CO CO lO — '
c<ii> Tf cocq
31 lO t^ in 35
o m t^ M t^
CO CO — ' m a:
CO O I^t^
— — 3: m t^
CD CD re re "^
— — t~ 3: re
— -T -r CO CO
t^TC — mm
j jaquinu
E3JV
^ — ' <N O ^
— . X —
-* 't X C0<
■*<<0:X
— XC) —
05 35 — t-X
^h<~^
sjajsaxiB
JO 3<iA%
5U3S3J(J
tt^dOOO
OfeJCQOIt,
fe,Ua5ISlK|
!O0ClK|k,Q
fe.te, Co 14.05
a5Cse5fe.O
Jv^S. j3d siino
-xunq 1U30
J3d 93BJ3AV
t^c^ eo coco
OOCNIV^Oi
d (N d t-I d
1.61
0.
0.73
0.
1.54
0.50
0.88
1.08
0.91
0.35
CO CD CO CO CO
Tf_ CO Tf OC Cl
0.35
0.80
0.88
1.66
0,31
0, '
0.34
0.75
0.86
0.
aiiuj ajEtibs
1 J3d SJ31S3JJB
JO XjisuaQ
CO CO CO CO CO
N CO coco -*
CO CO COCO CO
OCO CDt^^-
CO CO CO CO CO
r- X o ino
CO CO-M- Tf Tj-
X re^ CO X
'f in in in in
35 3; o re re
m m CO CD CD
sauit
Xq P3J3A0D
UllBniDB B3JV
t^ rf X C: t^
1^ (M CO CO as
Tf -• OC CO C-)
ddddd
ci o — CO X
to IC X " CO
ddddd
X CO CO 3; Ol
1^ C^ CD T in
ddddd
t- ^ X TT X
ddddd
0.39
0.75
0.43
0 19
U0HD3S UI
oiidiq sasrij
JO jsqranj^
^ ^co-^ O
00-HCOO5O3
■^rorooiM
t- WCOINOO
r^co-j-osw
uoi^oas ui
s^rio-uinq
JO jaquini^
Tf b- '^ I> »0
moT)< oco
(NCOCD-H-H
recoNre —
orecocoN
ON-^reo
1 U01133S UI
sjatujojsuBJi
JO J3qujni»j
OS CD O NO
Ol CO o: -#32
CD " O -^ CO
O X 3-. <N 35
XCOOIM-*
l^ CDX — CD
o coxt^t^
in m t^ ci X
— i^ re t^ ci
« 35 CO O O
CO — o i^ re
jsquma
uoi^oag
-^ »OC0 -^ Ol
-^ O^ o i^ to
X X l^ '-H W
— LO — CO CO
•^- O CO — o
c^ -r in X CO
>0 t^ 3: 3; ^
t^ CO O "-O —
!M 3: — X CO
t^ -- in in -^
t- O CO O -T
3-. X CO —
m m ci 3c t-
re I- re m re
-r m CO CO T
— 1~ t~
jaqomu
B3JV
« * *
CO cox Tf 00
* * ^
•oi--re t~<
t^co — < re
— — X
c^ w TOt^tr
* *
SJ33S3JJB
JO 3dXi
1U3S3J(I
^JblOQU
QUCQtjcq
lUKICibJcq
tiJb,OH)0
UajCitiJiJ
QOOOO
JB3A J3d s^no
-ujnq ijuao
J9d 33BJ3AV
ddddd
ddddd
O -H
N ONCJ d
reoo CO re
re' e<i IN d m'
1.59 :
1.82
1.11
0.
3.53
re wo
■^ "*. 9
c4 CO CO d d
3^1 m sjBnbs
J3d SJ35S3ilB
JO A^isuaa
O: — ■ — wco
"CO CO CO
0-. — o a c^
CO ^ «;co CO
M r~ t^ o -•
CO cocot^ t^
in« 3: re-s-
I^I^t^MX
in in CD o: o
X X X X ^
050 5 So
S3Ul{
Aq P313A03
0.45
0.09
0.13
0.03
0.09
X t- Ol ig x
ddddd
CO 3: O t^ t^
-; O CO W O
ddddd
X3CXX —
C C CO — CO
ddddd
-r CI — C-. 3:
i~ in ci CI —
ddddd
0.24
0.23
O.IO
0.03
0.09
U011D3S UI
UMOiq S3snj
JO jsquinf,^
OO'-OO
0.-«ON(N
NO— OO
— ■"IXSMTf
oworeej
TONcooe*
UOIpSS UI
s^rio-ujTiq
JO jsqiun^
OC O-H o
OO OOO
— 0(N(NO
— w^ore
«•* — o re
ro-J-reo o
uoi^oas ui
SJSUIJOJSUBji
JO J3quin>j
tT — . -t -H CO
t~ t~ —■m in
XC0005U5
CM —
CD t^ OOCD
re — N
re 'J- X o t-
CO •* — CJ — '
c^ re o re 05
J3quinu
uoi^oag
(N W^OO
CO OS O r^ — '
inro — ca t^
N ojcoin OS
o O
-H r~ — in 1^
inrot-cD'i'
l^NNO^
t>- lOo; mre
c;c^ 3: — m
CO 1^ X X 1^
— in in t^ in
inoicD — ■»■
CD rein 3s
m CO 03 3: —
oremoo
— — mxoj
jaqmn^
B3av
* « « « *
CO 00 X 00 CO
« * • * «
X re •-* 00 CO
• * • •
inreinx cR
• •
•t in 1* "I* in
Tp recoxt^
• • • •
rorex m t^
SJ3;iS3XIB
JO 3dAj
lU3S3Jd
OQCi:5Q
:>C;-oc;ci
^s-Ci^Ciii,
'o i^ til 'o it.
U Ci 'o aj ^
CiC;=Qi^i;
I
STUDIES IN LIGHTNING PROTECTION ON 4()00-VOLT CIRCUITS
991
—•(JlO-XJi
lO O ^ CO
-H C-1 CJS >o o
CO CD C'l CM CO
t^ coro cooo
■* -H ^ c^^ o
^ <33CM O
in -* cMt^
X 1^ X CO
rt ^ in in
CM ^
CO CO
in
CM
o oo o o
OOOOO
OOOOO
OOOOO
OOOOO
OOOOO
O'H
«D cno O (N
'^ Tt* ic »o »o
<M (M (M CM C-l
COCO lOCO 03
lO »0 lO lO »o
O) C^l CM CM CM
CO O t^ t^ 00
CD CD t~ t^ t^
C^l Oi C^l C^l CM
CDCD03— 1 00
00 00 00 c:3 03
CM CM CM C^l CM
OComcDCM
C O O -- IN
CO CO CO CO CO
CO in in X o
CM CM CO CO in
CO CO CO CO CO
oo
Of
Tf CO
CO t^ ^ lO ^
rt -* m a: O)
— 1 ^ o cooo
lO »C 00 CD t^
ddddd
c; t^ c: in CD
■^ in CO cj: c3i
ddddd
c3i -* o: CO in
r^ CD C^ CD CD
ddddd
CM in03 t^CM
o r^ CO r^ X
ddddd
in in i-o C3 CD
CO 03 03Cd o
d dddd
— in
oo
dd
(N-* O CO"*
rf OlTf — -.
OJCOTf CDOC
CM —1 r^ OS CO
CM --■ -H CM
omcD incM
XOt^CM CO
CM CM
•-< o
o CO — ' m ^
CM "* CO CM CO
CO ^ ^ CO ^
CD O t^ CM O
OCM ^t-I^
o in in o o
OCM
(MOO f~ CO CO
CO ^cj com
— CM Csl
CM CM O CDO
— ' -H CM — ' CM
OICO 00 COl^
CM in OCDCD
^ ^ '-H Ol C-J
T* CO t^ CO Tf
CM 00 00 00 00
CM " CM — ■ CM
coco 0> CO-*
CM ^ -*CD
CM -■ CM CM
CO inx X -<
^^rt C3CM
rf CO CO
TfCM
CO
.-1 CO CO ^ en
•* Csl ^ 00 CD
CD iOC<l <M00
-H ^ CK' Cn CO
lO ^ t^ O ^
CM CD -H c-) ^
O CO ^ Ci CO
t^ M COCK 00
-<* ^ 1-^ CO CM
t^ Oj O CO --H
CD CO Tt< (M CM
in -* CM CO CO
in 03 1^ m CO
CD inoxcD
CD CO t^ CM CO
._, ^ ^ t^ in
X O CDXCD
-* -* CO CD in
03 in
CM 0< "J< ^
Tt< 0< COO
Csl ^
* *
rt O^ CD Tti
03^-* COCO
CM CD CO -^ CD
*
.-- CD CD CO ■*
W 03
0'5;cqfe.!j
fe,(i,pqQ"5;
Q^OOfe,
fe,C|fe.OO
0<JK)fe,U
CqOOCQO
uqfe.
1^ in "* to CO
CM 00 -< X CO
ddddd
o: i^ ^ CO) t-
C-5 CM t^ C^l CO
^dddd
00 -- o
(N 00 00
-^dddd
0.77
0.38
0.
0.84
0.44
t^ t~in CO
X CD CM CM
d d-^'-^d
XCOOiCM
-* CM C^l -^
ddddd
0.
0.
0.52
0.21
0..31
CD CO CD CD :d
t^ CM ^ "* lO
CDt~ t^ t^I^
incO CD t^ CTJ
Oj o o o o
t^ 00 00 00 00
-< CO ■* X o
X XX X 03
r-i CO in o 03
C3 03 03 03 03
03 O C-5 C^l CO
03 0 00 0
— • 01 CM CM CM
lO CO "* CD (N
T)< Tl< CO OV t^
ddddd
CO CD cr^ 00 ^
r^ 00 •* c: CO
ddddd
00 in Ol C32 -*
q t- ^ CD --
ddddd
03 03 in f~ in
CM in O "* CM
ddddd
t^ X CM t^ CO
t^ X t^ — ' 03
ddddd
t- ^ CD CO CM
X 03 CO X CD
ddddd
03 — ' X -* C^l
O OCO O3C0
ddddd
CDOO lO Oi o
CM — 1 .-H
cDoo ^ lom
^ in 00 CO CM
t^ in^ Otc
X CD r~ ^ o
in X CO X t^
f CM t^ c^i —
.-^ CO "-1 t~ C<l
OOCM COCM •-<
CO O CO o ^
CM CM 0-* rH
Ot~ ^ Cd CM
■* C^] ^ r-< O
oo CM Cd -<
Tt^ ^ Ci T-H CD
t^ r^ CO CD ^
ooo lo ^^
ca '^ 00 r^ »o
^CCl "* CD in
T-H CO t^ r-H CM
r^ CD 03 in m
^ o 00-*
O ^ CM CM CD
-* CO CO CO t^
CO CD O CD CO
CD r- t^ CD oi
X CM t^ in in
— t- X CD
CO CO t^ CO o
CO o o t^ --<
t^ r- CI ^ CO
lO Oi ^ CO t^
Tt< O C^l "* CO
^ -H OOCM t^
03 1^ t^ com
O 00 O C-l •*
O CM 03 00 1<
t^ in r^ in in
O t-~ O CO o
W -* CO -* 03
t^ in in m r^
in t^ in CO ^
CO --I CO t^ CO
t^ in CO CO --<
t^ O CD O CD
CM t^ CD ^ a3
in t^ in in 03
— O — ' O;. o
O -* CM CO W
■* tH CO co<
w-H oo
<<; th CO"*
CM Cq —I —1
OOxf t^ -* C33
<; X CD 00 1-
CM -<
i-i i-H
in-* CM CDX
XCD<;< CO
— CM X —
i^oqcjcq
CQKlIoQO
Qte-KlUfe,
pqcqOCQKl
fe,qoote.
ttlOOUCQ
QUCQcqcq
CO in ^
00(N CD
CD CM 00 CD
t-CM CO-^O
CO-* 0>^CM
-H in CM CO
coco c^t^
(Ninooin
X 03 05 X 03
coco CO-* 03
cocDco— ' in
03 --H CO CO
xin Xt*
oro ^OCM
CO ^ rf OCM
rt oooco
O ^CM --iCM
rt ^COOO
CO"-! CO— — 1
O «0 CM CO
O C^ CO CO Ci
ooooo
oo o ^ --■
CM C^l CM CO "*
»i< in cDcoo
CD t^ t^ 03 O
^ ^ -H ^CM
oo O -H —
CM CM CM C-J CM
CM CO CO I^ O
CM CM C^l CI CO
Tf CO --CM "5
OTf CO CO m
ddddcD
03 OOOO 00
lO CO CM CM r-H
ddddd
CO C-1 t^ 00 CM
^ ■* t~ O C^l
ddddd
^ -* ^ in in
c^i in T* -* CM
ddddd
c:3 in in -- in
^ CO CO CM CO
ddddd
o o in cT. CO
CO — O CM t^
ddddd
I^ CO t^ — t^
CO -* CI -H O)
ddddd
^-*CMK3^
T— 1 rH .-H
C33 t^ CD <N '-'
CM ^
ino>cooco
1— t 1— t CD CD -^
co^ xcDin
— O CMt- CO
CO —
C3-* xm ci
CM
OOicq -H CD
(M (NCM ^CM
•* -1 ^ CM-*
o in CD in -*
CM-*X--iCM
CD .-1 — CM t-^
CM •* ocqcD
T*< t^ c<i CO o
Tt< cococo
lOCOCM — ' O
CD CO CM CO CM
00 1^ CD 03 in
^ ^ 00 CM
^ Cd ^ CM 03
c<i CD in in CM
CM ^ ^ in CM
Cd T* -* CM ^
CO CM com X
co-H cox
in CO CM f in
f in CO — CO
»0 05C0 1^ t^
o ^ t^ -r ^
" 00 CM C-. O
CO ^ X t^ lO
cr. O CD -H CD
00 CD " 00 00
in — ico^ 05
CO -* ino3 CO
C-l CO ^
03 r^ [^ w CO
CO CO in CM CD
CM rt X t^ «
in C33 03 C3 »-i
CO 03"* -• o
0-- —■ CO o t-~
[^ in CO CO 00
03 — CM CO O
X m t^ 03 CM
03 w -rii in in
03 o 03 1^ in
CO CM CO c:3 X
* « « *
00'* lO t^OO
* * * *
CO^rfCOCO
* * *
coco t~ in CO
T-4 »-( 1-H
t^co-* x^
t^ xm t~ CO
■* co< t^t~
qUttlKlC)
KlCQQCjto
Cj O O C5 Q
qQttife,^,
KlC;!oC)0
K!tC|k,bqQ
OCjCQKlKl
o
CD
d
o"
03
00
00
CM
03
cm"
o
[■-_
a-'
Cl
in
-CM
C3 03
C] —1
t- o
*"* *+J
ojjcn
1-. O
■" i_
■o S
c d
ts c
■3 ■(3
o o
992 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
r
■\
/
\
\
\
/
\
\
\
/
^
\
\
50
100
150
350
400
450
SOO
Fig.
200 250 300
DENSITY OF ARRESTERS
NUMBER PER SQUARE MILE
9. Diagram Showing the Area Actually Covered by the
Distribution System for Various Densities of Arresters
The number of arresters for various densities
and for each type of arrester is shown in Fig.
10. In this drawing it will be noted that
arrester ^4 was installed in sections with a very
narrow range in density. The section num-
bers given in the third column in Table I
are shown in Fig. 8. The stars preceding
the section numbers in Table I indicate the
sections in which a change in the type of
lightning arrester was made preceding the
lightning season of 1917 for the purpose of
permitting the installation of an additional
type of arrester and securing a better distri-
bution of the several types of arresters in
different portions of the city.
In Fig. 11 there has been plotted for each
section the density of arresters as shown
in the eighth column in Table I and the average
per cent of burn-outs as shown in the last
jOCO
5000
A
—
„
/
\,
5i
/
\
%
o
7^
'
\
/
\
g2000
/
\
/
'^-^
\
/
^-^
^^-^
^
\
1000
/
//■
\ \
"S \
L.
C-^
/^
f
^C^
^
N
1
<^/^
0
^
L^
■£ J
.ii^
200 250 300 350 400
DENSITY OF ARRESTERS
NUMBER PER SQUARE MILE
Fig. 10. Diagram Showing for Various Densities of Arresters
the Number of Each Type of Arrester and of all Types Con-
nected to the Distribution System
column. The final cun.-e for all arresters
showing the variation in the performance
of arresters with their density is also shown in
the same figure, but the cur\-e cannot be drawn
directly from the points shown in this figure
because these points, representing different
areas and different numbers of transformers,
are not of equal weight. Nothing in the tables
or records shows the wide variation in the distri-
bution and intensity of the lightning quite so
wellastheplottingof thesepointsinFig. 11. Out
of the 192 sections it will be noted that in
about one-sixth of them the points are on the
line of zero burn-outs, showing that there
were no burn-outs whatever in these sections
during the five-year period.
u 5
6
i
\\ .^' ^
200 3(M «00
DENSITY' OF ARRESTE.'^S
NI'MBFR projOKABP MILF
Fig. H. Diagram Showing for Each of 192 Sections the Average
Per Cent of Transformer Bum-outs Due to Lightning for the
Five-year Period Plotted Against the Density of Arresters.
The curve shows for all types of arresters the final de-
termination of the relation between density of
arresters and transformer burn-outs due to
lightning. The curve cannot be plotted
directly from the points shown ia
the figure as they are not of
equal weight
In order to secure points of equal weight for
the purpose of drawing the cur\-e, it was
decided to have each point represent the
experience with the same number of trans-
formers. At first trial it was agreed to
assemble the data so as to get IS points, each
of which would therefore include the data
from approximately 1000 transfomicrs. Data
for the first point were obtained by starting
at the top of Table I and including enough
sections to get a total of about 1000 trans-
formers. Then the figures showing the area
covered and the number of burn-outs was
totaled for these sections, from which could
be determined the average density of the
arresters and the average per cent of burn-
STUDIES IN LIGHTNING PROTECTION ON 4000-VOLT CIRCUITS
993
2.0
a
>■ 12
q:
0- 10
0 08
1 "06
0.2
.\
\\
\\
\
•
•\
•
\
•
•
\
•
«
•
V
"100 200 300 400 500
DENSITY OF ARRESTERS
NUMBER PER SQUARE MILE
Fig. 12. Diagram to Logarithmic Scale Show-
ing the Data in Table H and Fig. 8. Assem-
bled into Eighteen Points Each Covering the
Experience for the Five-year Period with
Approximately the Same Number of
Transformers
outs for this group of transformers. This
was equivalent to taking a vertical band of
Fig. 11 which would include enough points
to make a total of 1000 transformers and
finding one point to represent the average
experience for the entire band. In the same
way the other 17 points were calculated and
are shown plotted to logarithmic coordinates
in Fig. 12. The use of logarithmic coordinate
paper was adopted for the purpose as it was
found to greatly facilitate the work. There
1-2.0
z.
iij
"1.6
q:1.2
<
^1.0
q:
LLI
Q-0.8
-2.0
UJ
"10 \
a. \\
UJ
CL
Qrl.2
S
> 1.0
\
\
\
•^ 08
»—
o
o06
UJ
\
\
m
1/^04
0;
UJ
S
\
S
\
o
U-
z
<
1- n 7
\
\
_-
1
06
,0.4
O
-02
^
\
\
\
*\
\
•
• ^
\
•
\
\
100 200 300 400 500
DENSITY OF ARRESTERS
NUMBER PER SQUARE MILE
Fig. 14. Same as Fig. 12 Except That Data Are
Assembled Into Four Points of Equal Weight
was some-question as to whether the number
of points selected for assembling the data in
this manner had any effect on the resulting
cur^^e, but it appeared that if practically the
same line were obtained by using a different
number of points that there would be no
serious error in the method. The same data
were therefore assembled in a similar manner
in 7 points, 4 points and 2 points and the
results are shown respectively in Figs. 13, 14,
15. After a number of attempts to draw
curves through these points in the several
2.0
1.6
1.2
1.0
0.8
S 06
CD q:
or
O
100
500
200 300 400
DENSITY OF ARRESTERS
NUMBER PER SQUARE MILE
Fig. 13. Same as Fig. 12 Except That Data Are
Assembled Into Seven Points of
Equal Weight
0.4
02
1
\
\
'^
\
^
\
^
100 200 300 40O 500
DENSITY OF ARRESTERS
NUMBER PER SQUARE MILE
Fig. 15. Same as Fig. 12 Except That the Data
Are Assembled Into Two Points of
Equal Weight
<)94 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
figures, it was found that a straight line would
properly represent the results just as well as
any cur\'e which might be drawn; and it was,
therefore, assumed that the cun-e when
drawn on logarithmic coordinate paper was
a straight line, which is equi^■alent to assum-
\\l
1
I
w
^
^
sA
^All
-^
^
100
250
300
350
150 20O
DENSITY OF ARRESTS
NUMBER PER SQUARE MILE
Fig. 16. Diagram Showing the First Approximation of the Rela-
tion Between the Density of Arresters and the Percentage
of Transformers Burned Out by Lightning for the
Five-year Period, 1915-1919 Inclusive
The cur\'e for all arresters is plotted from the dashed line in
Figs. 12 to 15 inclusive. The curves for the indiWdual types of
arresters were derived in a similar manner from logarithmic dia-
grams which are not reproduced.
A point instead of a line is shown for arrester A as the records
for this type include only one transformer bum-out in the three
years in which the arresters have been in ser\'ice.
ing that the relation between the quantities is
an exponential function. In each of the four
figures the full line is determined by the
points in that figure and the dashed line is the
average of all of the four. It will be noted
that the variation of the points through
the straight line decreases as the number of
points decreases, or in other words, as the
number of transformers represented by one
point increases. The average curve repre-
sented by the dashed line in these four figures
transferred to arithmetical coordinates is
shown in Fig. 11.
While one engineer was engaged in the task
of assembling the data and drawing the lines
on logarithmic coordinate paper as above
described, another engineer was given the
task of assembling the data in a similar
manner except that he used for each point
the experience from an equal area covered by
the lines as given in column seven of Table i,
instead of an equal number of transformers.
This was done with the idea that any serious
personal errors or any error due to the
assumption made in drawing the curv^es or in
transferring them to arithmetical coordinate
paper would be indicated by differences in the
final cur\-es. After these two engineers had
independently drawn final cur\-es similar to
the one shown in Fig. 11 the two cur\-es were
then transferred to the same sheet and found
to be practically superposed. The equation of
the cur\-e in Fig, 11 is:
5450
where A' = the number of arresters per square
mile, and
V = the average per cent of trans-
formers burnt out by lightning per
year during the five-year period.
This equation means that the density of
arresters has a vers' important influence on the
results secured by lightning arresters. If we
assume for example that there are 1000
transformers installed in an area of 10 square
miles each protected by an arrester on the
same pole, and that later the number of
transformers in this area is doubled and at
the same time uniformly distributed, the
30
l^,
'15
, 1.0
s
o
\l
\l
\\
\\
1
s
Sf
t«^
1
d
1^
1 1
50
250
300
3SC
100 150 200
DENSnV Of ARRESTERS
NUMBER PER SQUARE MILE
Fig. 17. Diagram Showing the Final Determination of the Rela-
tion Between the Density of Arresters and the Percentage of
Transformers Burned Out by Lightning
These are the curves shown in Fig. 16 modified by the assump-
tion that the lines representing the data on logarithmic paper
should be parallel to the dashed line in Figs. 12 to 13 iLclusive,
showing the experience with all types of arreater.
results of the change are shown in Table II.
From this table it will be noted that although
the number of transformers in the area has
been doubled, the percentage of bum-outs
has decreased from 1.67 per cent to 0.5 per
cent and that the actual number of bum-outs
STUDIES IN LIGHTNING PROTECTION ON 4000-VOLT CIRCUITS
995
has decreased from 17 to 10 per annum. In
other words, the doubling of the number of
transformers and arresters in a given area will
not result in more transformers being burnt
out by lightning per year as might be supposed
but will result in an actual reduction of
about 40 per cent in the number of such
burn-outs per year.
The data for each type of arrester were
then plotted in a similar manner, a set of four
curves similar to Figs. 12, 13, 14 and 15 being
drawn for each type of arrester. As might be
expected with a smaller number of obser-
vations, the variation of the points from a
straight line when plotted on logarithmic
paper and the variation of the four lines from
their average was somewhat greater than in
the case of the corresponding lines for all
types of arresters. These straight lines for
the different types of arresters were not
parallel, and when transferred to arith-
metical coordinate paper as shown in Fig. 16,
the curves cross each other in a confusing
manner. While the curves thus drawn may
be mathematically accurate, they appear to
be physically impossible as there seems to be
no sufficient reason why one type of arrester
should be better than another at one density
and poorer at another density. It seems more
reasonable to suppose that if one arrester is
better than another at any particular density
of arresters, it will be better throughout the
entire range of densities. After giving this
subject considerable study it was decided
to assume that the straight line representing
the experience with any type of arrester,
when plotted on logarithmic coordinate paper
should be parallel to the line showing the
results for all types of arresters, that is,
the dashed line in Fig. 12. To make this
change: the midpoint of the line for each
arrester was found, which is a point so located
that there are an equal number of arresters
represented by the line on either side of the
point. The line which was finally taken as
representing the experience with this type of
arrester was then drawn through this mid-
point and parallel to the dashed line in Fig. 12.
The results of this assumption when trans-
ferred to arithmetical coordinate paper are
shown in Fig. 17. If these several assump-
tions are reasonably accurate, and they
appear to do no violence to the facts, then the
methods which have been used result in
curves which can be taken as representing
the performance of each of the arresters with
varying densities, and the most troublesome
variable, that is, the variation in the dis-
tribution and intensity of the lightning has
been eliminated by the method of assembling
the data and drawing the curves. From these
curves it will be noted that four of the
arresters designated as C, D, E and F are so
close together that the differences may be
considered as well within the possible errors
of observation.
In Fig. 17 an ordinate has been drawn to the
midpoint of each of the curves as above
defined or at the position corresponding to the
average density for that curve, that is, for
each type of arrester the number of arresters
to the right of the ordinate is the same as
the number to the left. These ordinates
represent the same values that were given
in the previous paper as showing the average
experience for each type of arrester, but it is
now seen that in the case of the four arresters
C, D, E and F, the curves are so close together
that the ordinates for these curves, instead of
correctly representing the relative merits
of the four arresters, are practically four
different ordinates of the same curve. The
four arresters are therefore of practically
equal protective value.
It will be noted that the ordinates for curve
B in Fig. 17 are about 40 per cent of the
corresponding ordinates of the average of
curves C, D, E and F. Arrester B is one of
the oldest types on the lines and the arresters
are fairly well distributed over a wide range
of density as shown in Fig. 10. It is, there-
fore, considered that this difference of about
40 per cent as compared with the other four is
a real difference due to the value of the
arrester as a protective device and is not due
to an error in the observations or calculations.
In the case of arrester A, Fig. 5 shows that
this arrester was installed in only three
contiguous sections and Fig. 10 shows that
these sections had a narrow range in arrester
density. In addition the arresters had been in
our service for only three years and in view of
all of these circumstances, it appears that the
data regarding this particular type of arrester
are not conclusive. For the purpose of securing
more conclusive data regarding this type of
arrester, additional arresters were installed
early in 1920 in the areas shown by the heavy
shading in Fig. 5. Thelight shadinginthesame
figure shows the areas in which an additional
type of arrester was installed early in 1920.
Comments on the Designs of Lightning Arresters
Covered by This Investigation
It is possible that the experience with
the several types of arresters covered by these
996 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
investigations, as well as the earlier types
which they replaced, may be best sum-
marized in the form of a tentative specifi-
cation for lightning arresters, which would
state some of the important points to be
included and to be avoided in such design.
Such a specification would read about as
follows ;
1. The arrester must consist of a number
of gaps in series with a resistance, with the
number of gaps and the amount of resistance
properly adjusted to the line voltage so that
the dynamic arc following a lightning dis-
charge will be quickly broken without
damage to the arrester.
2. The resistance rod must have the
resistance uniformh^ distributed throughout
its length, so as to prevent the progressive
short-circuiting of the rod with heavy light-
ning discharges and the destruction of the
arrester which will follow.
3. The amount of resistance in the resist-
ance rod should not be seriously affected by
repeated heavy discharges.
4. The leads for connecting the arrester to
the line should leave the arrester so that they
will form drip loops, and the leads should be
so arranged that the arrester can be con-
nected to a line wire on either side of the
arrester.
For low maintenance cost the following
features are desirable :
5. The enclosing case should be of fireproof
insulating material that is not affected by the
weather, and it should be constructed so as to
protect effectually the metal parts from the
weather, and to prevent accumulation of dust
on the gaps.
6. The gaps in the arrester should be
between parallel plates, disks, or rings
instead of between cylinders or spheres so as
to permit repeated heavy discharges without
seriously altering the length of the gaps.
7. The arrester should be constructed so
that in the event of the failure of the arrester
to interrupt the dynamic arc the enclosing
case will be shattered by the heat so as to give
some visual evidence of the trouble and result
in the opening of the circuit.
8. The arrester should be without moving
parts or parts which recjuire inspection,
renewal or adjustment and should preferably
be made in the form which cannot be inspected
or repaired without removing it from the pole.
The experience with the arresters covered
by this investigation indicates that several
types of arresters arc now available whicli
comply with all of these specifications.
Conclusion
The conclusions from the investigations
described in this paper, together with the
more important conclusions from the pre-
vious paper, some of which have been modi-
fied and extended by these investigations,
may be summarized as follows:
1. Transformer troubles during lightning
storms may be reduced (a) by the removal
of transformer primary terminal boards, (b)
by the installation of lightning arresters, (c)
by the use of larger bushings on the primary
leads of transformers where they enter the
case.
2. Lightning arresters installed on trans-
former poles are considerably more effective
than if installed on the line poles.
3. Even in the most severe lightning
storms, which apparently cover the given
territory quite completely, there will be
numerous extended areas within this territory
which will be entirely free from lightning
disturbances. Careful records extending
over a period of several years arc, there-
fore, necessary in order to determine defi-
nitely whether immunity from troubles due
to lightning is due to the efficiency of the
lightning protection or to the absence of light-
ning.
4. There is a vcr>- marked improvement
in the effect of lightning arrester protection
with an increase in density, that is, the
number per square mile, and this effect is such
an important factor in their performance
that no accurate comparison of the relative
merits of various types of arresters can be
made without giving this point proper
consideration.
.'). Where the number of transformers, each
of which is protected by an arrester on the
same pole, is large per square mile so that the
transformers and arresters are on the average
only a few hundred feet apart, the total com-
bined effect of all of the adjacent arresters is
greater than that of the arrester on the same
pole with the transformer.
0. In districts where transformers are
widely scattered, that is, where the local
density is materially below 100 per square
mile and where continuous service is imjior-
tant, it will probably be found desirable
to install arresters on line poles in addition to
an arrester on the same pole with each trans-
former; where the local density is of the order
of "lO per square mile, or lower, the installation
of such additional arresters will probably be
found to be warranted solely by the reduction
in operating expenses.
STUDIES IN LIGHTNING PROTECTION ON 4000-VOLT CIRCUITS
997
7. The increase in the density of lightning
arresters also results in a marked decrease
in the percentage of burn-outs due to light-
ning of underground cables connected to
overhead distribution circuits, and while the
exact figures for the early years are not avail-
able, the percentage has been reduced from
several per cent per annum with a very low
density to a figure running well below one-
tenth of one per cent per annum with the
density averaging about 200 per square mile.
8. In the case of high-voltage cables, that
is cables operating at voltages ranging up to
25,000, and where the present practice in this
country calls for a maximum of one arrester
at the point where the underground cable
connects with the overhead line, the installa-
tion of additional arresters in the vicinity
of the cable pole would in all probability cause
a marked reduction in the percentage of
burn-outs of such cables due to lightning.
9. The effect of density of arresters, of the
shielding effect of high buildings, trees, etc.,
and perhaps also other features, have such
an important eft'ect on the amount of trouble
from lightning that no accurate comparisons
of the results secured in different cities can be
made without giving due consideration to all
such features of the conditions under which
the lightning arresters are installed.
10. For use in the protection of trans-
formers in districts where each transformer
is protected by an arrester on the same pole
and where the density of arresters ranges
above 200 per square mile, the most economi-
cal arrester of the several types covered by
this investigation is probably the cheapest
arrester. It is entirely possible and even
probable that the local conditions will have
an important bearing in determining the best
type of arrester to be used in any given
locality, and that where the amount of shield-
ing from buildings, trees, wires of other
companies, etc., is very slight and where the
securing of adequate ground connections for
the arresters is expensive it would be pre-
ferable, even in areas of low density, to use
arresters whose discharge capacity is con-
siderably greater and whose discharge poten-
tial is considerably lower than the arresters
covered by these investigations and to
confine the installation of the arresters to
the transformer poles.
11. It is possible, by carefully distributing
the various types of lightning arresters over
a large area and by securing the results of the
performance of arresters over a period of
years, to place the several types of lightning
arresters used for the protection of trans-
formers under conditions that are practically
identical as regards the features which would
affect the relative performance of the various
types of lightning arresters, and to secure
data which will pennit a comparison of the
relative merits of the several types of light-
ning arresters as protective devices.
12. It is entirely possible to make light-
ning arresters of the self-contained type, that
is, of a type not requiring an external protect-
ing box and so constructed as not to require
or permit inspection. The annual mainte-
nance cost of such arresters is practically
limited to the replacing of damaged arresters,
and the total annual maintenance cost as
indicated by an experience of five years with
several thousand such arresters is well below
1 per cent of their original cost of installa-
tion. The adoption of such types of arresters
will result in a material reduction in the
annual maintenance cost as compared with
the older types.
13. A change in the form of lightning
arrester gap from a cylindrical or spherical
shape to parallel flat surfaces which was
adopted by the manufacturers when changing
from the wooden box type to the self-con-
tained type of arrester, appears to result in a
form of design which allows repeated heavy
discharges without requiring renewal or
adjustment of the parts, and has been an
important factor in changing the design from
a type requiring annual inspection, renewal
and adjustment to a type which does not
permit or require such annual attention.
14. The four types of arresters which have
been designated by the letters C, D, E and F
and which consist essentially of a resistance
in series with a number of gaps, together
with such additional features as antennas,
compression chambers, expulsion chambers,
and solenoids to vary the length of the gap
following dynamic discharge, all appear to be
practically identical in their value as devices
to protect line transformers.
15. The type of arrester designated by B,
which consists of a large ntimber of gaps in
series without any resistance, in addition
to two other paths through a high and a low
resistance shunting a large and a small
number of gaps, appears to be considerably
better protective device than arresters desig-
nated by C, D, E and F, and as far as can
be determined from present information,
this difference in its value as protective
device appears to be due to features of its
design.
998 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
16. With the aid of the data contained in
this paper it should be possible to make
estimates of the cost and results of lightning
protection in Chicago with the same degree of
accuracy as the estimates of cost of construction
or maintenance of overhead lines, when the
figures are averaged over a period of years.
17. The shielding effect of high buildings,
trees and other similar features which might
be considered as determining the exposure
of the lines to lightning have an important
bearing on the amount of damage that will be
caused by lightning. In local areas in a distri-
bution system which have for years shown a
high percentage of troubles caused by light-
ning and where the troubles have been
allowed to persist because of the thought that
some mysterious influence local to the
neighborhood attracted the lightning, it will
probably be found that a large percentage of
troubles is due to the lack of shielding from
the surroundings or a low density of arresters,
and that the trouble can be materially reduced
by increasing the density of the arresters in the
locality.
18. Great caution should beused in attempt-
ing to compare the results secured by light-
ning arrester protection in Chicago with
results secured in other localities without
giving due consideration to all of the factors
which might afTect lightning arrester per-
formance.
In conclusion the author desires to express
his appreciation to the General Electric Com-
pany and the Electric Ser\-ice Supplies Com-
pany for their many helpful suggestions and
hearty co-operation during the progress of the
investigations.
DISCUSSION BY CHARLES P. STEINMETZ
Chief Consulting Engineer, General Electric Company
For some 3^ears we have realized that the
conditions of lightning protection in primary
distribution networks are in some respects
materially different from those in high
voltage transmission lines. Many of the
phenomena, which are of serious danger in
the high potential transmission line, such as
steep wave front impulses, high frequency,
traveling or standing waves, recurrent and
cixmulative oscillations, etc., can not develop
to a dangerous magnitude in the primary dis-
tribution circuits. Dissipation due to leak-
age and the low voltage character of the
insulation, and interference within the net-
work of circuits and apparatus dampen
oscillations. Because of the relatively low
circuit voltage the electrostatic energy is
small, and the most serious source or aggravat-
ing cause of lightning trouble in high potential
circuits, the arcing ground or oscillatory
spark, cannot develop. On the other hand,
due to the low circuit voltage, the insulation
strength is low compared with the disruptive
strength of lightning voltages, and the trans-
formers distributed all over the circuits make
the system vulnerable throughout its entire
extent.
The material given in Mr. Roper's paper
is, therefore, the most valuable contribution
ever made to the study of lightning dis-
turbances in primary distribution networks,
as it contains the exact performance records
of nearly 90,000 lightning arrester years
comprising 529 apparatus failures; that is, an
amount of data greater than has ever before
been collected on lightning disturbances in
primary distribution systems.
I wish to say that all the phenomena
obser\'ed by Mr. Roper arc in complete agree-
ment with, and all the conclusions which he
drew from his experimental observ'ations
follow as theoretical conclusions the state-
ment:
In primary distribution circuits, lightning is
the discharge of a very high voltage (of the magni-
tude of hundred thousand volts) and corrcs pond-
ingly high electrostatic charge, instantaneously
produced over a large part of the distribution
system.
These voltages are far higher than the
insulation of the transformers can stand for
any appreciable time. It is thus a race
between the time lag of the transformer
insulation and the rate at which the lightning
arresters can discharge the excessive voltage.
Thence immediately follows the all-domi-
nant charactcrof thelightningarrestcrdensity,
that is, the number of liglUning arresters
per square mile or per lineal mile of circuit.
The rate at which the excess voltage decreases
is directly proportional to the number of
discharge paths, that is, the number of
arresters, and the time during which the
transformer is exposed to excess voltage is
STUDIES IX LIGHTNING PROTECTION ON 40()()-VOLT CIRCUITS
999
therefore inversely proportional to the nvim-
ber of arresters.
Also follows the explanation of why trans-
former terminal boards and transformer
bushings, though standing a higher sustained
voltage than the transformer windings, are
more vulnerable, since their insulation is air,
which does not have the high time lag of the
oil and solid insulation of transformer wind-
ings.
With 100,000 volts instantaneously im-
pressed upon a 2;5()0-volt lightning arrester,
differences in the number, length or shape
of the spark gaps, in the discharge voltage
or equivalent sphere gap, within the range
which may be expected between different
types of such arresters, can have little effect,
^s the excessive overvoltage causes the
discharge to begin instantly. An appreciable
difference in the protective value, however,
may be expected from the discharge rate of
the arrester. It is interesting to note that the
arrester (Type B) which shows a superiority
sufficiently great not to be overshadowed
by the effect of the arrester density — a 40
per cent decrease in transformer losses —
is the only one in which the discharge
capacity is not limited by a series resist-
ance.
An arrester not at the transformer, but at a
small distance from it, would have the same
effect in discharging the excessive voltage
of the circuit as an arrester at the trans-
former, and could thus differ in protective
value onl}' by the time lag required by the
charge to travel the distance from the trans-
former to the arrester — ^ about one ten-
millionth of a second per 100 feet. Aside
from this, all the arresters within the area
covered by the instantaneously produced
excessive voltage would equally share in
protective value.
The question which then arises is that of
the origin of such a very high voltage instan-
taneously produced over a considerable part
of the distribution system.
I have given the phenomena of the thunder
storm and the origin of the lightning flash
considerable study for a number of years and
find that such voltages must be produced on
lines as a result of the equalization of cloud
potential by the lightning flash.
Let L, Fig. 1, represent a wire of the
primary distribution circuit, 6 meters above
the ground G. Let C be a thunder cloud at an
elevation of 1000 meters above ground G,
having a potential difference of 20 megavolts
against ground. There is thus an electro-
static field between cloud and ground, of a
gradient of 20 kilovolts per meter. If the line
L were perfectly insulated by its position
in the electrostatic field 6 meters above
ground, it would have a potential difference
of 120 kilovolts against ground. It is, how-
I
^
Fig. 1
ever, not insulated for such voltages, and
while the cloud gradually builds up to 20
megavolts, a bound charge accumulates on
the line L, by leakage through the insulation,
corona, static sparks over the arresters, etc.,
and therefore keeps the line substantially
at ground potential. The cloud discharges
by a lightning flash, its voltage disappears,
and the electrostatic field between cloud and
ground collapses. The bound charge on the
line L then becomes a free charge. Since as
bound charge it kept L at ground potential,
though by its position in the electrostatic
field it would have had a potential difference
of 120 kilovolts, as free charge it now raises
the line L to 120 kilovolts above ground.
Hence instantaneously, that is, with the
rapidity w4th which the lightning flash dis-
charges the cloud, a voltage of 120,000 volts
is produced over that part of the distribution
system which was in the electrostatic field of
the thunder cloud.
1000 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, Xo. 12
This is the origin of the \-ery high voltage
instantaneously produced over a large part
of the distribution system.*
This also explains why the impedance
of thegroundwire — which should be extremely
high at the extreme rapidity of the dis-
charge— seems to have so little effect, while
even a small series resistance in the lightning
arrester (small compared with the surge
impedance of the line) — has a marked effect.
The ground wire also is in the electrostatic
field between cloud and ground, and thus
charge of the ground surface) at the bottom
to equality with the charge of the line at the
top. This charge and the voltage produced
by it are shown by the shaded area in Fig. 2.
This, however, is the distribution of voltage
and thus electrostatic charge (or dielectric
field) existing on the ground wire during the
discharge of the lightning arrester; that is,
there is no transient retarding the starting
of the discharge current in the ground wire,
since the energv' which the transient stores is
already present in the free charge left on the
Fig. 2
accumulates a bound charge which becomes
free charge by the lightning flash. This,
however, is a tapering charge, increasing
from zero (or rather equality with the bound
*In reality, the phenomena in the cloud are not as simple. As
the result of rain formation, potential differences against
ground build up in the cloud, vaonng in magnitude probably
between 10 to 100 megavolts in the various parts of the cloud,
depending on the moisture content and thus the rate of rain
formation. These potential differences between different areas
of the cloud are equalized by the lightning flash, so that in some
parts of the cloud the potential difference against ground is
instantaneously lowered, in others probably raised. Thus if in
some part of the cloud the potential difference against ground is
lowered by the equalizing lightning flash from 60 megavolts to
40 megavolts. the bound charge on the line under this part of the
cloud decreases from that corresponding to 60 megavolts to that
corresponding to 40 megavolts, and a free charge corresponding
to 20 megavolts thus appears. In other parts of the cloud, by
the same lightning Hash, the potential difference against ground
may be raised from 20 to 40 megavolts. setting free on the line
under this part of the cloud a charge of opposite polarity.
Fig. 3
wire. The discharge current thus starts
simultaneously throughout the length of the
ground wire, at a rate depending on the
initial potential gradient, viz., 20 kilovolts
per meter. Its rate of rise is given by:
,di
or, with L= 1.34 X 10-»; r = 200 volts per cm.
this gives:
-J- = 150 X 10* amperes per second.
With a surge impedance of the distribution
lines of 400 to 500 ohms, and the lightning
arrester connected into the line so that the
discharge current can reach it from two
STUDIES IN LIGHTNING PROTECTION ON 4000-VOLT CIRCUITS 1001
wires, giving a surge impedance of 200 to 250
ohms, a voltage of 120 kilovolts would give a
discharge current of 4S0 to GOO amperes. It
would thus require about one three-hundred-
millionth of a second for the current in the
ground wire to build up. That is, the time
lag of the ground wire would be of the magni-
tude of one three-hundred-millionth of a
second.
Suppose, however, a series resistance is
used in the lightning arrester. The dis-
tribution of the bound charge (set free by the
lightning flash) along the ground wire would
still be the same as shown in Fig. 2, or by the
shaded area in Fig. 3. The distribution of
voltage during the discharge of the lightnmg
arrester, however, would be as shown by the
heavy drawn line in Fig. 3, having a break
equal to the voltage drop across the series
resistance at the point P, where the arrester
is located. That is, a rearrangement of the
charge and voltage distribution in the ground
wire becomes necessary, resulting in a tran-
sient retarding the discharge, that is, a time
lag which limits the protective value in this
case, though the resistance may be far below
the surge impedance of the lines.
From this explanation of the phenomena
we can realize the limitations within which
the conclusions of Mr. Roper's paper apply.
They probablv apply to all extended primary
distribution systems, that is, networks of
relatively low voltage, with about the same
magnitude, and the numerical values are
modified only by the climatic conditions, that
is, by the frequency and severity of thunder
storms, and in this respect Mr. Roper's state-
ment is rather too modest. They would not,
however, apply to circuits of materially high
voltage, in which the insulation strength of
the circuits and the discharge voltage of the
arresters are not negligible compared with
the instantaneous voltage of the free charge
produced by the lightning flash. Also_ they
would not apply to high voltage transmission
lines, in which the apparatus is localized at
the terminals, where the area affected by the
free charge is only a part of the line, and where
dissipation through leakage, interference, etc.,
is small and secondary effects such as sparks
produced by the charge predominate; and
where oscillatory waves piling up the voltage
by reflection, etc., and secondary effects pro-
duced by the discharge, such as oscillatory
arcs, make available for destructive action the
engine power back of the generators.
DISCUSSION BY J. L. R. HAYDEN
General Engineering Laboratory, General Electric Company
The large amount of data given in Mr.
Roper's paper enables us to investigate some
further features. Some information on the
protective screening effects of buildings, trees,
etc., may be expected from the following
reasoning: Column 7 of Mr. Roper's paper
gives the area covered by the lines in each of
the 192 sections. As most of the sections are
one square mile, these values represent the
area covered by the lines (except in a few
smaller sections, where correction is easily
made). In general, where all or a large part
of the section is covered by the lines, it may
be expected that the section is well built up,
and the screening effect of buildings, etc.
therefore a maximum. Inversely, sections of
which only a small part is covered by the
lines probably are sparsely built up, and the
screening effect therefore a minimum. By
dividing the data into two parts, for small and
for large area of section covered by the lines,
and working up the two separately, a dif-
ference in the results should indicate the
difference between low and high screening.
^ 1 Steinmetz, Engineering Mathematics. Chapter VI, C.
The material was divided into eight groups
by the arrester density, so that each group
contained about the same number of failures.
Then each group was divided into two sub
groups of about the same number of failures,
one comprising the sections of small area
covered by the lines, that is, probably low
screening, the other the sections of large area
covered by the lines, that is, probably high
screening. The total material, and the two
subgroups separately, were then worked up
into empirical curves of the form proposed
by Mr. Roper, by the 2 A method' and gave
the three curves shown in Fig. 1, of the respec-
tive equations:
T.- Total data y =
1.62
S • Small part of sections covered by the
1.64
lines; probably low screening y= ^
G- Great part of sections covered by the
1.59
lines, probably high screening y = -~,
1002 December, 1920 GENERAL ELECTRIC REVIEW
where
J = percentage of failures per j'ear;
-v = arrester densitv, hundreds per square
mile.
It is interesting to note the difference of the
exponents, which means that curve G is much
Vol. XXIII, Xo. 12
20 40 60 so roo ,20 .40 >60 ISO eoo KO 2^0 260 280 300 32."
Arrester Density PerSqMile
Fig. 1
Steeper than 5. At low arrester densities
the three cur\-es come together, but in-
creasmgly separate with increasing arrester
density, so that with 120 arresters per square
mile there is a difference of 12 per cenf 3S oer
cent at 200 arresters; and 59 per cent dif-
ference in the percentage of failures at 300
density.
We may account for the increase of screen-
ing with increasing arrester densitv thus-
At low arrester density each arrester has to
drain a considerable length of line, and the
treedom from charge of its immediate neigh-
borhood, due to the screening, has little
effect on the total charge which the arrester
has to carry off. With high densitv of
arresters, however, each arrester drains onlv
a small area, and the reduction of the volume
ot the discharge by the screened area is much
more appreciable.
The exponent 1.0 differs slightlv from the
,,t ^•1:^.^°""^ by Mr. Roper, prbbablv due
to the different grouping of the data' here
used. This suggests a change of the curx-e
shape between high and low arrester densitv
Theretorc the data were worked up separateiv
Au"" u^?^"" °^ '°''' density, medium densitv
and high density. This gives the three curves
S^Tt" f'fu^' *°^^*'^^'" ^^"'th the average
cur\^e T, of the respective equations:
L: Low arrester density y =
L71
1.4S
M: -Medium arrester densitvr = —
H: High arrester densitv v = i^
- -' J-I.07
As seen, the low density cun-e is verv much
steeper, about twice as steep, as the hi-h
density cur^.e. In other words, increas nj
the number of arresters has much morS
At low density a 1 per cent increase of arrester
decreases the failures by 2 per cent. wSle S
high density it reduces the percentage of
failures by 1 per cent onlv.
When using all the data. Mr. Roper's
equation of failures gives: • ^
.4
the constants
a = 1.6 and A = 1.02
When using the exponent L6, but usine
only a portion of the data for the calculation
-■1, the value of .4 so derived compared with
"''"'"" »0 .» .40 ,M «I0 200 £20 2« !M :„lir5o
Mrnesttr Otnsitv ffa- 5(j Mile
Fij. 2
the average .4 = l.(;2 shows how the failures of
this group compare with the average.
In Table I arc given the values of .4 for
the 9 conditions, viz., low densitv high
density and total, low screening, high screen-
ing and total. While the numerical value-^
STUDIES IN LIGHTNING PROTECTION ON 4000-VOLT CIRCUITS
1003
y ^1.6
A =
Low Density
Total
1.86
1.62
1.41
High Density
Small Area
Total
Great area
1.69
1.61
1.54
.15
9.. 3
2.05
1.63
1.29
Max. dif.
Per cent:
.45
27.8
.76
46.7
themselves have little meaning, their general
trend seems to be decidedly significant in
indicating the relative increase of failures
with decreasing arrester density and increas-
ing screening, and the increased effect of
screening at higher arrester density as shown
by the percentage difference given in the table.
At high arrester density, the exponent a in
Mr. Roper's equation approaches 1. That is,
the percentage of failures decreases inversely
proportional to the number of arresters; or
in other words, the total number of failures
approaches constancy. This suggests plotting
not the percentage of failures but the total
number of failures as function of the number
of transformers or arresters per square mile.
This is done in Fig. 3. Approach to con-
stancy suggests the exponential function.
This is the more clearly indicated, as the
phenomenon is one of probability, and the
probability function is exponential.
If then t = number of transformers lost per
square mile per year, by the ^ A method the
equation is derived :
i = 6.8 XIO-'S-* ^-1-0.92
iz
to
a
?,6
I'"
£ 1,2
o-
u 1-0
i:
so.a
OA
02
to 40 fcO 80 100 120 MO leO ISO 200 220 210 2fcO 280 300 320
Arrester Density Per 5q Mile
Fig. 3
that is, for extremely high lightning arrester
densities the average failures approach a
minimum of 0.92 transformers per square mile
per year. For very low arrester densities
they approach 7.7 transformers.
This equation is given by the curve in
Fig. 3, and the 8 groups of data marked by the
circles.
r
\
V
2.5
S.
iL ^^0
\
^^
ji
^.
-^
^^
__-
E
— -
—
—
IL
f-
DISCUSSION BY V. E. GOODWIN
Lightning Arrester Engineering Department, General Electric Company
The art of protecting electrical apparatus
against voltage disturbances has made
material progress during the past ten years.
This progress has been due, not only to the
development of new protection methods, but
also to a wider knowledge of the nature and
character of the effects of lightning on electric
circuits. We have had a good conception of
these effects on high voltage circuits, but until
recently have had little accurate data on
low voltage distribution circuits. Low voltage
arresters have therefore been designed to
handle a wide range of impulse and high
frequency conditions. These arresters must
have low cost and reliability; hence it is
difficult to incorporate all the best protection
features for this entire range of conditions
and still have an arrester which is cheap
enough for the sen.nce.
In this paper, Mr. Roper has given us the
most complete operating record which has
ever been collected. This paper is of greatest
value since it shows the failure of trans-
formers and fuses blown during lightning
storms covering a period of five years and
includes an average of some fifteen thousand
installations. This paper clearly shows the
futility of trying to draw conclusions on the
relative merits of protective schemes without
the most careful study of operating data
including several thousand installations and
comparing each year's operation with each
successive year. A study of this report shows
that with high density of arresters, trans-
former failures are reduced to a fraction of a
per cent per year. These and other data show
the prevalence of a certain class of dis-
turbance having high rates of change of
potential and large destructive charges. Such
disturbances as these require the use of
either a few arresters having high discharge
rates or the use of a larger number of low
discharge rate arresters in parallel. This
point is further brought out by the fact that
the Type B arrester is shown by Mr. Roper's
paper to be superior to all the other types.
1004 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
As these tests progressed we have been
able to better understand the nature of
disturbances on these circuits and to work
on the development of a protector which will
have even greater discharge rates and at the
same time incorporate the best features of the
all-porcelain enclosed type.
By studying the transformer losses by
storms and by years, it is noted that the
losses seem to be confined to certain storms
and that these losses for a given storm are
grouped into a few square miles. .The
thought naturally comes to mind as to
the possibility of many of these failures
having been caused simultaneously by one
unusually heavy lightning discharge. Such
a discharge would release a very large
bound charge on a system as large as
the Commonwealth Edison Company. Such
a condition would suggest the application
of a few additional arresters having a high
discharge rate, as, for example, the alumi-
num or oxide film types, these arresters
to be distributed about the city in the
most important points. The same result
could be obtained by the use of a greater
number of arresters having a discharge
rate intermediate between the aluminum
and the multigap types.
The data presented in this paper, while
collected on a four-wire grounded neutral
system, probably represent conditions com-
mon to most low-voltage distribution circuits.
However, non-grounded circuits may present
slightly different results and it would be
most interesting if some of the large com-
panies operating non-grounded systems would
tabulate their results.
New Direct-current Reversing Motor for
Steel Mill Drive
The motor illustrated below has just been
completed by the General Electric Company
for the Tata Iron & Steel Company, Sakchi,
India. It is a double unit machine rated
6300 h.p., 80 r.p.m. with a speed range from
60 to 100 r.p.m. Speeds from 60 to 80 r.p.m.
are secured by means of generator field con-
trol and from 80 to 120 r.p.m. by motor field
control. Power is supplied by a flywheel
motor-generator set, consisting of a 6500 h.p.,
375 r.p.m. induction motor and two 2500-kw.
generators. The flywheel weighs 50 tons and
the motor is normally operated non-reversing,
but may be quickly stopped and reversed
when necessary-. The armature of the motor
weighs 132 tons and was the heaviest crane
lift on record in the shop in which it2;was
built.
6300h.p., 80 r.p.m.. Direct-current Reversing Motor for Steel Mill Drive; 22.aOO-h.p. Momentary
1005
I
The Bowl-enameled Mazda C Lamp
A NEW DEVELOPMENT IN ILLUMINATION
By Ward Harrison
Illuminating Engineer, National Lamp Works of General Electric Company
The new lamp described in this article was developed to provide a high-powered light source for indus-
trial plants which would be free from the objectionable glare that is common with existing types of lamps.
The enamehng on the lower half of the bowl effects almost complete diffusion of the transmitted light, and
when the lamp is used with special steel reflectors the resulting illumination is noticeably free from glare and
sharp shadows. The enamel forms a smooth surface which does not collect dirt; it withstands the action of
water, acid fumes, and ordinary mechanical abrasion incident to shipping and handling. — Editor.
Recent tests and experience have shown
the desirability of much higher illuminations
in industrial processes than were considered
necessary a few years ago. Increases in pro-
duction of from 8 to 25 per cent have been
registered in specific cases where improved
lighting systems pro^dding more foot-candles
on the work have been installed. However,
in nearly every case, the greater illumination
has made necessan,' the use of higher powered
lamps. This has brought about a serious in-
crease in glare in those instances where the
larger lamps were used in the older styles of
open reflectors designed for smaller lamps of
a less brilliant type. In fact, in extreme cases
the change has so increased the glare as to
have actually resulted in an installation of
reduced effectiveness.
In the early days of tungsten-filament lamps,
practically half of those sold were frosted to
reduce the amount of glare. During recent
years, however, this proportion has graduallv
dwindled with the result that today the pro-
portion of frosted lamps used is very low in-
deed, even though the size and power of the
lamps used have increased materially.
Various other means have been employed
in industrial lighting to secure better diffusion
than that provided by clear lamps in open
reflectors. Diffusing globe units, reflecto-cap
diffusers, opal-cap units and even semi-in-
direct and totally indirect lighting systems
have been adopted in industrial locations
suited to their use. However, even though
these installations were often highly success-
ful where intelligent super^'ision was given to
the use of the equipment, no one of these
types has appeared to be sufficiently desirable
from all the different standpoints of diffusion,
efficiency, cost of maintenance, and adapta-
bility to have become recognized as a standard
type for general industrial lighting. By far the
greatest percentage of indtistrial lighting was
still done by clear lamps and open reflectors.
The recent RLAI standardization effected
the production of a steel reflector more suited
to the Mazda C lamps than the previous
types of shallow and deep-bowl reflectors.
Used in conjunction with the RLM dome
reflector, the newly developed bowl-enameled
lamp presents a lighting unit which has great
possibilities as a standard unit for a wide
Fig. 1. The Bowl-enameled Lamp
variety of industrial locations. In fact, it is
estimated that this combination meets the
lighting requirements of at least 85 per cent
of industrial plants.
In appearance the bowl-enameled lamp
differs from a bowl-frosted lamp in that the
bowl is decidedly white and might be de-
scribed as having an egg-shell finish. When
1006 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
lighted, the lamp can be viewed end-on at
close range without discomfort; there is a
decided contrast in this respect between a
frosted lamp of a given wattage and a bowl-
enameled lamp of the same wattage.
The enamel is superficially applied. Its
edge is vignetted or shaded off as indicated
Fig. 2. This Combination Will Meet the Requirements of at
Least 85 Per Cent of Industrial Plants
in Fig. 1. thus avoiding the possibility of a
sharp line of cut-off on the reflector. As
regards durability, it will resist practically
any mechanical abrasion it is likely to en-
counter; it can be scratched or scraped only
by deliberate effort with a sharp knife or
similar tool. It will not chip off. Repeated
tests have shown it to be proof against deterio-
ration by acid fumes. Lamps have been im-
mersed for a considerable period of time in
boiling water without any damage resulting
to the enamel. From all test data available,
it is safe to conclude that the enamel will not
discolor whatever during the life of the lamp.
The lamp can be readily washed. It differs
from a bowl-frosted lamp in this respect. A
frosted lamp when placed under water becomes
almost transparent, with the result that it
is very difficult to detect the ])resence of dirt
or grease, which will show up only after the
lamp is dr}-. The bowl-enameling, on the
other hand, appears decidedly white even
under water, and dirt is casih- detected and,
also, easily removed. Because of its smoother
surface, the new lamp docs not collect dirt as
readily as a frosted lamp.
The bowl-enameled lamp was designed
principally for use with ojjcn reflectors. When
used in this manner, the lower j)art of the
lamp acts as a reflecting and difl'using surface,
serving the same purpose as the opal cap. Its
advantages over the opal cap are greater case
and decreased breakage in cleaning, and the
absence of any sj^ace iDctwccn the lamp and
the cap in which dust may collect.
Bowl-frosting of MazdaC lamps, particularly
n the larger sizes and in industrial ligliting.
has not sufficiently reduced the brightness of
the lamps as to fully meet the requirements in
many locations. Bowl-frosted lamps of the
100-watt size have a maximum brightness of
something like 75 candle-power per square
inch. The bowl-enameled lamp has a bright-
ness of about 10 or 12 candle-power per square
inch. If one looks at a lighted bowl-frosted
lamp, he will observ^e that the diffusion is by
no means complete, for at the center of the
frosted area there will be seen a brighter spot
an inch or so in diameter, whereas with com-
plete diffusion, as in the case of the bowl-
enameled lamp, the entire area is of the same
order of brightness.
The distribution cun-es for the clear, bowl-
frosted and bowl-enameled lamps form an
interesting comparison of their light-directing
properties. In Fig. 3 are shown typical dis-
tribution curves for these lamps, each of the
same size. In the clear-bulb lamp, the amount
of upward and downward light is. of course,
practically the same, so that when it is used
in an open reflector a large amount of the il-
lumination comes directly from the concen-
PERCtNT TOTAL LUMENS Of CLOIR LAMP
ZONE.
0-60
0-9O
90-IS0
0-iao
A
(gCtH LAWP)
B
21
JL.
ji3_
lOO
tf go'
n
40
_5§_
96
c ^
.2S-
69
_^
Fig. 3. Comparative Candle-power Distribution Curves
foi Mazda C Lamps. I A) Clear, (B) Bowl-frosted,
iC) Bowl. enameled
trated lamp filament. Shadows arc therefore
comparatively sharp and reflected glare from
]iolished surfaces is likely to he serious. Bowl-
frosting partially diffuses the downward light
from the filament, but redirects upward only
a very small proportion of the light flux.
THE BOWL-ENAMELED ^L\ZDA C LAAIP
100^
Bowl-enameling on the other hand not only
diffuses the downward light coming from the
bowl of the lamp, but what is very important,
ser\-es to redirect a large portion of the
light from the filament against the upper
reflector.
For a given light flux, the brightness is, of
course, inversely proportional to the area of
the source. In the combination employing
the bowl-enameled lamp with the RLM dome
reflector, the major portion of the effective
light comes from the surface of the large dia-
meter reflector instead of directly from the
lamp filament. The larger light source of lower
brightness has the direct result also of soft-
ening the reflections of the source from pol-
ished or oily surfaces. These specular reflec-
tions, resulting in what is termed reflected
glare, are often the cause of more serious eye-
strain than direct glare. Even where the
lamps are properly shielded by reflectors or
shades so that a workman does not encounter
the direct glare of the source, the lamp's
—RLM STAhDARD DOMKOO VATT
BOlVL-CriAMELCD MAZDA CLAMP
^00 VATT BOH/L-mAMCLCD MAZDA
CLAMPVITMOUTRErLCCTOR
ZONE
O-60
0-90
LUH£MS IXTomcmiiLmp
IB49
2055
53
66
Fig. 4. Candle-power Distribution Curve of the
Bowl-enameled Lamp in an RLM
Standard Dome Reflector
position may often be such that he will be
greatly hampered b>' the reflected glare, which
comes from his work or tools and is in his line
of vision throughout the day. This can only
be insured against by some means such as the
bowl-enameled lamp, which well diffuses the
downward light rays from the filament of the
lamp.
The comparatively large light source of low
brilliancy provided by this unit has an added
important advantage in softening shadows
and in avoiding the characteristic denseness
ZOOmJT.CLlAR MAZDA C LAMP IN
INDIRLCT LIGhTlhG FIXTURL
200WATT,B0WL LMAMLLLD MAZDA C
LAMP IN SAME TIXTURL
PERCENT TOTAL LUMLN5 OF CLLAR LAMP
zone:
~P~-60
[CLEAR LAMP
BOWL-CMAHElLtD
0-90
-f^
Fig. 5. Comparative Candle-power Distribution
Curves of a Semi-indirect Lighting Fixture.
(a) Fitted with Clear Mazda C Lamp,
(b) Fitted with Bowl-enameled
Mazda C Lamp
and sharp edges which come from concen-
trated light sources, and which are often both
annoying and dangerous in industrial loca-
tions. Dark sharp shadows interfere with
fast work and increase spoilage. In some
instances the shadows cast by moving parts
may be so sharp and dense as to cause con-
fusion between the object and the shadow,
with consequent likelihood of injury to the
workman. The softer shadows with shaded
edges, characteristic of larger effective light
sources such as the bowl-enameled lamp RLM
dome combination, avoid these undesirable
results and are important factors in populariz-
ing the use of this unit.
From the standpoint of absorption, the
total output of an R.LM standard dome re-
flector equipped with a bowl-enameled lamp
will be of the order of 10 to 12 per cent less
than for the same reflector equipped with a
clear lamp. This increased absorption is
about the same as that obtaining with the
opal cap and about twice that obtaining with
1008 December, 1920
GENERAL ELECTRIC REVIEW
Vol. XXIII, No. 12
a bowl-frosted lamp. A reasonable amount
of light can well be sacrificed for a major gain
in avoidance of glare effects and in softness
of shadows. It is important to note, however,
that with a deep-bowl steel reflector, a bowl-
frosted lamp results in a loss of about 10
per cent of the light, while a bowl-enameled
lamp results in a loss of from 25 to 30 per cent
of the light. This is due in both cases to the
bottling-up of the light in the narrow reflec-
tor by the frosted or enameled area of the
bulb. Distribution cur\'es show, however,
that most of this loss occurs above the angle
of 25 degrees with the vertical, so that there
will be exceptional cases where the bowl-
enameled lamp can be used to good advan-
tage in a deep-bowl steel reflector, but only
in such cases is this combination to be recom-
mended.
Bowl-enameled lamps are now available in
seven sizes; viz., the 100, 150, 200, 300, 500,
750 and 1000-watt lamps.
While, as has been suggested above, one of
the most important fields of application of the
bowl-enameled lamp is its use in conjunction
with steel reflectors for industrial lighting,
its possibilities are by no means limited to
this field. When used in place of clear bowl-
frosted lamps in direct lighting opal glass
reflectors, it provides illumination character-
ized by the same order of improvement in
diffusion and avoidance of glare that is ob-
tained when used with steel reflectors.
Another interesting possibility is the use of
bowl-enameled lamps in semi-indirect fixtures
which have a rather large downward compo-
nent, such as is found in the case of light-den-
sity glass equipment. Since the bowl-enamel-
ing in itself redirects upward a considerable
amount of the light flux from the lamp, the
result of placing a lamp of this kind in a light
density semi-indirect bowl is to change the
distribution by throwing more light on the
ceiling, thus lessening the direct light and the
apparent brightness of the fixture. Fig. 5
illustrates the distribution cun.-e of a popular
semi-indirect unit used with a clear lamp and
the same fixture when equipped with a bowl-
enameled lamp. In installations such as
offices and drafting rooms, where the max-
imum diffusion and low brightness of the
fixture are often considered highly desirable,
existing equipments considered lacking in
these characteristics can be readily adapted
by the use of the bowl-enameled lamp, to give
an improved light distribution.
Fig. 6. A Modern Lighting Installation in an Armature-winding Room. RLM Standardp^ome
Reflectors and 300-watt Bowl-enameled Lamps. Spaced 10 by 13 '^ feet. Mounted
11 feet above floor. Average illumination obtained. 17 Foot -candles
General Electric Review
INDEX TO VOLUME XXIII
January, 1920— December, 1920
INDEX TO VOLUME XXIII
Jan.^ 1-76
July 555-644
Feb. 77-181
Aug. 645-719
INCLUSIVE PAGES
Mar. 183-234
Sept. 721-790
Apr.
Oct.
235-354
791-866
May 35.5-464
Nov. 867-933
June 465- 553
Dec. 935-1008
Activity, unprecedented; from uncertainty to (Ed.) 357
Altitude, airplane; measurement 482
Ampere. Oersted and Arago. 100 years since 971
Arago. Ampere, and Oersted. 100 years since 971
Automobile
electricity on the (Ed.) 185
lamps, headlight. Mazda ;improved methods manufacture. 67
starting systems, electric 186
Bearings
generator, a-c 13o
motor, synchronous 135
spring thrust
generator, a-c. vertical 162
marine; developments 1919 10
Boilers, exhausted generator air used under , 101
Bushings, high-voltage; developments 1918 30
Busses
generating station, superpower 408
test . 418
Calculating table, short-circuit, new 669
Circuit breakers
high-speed 394
developments 1919 15
locomotive, passenger. C. M. & St. P. Rwy 282
new type 286
Condensers
-resistance protective device 774
static; developments 1919 34
synchronous 143. 147
capacity 143
design, electrical 144
developments 1919 20
highest voltage (22.000 volts) 152
largest capacity (30.000 kv-a.) 150
noise 146
speed 143
starting 146
ventilation 146
Connections, high-voltage; employed in generating stations,
relative merits 386
Control
car equipments. 1200 and 1500-volt 314
generating stations in parallel 647. 688
hoists, mine, electric 778
locomotive passenger, C. M. & St. P. Rwy 275. 278
air compressor 284
blowers 284
circuit breaker, high-speed 282
contactors 283
disconnecting switches, knife-blade 282
motoring 278
motor-generator for control and lighting 284
pantograph trolleys 279
regenerative braking 278
storage batter>' . . . . 285
Converters
synchronous. 60-cycle
brush rigging 393
commutating poles 392
commutating regulator 395
compared with motor-generators 392
cost 398
efficiency , 396
flexibility 397
floor space 398
reliability 397
flash barriers .■ 394
voltage control 395
Corona, loss tests 419
Crane, fitting out; hammer head, 350-ton. , .550
Distribution
C. M. & St. P. Rw>-. (Dia.) 726
stations, automatic; developments 1919 17
Dynamometers, developments 1919 27
Eden, T. S.. (In Mem.) 181
Edison
birthday comments on work 233
medal, 1919. awarded William LeRoy Emmet (Ed.) 3
Efficiency
measurement by temperature rise ventilating air. ...... 153
motor, synchronous 112
reward for 715
Electrical
development, self interest will solve problems confronting 451
engineering, co-operative course, new. Mass. Inst. Tech. . 784
industry
developments 1919 4
importance of, in foreign trade of United States . . . 752
Electrification, railroads
advancement (Ed.) 237
C. M. & St. P. Rw>-.
Coast and Cascade Divisions 263
input (Tab.) 728
operating statistics, 1919 (Tab.) 729. 730
profile 250. 724
results (Ed.) 723
choice of system, conclusions French Mission 244
direct -current high-voltage 243, 313
economic considerations 245
French Mission's report, summary 239
frequencies 255
Hershey Cuban Rwy 307
Loetschberg Rw>- 256
Melbourne (Australia) suburban 662
distribution, direct-current 668
equipment, electrical 664
features, general 662
power station 667
rolling stock 664
substation equipment 667
monophase 24 1
single-to-three-phase 242
superpower rone 246
United States 249
coal sa\'ing (Tab.) 253
power required (Tab.) 254
year 1919 12
Electro-magnetic waves. Maxwell; Professor EUhu Thom-
son 's early discovery 208
Electron tubes
applications 514
power 514
tungsten-cathode, fundamental phenomena 503. 589
tungsten-filament
anode 843
bulb and glass 843
filament 840
grid 843
operating features, some practical 840
plate 843
power supply 846
tube circuits and their operation 844
vacuum conditions. . 844
Emmet, William LeRoy; Edison medal 1919 awarded (Ed.). 3
Engineer's part in readjustment present high prices 222
Excitation
generating station, superpower 410
systems 558. 566
common plant 566. 568
individual exciters 568, 572
storage batteries with exciter . 571
various, discussion 575
Exciters 558. 566
circuit breakers 570
com pound -wound vs. shunt 57 1
commutating-pole 573
driving methods .S68
field switches . 570
polarity reversed . . 572
regulators, with . . . . 573. 574
rheostats for 570
shunt vs. compound- wound 57 1
voltage 569
range 569
Flywheel effect, synchronous motors, reciprocating com-
pressors 653
Foreign trade, importance electrical industry United States 752
Frequency
absorber, high 432
commercial 88
converters 136. 140
lag angles 136
parallel operation 137
phase adjustment, parallel operation . 139
reversibility 138
synchronizing 138
Fuel
petroleum (Sec Petroleum)
-ail way
coal
C. M. St. St. P. Rw>-.. Rocky Mt. Div. (Tab.) 251
pounds per 1000 ton-miles, steam locomotive
(TaV) 883
roundup, analysis (Tab.) 252
saving, electric locomotive 880
U. S. 191S (Tab.) aSl
PAGE
Furnaces
electric, heat treating, metallic resistor 433
thermal economy relative to fuel-fired 768
fuel-fired, thermal economy relative to electric 768
Gases, kinetic theory 495
Gasolene plants, casing-head 628
Generating stations
automatic, developments 1919 16
capacity in U. S. 1918 (Tab.) 254
connections, high-voltage, relative merits 386
control in parallel 647. 688
Hershey Cuban Rwy 309
hydraulic; excavated in mountain, Norway. 166
load in oil fields 616
outdoor
construction 194
design 194
Muscle Shoals. Ala., proposed 196
Melbourne (Australia) electrification 667
stability in parallel 647. 688
superpower
apparatus arrangement, electrical 417
ash handling equipment 401
benchboard 410
blowers and fans 402
boilers _ 402
boiler room auxiliaries 416
busses 408
test 418
circulatmg water tunnels 404
coal handling equipment 399
condensers 406
economizers 404
electrical design 407
excitation 410
outdoor structures 418
piping 404
power supply, auxiliary 406, 411, 412
preheaters 404
reactors 407. 417
station lighting 412
stokers 402
switches 408
switchboard
central auxiliary 411
turbine auxiliary 413
turbines, house 405
Generators
alternating-current
bearings 135
belted; design and construction 171
charging transmission line, behavior 109
compensated 87
design, early days 80
development 21, 79
engine driven 90
excitation kilowatts required 570
field spiders 134
heat
flow 564
storage 564
mechanical features 130
monocyclic 85
oscillating frequency between two dissimilar 125
outdoor; alternative for 360
periodicity and near periodicities 88
polyphase 85, 86
rotor spiders 134
shafting 135
sine wave; for testing 177
speed, peripheral, (rotor losses) 562
stator frames 133
temperature in large 557. 560
high vs. low 565
turbine ; early 87
ventilation 135
vertical shaft, bearings and lubrication 162
water cooling 561
waterwheel
excitation kilowatts required 570
horizontal, large 147
unique design (Norway) 166
welded construction 131, 132
automobile starting and lighting 190
direct-current
arc welding; new type 442
exciters (See Exciters)
turbine
10.000 kv-a. short-circuit tests 214
excitation kilowatts required 570
light and power; U. S. Navy, developments 1919. . 8
mechanical design, large 105
Heating, industrial; developments 1919 38
Helium, substitutes for hydrogen in balloons and dirigibles 22
Hoists
mine, electric
developments 1919 22
installations, typical. South Africa 775
number above 250 h.p. South Africa (Tab.) , . .781
one-ton, Sprague; adoption of .roller bearings 24
Hydrogen, penetration of iron 702
PAGE
Illumination
color walls and ceilings
economies 211
effect 209
industrial plants 209
offices 209
paint, methods of applying 211
residences 210
schools 209
stores 210
wall finishes, permanency of various 210
Integration, curve areas 917
Iron, penetration by hydrogen 702
testing; sine-wave generator for ISO
Lamps
accessories, new 50
arc
enclosed carbon, vs. Mazda novalux 534
luminous 51
Cooper Hewitt
mercury vapor
application 858
development 858
operation 741
theory 741
quartz 909
fixtures, new 50
incandescent
automobile, headlight Mazda; improved methods
manufacture 67
bowl-enameled. Mazda C 1001
Novalux (Mazda), vs. enclosed carbon arc 534
sales, amount 48
white, Mazda 711
Moore, gaseous conduction; light from low-voltage
circuits 577
Light
ultra-violet 909
effect on eye 893
Lighting
codes, typical 51
daylight, artificial; for merchandising and industry 527
developments 1919 48
displays 52
exhibits 52
street
arc, enclosed carbon vs. Mazda novalux 534
incandescent (Mazda) novalux vs. enclosed carbon
arc 534
intensive 362
Chicago 373
Los Angeles 365
New Orleans 365
Salt Lake 364, 365
San Francisco , 365
Saratoga Springs 373
transformer
constant -current 543
series; low-voltage 545
Lightning arresters
alternating-current 429
aluminum cell . . 429
developments 1919 35
locomotive passenger. C. M, & St. P. Rwy 285
oxide film 430
capacity current discharge 929
life 929
performance 928
sensitiveness 928
tests 929
Lightning protection, studies oa 4D JO- volt circuits 981
Locomotives
electric
advantages of modern 878
C. M. & St. P. Rwy.
maintenance cost (Tab.) . .880
passenger 272
air compressor 284
auxiliary apparatus 284
blowers 2S4
circuit breaker, high-speed , , 282
contactors . . 283
control (see Coatrol)
dimensions 277
disconnecting switches, knife-blade 282
main circuit apparatus 279
mechanical construction 276
motoring 278
motors 274
motor-generator for control and lighting. .284
pantograph trolleys 279
regenerative braking,, 278
repairs (Tab.) ,880
storage battery ... . 285
performance (Tab.) 883, 884
compared to steam (Tab.) , 879
construction 258
cost, comparative 884
design
possibilities 878
simplicity of. importa.ice 341
PAGE
Locomotives
electric
developments 1919
fuel saving
gathering new type .
Hershey Cuban Rwy.
maintenance costs .
B. A. & P. (Tab. I
N. Y. C. (Tab.) . .
regenerative braking
steam
C. M. & St. P. Rwy.
performance (Tab.) . .
repairs (Tab.)
coal per 1000 ton-miles (Tab,
compared to electric (Tab.)
last stand
Losses
corona, tests 419
measurement by temperature rise ventilating air . ! 153
Lubrication, alternators, vertical shaft 152
Manometers 731 qa-v
ionization gauges ' 554
!' .734
731
847
851
735
. 15
880
446
.308
. .879
.880
59. 880
. .879
883. 884
880
.883
879
.249
mechar.ical.
mercury . . . .
radiometers .
resistance. . .
viscosity. ,
Marine propulsion
bearings, spring-thrust, developments 1919, , IQ
electric
battleships, developments 1919 10
cost . . 65
developments 1919 '/..'.'.... 9
economy ] g9
efficiency of transmission 64
merchant ships gO
Mariner, first electrically-operated trawler. . .. 455
New Mexico, battleship, two years' service 615
reliability 63
space saving over reciprocating-engine 6'
gear
efficiency of transmission ... 65
reliability .'.'..'. 63
torpedo boat destroyers, developments 1919 7
turbines, developments 1919 6
operating force * gg
propeller speeds g-
Massachusetts Institute Technology.' electricai ' engineering
co-operative course 784
Meter, testing; sine-wave generator for . . 180
Motors
automobile starting Igg
fractional horse-power; developments 1919 27
induction (large); control speed and power factor by
neutralized polyphase commutator machines. .559, 630
railway
design and construction, improvements in . . 335
locomotive, passenger C. M. & St. P. Rwy.. 274
synchronous U^
amortisseur winding .'!.*!." 132
application !!."!!!""•"'"' 113
bearings .^IW .[.... .\.][ 135
dependability 112
developments 1919 ' .\... ...... .[ 22
efficiency iTo
field spiders I34
flywheel effect .. .121. 653
limitations II3
magnetomotive force diagram '.'.'.' 12'*
mechanical features [.[ 130
power-factor ." . ...".' .'.',".'112, 119 1->1
rotor spiders ' 134
shafting 1.35
starting
ability II3
systems, oil . !!!.'!!!! 133
stator frames.
133
^o^<ife. :;:::;::: 1 15
ventilation
.135
.653
welded construction '. 131 1T>
Motor drive ""• '■*"
compressors, reciprocating, flywheel effect for synchro-
nous ' g
printing presses, developments 1919 07
steel mill ~
developments, 1919 24
reversing, tonnage '. oob
Motor-generators ^^
brush rigging 303
compared with synchronotis converters, eo^cycle. 392
"^S?'- .'.'.'!' .398
efficiency .,ha
flexibility ■'■'■'■'.'.'.'.'.'.'.'.]'.'.[['. 397
floor space oqc
reliability '.'.'.'.'.'.'.'. 397
excitation kilowatts required '. t7ii
flash barrier.
394
locomotive passenger. C. M. jb St'.Pi RWy. (for con-
trol and lighting) .,0 .
="= :::::::;:::::::;:::;::uo
N. E. L A., proposed changes, conducting 350
Network protector, alternating-current 427
Noise, condensers, synchronous fjg
Oersted. Ampere, and Arago. 100 years since 971
Umce, opportunities, work. 700
Oil field '82
electric load ria ait
dehydrators 6^9
gathering pumps rSJ
lighting .".... 628
line pumps ggo
machine shops g|g
Oil pump, lubricating, vertical generator. ....'..'. 155
Oil well
drilling
electric power, advantages go?
motor equipment .'.'.'.'.'. 625
power consumption . . [[[ gOo
powers. * ■ motor drive . ' goo
pulling, motor capacity required 621
pumping
electric power, advantages . 623
motor capacity required. ... gon
motor equipment gjo
power consumption goi
'"^^^^ : 627
Parallel operation
frequency converters. j3g
generators, a-c.. two dissimilar, oscillating freauencv 12=i
generating stations
control and stability g^y ggg
electromotive forces .../.. 756
equations [ yQ,
oscillation ...'..['.'..'...'. 695
power limitation gag
pulling in step ggg
reactors
busbar gg^
^e«der : ; .':.■.■;;.' .'.■.■;.*;:;;6yi. 758
generator gg j
slipping ; ' Qgj
steady strain 693
synchronous machines
Pasadena convention, electricity at (Ed.)
Patterson, Charles E.; Vice-President Gen
pany. elected a 70^
Petroleum ,90
distribution. . ..yg 007
"^^"'•e " 198
198
; General Electric Com
693
358
occurrence .
origin jgg
production qoI
refining ... !'!!!!."!!!! ''05
transportation * oru
Phosphorscope. special form .'.'.'..'/.'.[].'. 856
Photo-elasticity, determination of stresses. . ..,'.'." 869 870 962- "^
Photometry, commercial '/. , g = * '
Position indicators .'.....'.'..... 45
Power-factor, motor, synchronous ijo j jg loi
Power indicating system. C. M. & St. P. Rwy "' ' 29>
Power limiting
generating stations, parallel operation. . . 688
system, controlled by train dispatcher. C. M. & St. P jlwy 7*'7
Prices, high, present; engineers' part in readjustment of *>2-'
Protective devices - - -
condenser-resistance .
hoists, mine, electric ,
network, alternating-current ...
Pumping plant, electric. Sutter Basin RecUmation.
Radio communication
Alexanderson system. g^g
alternator .795.815
developmen: g| .
speed regulator .796^ 818. "833 835
vibrator regulator 7gg
amplifier, magnetic '821 836
antenna, multiple tuned. . .797, 8 Irf, 8^3. 824. 825," 831
efficiency radiation * ' g32
827
831
778
774
427
684
feed ratio.
ground inductances.
phase difference
resistance, mult i pit-
support ....
theory
tuning
voltage,
earth system .
capacitive . .
wire
modulating system
radiophone duplex
New Brunswick arrangements
requirements
receiver
barrage gQj
duplex on U.S.S. Gtorgt Waskingion ... , .90/7. S\0
standard equipment .... g]^
station circuits, fundamental ^22
transformer, alternator-antenna Jji?
transmitter, duplex on U.S.S. C#or«» 'Waiki^t*<,H . . .804
827
829
823
798
S26
828
820
831
S31
797
808
807
PAGE
development 17^ 793
generating systems. '. 795
radiation, directive !!!!!!!!!!!!!!!!!! 801
receiving system 801
telephony 838
transmitting sets, electron tube 523
transmitting systems 794
transoceanic 794
Railway (See Electrification and also Traction)
Reactors
busbar 689
feeder 691
generating station, superpower. . _ 407, 417
generator, . ; . .689
Reflection
coeflicients. . 213
diffuse, coefficient; absolute determination 72
factor, measurement 212
Relativity CEd.) 467
theories in physics 486
Relays
differential, mechanically balanced 44
hesitating control 45
Research. Laboratory, G.E., developments 1919. 20
Safety cars .... 397
air brake 603
air compressor 604
governor. ..... ... .605
developments 1919. . 14
installations ... .600
maintenance costs 600
motor equipment 602
power consumption 599
safety features . 603
Short circuit
calculating table, new . , 669
tests on 10.000 kv-a. turbine alternator. 214
Shovels, electric
300-ton in open-cut quarrying . 937
developments 1919 ... 24
open-cut mining , , 938
Sine wave testing sets ... 177
Spokesmen, silent; factory (placards, etc.) 229
Stability, generating stations in parallel 647, 688
Statistics, commercial; and value to executives 648
Storage batteries
automobile starting and lighting, taper charging 191
exciter, use of 571
locomotive, passenger, C. M. & St. P. Rwy., auxiliary
low-voltage supply 285
Stresses, photo-elasticity of 869, 870. 962
Substations
automatic
developments 1919 14
railway
control and devices, modern. . 342
Sacramento Northern Rd. . . . .894
signal power supply, a-c 902, 949
C. M. & St. P. Rwy. equipment (Tab.) , . . .725
- Hershey Cuban Rwy 309, 311
Melbourne (Australia) electrification 667
Superchargers
airplane 468
developments 1919. . . 11
engines 471
engine power, maintaining at great altitudes 474
turbo, General Electric 476
Switches
developments 1919 40
generating station, superpower. . 408
lever enclosed safety 43
Switchboard
benchboard, generating station, superpower 410
central auxiliary; generating station, superpower 411
safety panels
swingout type 41
truck type 40
turbine auxiliary; generating station, superpower 413
Synchronizing frequency converters 136
Tapalog. kilowattmeter, totalizing, curve-drawing 294
Terminals, automobile electric starting apparatus 192
Tests
corona loss 419
short-circuit, 10.000-kv-a. turbine alternator 214
Testing
iron; sine-wave generator for 180
sine-wave generator for 177
Thermometers, lubrication; thrust bearing .164
Thermostatic metal, manufacture, characteristics and appli-
cations 57
Thomson, Elihu 970
early discovery Maxwell electromagnetic waves 208
Traction
busses, motor vs. trackless trolleys 331
C. M. & St. P. Rwy.
kw-hr. to ton-miles, 1918 (Tab.) 252
load curve 725
load factor. 1919 (Tab.) 255
power consumption
demand controlled by train dispatcher 727
limiting and indicating system 292
Rocky Mountain and Missoula Divisions 724
train sheet 725
standby losses (Tab.) 251
water used, Rocky Mt. Div. (Tab.) 251
city cars, various types, operating costs 326
coal equivalent per kw-hr. (Tab.) 252
control (see Control)
developments 1919 12
equipment
metropolitan; developments 1919 15
miscellaneous; developments 1919 15
European conditions 253
line work car. Hershey Cuban Rwy 309
locomotives (see Locomotives)
motors, design and construction, improvements in 335
public trusteeship, Boston Elevated Rwy 319
safety car (see Safety Cars)
substations (see Substations)
ton-mile movement. U. S., 1918 (Tab.) 250
trackless trolleys vs. motor busses 331
Transformers
developments 1919. . . . 28
phase rotation 374
polarity 374
street lighting
constant-current 543
series, low-voltage 545
voltage diagrams 374
Transmission
C. M. & St. P. Rwy. (Dia.) 726
charging line, behavior a-c. generators 109
Hershey Cuban Rwy 312
Turbines
development 1919 5
marine geared; developments 1919 6
thermal efficiency; instances of high 5
Vacua, high
exhaust procedure 678
manometers (see Manometers)
measurement 731
production methods 605, 672
pumps
mechanical 607
mercury vapor , . . 672, 677
theoretical considerations. . . 605
Ventilation
air washers 1 00
boilers, exhausted generator air used under 101
classification machines with respect to 561
closed air circuit system 102
air, freeing from water 103
applications 104
danger signal 103
heat regulating 103
heat run on a turbine generator 104
water required, amount 102
condensers, synchronous 146
generators, a-c 99, 135
closed and semi-closed 561
dampers 100. 103
direction, cooling air,. . . 563
double air flow system. . 101
ducts
armature cores 563
external innovation of 100
fans 100, 101
fire protection. . , 103
hoods 99
motors, synchronous . 135
turbine generators (see Ventilation, and Generators A-c.)
vertical to horizontal units (turbine-generator) changing
from 100
water-air radiators
development 1919 ^. . 11
generators * 91
motors 91
Voltage regulators
developments 1919 34
exciter, use with 574
feeder; developments 1919 . 33
Wakefield, William O., biography 232
Welding
arc
applications to construction
generators, a-c 131
motors, synchronous 131
generator, new type 442
electric, developments 1919 36
spot
applications to construction
generators, a-c 132
motors, synchronous 132
Winch, electric; developments 1919 , 24
Woolley, Geo. A. (In Mem.) 719
INDEX TO AUTHORS
Addison. Thomas
Proposed Changes in Conducting X. E. L. A. (Ed.) . .
Alexanderson. E. F. W.
Transoceanic Radio Communication
Anderson, Earl A.
The White Mazda Lamp
Andrews, H. L.
Motor Busses or Trackless Trolleys
Andrews. W. S.
A Special Form of Phosphoroscope
Helium, the Substitute for Hydrogen in Balloons and
Dirigibles
Armstrong. A. H.
The Advantages of the Modern Electric Locomotive .
The Last Stand of the Reciprocating Steam Engine. ,
Axtell. C. J.
Control for 1200 and loOO-volt Car Equipments
Bachman, E.
Five Thousand Horse Power Electrically-operated
Pumping Plant
Baldwin. G. P.
Commercial Statistics and Their Value to the Execu-
tive
Ballard. R. H.
Electricity at the Pasadena Convention. (Ed.)
Barton. F. C.
Electric Starting Systems for Automobiles
Batchelder. A. F.
Importance of Simplicity in Locomotive Design
Passenger Locomotives for Chicago, Milwaukee & St.
Paul Rwy
Bearce, W. D.
Melbourne Suburban Electrification. Australia
Summary of High-voltage Direct-current Railways.. .
The Safety Car
Beers. R. S.
Control for 1200 and 1500-volt Car Equipments ....
Beeuwkes, Reinier
Electric Power Consumption on the Rocky Mountain
and Missoula Divisions of the C. M. & St. P. Rw>-.
Bell. E. B.
Typical Installations of Electric Mine Hoisting in
South Africa
Benford. F. A. ,
An Absolute Method for Determining Coefficients of
Diffuse Reflection
Bergman. S. R.
A New Type of Arc-welding Generator
Berkshire. W. T.
Synchronous Motors
Beverage. Harold H.
Duplex Radiophone Receiver on U. S. S. George
Washiiigloyt
Boyajian. A.
Fundamental Principles of Polarity. Phase Rotation,
and Voltage Diagrams of Transformers
Bradley, C. D.
The 300-ton Electric Shovel in Open-cut Quarrying
(Ed.)
Bucher, Elmer E.
The Alexanderson System for Radio Communication.
Buck. H. W.
The Outdoor Generating Station
Burnham. E. J.
Sine Wave Testing Sets
Sixty-cycle Converting Apparatus
Butler. H. E.
Enclosed Carbon Arc Lamps vs. Novalux Mazda
Units
Buttolph. L. J.
The Cooper Hewitt Mercury Vapor Lamp
Part I — Theory and Operation
Part II — Development and Application
The Cooper Hewitt Quartz Lamp and Ultra-violet
Light
Case, F. E.
Control Equipments of the New Locomotives for the
C. M. &St. P. Rwy
Clark. V. E.. Lieut-Col.
Maintaining Airplane Engine Power at Great Alti-
tudes
PAGE
359
794
711
331
856
227
878
249
314
684
648
358
186
341
272
662
313
597
314
724
775
72
442
112
807
374
937
813
194
177
392
534
741
858
909
278
474
P.\GE
C'aike. C. L.
Step-by-step Integration of Curve Areas of Phase
Significance. Correct and Incorrect Methods. . . 917
Coker. E. G.
Photo-elasticity for Engineers
P^rt I 370
Part II 962
Collins. E. F.
Metallic Resistor Electric Furnaces for Heat Treating
Operations 433
Relative Thermal Economy of Electric and Fuel-fired
Furnaces 75g
Dana. Edward
The Public Trusteeship of the Boston Elevated Rail-
way 319
Davis. C. M.
Modern Devices and Control for Automatic Railway
Substations 342
Dawson. W. F.
Measurement of Losses and Efficiency by Temperature
Rise of Ventilating Air 153
Delehanty. W. J.
Five Thousand Horse Power Electrically-operated
Pumping Plant 684
Dodd, S. T.
Passenger Locomotives for Chicago. Milwaukee & St.
Paul Rwy 272
Doherty. R. E.
Flywheel Effect for Synchronous Motors Connected
to Reciprocating Compressors 653
Oscillating Frequency of Two Dissimilar Synchro-
nous Machines 125
Dushman. Saul
The Production and Measurement of High Vacua
Part I — Kinetic Theory of Gases 493
Part II — Methods for the Production of Low
Pressures 605
Part III — Methods for the Production of tow
Pressures (Cont'd) 672
Part IV — Manometers for Low Gas Pressures 731
Part V — Manometers for Low Gas Pressures
(Cont'd) 847
Edgerton. E. O.
The Reward for Efficiency. . ... 715
Emmet. W. L. R.
Electric Propulsion of Merchant Ships 60
Evans. W. H.
Automatic Substation. Sacramento Northern Rail-
road 894
Fisher. C. R.
The Electric Shovel in Open-cut Mining 938
Foster. W. J.
Early Days in AUcrnator Design 80
Temperatures in Large Alternating-current Generators 560
Freiburghouse. E. H.
Investigation of Water-air Radiators for Cooling Gen-
erators and Motors 91
Fuller. T. S.
The Penetration of Iron by Hydrogen. . . 702
Gilbert. C. G.
Methods for More Efficiently Utilizing Our Fuel Re-
sources.
Part XXXI— Petroleum 198
Glass. A. E.
A Unique Design of Waterwhcel-driven Alternator. . . 166
Goodwin. H.. Jr.
Design of a Superpower Station 399
Goodwin. V. E.
Alternating-current Lightning Arresters 429
Discussion of Paper, Studies in Lightning Protection
on 4000-voU Circuit 999
Gordon. T. W.
Bearings and Lubrication for Vertical Shaft Alter-
nators 162
Hadley. A. L.
Belted Alternating-current Generators. . . 171
Hallett, George E. A.. Major
Superchargers and Supercharging Engines 468
Harrison. Ward
The Bowl-enamcled Mazda C Lamp, A New Develop-
ment in Illumination 1001
PAGE
Hayden, J. L. R.
Condenser-resistance Protective Device 774
Discussion of Paper. Studies in Lightning Protection
on 4000-volt Circuit by D. W. Roper 997
Head. H. G.
The Electric Shovel in Open-cut Mining 938
Henningsen. E. S.
Magnetomotive-force Diagram of the Synchronous
Motor 122
Short-circuit Tests on a 10.000-kv-a. Turbine Alter-
nator 214
Herrman, Henry
Thermostatic Metal 57
Hull. John I.
Theory of Speed and Power-factor Control of Large
Induction Slotorsby Neutralized Polyphase Alterna-
ting-current Commutator Machines 630
Jackson, D. C.
Dr. Elihu Thomson 970
Jackson, J. A.
350-ton Hammer Head Fitting Out Crane 550
Jacobs. H. M.
Automatic Substations for Alternating-current Rail-
way Signal Power Supply
Part 1 902
Part II 949
Johnson, E. S.
Electrification of the Coast and Cascade Divisions of
the C. M. & St. P. Rwy 263
Langmuir, Irving
Fundamental Phenomena in Electron Tubes Having
Tungsten Cathodes
Part 1 503
Part II 589
Lewis. W. W.
A New Short-circuit Calculating Table 669
Some Corona Loss Tests 419
Linebaugh. J. J.
Power-limiting and Indicating System of the C. M.
& St. P. Rwy 292
Liston, John
A New Type of Gathering Locomotive 446
Some Developments in the Electrical Industry During
1919 4
The Mariner: The First Electrically-operated
Trawler 455
Lougee. N. A.
Performance and Life Tests on the Oxide Film Light-
ning Arrester 928
Lovejoy. J. R.
From Uncertainty to Unprecedented Activity (Ed.) 357
McCann. Anna
Opportunities in Office Work 782
Mauduit. A.
Summary of French Mission's Report on Railway
Electrification 239
Monson, George
Steam Turbine Generator Ventilation 99
Moore, D. McFarlan
Gaseous Conduction Light from Low-voltage Circuits. 577
Morse. W. O.
The Behavior of Alternating-current Generators When
Charging a Transmission Line 109
Moss, Sanford A.
The General Electric Turbo-Supercharger for Air-
planes 476
Olson. M. C.
Large Horizontal Alternating-current Waterwheel-
driven Generators and Synchronous Condensers. . . 147
Oudin. M. A.
The Importance of the Electrical Industry in the
Foreign Trade of the United States 752
Pauly. K. A.
The Electric Reversing Mill Considered from the
Standpoint of Tonnage 886
Payne. John H.
Radiophone Transmitter on the U. S. S. George
Washington 804
Peters, F. W.
Electrification of the Hershey Cuban Railway 307
Plenge, E. B.
Synchronous Condensers - 143
PAGE
Pogue. J. E.
Methods for More Efficiently Utilizing Our Fuel
Resources.
Part XXXI— Petroleum 198
Porter, L. C.
Improving the Mazda Automobile Headlight Lamp, . 67
Potter. W. B.
Railway Electrification in the Superpower Zone 246
Powell. A. L.
Commercial Photometry
Part 1 954
Effect of Color of Walls and Ceilings on Resultant
Illumination 209
Pragst, Ernest
Relative Merits of Connections Employed in High-
voltage Generating Stations 386
Priest. E. D.
Improvements in the Design and Construction of
Railway Motors 335
Reist. H. G.
An Alternative for Outdoor Generators 360
Investigation of Water-air Radiators for Cooling
Generators and Motors 91
Roper. D. W,
Studies in Lightning Protection on 4000-volt Circuit 981
Ryan. W. D'Arcy
Intensive Street Lighting 362
Savage. M. A.
Mechanical Design of Large Turbo-generators 105
Scott. Roscoe
Silent Spokesmen in the Factory 229
Shirley. O. E.
Parallel Operation and Synchronizing of Frequency
Converters 136
Smith. A. R.
Design of a Superpower Station 399
Snyder. Monroe B.
Professor Elihu Thomson's Early Experimental Dis-
covery of the Maxwell Electro-magnetic Waves.. . . 208
Steinmetz. C. P.
Discussion of Paper, Studies in Lightning Protection
on 4000-volt Circuit by D. W. Roper 994
Power Control and Stability of Electric Generating
Stations
Part 1 688
Part II 756
Stewart. H. C.
The Alternating-current Network Protector 427
Stickney. G. H.
Artificial Daylight for Merchandising and Industry. . . 527
Summerhayes, H. R.
Exciters and Systems of Excitation 566
Summers. J. A.
Commercial Photometry.
Part I 954
Tappan. G. H.
Motor-generator Sets 140
Taylor. W. G.
Electric Power in the Oil Fields as a Central Station
Load 616
Thirlwall. J. C.
Operating Costs of Various Types of City Cars 326
Thomson. Elihu
One Hundred Years Since Oersted. Ampere and Arago. 971
Timbie. W. H.
A New Co-operative Course in Electrical Engineering. 784
Tolman, R. C.
Relativity Theories in Physics 486
Townley, Calvert
The Engineer Can Do More About It Than
Pay and Grin 222
Tritle. J. F.
New Type of High-speed Circuit Breaker 286
White. W. C.
Electron Power Tubes and Some of Their Applications . 514
Some Practical Operating Features of Tungsten-fila-
ment Electron Tubes 840
Wishon. A. Emory
Self-interest Will Solve the Problems Confronting
Electrical Development 451
Wood. A. P.
Some Mechanical Features of Synchronous Machines. 130
'i.t-S3
, . M- ■- -
- rv ' •'. '
■■'. -~rV-- -
::r'J-
K^'C' ' \
■ V -^-i ,
'»\
' V- V "
. . . '- t'
^
-.V''''''-:-^''-
■''■i
■
.-' ,'■-.
:''• - '-"'.' .-'
■ ■■ ■ .:)fl
H
' -^- » , - •,
I .-,*.. --■■.
■ ■ ' -I
M
" ->
■ :";"■.■"■■' .iSfv-, ,■'■■,'-?.
-LX)
1= ^
= CO
= C\J
,= o
= — O)
^ — CO
= — N
IE — CD
= — in
E — -^
E — CO
E — ^
S_ E
LO-
CO-
C\J-
— E
C) -1
r
1
TC| tOINCHPl 0| t J
■ma om:.
•ysib ffits
■••^,"
-N
■00
CD
0)
■nI =1
00"
@
o I
z
o
o
CD _^|
O J
OJ
(V)
N)
CJi
fV)
G)
IV)
"■^ w
IV)_
00
— ;
_i o
— 1^
n
— '
X
z
l\)
'~~',
>
CD
^
o
— ^,
_^
^
IV)
6
7 8 9l lib ill 12 i|3 1
^ivy^JK/-'. ■/•■"*■" ■
««» N:«e
Live Music Archive
Librivox Free Audio
Metropolitan Museum
Cleveland Museum of Art
Internet Arcade
Console Living Room
Books to Borrow
Open Library
TV News
Understanding 9/11