RADIO TELEPHONY
FOR
AMATEURS
BY
STUART BALLANTINE
CONSULTING ENGINEER AND FORMERLY EXPERT RADIO AID, UNITED STATES NAVY
PHILADELPHIA
DAVID McKAY COMPANY, PUBLISHERS
f
Copyright, 1922, by
DAVID McKAY COMPANY
CONTENTS
CHAPTER PAGE
I. PRINCIPLES OF RADIO TELEPHONY 11
II. THE AUDION 44
III. ANTENNA CONSTRUCTION 58
IV. CONSTRUCTION AND OPERATION OF THE TRANSMITTER 96
V. SOURCES OF POWER 156
VI. RECEIVING APPARATUS 191
APPENDIX
A. UNDERWRITERS' SPECIFICATIONS GOVERNING INSTALLATION OF RE-
CEIVING ANTENNA 281
B. RADIO CLUBS. THE AMERICAN RADIO RELAY LEAGUE 284
INDEX.. 289
504606
ERRATUM
Due to an unfortunate choice of words, the statement on page 91
(Art. 37) relative to the amount of resistance to be employed with the
Beverage receiving antenna, may easily be construed to mean that
the surge impedance of a line 10 feet high of No. 16 A. W. G. copper
wire at high frequencies, is from 200-600 ohms. There is fortunately
no such uncertainty about the surge impedance of a line of these
dimensions, 550 ohms being the correct value. The range, 200-600
ohms, is indicated to allow for effects which the simple theory, giving
the correct resistance as approximately equal to the surge impedance
does not embrace.
PREFACE
Truly "of the making of many books there is no end"
and in view of the large number of radio books already
existant and advertised as in the process of preparation or
printing, perhaps some justification is necessary for the
appearance of another. My apologia is a simple one, and
possibly one engendered by ideas and aspirations not un-
familiar to all other radio amateurs. Since connecting up
my first receiving apparatus in 1908 I have longed for the
appearance of a certain type of radio book; a book issuing
preferably from the pen of an amateur, radiating the true
amateur spirit of inquest and investigation, a book chock-
full of practical information and suggestions for new things
to do and instructions for going about it, in a word, a
book which would at once ignite the spark of my enthusiasm
and furnish the material for its combustion. Such a book
has never appeared, nor ever shall for the great majority
of us, for we have all our own ideas as to what the ideal
radio book for amateurs should be like. But intercourse
extending over a number of years with kindred spirits and
frequent discussions with them on this subject has induced
the belief that in the main our ideas converge in a common
channel. This book represents my conception of the direc-
tion and breadth of this channel.
I have always looked upon the book-making business
with a great deal of fear and hesitation, clearly recognizing
5
6 PREFACE
the seriousness of addressing such a vast audience and the
power of the pen for the dissemination of good or evil
information. Particularly is this true of a book of this
nature, whose readers are for the most part non-technical
and in no position to judge the theoretical soundness of the
suggestions and methods exploited, yet who are at the
same time full of enthusiasm and ready to spend time and
money upon almost any suggestion offered. My original
plan therefore was to defer the book-writing for the growth
of a few gray hairs, but this plan has aborted due to the
encouragement of friends, the interest of the publishers, and
my own feeling that in view of the present widespread pop-
ular enthusiasm a more auspicious time for publication will
probably never arrive.
With elementary treatments of the theory of radio com-
munication on the one hand, and with systematic engineer-
ing texts on the other, my book obviously enters into no
competition. It is addressed mainly and first of all to the
amateur; to the amateur in both the commonly accepted
meaning of the word and in the true French sense of a lover,
or admirer. This latter may be regarded as the legitimate
amateur; a serious-minded individual with perhaps some
technical propensities. My primary aim has been to include
a maximum amount of practical information between its
covers; secondarily to furnish an elementary theoretical web
for this information, and to indicate in detail consistent with
its restricted scope, the reasons for the suggestions and
recommendations that have been made. No discussions
of the "wonderfulness" of radio will be found, these being
regarded as quite valueless and already furnished in super-
sufficient quantity (considering the intelligence and imagina-
PREFACE
7
tion of the average individual), in radio books more obvi-
ously designed for exploitation on a favorable market. I
have preferred to deal with hard facts rather than with such
soft fancies.
In a book of restricted theoretical scope it is very hard
to avoid a certain looseness of expression, and a consid-
erable sacrifice of accuracy is sometimes necessary in order
to leave the real issue free from obscurity. To take an
example, on page 161 will be found a statement that the
temperature of the audion filament is proportional to the
rate at which heat is developed by the electric current, and
I can picture in my mind a dozen academic purists and super-
sophisticated electricians rising and scornfully tossing the
book into a corner, molested by thoughts of the changing
specific heats at high temperatures, the Stefan-Boltzman
radiation law, heat of vaporization of the electrons, and a
myriad other technical objections to this simple statement,
the modification of which would not only be the rankest and
most offensive kind of pedantry, but would also obscure the
real point to be emphasized. I have also avoided mathe-
matical formulae, not because I did not feel that many of
the arguments could have been considerably illumined by
their use, but because I preferred to use the formulae myself
and state the results concisely in wire sizes, number of turns,
and other terms of greater meaning to the average amateur.
Most of the designs and specifications for construction rest
upon such calculations, and in most cases I have at one time
or other verified their correctness by experimental test. I
have also had no hesitancy in making use of the results and
experiences of others, as the references in the footnotes and
the acknowledgements at the end of this foreword will show.
8 PREFACE
The first two chapters undertake an exposition of the
fundamental principles of radio communication, particu-
larly as applied in the present-day carrier-wave system of
radio telephony. While I have had the beginner in mind
and have endeavored to present this in a clear and under-
standable manner, yet lack of space prevents a detailed treat-
ment of the elementary electrical theory a knowledge of
which is desirable for the best comprehension of the more
difficult theory of radio communication and the complicated
electrical processes in the radio circuits. The reader who
finds his interest in these matters stimulated is urged to
consult works devoted to them, of which the best in the
language are that admirable little book of Prof. J. J. Thom-
son's, The Elements of the Mathematical Theory of Electricity
and Magnetism (Cambridge University Press) ; and another
dealing more^particularly with radio theory and no less
carefully prepared and written by a staff of experts of the
Bureau of Standards, Principles Underlying Radio Com-
munication, copies of which latter may be obtained from
the Supt. of Documents, Government Printing Office,
Washington, D. C., for a dollar.
In the third and succeeding chapters all mistaken at-
tempts at a popular treatment have been frankly abandoned.
This I felt was necessary in order to get something said and
to avoid the repetition of much subject matter adequately
covered in other books. But the practical man will have no
trouble, I think, in following the discussion, since the
statements are explicit and to the point. I have tried to
avoid the expression of opinions which are not defensible
by rigid mathematical and experimental tests, and on con-
troversial matters, feeling that in view of the advanced stage
PREFACE 9
to which the experimental technic has advanced and the
rational theory which is now available respecting many of
the radio-electrical processes (except radiation and propa-
gation over the earth's surface) there is no excuse nor room
in radio for differences of opinion. A number of pernicious
and prevalent fallacies I have tried to expose and condemn
and I hope the reader will overlook the impassioned way in
which in many cases I have gone about it. Strong and
forceful language is necessary when one is not gifted with
the power of lucid expression.
While the preparation of this book may be said to date
back to 1908 the actual writing was undertaken on short
notice and I have not had the opportunity to look into the
literature of amateur radio to the extent that I should de-
sire, or to give to them the thought that many of the matters
deserve. But I have had the advantage of advice, assistance
and many helpful suggestions from numerous friends, to
whom it remains to make the following acknowledgments:
To conversations (1920) with Dr. J. M. Miller, of the
Bureau of Standards, on the subject of dielectric losses in
antennae; to the Harvard lectures of Prof. G. W. Pierce on
the radiation from flat-top antennae, with the assistance of
which I was able to demonstrate the advisability of operat-
ing the transmitting antenna at its fundamental wavelength
in this case; to Prof. L. A. Hazel tine, of Stevens' Institute
of Technology, for information on designing power audion
oscillators; to Mr. L. M. Hull, Whiting Fellow in Harvard
University, for the use of material from his unpublished
doctorate thesis, dealing with the calculation of the power
output of the separately-excited audion oscillator; to Mr.
E. S. Purington, formerly Associate Physicist in the Bureau
io PREFACE
of Standards, who kindly supplied me with numerous
oscillograms and data from his study of radio telephone
modulation; to Mr. W. D. Loughlin, engineer of the U. S.
Navy, for assistance in securing the electrical characteristics
of certain commercial types of audions; to Mr. L. R. Damon,
of the Federal Institute of Radio Telegraphy, and to Mr.
W. G. Ellis, engineer of the U. S. Navy, for many practical
suggestions; to Mr. K. B. Warner, Secretary of the Ameri-
can Radio Relay League for information concerning the
League and the use of material and illustrations from its
official publication, QST', to Mr. W. S. Fogg for most of
the illustrations; and to the publishers for their encourage-
ment and constant consideration shown in matters connected
with the printing.
In the first impression freedom from error is hardly to be
hoped for and I shall be grateful for the report of any which
the reader may discover; other minor errors of typography
and circuital infelicities he may magnanimously overlook or
silently emend.
STUART BALLANTINE.
PHILADELPHIA, PA.,
May 17, 1922.
CHAPTER I
PRINCIPLES OF RADIO TELEPHONY
1. Radio Telephony and Radio Telegraphy. Radio tele-
phony may be designated as the art of transmitting spoken
words through space without the means of connecting wires.
Just as its sister art, the now very familiar wire telephony,
constituted a great improvement over the cruder, slower,
and more laborious wire telegraphy and supplanted it to a
large extent, so also is this newest of communicational
means destined soon to share the rapidly increasing burden
of radio telegraphy. From a technical point of view the
underlying principles of radio telegraphy and radio tele-
phony are the same, the chief difference in detail being that
in telegraphy the radiated energy is inelegantly chopped up
into lumps to form the dots and dashes of the telegraph
code, whereas in telephony it must be moulded to corre-
spond to the modulations of the sounds to be transmitted.
Thus radio telephony, while obeying the same funda-
mental principles as radio telegraphy, is yet a more com-
plicated mechanism, as might be expected since the finest
modulations of the voice are to be transmitted instead of
the characterless dots and dashes of the telegraph. It was
just this refinement of modulation which radio telephony
demands, this necessity, so to speak, of working in a plastic
clay rather than in mosaic, that prevented it for many years
from assuming its proper place with radio telegraphy in
12 RADIO TELEPHONY
communication. Commenting upon the difficulty of the
problem and the relative ease of the two methods of sig-
nalling, Goldsmith remarks (Radio Telephony, New York,
1918) with singular justice: "The difference in degree is
not far from that between ruling a dot and dash line and
making a dry-point etching of an autumn landscape."
The problem is further complicated by another element,
the size of the job. Relatively enormous quantities of elec-
trical energy are required for radio transmission as com-
pared with wire communication because the radio waves
are not guided as in the latter case by wires, but spread out
over the earth's surface in all directions; consequently but
a small fraction of the original radiation reaches the receiv-
ing station. This greatly increased energy of the trans-
mitter must be controlled by the relatively feeble voice
waves generated when we speak into the transmitter. It
is very much like prodding an elephant with a toothpick.
It was, therefore, not until DeForest invented that remark-
able amplifying device known as the "audion," and the
application of it to magnifying the feeble voice effects, that
the art of radio telephony came into its own.
Most of the development of radio telephony has taken
place since about 1914, and received great stimulus from
the War, chiefly on account of its great military importance
as a means of rapid communication between aircraft and
their bases and between ships at sea. A great portion of
this work was done in this country by DeForest, and in the
research laboratories of the Navy and Army, and of the
large electrical organizations such as the Western Electric
Company and the General Electric Company. As in the
case of the radio telegraph, no man invented it in spite
PRINCIPLES OF RADIO TELEPHONY 13
of a popular view and the credit for the beautiful mechan-
ism as we know it today must be very widely distributed.
2. What is Electricity? Perhaps the most natural and
therefore, the first question most frequently asked by the
layman is this: What is electricity? A harder one can hardly
be imagined. So far as I am aware the question is un-
answerable, since electricity is a something which enjoys a
unique existence and cannot therefore be described in terras
of things more familiar to the inquisitor. We might reply,
enigmatically, that "Electricity is electricity," or that it is
a perfectly good word used in discussing something which
we can neither see, hear, smell, taste or feel, and whose
presence and actions are evident to us only through other
physical effects, such as heat, light, mechanical forces, etc.
But in spite of the enormous handicap of this indirect method
of dealing, our knowledge of it is by no means scant, and
indeed, there has been amassed in the past thirty years an
astonishing amount of information concerning it and its
ultimate relation to material things. From the study of
this we now believe that matter, instead of being itself
fundamental, as was formerly supposed, is. entirely made up
of electricity. This view is embodied in, and forms the
substance of, what is called the electron theory of matter.
3. Fundamental Ideas and Experimental Facts of Elec-
tricity. In this section it is proposed to review very briefly
only so much of the material included by the above caption
as is considered essential to a simple explanation of the
principles of radio telephony.
(a) The Electric Current. When electricity moves or flows
from one place to another there is said to be an electric
current. Of such currents we shall be interested in two
i 4 RADIO TELEPHONY
types: those which flow in conducting substances, called
conduction currents; and those which flow through free
space like a flock of birds, and are called convection currents.
The first kind of current is familiar to all of us; the second
class we meet with in the "audion." The conduction cur-
rent alone will be considered here, and will be referred to as
the current unless otherwise stated. Such a current flows
through a metallic wire connected to the terminals of an
electric battery, as shown in Fig. 1. In this simple case the
terminal voltage of the battery (meas-
ured by the voltmeter, V) is the driv-
ing force, and the quantity of current
which flows (measured by the ammeter,
A) will be determined by the conduct-
ing qualities of the wire circuit. Sub-
stances differ in their ability to con-
duct the electric current. (Some of
them are, in fact, regarded as insula-
Fig. L-Eiectric current tors > \ non-conductors, because they
in simple circuit. permit its passage in such small quan-
tities.) This conducting property of
an electrical circuit, or device, when steady currents are in-
volved, is called its conductance, or reciprocally, its resistance,
and an important relation involving it exists between the cur-
rent which flows and the impressed electromotive force (e.m.f.).
(b) Ohm's Law for Steady Currents. This relation is known
as Ohm's law and may be stated as follows: The current
which flows in a circuit in response to a steady impressed
e.m.f., , is equal to this e.m.f. divided by the resistance of
the circuit; or algebraically stated, / = E/R (where / is
the current, and R the resistance). If in the above law the
PRINCIPLES OF RADIO TELEPHONY 15
e.m.f. is given in wits and the resistance in ohms, then the
current will be expressed in amperes.
(c) Resistances Connected in Series and in Parallel. In the
case of a more complicated form of circuit in which there
are different kinds of conductors connected in series, or
where there are two or more paths for the current, the
value of resistance to be inserted in this formula is deter-
mined by the following rules:
A/VW - WW^ R = Ri 4- Rz; (resistances in series),
R\
/-AMAA-x & R-2
< > R = -- ; (resistances in parallel).
With the aid of these relations the current flowing in any
circuit or branch of any circuit which obeys Ohm's law, can
be computed.
(d) Effects of the Current. We are able to detect the pass-
age of current through the wire in Fig. 1 by two principal
effects :
1. Heating effect.
2. Magnetic effect.
The heating effect is caused by the friction of the electricity
in flowing through the wire, by means of which some of the
electrical energy is converted into heat and therefore from
the electrical point of view, lost. This is called the Joulean
effect, or the Joulean loss. The rate at which heat is gener-
ated in this way is equal to the resistance multiplied by the
square of the current, or algebraically: Heat = PR. This
effect is a familiar one on our every-day life, being utilized
in the electric incandescant lamp, arc lamp, electric heaters,
and so forth.
i6
RADIO TELEPHONY
The magnetic effect may be described as follows: Fig. 2
shows a solenoid carrying a steady current due to the bat-
tery, E. If a magnetic compass, C, is placed at various
positions in the neighborhood of the solenoid its needle will
be found to be affected and to take up approximately the
positions shown by the dotted lines of this figure. The force
which deflects the needle is a magnetic force, and by this
exploration with the compass we have mapped out the
magnetic field of the solenoid. The magnetic force (desig-
nated by the letter, H) has the direction of the dotted lines
Fig. 2. Magnetic
field of solenoid carrying
steady current.
Fig. 3. Showing re-
lation between directions
of current and magnetic
force.
Fig. 4. Illustrating
right-hand rule.
in the diagram. The lines of magnetic force are always
closed when there are no magnetic substances, such as iron,
in the field. Their direction is simply related to that of the
current by Ampere's Rule, which is illustrated by Fig. 3.
Here the current flows in the straight conductor and gives
rise to the circular lines of magnetic force, shown dotted.
The arrows indicate the relations; or they may be kept in
mind by the device of supposing that if on the right hand
the thumb points in the direction of the current, then the
fingers will give the direction of the magnetic force, and
vice versa (see Fig. 4). . This remarkable relation be-
PRINCIPLES OF RADIO TELEPHONY 17
tween electricity and magnetism is of great importance in
radio communication, as in many other electrical applica-
tions.
(e) Electromagnetic Induction. Faraday's Law of Induction.
We come now to the discussion of an effect which is the
converse of the above. Instead of inquiring as in this case
into the creation of a magnetic field, let us assume that one
is already created, either by means of an electric current
or by a steel magnet. In this field place a solenoid whose
circuit is closed through the current indicating instru-
ment, A, Fig. 5. Now so long as the field is undisturbed, or
constant, the ammeter will register no current; but as soon as
the number of lines of force embraced by the solenoid is
changed, either by moving the source of the field or by mov-
ing the solenoid itself through the field, a current will be indi-
cated. Indeed, this we would naturally expect, for if a cur-
rent is capable of producing a magnetic field (which represents
energy), so conversely should the change or disappearance
of this field be capable of delivering back to the circuit, or
any other circuit, some of this energy. If the circuit is
closed, a current flows; if the circuit is not closed there is
still generated an e.m.f. capable of driving a current when
a path is provided. This phenomenon is known as electro-
magnetic induction, and also has important applications hi
the electrical world, in dynamos, etc., and in radio commu-
nication.
It was first extensively studied experimentally by Fara-
day, who disclosed the following fundamental law regarding
it: The e.m.f. induced in a closed electrical circuit by a vary-
ing magnetic field is equal to the rate at which the total flux
of magnetic induction linked with it positively is decreasing
i8 RADIO TELEPHONY
with respect to time. This is known as the first law of induc-
tion.
(f) Magnetic Coupling Between Two Circuits. Mutual Induction.
Instead of producing the variation of the magnetic field
by moving the circuits as in the last paragraph, let this be
accomplished by varying the current in the first (1) circuit
(Fig. 5). An e.m.f. will be induced in the second (2) cir-
cuit as before, and since the circuits are fixed in position
its value will depend upon the time rate of decrease of the
current in circuit (1), and the geometry of the circuits; thus
e 2 = M X rate of decrease of current in (1). M is a con-
stant, called the coefficient of
mutual induction and depends
upon the linkage of the flux
from (1) through (2), or upon
the "coupling" of the circuits.
Two circuits related in this
Fig. 5. Illustrating electromag- mcmT , pr ~ rp ^J trk UP wnowptir
. . , .. mcLimer die sdiQ to ue /riu>}'riei'i/L-
netic induction.
ally coupled.
(g) Self-induction. It has been explained that the varia-
tion of the magnetic flux through a circuit induces therein
an e.m.f. Such a variation of magnetic flux will be pro-
duced when, after connecting the coil in Fig. 2 to the bat-
tery, the current begins to rise and tends to establish the
steady value given by Ohm's law. The circuit reacts upon
itself, inducing an e.m.f., called the back e.m.f., opposed to
that of the battery and acting to restrain the swift rise of
the current which would otherwise take place. This phe-
nomenon is known as self-induction and is very similar to
the inertia of mechanics which is manifested when a massive
body is acted upon by a suddenly applied force (Fig. 6).
PRINCIPLES OF RADIO TELEPHONY 19
In this case the motion of the body lags behind the appli-
cation of the force just as in the electrical circuit the cur-
rent lags behind the application or change of the electric
force.
This property of an electric circuit, by virtue of which it
opposes changes in the current, is termed its coefficient of
self-induction, or simply its inductance, and is denoted by
the letter L. It is denned in terms of the back e.m.f. in-
duced as follows: e = L X rate of decrease of current, and
depends only upon the geometry of the circuit, that is in
the case of the solenoid, upon its dimensions, number of
x^/jvfey
Fig. 6. Mechanical analogy of self-induction in an electric circuit.
turns of wire, etc. The practical unit of inductance is the
henry, defined as the inductance required to produce a back
e.m.f. of one volt when the current is changing at the rate
of one ampere per second. For radio purposes where small
coils are used, subdivisions of this unit are convenient and
customarily employed as follows:
1 milli-henry (m.h.) = .001 h.
1 micro-henry (p.h.) = .000001 h.
1 centimeter (cm.) = .000000001 h.
Inductance has the dimensions of a length and might also
be expressed in feet or miles; but this practice has fortu-
nately not yet been established.
20 RADIO TELEPHONY
The energy stored in a magnetic field associated with a
circuit carrying a steady current /, is equal to \LP.
Inductance coils (inductors) connected in series and paral-
lel in such a way that there is no coupling between them,
are computed by rules given in Art. 2 (c) for resistances.
(h) Forms of Inductors for Radio Telephone Circuits. Induct-
ance is one of the most important properties of a radio cir-
cuit, as we shall presently see, and many forms of "induct-
ances" (more properly inductors) or "load-coils" are em-
ployed. These may be divided into two classes: those of
fixed value, and those whose inductance can be varied.
Fig. 7. Some types of fixed inductors for radio receiving circuits.
Generally these inductors take the form of coils wound in
various ways with a conductor having a low resistance for
the radio frequency (r.f.) currents. A number of such coils
of fixed values designed for use in radio frequency circuits
are shown in Fig. 7. The inductance for a coil of given di-
mensions may be increased by inserting an iron core; this is
often done for low frequencies or direct current (d.c.), but
is not advantageous for the very high frequency currents
used at amateur wavelengths on account of large heat
losses in the iron core. Typical forms of variable inductors
are shown in Fig. 8. This depicts a very convenient and
PRINCIPLES OF RADIO TELEPHONY 21
popular type suitable for use in receiving circuits where low
currents are encountered. Technically this inductor is
called a "variometer" and the principle of its adjustment is
as follows:
Fig. 8. Types of variable inductors for radio receiving circuits.
Figure 9 shows at (a) two coils LI and L 2 wound in the same
direction so that their fields are additive and the mutual
inductance between them is positive. The effective induct-
L = Li + U + ZM. . . . (/)
-2M.... (J)
Fig. 9. Illustrating principle of the variometer.
ance of the arrangement is larger than the sum of their
separate inductances by twice the mutual inductance.
[Equation (1).] If they are placed so that there is no
coupling between them, at right angles for instance, as
22 RADIO TELEPHONY
shown at (&), the effective inductance will be simply Li + L 2 .
But if one of the coils is reversed, so that their mutual in-
ductance is negative, the inductance will be less than the
sum Li + 1.2 by twice the mutual inductance. [Equation
(3).] Thus by simply varying the coupling between two
coils connected in series a variation of the total inductance
may be produced. This is the principle of the variometer
and in the types illustrated above the coupling is varied
from full positive through zero to full negative by rotating
one of the coils through 180 degrees. This gives a smooth
variation of the inductance difficult to obtain in any other
way and for this reason the
instrument is a very valuable
one in radio circuits requiring
21 close adjustment. They are
rather well developed and
-'40Z'' X 'T.'\ '\^^ ^ ^C many different types are avail-
' x ^/^///!\\\\ a ble on the market at very
Fig. io.-Eiearic 'field of a reasonable prices.
charged conductor. (i) The Electric Field. We
have considered the magnetic
field, particularly as produced by an electric current; let us
now discuss the electric field. Figure 10 shows a conductor
upon which an electric charge has been placed. In the
magnetic case the field was explored by means of a magnetic
compass needle; here we shall employ a small particle carry-
ing a unit charge of positive (+) electricity. As this ex-
ploring charge is placed in the different positions it will
be urged in various directions with various intensities. The
mechanical force thus observed is a manifestation of the
electric force and serves as a means of measuring and defin-
PRINCIPLES OF RADIO TELEPHONY 23
ing it. If the direction in which the particle is urged is
plotted for its various positions, the lines of electric force
shown in Fig. 10 will be obtained, analogous to the lines of
magnetic force of Fig. 2. But unlike the latter these lines
are not closed (in the steady field), but terminate upon elec-
tric charges.
(j) Electrostatic Capacity. Condensers. The ability of a con-
ductor to become charged as above is termed its electrostatic
capacity, or simply its capacity. This ability depends among
other things upon the size of the conductor and upon the
proximity of other conductors in the field. It may be in-
/ I ^=^r
Fig. 11. Parallel plate electric condenser.
creased for a conductor of given dimensions by means of
an arrangement shown in Fig. 11. The parallel plates A
and B are preferably of large area and are separated by a
small distance. The field and the electrostatic energy it
represents is thus largely concentrated in the space between
the plates, and on account of this the device is called a
condenser. Its capacity is directly proportional to the area
of the plates and inversely proportional to their separation.
When the plates are connected to the battery, E, a current
flows through the circuit in the direction of the full line
arrows, conveying charges which accumulate upon the op-
posing surfaces of the plates. The condenser is then said
24 RADIO TELEPHONY
to be charged. The electrostatic energy stored in the field is
JCE 2 , where E is the voltage between the plates. If the
battery is now removed and replaced, let us say, by a re-
sistance, the condenser will discharge, that is, a current will
flow (dotted arrows) until the charges on A and B are
equalized.
Fig. 12. Mechanical analogy of the charged condenser.
Thus in the charged condenser there is a state of tension
which has its mechanical analogue in the state of a stretched
spring (Fig. 12), and the electrical energy corresponds to
the potential energy of the stretched spring. Each will do
work if the opportunity is provided.
Condensers connected in series and in parallel are com-
puted by the following rules:
C| c * Ci G
1 1 1 1 ; C = c * V., (capacities in series),
/HK
/ \ ; C = Ci + C 2 , (capacities in parallel).
HK
<a
It will be noticed that the effect of these connections is
opposite to that in the case of inductance and resistance;
connecting condensers in parallel, for example, increases the
total capacity, whereas in the case of resistance and in-
ductance the composite value is diminished.
The practical unit of capacity is the farad, defined as the
PRINCIPLES OF RADIO TELEPHONY 25
capacity required to pass a current of one ampere when
the voltage across it is changing at the rate of one volt per
second. (Cf. definition of unit of inductance, Art. 2 (g).)
This unit is, however, much too large for practical use (the
capacity of a sphere the size of the earth is only .005 farad),
and subdivisions are convenient and generally employed in
practice as follows:
1 micro-farad (mfd.) = .000001 f.
1 micro-micro-farad (mmfd.) = .OOOpOOOOOOOl f.
1 centimeter (cm.) = .00000000000111 f.
(k) Specific Inductive Capacity. Dielectrics. It is found ex-
perimentally that the capacity of the parallel plate con-
denser is increased by inserting a piece of glass, mica, par-
affined paper, etc., between the plates. This discovery is
often utilized in the construction of condensers and permits
a great saving in space and materials in building a con-
denser of given capacity. The voltage at which the con-
denser breaks down, that is, at which disruptive discharge
takes place between the plates, is also often considerably
increased by the use of these materials. The ratio of the
capacity with and without the substance between the plates
is called the specific inductive capacity, or simply dielectric
constant of the material; and is a positive number greater
than unity for all materials at radio and lower frequencies.
Non-conducting substances which possess this property are
called dielectrics. The dielectric constant of air at ordinary
pressures and temperatures is not appreciably greater than
unity.
(l) Forms of Condensers for Radio Circuits. With inductance,
the capacity of or in a radio circuit is a property of funda-
26 RADIO TELEPHONY
mental importance. Condensers for radio work take a
variety of forms, with both fixed and variable capacity
values. For transmitting purposes where high voltages and
large amounts of power are involved, the condenser must be
designed to withstand the voltages and to dissipate the heat
lost in the dielectric. For such condensers glass and mica
dielectrics are generally used; a well-chosen grade of mica
makes a particularly good condenser because its dielectric
constant is moderately high, and its great dielectric strength
Fig. 13. Types of condensers for radio circuits: (a) Mica condenser for trans-
mitting circuits; (b) and (c) variable condensers for receiving circuits.
permits thin sheets to be used thus yielding a condenser
of large capacity and small dimensions. For receiving
purposes variable condensers are extensively employed, of
which there are quite a variety of types on the market.
Two of these are illustrated in Fig. 13. The dielectric is
generally air because there is no necessity for insulating
high voltages, as in transmitting circuits, and an air
dielectric has very little loss. The capacity is varied
by rotating the plates of one set between those of the other
set.
PRINCIPLES OF RADIO TELEPHONY 27
4. Alternating Currents. Leaving now the province of
direct or steady currents and voltages, let us take up the
discussion of currents which vary; this is, of course, the
class of greatest interest in radio communication as in most
other applications of electricity. The simplest and most
frequently appearing type of variable current is that known
as the sinusoidal or simple harmonic (s.h.) type. This kind
of variation is illustrated in Fig. 14. The horizontal dis-
tances represent the time and the vertical distances repre-
sent the value of the current or voltage at that instant of
time. The term "sinusoidal" follows from the fact that the
current value at any instant is a trigonometric sine function
of the time. As shown, the current undergoes a complete
-TifiC-*- \
Fig. 14. Illustrating sinusoidal, or simple, harmonic type of alternating current.
reversal from A to J5, and a complete cycle from A to C.
The number (n) of such cycles passed through per second is
called the frequency; and the time, T = 1/w, required for
the completion of one cycle is called the periodic time or the
period. Alternating currents of this type are used ex-
tensively in many every-day electrical applications; for
power and lighting purposes, where the frequency is rela-
tively low (60 cycles per second); for wire telephony, in
which the frequencies may vary from 100 to 5000 cycles;
and for radio communication where they are much higher,
from 30,000 to 2,000,000 cycles. We shall consider first
the effect of impressing an a.c. voltage of this type upon sim-
ple circuits containing inductance, capacity and resistance.
28
RADIO TELEPHONY
(a) Circuit Containing Resistance. Figure 15 (a) shows an a.C.
generator connected to a circuit containing resistance. In
this case the current flowing at any instant will be given
by Ohm's law, that is, it will be proportional to the volt-
(b) (c) (d)
a
Fig. 15. Simple a.c. circuits containing inductance, capacity and resistance.
age, inversely proportional to the fixed resistance. The
current variation may be shown graphically, as at (a)
Fig. 16, where the full line represents the impressed voltage
\ \ \
~
(Resistance)
(Inductance and resist-
ance]
^Ap (Capacity and resistance)
(d) X*~*\ X^\ /^~*\ /^\ (Resonance with induct
-r- V^ -f- V TT >. f V ance, capacity and re-
Vx \-/ x^/ sistance)
Fig. 16. Currents and voltages in a.c. circuits of Fig. 15 containing inductance,
capacity and resistance.
and the dotted line the resulting current. The current keeps
time with the voltage, reaching its zero and maximum
values simultaneously with it, and on that account is said
to be "in-phase"
PRINCIPLES OF RADIO TELEPHONY 29
(b) Circuit Containing Inductance and Resistance. We noticed
in a previous paragraph that self-induction had a tendency
to suppress variation of the current. This effect is impor-
tant here, and an a.c. circuit containing inductance is found
to offer a greater resistance to the passage of current than
can be accounted for by Ohm's law and the resistance of
the circuit alone. This impedance depends, moreover,
upon the inductance and frequency. The current is no
longer given by Ohm's law, but must be calculated from the
new law: / = V^ 2 + " 2 L 2 . The factor V# 2 + " 2 L 2 is
called the impedance of the circuit; G> = In X frequency is
called the angular velocity corresponding to the frequency n.
Note that the choking effect of the inductance increases
with the frequency. In addition to the limiting effect the
inductance also causes the current and voltage to be out of
step and the current "lags" behind the e.m.f. as shown at
(b), Fig. 16. The current is still sinusoidal in form.
(c) Circuit Containing Capacity and Resistance. A condenser
prevents the flow of a steady current through a cir-
cuit, but permits the passage of a.c. in proportion to its
capacity and the frequency of the a.c. In series with a
resistance, as at (c), Fig. 15, the impedance offered to the
a.c. voltage is \/R 2 + l/o 2 C 2 . In this case, however, the
current instead of lagging behind the e.m.f. as in the case
of inductance, actually precedes it, or "leads," as shown at
(c), Fig. 16. The amount of this lead depends upon the
product of resistance and capacity.
(d) Circuit Containing Inductance, Capacity and Resistance.
Resonance. Tuning. We come now to the most interesting
and important case, that of a circuit with all three con-
stants Lj C and R. Comparing cases (b) and (c) it will be
30 RADIO TELEPHONY
observed that inductance and capacity have opposite effects,
the first causing the current to lag, and the latter to lead.
Thus when combined the impedance of the circuit is
+ (uL - l/uC) 2 . It is evident that if L and C are
properly chosen an exact neutralization of their effects maybe
had. To bring this about the simple relation, oZ,= 1/coC, must
be satisfied, and when this adjustment is made the current
is a maximum and is given simply by / = E/R, as in the
case of a simple resistance; the circuit then acts in fact as
if resistance alone were present and the current is in phase
with the voltage as shown at (//), Fig. 16. (Compare with
(a), Fig. 16.) This condition is known as resonance and the
adjustment of the inductance and capacity to bring it
about is commonly called tuning. The operation of tuning
is very important in radio work, its purpose being gen-
erally in this case to obtain a maximum flow of current in a
circuit, as for example, in the case of a receiving antenna.
Obviously instead of tuning the circuit we can also produce
resonance for a given L and C by adjusting the frequency.
5. Electric Waves. When the smooth surface of a pond
is disturbed by throwing a stone into it, ripples are formed
which spread out along the surface in a series of miniature
circular water waves. This is one of many familiar examples
of wave motion. The waves so formed will travel with a
definite velocity of a few meters per second, and carry
with them energy capable of setting into motion a small
distant floating object upon which they may be inci-
dent. Replace the stone by the radio transmitting sta-
tion, the floating object by the radio receiver, the water
by the "ether," and conceive of ether waves or electric waves
instead of the waves in the water, and an idea of the funda-
PRINCIPLES OF RADIO TELEPHONY 31
mental scheme of radio communication will be secured. In
the radio case, however, the waves travel with the rela-
tively enormous velocity of approximately 186,000 miles per
second.
The exact structure and nature of the electric waves is
somewhat harder to describe and to understand. In the
case of water waves we have to deal with the propagation
of a vertical displacement of the water; in that of electric
waves with the propagation of electric and magnetic forces,
a physical picture of which is not easily drawn. The
medium through which these forces are propagated is
called the ether and is not ponderable, and cannot be seen
or sensed in any way, like the water medium. The ether is
a very subtle thing and appears to defy physical descrip-
tion; there are, in fact, a number of reputable scientists who
refuse to have anything to do with the idea, preferring to
regard it as being sufficiently described by the mathematical
equations. So we shall not attempt the impossible here,
avoiding this pitfall of so many works on the subject of
electric waves; but will be content to remark that the
electric waves travel through something and that some-
thing is called the ether. The electric and magnetic forces
are states of it, and it occupies all space, permeating even
the interior of solid substances. Hence electric waves have
little difficulty in passing through insulating substances like
wood, but do not, of course, pass thus unrestrained through
conductors. It is understood that we are speaking of long
waves, as in radio communication.
Electric waves play a very important part in our every-
day life; heat, light and the X-ray are all propagated by
their means. The velocity in all cases through free space
32 RADIO TELEPHONY
is the same, viz., 186,000 miles (300,000,000 meters) per
second, and because it was first measured in connection
with light rays, is popularly referred to as the light-velocity.
The various radiations differ, however, in their wavelengths.
In the water waves the wavelength is easily measured as the
distance from the crest of one wave to that of its successor;
in the electric case it is the distance from one maximum of
electric or magnetic force to the next maximum. There is
a simple relation between the wavelength and the frequency
of vibration of the source, or with which the floating object
(in the water case) or a receiving antenna (in the radio
case) vibrates when the wave acts upon it. This may be
explained with the aid of Fig. 17 as follows:
A B
Fig. 17. Illustrating relation between wavelength and frequency in a sinusoidal
Here AB represents the constant distance the wave travels
in one second, regardless of its wavelength, frequency, etc.
Now if the waves are generated at the rate of n per second
it 'is evident that precisely n of them will be between A and
B at any given time. The length of one wave (usually de-
noted by the Greek letter, /I) will therefore be AB -*- n, or
velocity /n. It should be noticed that the wave length is a
characteristic of the wave and has nothing whatever to do
with the distance it travels.
The chart of Fig. 18 gives a graphical comparison of the
waves of X-ray, light, heat and radio. The enormous dif-
PRINCIPLES OF RADIO TELEPHONY
33
ferences in wavelength, i. <?., from .000,000,000,01 meter to
10,000 meters, cannot be represented to scale, but are
marked in the diagram. These differences in wavelength
account for the differences in their phenomena; funda-
mentally they are all electric waves.
0.1 x to'jpn r3oooxio~'m.t-zooo>eio~ lo m.r e>x io' 3 m. r 100 m.
laoox io"Th-) j-eooox f 1 * io'4-mj- loom. [ 10,000 rn.
xNNv
RAOI
\\Vv\\:
WAVE LENGTH
Fig. 18. Chart comparing wavelenghts of X-radiation, light, heat, Hertzian and
radio waves.
6. Production, Transmission and Reception of Radio
Waves. The special electric waves of radio communica-
tion, called radio waves, will now be described, and their
radiation from the transmitting antenna, their propaga-
tion over the earth's surface, and finally their effect upon
the distant receiving antenna will be briefly explained.
(a) Production. As before let us employ a mechanical
analogy; the water waves of the last section will do. Im-
agine then, a cork floating upon the water to which, by
hand or otherwise, an up-and-
down motion is imparted.
This will produce a series of
waves, and if the motion is a
regular one, a series of waves
of uniform wavelength. In
the radio case the transmit-
ting antenna replaces the cork
and the current therein, flowing up and down, corresponds
to the latter's motion. Figure 19 shows a simple type of
antenna connected to a generator of a.c. of radio frequency.
3
Fig. 19. Illustrating radiation of
radio wave from simple vertical
antenna.
34 RADIO TELEPHONY
The antenna has both inductance and capacity, hence
according to the paragraph on "resonance," either the fre-
quency of the a.c. voltage is to be adjusted for resonance or
the LC constants adjusted by means of additional coils and
condensers. When the adjustment has been made a maxi-
mum a.c. current will flow in the antenna. This current will
produce a magnetic field (shown by the dotted lines) and an
electric field (full lines) which travel out from the antenna
with the speed of light and constitute the emitted radio
Fig. 20. Airplane view of electric wave radiated by simple vertical transmitting
antenna, showing direction of electric and magnetic forces.
wave. Figure 20 shows the wave just after emission as it
might appear to an airplane observer able to see the elec-
tric (E) and magnetic (H) forces. Figure 21 is a cross-section
through Fig. 20 and shows in detail the electric lines for the
first few waves. These, it will be observed, are vertical
near the earth, while the magnetic lines (shown in section
by the dots and crosses; dots pointing toward the reader,
crosses away from him) are horizontal and circular with the
PRINCIPLES OF RADIO TELEPHONY
35
axis of the antenna as their center. The electric and mag-
netic fields just described are inseparable in such a wave
and proceed, so to speak, hand in hand.
The wavelength, /I, is indicated as the distance between
successive crests and is related to the frequency of the
antenna current by /I = v/n, as already explained. Theo-
retically, if the vertical antenna contains no extra induct-
ance or capacity in the form of a load in series with it, and
Fig. 21. Cross-section through Fig. 20 showing lines of electric force () in
Jirst few waves. Magnetic lines (H) shown in section by arrows () toward
reader, (+) away from reader.
the generator frequency is adjusted for resonance, the
wavelength of the emitted waves will be very nearly four
times the length of the wire; for example, a wire 50 meters
high would radiate waves of approximately 200 meters
wavelength. This is termed the fundamental wavelength of
the antenna and is the wavelength at which most efficient
radiation takes place. If it is desired to radiate waves
longer than the fundamental, an inductance is inserted in
RADIO TELEPHONY
series with the antenna as shown at (a), Fig. 22. Waves
shorter than the fundamental are produced by connecting
a condenser in series as shown at (b). This process is
termed loading the antenna.
(b) Transmission. The hemispherical wave whose genesis
has just been described expands at the velocity of light,
retaining its hemispherical shape, and being guided at its
base by the conducting surface of
(I || || the earth. Actually the earth is not
a perfect conductor, consequently
^ the wave, instead of gliding along its
surface as it would in the case of
perfect conductivity, penetrates it to
some extent and induces currents
which produce heat loss. Some of
the original energy of the wave is
lost in this way. Moreover, since
the energy is distributed over the
entire surface of the wave, the
amount contained in a unit surface
2 2
(b)
(c)
Fig. 22. Loaded anten-
na; arranged to produce: (a)
Waves longer than funda-
mental, (b) -waves shorter i a y er continually diminishes as the
than fundamental, (c) waves j , -, > , . j
of fundamental wavelength. wave expands; the electric and mag-
netic forces diminish, in fact, very
nearly inversely as the distance from the source, and the
energy diminishes inversely as the square of this distance as
in the case of other radiations, heat, light, etc. In addition,
there are other sources of energy loss, such as scattering by
upper atmospheric layers, absorption by terranean objects,
trees, etc., which we have no space to consider in detail.
(c) Reception. Consider now the effect of this wave upon
an antenna used for receiving purposes, Fig. 23. In ac-
PRINCIPLES OF RADIO TELEPHONY
37
cordance with the principles of electromagnetic induction
of Art. 2 (e) there will be induced therein an e.m.f. whose
frequency will be that of the wave and consequently that
of the transmitting current. From an electrical point of
view the antenna circuit is equivalent to that shown at (b)
Fig. 23, where L a , C a and R a represent respectively the
effective antenna inductance, capacity and resistance; and
L and C the inserted tuning inductance and capacity.
The e.m.f. due to the passing wave is represented by e.
The current in the antenna
is used to activate the de-
tecting instruments (to
be subsequently explained)
and in order that it shall
be as large as possible,
L and C are adjusted so
that the circuit is in reso-
Fig. 23. Receiving antenna (a), and its
equivalent electric circuit (b).
nance with the induced
e.m.f. When the circuit
is tuned in this manner, the current is a maximum, being
limited only by the resistance.
7. Resume; the Mechanism of Radio Transmission.
The fundamental processes just described may be summed
up as follows: A generator, or source of radio frequency
energy, is connected to the transmitting antenna and pro-
duces therein an alternating current. The oscillation of
the current produces in turn varying electric and mag-
netic states of the ether which travel outward in hemi-
spherical electric waves at the velocity of light. Impinging
upon a distant receiving antenna properly tuned these
waves produce in it an alternating current of the same
38 RADIO TELEPHONY
type and frequency as that in the transmitting antenna.
The presence of these currents is made evident to us by a
detecting means to be presently described. Thus we have
a system whereby radio frequency currents at one place are
capable of producing identical currents (of much dimin-
ished amplitude) at a distant place. The radio telephonic
utilization of this scheme for the transmission of sounds
will be considered.
8. Principles of Telephony Over Wires. It will be help-
ful, before attempting to explain the principles of radio
telephony, to briefly discuss the theory of ordinary tele-
Fig. 24. Illustrating simple wire telephone circuit.
phony over wires, since between the two interesting anal-
ogies may be drawn and there is besides some didactic
advantage in progressing from the familiar to the unfamiliar
and from the simple to the complex.
The scheme of a simple wire telephone is illustrated in
Fig. 24.- Here T represents the transmitter, into which the
words are spoken at the sending end; E is the source of
power, or electrical energy, and usually consists of a battery
of constant voltage; L is the metallic line; and R is the
telephone receiver. The operation of this circuit as a tele-
phone is somewhat as follows :
When words are spoken into the transmitter the dia-
PRINCIPLES OF RADIO TELEPHONY 39
phragm A is set into vibration in accordance with the
speech sounds. The space between the electrodes A and B
is loosely packed with carbon granules which offer to the
passage of current
a resistance which
depend upon the de-
gree of compression
to which they are
subjected. Obvi-
ously by this means
the current is varied
in accordance with
the diaphragm's vi-
brations and the
speech sounds, or
technically we say
that the current is
modulated. The
modulated current
flowing through the
windings W of the
telephone receiver
produces a similar
vibration of its iron
diaphragm D,
"ah"
"u"
("00")
Fig. 25. Photograph of current vari-
ations in wire telephony, produced by
three vowel sounds spoken into the tfftns-
mitter of the circuit, Fig. 24.
re-
producing in this way
the original speech.
In Fig. 25 are shown some interesting photographs of the
current variations which are produced in a circuit of this
type by various vowel sounds spoken into the transmitter.
These irregular vibrations may be regarded as being com-
4 o
RADIO TELEPHONY
pounded of a number of pure sinusoidal currents of various
amplitudes and frequencies. From this point of view the
principle frequency of speech is often said to be 800 cycles
per second; although the entire range from 100 cycles to
5000 cycles is necessary for the transmission of speech of
high quality. Thus in Fig. 25 the modulation in the vowel
sound "e" takes place with a frequency of 2250 cycles, of
which a chief modification occurs at a lower frequency of
about 170 cycles. In the letter "s" the frequency of the
sound is still higher.
These alternating currents are the carriers of the speech.
In flowing through the line they are accompanied by elec-
ill _ui
Fig. 26. Simplest scheme for wireless telephony, using voice currents directly.
trie and magnetic forces in the ether, these states traveling
along with them to form an electric wave gliding along the
wire. Such an electric wave is said to be guided and in the
case of the usual two-wire transmission, the waves are
called double-cored waves. Thus a message conveyed by this
kind of wave goes right to its destination and does not, as
in the case of radio waves, spread out in all directions.
PRINCIPLES OF RADIO TELEPHONY 41
It is natural to suggest at this point that the wire be
dispensed with and a radio telephone system produced by
connecting the transmitter and receiver to antennae as
shown in Fig. 26. This is quite possible, but not feasible
for the following reasons: As already mentioned the tele-
phone current frequencies range from 100 to 5000 cycles.
The wavelength emitted at a frequency of 1000 cycles would
be 300,000 meters. To secure most efficient radiation at
this wavelength it would be necessary to build an antenna
300,000 -v- 4 = 75,000 meters, or about 46.6 miles long!
This would indeed be an ambitious project. Satisfactory
work might be done with antennae from 10 to 20 miles in
length, but the radiation from the average antenna as we
know it would be neglibible at these frequencies. Thus it
is apparent why this most obvious system of radio tele-
phony is impracticable and recourse must be had to the more
complicated methods with which this book is concerned and
which will now be described.
9. Principles of Radio Telephony. We have just pointed
out how difficult radio telephony would be if attempted with
the low frequency currents directly. What we require then
is a system wherein higher or radio frequency currents are
utilized in the wireless transmission; then the energy can
be effectively radiated and received by antennae of prac-
tical size, and the transmission greatly improved. This
radio frequency energy must, of course, be controlled and
modulated in accordance with the speech sounds to be
transmitted; and at the receiving station must be demod-
ulated or translated back into the low frequency currents
with which the telephone receivers are to be activated.
This is the essential principle of the radio telephone of
RADIO TELEPHONY
the present day. The schematic diagram of Fig. 27 will
perhaps make it clearer. The RADIO FREQUENCY POWER
SOURCE produces suitable radio frequency current in the
transmitting antenna. The amplitude of this current is
Ill 111
=*=<l -
Fig. 27. Illustrating the scheme of radio telephony using radio frequency carrier-
waves.
caused to vary by means of the microphone assisted by
the MODULATOR, to correspond to the speech sounds. The
actual current in the antenna or the forces in the radiated
Fig. 28. Type of signal used in radio telephony, showing high frequency "carrier-
wave" modulated by speech current.
wave then appear about as shown in Fig. 28; and consist
of a radio frequency wave, called for obvious reasons the
carrier-wave, whose intensity undergoes variation at the
much lower frequency of speech, viz., 100 to 5000 cycles.
PRINCIPLES OF RADIO TELEPHONY 43
This wave produces in the tuned receiving antenna a cur-
rent of approximately the same form, which is decomposed
by the DETECTOR into the original speech currents of low
frequency, or a more or less faithful copy of them.
Comparing this with wire telephony we have :
Radio Telephony Wire Telephony
Production of r.f. power; Production of d.c. power;
Modulation ; Modulation ;
Radiation, transmission and Transmission by wire;
reception ;
Demodulation;
Reception by telephone re- Reception by telephone re-
ceivers, ceivers.
Radio telephony thus appears to be a more complicated
mechanism, at least in the broad outlines of the processes.
CHAPTER II
THE AUDION
10. The Audion in the Radio Telephone System. The
device to which this chapter is devoted occupies the key-
stone position in the radio telephone system. Its impor-
tance and versatility may be appreciated from the fact that
in the radio telephone scheme. Fig. 27, it plays the parts of:
(1) modulating device, or amplifier for the voice currents;
(2) generator of radio frequency power for transmitting;
(3) amplifier for the received radio frequency signals; (4)
demodulator, or detector; and (5) is often employed for
amplifying the signals after detection, either to secure
greater intensity in the telephone receivers or for the opera-
tion of a "loud-speaker." In these various applications it
functions in two different ways: (1) as an amplifier, and
(2) as a detector. Both of these functions will be explained
in subsequent paragraphs.
11. Description of the Audion. The name "audion" was
originally given to the device by its inventor Dr. Lee De-
Forest. It has since become fashionable to employ other
designations such as, three electrode vacuum tube, or simply
vacuum tube (V.T.); electron tube, thermionic amplifier,
triode, and while very high-sounding and smacking of great
erudition on the part of their users these terms are no more
descriptive of the actual apparatus than the term "audion";
hence recognizing the prior right of the inventor in the
44
THE AUDION
45
naming of his child we shall employ the term audion in this
book.
The audion consists of three electrodes inclosed by a
glass container which has been very carefully exhausted to
a high vacuum. The first of these electrodes is an ordinary
filament similar to those used in electric light bulbs and
arranged to be heated in the same way, viz., by means of
an electric current. The second
electrode, called the plate, is sim-
ply a piece of metal (usually nickel)
in the shape of an elliptic, rectang-
ular or circular cylinder which sur-
rounds the filament. The third
electrode is called the grid and is
interposed between the filament
and plate. Its construction is
such that small particles of elec-
tricity coming from the filament
may pass through its meshes on
their way to the plate. This is
sometimes called the control elec-
trode.
The construction of a typical de-
vice with plane electrodes is shown
in Fig. 29, in which parts of the grid and plate have been
cut away to show the relation between the three electrodes.
Electrical connections are made in the usual way by bring-
ing out leads through the glass mash to the base upon
which four terminal pins are mounted; two for the filament
and one each for the grid and plate. The bases of most
commercial audions have been standardized, and are of the
Fig. 29. View of audion
with plane electrodes in which
the glass container and part
of the electrodes have been re-
moved to show the relation be-
tween grid, plate and filament.
46 RADIO TELEPHONY
"bayonet" type. The connections to the terminal pins are
shown in Fig. 30.
12. How the Audion Works. It is found experimentally
that when a piece of metal is heated to the temperature of
incandescence small particles of negative electricity, called
electrons, are thrown off from it. This evaporation of elec-
trons from the metal is called thermionic emission and is a
phenomenon very similar to the evaporation of water par-
ticles from water which occurs at ordinary temperatures.
The function of the filament in the audion is to send out
such particles, or as often stated, "to
supply electrons." We care very little
about the nature of these electrons, what
they are or how they are emitted; the
important thing to be noted is that they
carry electricity negative electricity and
Fig. 30, Illus- / . J
(rating connections that moving from one place to another
on standard "bay- they constitute an electric (convection)
oner type audion ^^
base.
With this in mind, look at Fig. 31. This
shows a filament which is heated by a current from the battery
"A" and sends out electrons; and an electrode designated as
the "PLATE" which is maintained at a positive potential with
respect to the filament by means of the battery "B." Both
electrodes are to be regarded as being in a vacuum. Upon
emerging from the filament the electrons find themselves in
the strong electric field between the electrodes, and being
negatively charged, move toward the plate under the action
of the electric forces. This kind of attraction has already
been noticed in Art. 2 (e) when the electric field was ex-
plored by means of a small charged particle. Or we may
THE AUDION
47
account for this movement of the negatively charged par-
ticles toward the positively charged plate by simply re-
membering the teaching of our school-days: like charges
repel, unlike charges attract.
The stream of electrons to the plate, as already stated,
constitutes an electric current; and since the charge con-
veyed is negative, the direction of the current is opposite to
that in which the particles are moving. The current flows
ELECTRONS
FILAMENT
PlRCCTIOM
of CURRENT
PLATE
Fig. 31. Current of electrons from filament to plate in a two-electrode tube.
around the plate circuit as shown by the arrows (Fig. 31)
and is registered by the ammeter A.
Now if the plate potential is made negative with respect
to the filament, that is, if we reverse the "B" battery, the
electrons are repelled and very little or no current flows.
The device thus acts as a rectifier permitting the passage of
current in one direction and not in the other, and may be
called a thermionic rectifier. These rectifiers have impor-
tant applications in radio telephone circuits which will be
described in Chapter V.
4 8
RADIO TELEPHONY
Consider now the effect of inserting between the posi-
tively charged plate of Fig. 31 and the filament, a third
electrode in the form of a grid (Fig. 32). The construction
of the grid is such that the stream of electrons is not com-
pletely hindered mechanically; a few of them do, of course,
strike it, especially when its potential is positive. It has
been observed that the electrons are attracted by a posi-
tively charged electrode and repelled by one negatively
GRIP
PLATP
Fig. 32. Illustrating insertion of grid to control the electron current to the plate.
charged. This principle is used to control the stream of
electrons to the plate.
If the potential of the grid with respect to the filament
is made positive by means of the battery "C," the electrons
will be attracted, or rather their attraction will be in-
creased, and an increase in the current to the plate will be
indicated by the ammeter A. Conversely, if the grid poten-
tial is negative, the stream will be repelled and the plate
current will be diminished. Thus the grid acts as a throttle
to regulate the flow of electrons to the plate, and it can be
THE AUDION
49
so designed that its throttling action is very great, that is
to say, very small variations of its potential are sufficient
to produce very marked changes in the plate current. On
account of this the device is called an amplifier, and ampli-
fies the small grid voltage variations.
13. The Audion as an Amplifier. It is sometimes helpful
to represent this action graphically, as is done in Fig. 34.
Here we have a curve, called the characteristic curve, which
depicts the relation between the current in the plate cir-
cuit, 7 P , and the voltage be-
tween the grid and filament,
E g . The plate battery volt-
age and filament heating
current are held constant.
The greatest change in the
plate current for a given
change in the grid voltage
I -5
P- Opcrttinf Point for
Amplification
0,4 Pj- Operating Ibint
for Perec t ion
CRIP VOLTA-
Fig. 34. Typical characteristic curve
of the audion showing relation between
the plate current and the grid voltage, for
constant plate battery voltage and fila-
ment current.
occurs at the point marked
"P"; consequently if the
device is to be used as an
amplifier the grid should be
maintained at the voltage
(= 1.5 volts) corresponding to this point; then when the
small variations are superposed they will have a maximum
effect. A typical circuit in which the audion is employed as
an amplifier is shown in Fig. 35. The a.c. voltage to be ampli-
fied is introduced by coupling through the input transformer,
Ti, and impressed upon the grid. The grid is maintained
at the point "P" in Fig. 34, that is, at the best part of the
curve, by means of the battery "C." The magnified cur-
rent is led out of the system through the output transformer,
4
50 RADIO TELEPHONY
T 2 . Modifications of this scheme, wherein the transform-
ers Ti and T 2 may or may not appear, are employed; some
of these will be described in later chapters.
14. The Audion as an Oscillator. In virtue of the am-
plifying properties just discussed the audion may be used to
generate alternating currents. This application is of great
importance in radio telephony, both for sending and re-
ceiving; and its principle may be explained as follows:
Consider the case of the audion connected in Fig. 35 as
an amplifier of a sinusoidal a.c. voltage impressed upon its
grid. Corresponding to the sinusoidal variation, of grid
Fig. 35. Elementary connections for using the audion as an amplifier.
voltage, the current in the plate circuit will undergo approxi-
mately sinusoidal variations and the amplified a,c. power
which they represent may be made available in a load con-
nected in the plate circuit (Fig. 36). In such a system we
might put in one watt from the generator E and take out
ten in the load. There is no violation of the principle of
conservation of energy we are not "getting something for
nothing" but the extra power is derived from the "B"
battery hi the plate circuit. From this point of view the de-
vice may be looked upon, not as an amplifier, but as a con-
verter of the d.c. energy supplied by the plate battery into a.c.
THE AUDION 51
energy. The generator E is used for excitation in the same
way that the small d.c. generator connected to the field of
a large alternator is used for its excitation. In the case of
the alternator the power is derived from the shaft which
drives it, in the audion case from the "B" battery; in each
the exciter plays a necessary but subordinate role. Any
audion amplifier may be thus regarded, whether used for
power amplification as supposed above, or for the mag-
nification of the feeble currents in the receiving set. The
point of view is more convenient and appropriate, however,
when the generation of power is under consideration.
Fig. 36. The audion in use as an a.c. power amplifier, or separately excited
oscillator.
The power amplifier discussed above is often referred to
as a separately excited oscillator and finds application for the
generation of radio frequency currents for radio telephony.
The details of this will be presented in Chapter IV under
"Master Oscillator Systems."
If in the above system we put in one watt of power and
can take out ten, let us say, it is perfectly obvious that one
watt of this ten may be brought back into the grid circuit
and used for excitation purposes, thus eliminating the
exciting generator, E. The system then becomes self-
exciting and may be used for the generation of a.c. power
RADIO TELEPHONY
directly. The modification of the connections of Fig. 36 for
this is shown in Fig. 37, where the power for the excitation
is brought back and introduced into the grid circuit by mag-
netic coupling. An audion connected in this way is said to
LOAP
CIRCUIT
Fig. 37. Method of bringing back power for the excitation of the grid circuit in
self-excited oscillator.
be retroactively coupled, or back-coupled, and the process of
bringing the energy back into the grid circuit is commonly
called feeding back. Various methods of feed-back are em-
ployed, and the load circuit also takes
a variety of forms, the most important
of which will be described in Chapter
IV. The essential principle in all of
these arrangements is the same and
the above explanation of it may be
made clearer diagrammatically as in
Fig. 38. In this diagram the flow of
power is traced. Coming originally
from the "B" battery, or d.c. source
in the plate circuit, part of it is absorbed by the load (for
instance, by a, radio transmitting antenna), the rest is dis-
sipated in the audion itself. Once started, the energy travels
around the vicious electrical circle shown, and the oscilla-
tions build up until the limit of the tube is reached.
Fig. 38. Diagram rep-
resenting the flow of power
in the oscillating audion.
THE AUDION 53
The load circuit usually contains an inductance and
capacity (as in Fig. 37) which determine the frequency of
the generated oscillations. By properly proportioning these
constants, currents of frequencies from a fraction of a cycle
per second to thirty million or so cycles per second, may be
generated. The function of the oscillator in the radio
telephone system is primarily to supply radio frequency
power to the transmitting antenna; it also has important
applications in receiving.
15. The Audion as a Detector. We shall now consider
the operation of the audion in its other important role, viz.,
as a detector. The function of the
detector in the radio telephone sys-
tem is to demodulate the received
signals (of radio frequency) and to
translate them back into the low
frequency currents which represent
/ Fig. 39. Showing audion in
the original speech. This may be use as a detector.
accomplished with the audion in
three different ways, two of which will be described; the
third is too complicated for inclusion in a book of this scope.
Let the device be connected as in Fig. 39. A sinusoidal
a.c. voltage of radio frequency is impressed on the grid by
the generator E, but now the "C" battery will be adjusted
so that the grid potential is normally at either of the points
DI or D 2 of the characteristic curve (Fig. 34). At these
points the curvature of the characteristic is large and good
operation as a detector will be secured. The reason for this
will be seen from the following :
Figure 40 shows the effect an a.c. voltage on the grid has
upon the plate current. The plate current variations are
54
RADIO TELEPHONY
no longer even approximately sinusoidal, as in the case of
operation at the point "P" for amplification purposes, but
are asymmetrical with respect to the normal current, 7 .
It will be observed that the current loops are larger above
than below, so that the average plate current has increased.
The radio frequency variations have no effect upon the am-
meter A, or upon a pair of telephones connected in the
plate circuit, but the change in the mean plate current
I .Mean Plate Current without A.C.
I with A.C.
Fig. 40. Showing variations of plate current caused by sinusoidal variation of
grid voltage; rectification effect in plate circuit.
( A/ g = /| 7 ) which they produce is immediately recorded.
If the generator E is separated from the grid by a telegraph
key, and its e.m.f . impressed in the form of dots and dashes,
the ammeter A will be deflected every time the key is closed
and the system could be used for telegraphy. But of
greater interest is the effect of the more complicated radio
telephone signal of the type of Fig. 28. This case is car-
ried through graphically in Fig. 41 using the signal of Fig. 28
as a model. The mean plate current is now no longer
constant, but varies roughly according to the modulations
THE AUDION
55
of the radio frequency wave. This is shown by the heavy
line curve in Fig. 41. This current, labelled "MEAN PLATE
CURRENT (TELEPHONE CURRENT)," flows through the tele-
phone receivers, whereas, on account of the high impedance
offered to them, the radio frequencies do not. This is
shown in Fig. 42, which is a photograph* of the currents
Fig. 41. Illustrating detection of radio telephone speech signal (Fig. 28) by
asymmetrical audion characteristic.
and voltages in a case of the detection of a signal of ap-
proximately sinusoidal modulation. The middle curve
represents the modulated wave (the voltage impressed upon
the grid); the lower curve shows the plate current varia-
tions; and the upper curve gives the current through the
* For this photograph I am indebted to Mr. E. S. Purington of the
Hammond Radio Research Laboratory, Harvard University.
56 RADIO TELEPHONY
telephones. The action of the telephone receivers in chok-
ing out the high frequency variations, and being activated
by the lower frequency changes in the average plate current,
is clearly brought out
Thus the process of demodulation, or detection, consists
in impressing upon the detector the modulated radio fre-
quency wave which is rectified by the asymmetrical char-
acteristic of the device giving a variation of the mean plate
current corresponding in form to the modulation of the
Fig. 42. Photograph of the currents and voltages (circuit of Fig. 39) in a case
of detection with the audion of a sinusoidally modulated wave. I T = Telephone
current; E = modulated wave (grid voltage); I p = plate current. (Purington.)
impressed wave. These low frequency variations of the
plate current activate the telephone receivers, which then
reproduce (somewhat distorted) the original speech.
The scheme of connections for detection by this method
is shown in Fig. 43. Here the detector is connected across
the inductance L of the antenna circuit and is activated
by the voltage drop across it. Operation by this method is
sometimes referred to as detection without grid condenser,
with grid bias or grid polarization.
The second mode of operation is somewhat harder to
THE AUDION
57
explain and to understand ; in fact, a completely satisfactory
explanation of the action has yet to be given. But in gen-
eral the following may be said of it : This mode of operation
is frequently referred to as operation with grid condenser, or
more correctly operation with grid impedance, and the ele-
mentary connections for it are shown in Fig. 44. Here Ci
is the grid condenser, usually of small capacity of the order
of .0002 mfd. ; R is the grid leak resistance (or an impedance)
of the orders of from 500,000 to 2,000,000 ohms (at speech
frequencies). By means of the battery "C," or by tapping
Fig. 43. Simple receiving circuit using Fig. 44. Simple receiving circuit
audion as detector with grid bias. using audion as detector with grid
condenser.
in at some suitable point of the filament circuit, the grid
is maintained at a positive potential. The reason for this
is that operation now takes place by grid circuit rectification,
that is, the rectifying effect just explained for the plate
circuit, occurs in the grid circuit, and the rectified current
flows through the grid circuit impedance, producing a voltage
drop across it and thus lowering the average grid potential.
This in turn affects the plate current, and produces in it
variations characteristic of the original modulation. The
telephone receivers reproduce the speech in the usual way.
CHAPTER III
ANTENNA CONSTRUCTION
16. Antennae for Sending and Receiving. From one
point of view the antenna may be looked upon as the con-
necting link between the radio instruments and the ether.
At the transmitting station it converts the radio frequency
currents into ether waves; at the receiving station it inter-
cepts such waves and translates them into radio frequency
currents. It has therefore two principle functions: (a) to
radiate r.f . energy, and (b) to receive it.
While the same antenna may perform both functions, its
actions in the two cases differ, \and in general the require-
ments for efficient radiation do not correspond exactly with
those for efficient reception. The design and construction
of the antenna depends therefore largely upon whether it is
to be used for both sending and receiving, or for receiving
alone. For instance, an efficient antenna for transmitting
demands careful design and may be expensive to build,
while for receiving purposes almost any kind of elevated
conductor fairly well insulated, a bed-spring, tin roof, or a
small coil, may be used with fairly satisfactory results. In
other words, a good transmitting antenna will generally
make a good receiving antenna, but the converse is not true,
and an antenna that may give good results for receiving may
be poor, or impossible, for transmitting.
Many amateurs, upon the threshold of their radio inter-
58
ANTENNA CONSTRUCTION 59
est, will want to make a modest beginning and equip them-
selves for receiving only. Naturally in this case a suitable
antenna may be rigged up with little trouble and very
cheaply; and detailed descriptions of such antennae will be
given later. On the other hand, there is a group who have
perhaps served their apprenticeship as receiving station
operators and have become sufficiently interested to want
to do some "talking'' of their own; for the benefit of these
the paragraphs dealing with the design and construction of
antennae for transmitting have been included. It will be
understood that no very detailed instructions can be given,
for the choice of location and even of the form of the antenna
itself, are questions which must be decided on the merits of
each situation. All that can be usefully attempted here is
to exhibit some representative forms, to indicate the desir-
able features to be secured and to give such details of con-
struction as are independent of the choice of the antenna's
location and other local conditions.
17. Description of the Various Types. It can be shown
theoretically, by means of the principle of similitude, that
the best type of radiator is a highly conducting sphere, so
the amateur who installs a silver hemisphere 90 feet in radius
can boast of having the utmost in antennae. This highly
efficient but very impractical type of antenna is shown at
(a), Fig. 45. The next type is derived geometrically by
pulling out the sphere from the pole and making an ellip-
soid of revolution as at (&). This is virtually the simple
vertical wire antenna with which we have already become
familiar, and may be considered the most efficient practical
form. It is, however, not too practical, for in order to secure
most efficient radiation an antenna should be operated at
6o
RADIO TELEPHONY
its fundamental wave length (equal to four times the height
in this case), and must therefore be high, demanding sup-
porting masts which are both inconvenient and expensive.
An antenna of this type designed to radiate most efficiently
a wavelength of 200 meters would be about 164 feet high!
Hence for practical reasons it will be desirable to modify
(b) so that shorter masts can be used. The best modifica-
tion is shown at (c) and consists hi capping the antenna with
a circular "top" made out of a number of wires radiating
from the down-lead and having their ends joined by a cir-
A/ = 7.255x1
W
L - 4 f o co
tt
Fig. 45. Illustrating evolution of practical form of antenna (d) from most
efficient radiator (a).
cular jumper. In this form the fundamental wavelength is
considerably greater than four times the height, so that
shorter masts can be used. There is, however, this objec-
tion to a circular topped antenna: that it requires a circle
of masts, so that while they are shorter, many of them are
required and the question arises as to whether much has been
accomplished in point of economy by this change.
(a) Triangular Flat-top. A more economical form is shown
at (d), Fig. 45, which represents a triangular flat-top that re-
quires but three masts for its support. This form is not
much inferior electrically to its predecessor and is undoubt-
ANTENNA CONSTRUCTION
61
edly the best practical structure, giving high radiation
efficiency (when properly designed) with a low cost. The
merit of this antenna has been recognized for years by the
engineers of the United States Navy and has been adopted
by them as a standard type for medium and high power
stations. It has also been used extensively abroad. The
/\ it
r /\l
TU
7
(a)
(b)
Flat-top types
(0
(d)
Cage construction
Fig. 46. Illustrating inverted "L" and "T" type antenna.
amateur who has plenty of space available and who is willing
to spend a little money in pursuit of the utmost in efficiency
will find this antenna of great interest.
(b) "T" and Inverted "L" Types. The next important modi-
fication of the flat-top to secure still greater economy is
shown at (a) and (6), Fig. 46. Two types are exhibited, in
both of which the flat-top is rectangular and can be sup-
62 RADIO TELEPHONY
ported by two masts. At (a) the down-lead, or lead-in as it
is called, is taken from the middle; at (6) it is taken from the
end. The form (a) is slightly superior to (b) electrically.
These are commonly designated as the "T" and inverted
"L" types respectively, and have found great favor for
amateur as well as for general low power use, chiefly on
account of their convenience and low cost of construction.
(c) Cage Construction. In flowing along the multi-wire flat-
top of this type of antenna the high frequency currents have
a tendency to crowd toward the outside wires, and these
wires carry more than their share of the current. This is
termed the edge effect and increases the resistance of this
portion of the antenna. While the increase is not extremely
important on account of the smaller current densities that
exist normally at the free end of the antenna system, the
effect may be avoided entirely by making all the wires out-
side wires by a form of construction shown at (c) and (d),
Fig. 46. This is called the squirrel-cage, or simply cage,
type of top (and lead-in as well) and while giving in the
usual form slightly less capacity than the flat-top is yet to
be preferred for its low resistance properties. In cases where
a large capacity is desirable, as for instance with the higher
powered sets, two of these cages may be mounted in parallel
as shown in Fig. 47. If the cages are not too close together
this will yield a large increase in capacity without appre-
ciably spoiling the uniform current distribution among the
separate wires of each cage, and is a very desirable design.
The lead-in in these, as well as in any of the antennae shown,
may advantageously be likewise in the form of a cage of
possibly small diameter (about 8 inches) containing 6 or 8
wires (see Fig. 50).
ANTENNA CONSTRUCTION 63
(d) Special Types. We have traced above the evolution
from the most efficient but most impractical antenna to the
cheaper and more practical forms of Fig. 46. There are, in
addition, certain hybrid types which do not fit in well in
Fig. 47. Showing cages mounted in parallel to secure greater antenna capacity.
this developmental scheme, but which are important and well
worth description.
At (a), Fig. 48 is shown a modification of the hemispherical
antenna of (a), Fig. 45. It consists of a plurality of wires
and is excited by exciting all of the wires, or by exciting the
(a)
(b)
(0
Fig. 48. Special types of antenna: (a) Symmetrical multiple antenna; (b)
Alexander son's multiple antenna; (c} loop antenna.
central lead-in as shown. The tuning and frequency are
adjusted so that the current node resides at the pole P,
otherwise the efficiency is lowered. Even if erected over an
imperfectly conducting earth this antenna may still be made
very efficient by burying a circular metal cylinder directly
64 RADIO TELEPHONY
under the bases of the component wires, the diameter of
the cylinder being the same as that of the antenna, and the
ground connection of each wire is made to the cylinder.
The central lead-in is omitted and the system excited by
exciting each of the load coils. In this form the antenna is
very nearly as efficient as the spherical antenna from which
it is derived.
Mr. E. F. W. Alexanderson has proposed an arrangement
similar to this in which the wires are arranged in a straight
line at (b), Fig. 48. This is called the multiple-tuned an-
tenna, and is less efficient than the symmetrical form (a) for
two reasons: (1) the ground currents cannot be distributed
so uniformly as in the case of (a), and (2) the system will
develop current vortices (i. e., circulating currents in the
flat-top which radiate in a useless direction) unless very
carefully tuned. In fact, it can be shown theoretically that
this antenna when erected over an imperfect earth is no
more efficient than any one of its component down-leads
regarded as a simple antenna. Thus notwithstanding the
extravagant claims that have been made, it is no better as
a radiator than the simpler structures illustrated in Figs.
45, 46, 47 and is, besides, more expensive to build. Its
principle advantage, that of yielding a low input resistance,
is realized only with special forms of generators such as the
high frequency alternator; for amateurs it has little more
than a general interest.
If we remove all of the wires in (a), Fig. 48 but two, there
is obtained a simple type of antenna which is sometimes
referred to as a loop antenna. This is shown at (c), Fig. 48
and has no particular advantages for amateur transmission,
but is sometimes useful for receiving (see Art. 38). The
ANTENNA CONSTRUCTION
system may be so excited that it has remarkable directional
properties.
A type of antenna which is becoming very popular with
amateurs for 200 meter transmitting is called the fan antenna
and is illustrated at (a), Fig. 49. By the use of the stay
A-B the system may be supported by two masts. This is
a very good antenna, but not inherently so good as the
types previously recommended, Figs. 45, 46 and 47. It does
offer a low resistance and high capacity; the trouble is that
the capacity is too near the earth. However, a greater
efficiency with less care can probably be obtained with this
antenna than any of the types yet described. We have no
I
Fig. 49. Illustrating "fan" and "umbrella" types of antenna.
space to give reasons for this statement, preferring to use
what we have in showing how the better qualities theoret-
ically inherent in the other types can be practically secured.
Still another type is shown at (), Fig. 49 and requires
but one mast for its support. On account of its shape it is
called the umbrella antenna and while successfully used by
commercial companies, notably the Telefunken Company
of Berlin, for high power work, is not considered par-
ticularly suitable for amateur use. It is obviously less
efficient electrically than the triangular top type (d) of Fig.
45, and unless a lot of space is available for its erection will
generally be inferior also to the "T" and inverted "L" types.
66 RADIO TELEPHONY
The above antennae have been discussed particularly from
the point of view of radiation. The special types for receiv-
ing will be taken up in Arts. 35, 36, 37, and 38.
18. Requirements for Transmitting. All of the energy
supplied by the power source to the antenna is not radiated
by it. There are a great many ways in which some of this
energy is dissipated and frittered away. The merit or the
efficiency of an antenna system as a radiator may be taken
as the ratio of the power radiated to the power supplied and
the object of antenna design is to secure a system in which
this ratio is as large as possible. The various losses, in-
cluding that due to the radiation itself, may be represented
by resistances, the values of these resistances being equal to
the power lost divided by the square of the current at the
point of the antenna (usually the base) at which the power
is applied. The ratio just mentioned is then equal to the
ratio of the radiation resistance to the total resistance of the
antenna.
The most important sources of loss other than the only
desirable loss, radiation, are as follows: (1) loss in the
antenna conductors due to their resistance; (2) loss in the
tuning apparatus (load coil and condenser in series with the
antenna) due to its resistance; (3) loss due to heat developed
in the earth (earth resistance) by currents returning to the
lead-in; (4) loss due to imperfect dielectrics in the electric
field of the antenna; (5) loss due to direct leakage through
faulty insulators; (6) loss due to the induction of currents
in neighboring conducting structures such as metallic poles,
towers, guy wires, etc. ; and (7) loss due to the formation of
"corona" (brush discharge) on the wires. All these, with
the possible exception of the last, are important in amateur
ANTENNA CONSTRUCTION 67
antennae, and it is proposed to consider them in turn and to
give practical instruction for keeping them as small as
possible.
19. Best Operating Wave Length in Transmitting. The
resistance which represents the useful loss from the antenna
by radiation varies inversely as the square of the wave-
length, and directly as the square of the effective height.
In a well designed antenna the undesirable losses enumer-
ated above remain practically constant from the funda-
mental wavelength to a wavelength two or three times this,
and the ratio of the useful loss (radiation) to the total power
supplied is greatest at the fundamental wave length. Hence
this is the wavelength at which best radiation takes place,
and should be selected for transmitting. It will usually be
necessary to insert in the antenna a load coil for the purpose
of coupling the power circuit to it, but this should be kept
as small as possible and its effect hi raising the wavelength
above the fundamental may be compensated for by insert-
ing a series condenser having low losses. The fundamental
wavelength is not the wavelength at which maximum current
will be secured, but at which maximum PR & (R a = radiation
resistance) is obtained. The operator is warned therefore
not to be deceived by the antenna ammeter reading in
estimating how well his station is radiating; this tells only
a part of the story.
20. Losses in the Antenna Conductors. The heat loss in
the conductors of the antenna is the first source of loss listed
above and while not particularly important unless the an-
tenna is otherwise well designed and the other losses are
small, nevertheless is additive and well worth taking the
trouble to minimize. This can be done by using fairly
68
RADIO TELEPHONY
heavy conductors; a flat copper strip is best, stranded an-
tenna wire being a good second. The latter is somewhat
superior mechanically. Cage construction may be used hi
the top, the form of antenna shown at (c) and (d). Fig. 46
being a particularly good form to try this with; and should
certainly be used hi the entire lead-in. In this case the
cage may be tapered as shown in Fig. 50. The object of
this is to keep the capacity of the system as far away from
the earth as possible; the capacity of the cage per unit
length will increase rapidly with its
diameter. We want the low resist-
ance properties of the cage; but in
the lead-in do not want its capacity.
In the top, of course, the capacity is
desirable, in fact more desirable pro-
portionately than the low resistance
properties, although both are needed.
The portion of the lead-in, B, Fig.
50 (sometimes called the ground-
lead) should not be neglected as it is
here that the currents are heaviest.
A stout piece of copper strip, or sev-
eral strands of the antenna wire hi parallel, should be
used.
21. Losses in the Tuning Apparatus. While methods for
reducing the resistance of the tuning apparatus will be de-
scribed later when we come to consider the power unit,
its importance is fittingly emphasized here. It would obvi-
ously be foolish to go to a lot of trouble to improve the con-
ductivity of the antenna system unless as much attention
were paid to the inductance coils and condensers in series
Fig. 50. Tapered cage
lead-in of low resistance.
ANTENNA CONSTRUCTION
69
with it. And the importance of this increases with the im-
provement of the antenna.
22. Losses Due to Earth Currents. The third source of
loss, usually the most prolific in antenna systems, especially
at short wavelengths, is the heat generated in the earth
by the currents returning to or coming from the lead-in.
Remembering that the heat loss is equal to PR, it is
clear that the loss in any unit cube of the earth material
goes up as the square of the current density at that point;
consequently in order to keep down the whole loss the
concentration of current at any
point is to be avoided. The dis-
tribution of current depends upon
the wavelength, conductivity and
dielectric constant of the earth,
as well as upon the geometry of
the antenna. A symmetrical an-
tenna will give a better distribu-
tion and consequently a lower typic ? l 7 case f direct ground -
ing (Zenneck).
earth resistance than one which is
not symmetrical ; in other words, the circular topped type of
(c), Fig. 45 will be superior to either of the types (a) and
(&), Fig. 46, from this point of view. The distribution in
the earth is determined at long wavelengths mainly by the
earth conductivity; at short wavelengths, by the dielectric
constant (unless the conductivity is too high). For earth
of average conductivity, the penetration into the earth de-
creases as the wavelength is shortened. The flow of current
in a typical case is shown rn Fig. 51. The current converges
toward the lead-in and the current density is therefore
greatest at this point. In the antenna system, the current
Fig. 51. Showing flow of
currents in the earth in a
70 RADIO TELEPHONY
flows by a conductive path up through the antenna con-
ductors, thence by capacity paths to the earth, and finally
through the earth to the lead-in. It is precisely the con-
centration of current here that causes most of the loss in the
average grounding system. The loss may be diminished by
reducing the current concentration, and this may be ac-
complished by providing a generous surface hi the grounding
electrode.
23. Direct Ground. Such an electrode is shown in Fig. 52
and consists of a short circular cylinder of large radius. The
Fig. 52. Illustrating a very good type of direct ground with low earth resistance.
proper depth is regulated by the penetration of current into
the earth and for earth of not too poor conductivity and at
200 to 300 meters wavelength, a depth of 2 or 3 feet will be
ample. The connection is made by means of a number of
wires which converge to the approximate center of the circle
and unless the cylinder is very deep, and in any case if it is
more than 15 feet in radius, should be supported above the
earth and not buried in it or laid upon the surface. The
cylinder itself may be made up of galvanized-iron sheets,
such as are used in the construction of small temporary
ANTENNA CONSTRUCTION 71
shacks. These need not be soldered together, but should
overlap with no sharp edges protruding any distance from
the body of the cylinder. A connection should be made to
each sheet. They are better soldered, however, if it can be
done. .The ground is installed by digging a narrow circular
trench under the antenna, not too far from the point at
which the lead-in in the antenna diagrams, Figs. 45, 46 and
47, enters the earth. While it is not strictly necessary that
the cylinder should be circular, this is the best form and no
extreme departures which are likely to introduce sharp
corners should be made.
This ground has recently been redescribed by Capt. H.
J. Round, and is referred to in some circles as "Round's
Round Ground." A better name would perhaps be the
" Common Sense Ground, " for it is older than radio itself,
and was used by Fessenden* as early as 1910, and has also
been described by the Germans prior to that time and since.
In modern times it has received its greatest support by the
experiments of Dr. John M. Miller of the United States
Navy Radio Laboratory.
24. Counterpoise. There is another method for securing
a more uniform distribution of the earth currents which is
even superior and often cheaper and easier to install than
the excellent ground system just described. This may be
explained with the aid of Fig. 53 as follows:
Lay upon the earth a large metal disc and connect this to
the lead-in, [Fig. 53 (a)]. The currents will now find a large
conducting surface and on account of the large area and
circular shape of the plate a fairly low resistance ground
* According to Mr. J. W. Lee this ground was installed at the Brant Rock
station, in the "Cut River Experiments."
72 RADIO TELEPHONY
will be obtained. There will probably be a slight concen-
tration of current at the edges. This would make a very
good ground and could be still further improved by ex-
tending its edges down into the earth as in the cylindrical
ground system just described. The plate need not be on
the surface, but may be supported above it as shown at (b) ,
Fig. 53. The current flow is practically unaffected by this
change and is completed through the condenser formed by
the disc and the earth's surface. This system is found ex-
perimentally to yield a very low ground resistance, as the
(a)
(b)
Fig. 53. Illustrating equivalence of counterpoise (&) and surface electrode (a)
in securing more uniform distribution of earth currents.
above reasoning would lead us to expect. From a practical
point of view, however, a metal plate of this size is incon-
venient and expensive. The advantages of the arrangement
are not lost nor materially diminished if a net of wires is
substituted for the plate, provided the wires of this network
are sufficiently plentiful and they are not too far apart com-
pared with the distance above the earth. Such an arrange-
ment (b\ Fig. 53, is called a counterpoise or capacity ground,
and if properly designed and installed, is the most desirable
and satisfactory type of ground for the amateur, especially
in localities where the earth conductivity is poor. Con-
ANTENNA CONSTRUCTION 73
cerning the proper design and construction the following
general remarks may be made:
25. Design and Construction of the Counterpoise. The
area of the counterpoise should be as large as possible since
the distribution of earth currents is directly affected thereby.
The exact shape is not generally important, but the best
forms are the circular, elliptic, square and rectangular, in
the order given. It should be placed as nearly under the
antenna as possible and should extend well out beyond the
antenna's projection on the earth. The number of wires
should be as large as possible and the wires should be fre-
quently bound together with cross jumpers. This will
reduce to a minimum the generation of heat due to current
vortices which form as a result of its possibly irregular shape
and situation. The height of the counterpoise is governed
by several considerations, the most important of which are
the separation of the wires in the network, the evenness of
the ground, the character of the vegetation with which it is
covered, its conducting qualities, and the possible presence
of ground water near the surface. If the height is small
compared with the distances between the wires in the net,
there will be a tendency for concentration of the current
immediately under the wires. Thus the effect of laying the
whole net right on the ground will be a decidedly poor
ground system (at short wave lengths). The same thing is
true if the net is buried. Bushes, grass, and other flora
under the counterpoise constitute poor dielectrics and in
order to make the volume of dielectric which they rep-
resent as small as possible compared with the total dielec-
tric, the height of the counterpoise should be increased when
they are present. A similar remark holds for any type of
74 RADIO TELEPHONY
poor dielectric. Also the height of the counterpoise should
be uniform, that is to say, if there is a small eminence on
the earth's surface the capacity at this point will be larger
per unit area than in any other point of the area and the
currents will head toward the path of lowest impedance;
there will then be a concentration with corresponding in-
creased loss in this section. The effect of small irregularities
is reduced by increasing the height. From this point of view
and also from that of poor dielectrics in the field, a counter-
poise with a radio shack installed under its active area is
decidedly poor design unless the method described later,
Art. 25, or some similar device, is used to rectify the situa-
tion by redistributing the earth currents. Perhaps the most
important single thing, and the most prolific cause of medi-
ocre results in many counterpoises, is its support. The coun-
terpoise is a condenser and should be treated with the same
respect. Nondescript wooden stakes of all degrees of con-
ductivity, wet rope, poor insulators, etc., should not be per-
mitted in the intense electric field which exists between the
counterpoise and the earth and extends some distance
beyond its margin. The short poles used for its support
should be placed at a distance from the margin equal to
several times its height, and a good grade of porcelain or
glass should be used for the insulation. The 17-inch por-
celain insulators described later (Art. 29) in connection with
the antenna insulation are recommended here.
The above precautions and desirable features have been
incorporated in the design of a typical counterpoise system
shown in Fig. 54. This may serve as a model in planning
a counterpoise for any special situation. More detailed
specifications would be of no particular value on account
ANTENNA CONSTRUCTION
75
of the wide variation of conditions likely to be encountered
by the readers of these pages. For fairly even ground
covered with short grass, a height of 2 or 3 feet will be
adequate; for uneven ground or ground covered with bushes
24" HOP
INSULATOR
STAKE
Fig. 54. Model counterpoise system in which the design suggestions of this
article have been incorporated.
and undergrowth, heights two or three times this will be
necessary for best results.
In no circumstances should the counterpoise be directly
grounded at one or two points, for there will then be a
rush of current to these places which will defeat its whole
purpose. The only way to ground the counterpoise would
be to bury a circle of plates at its periphery, thus making
7 6
RADIO TELEPHONY
a very large direct ground of the type described in the last
section. A direct ground may, however, be associated with
a counterpoise by other means, as follows :
26. Combination of Direct Ground with Counterpoise.
The combination of the counterpoise described above with
the single direct ground is frequently desirable. Cases of
this arise from the necessity for: (1) grounding the d.c.
high voltage source with certain types of audion trans-
Pig. 55. Showing a method of combining direct earth and counterpoise to reduce
earth resistance.
mitting circuits when using the counterpoise, and (2) for
redistributing the earth currents in the case of a counter-
poise which either through unfavorable physical conditions,
lack of space, etc., or faulty design, is not operating with a
sufficiently uniform distribution. A very convenient method
for this, which has frequently been reinvented and re-
described, but which dates back to the radio station at
Spanishtown, Gibralter (1911), is shown in Fig. 55. The
ANTENNA CONSTRUCTION 77
direct ground should be in that part of the earth where there
is either not enough current or too much, and should be
constructed with the same care that would be taken if it
were to be the only grounding means employed. If the
ground tap 5 is placed at the point of the system where
the voltage with respect to earth is a minimum (called a
voltage node), it will have very little effect. This point can
be determined experimentally by adjusting the tap until the
effect of touching it to the antenna coil and removing it is
a minimum, as indicated by the reading of the ammeter in
"Position (1)" and the wavelength. If the antenna is
operated above its fundamental the voltage node will reside
upon the coil between the antenna tap and the counterpoise;
if it is operated at the fundamental the voltage node will be
found right at the antenna tap. Starting from this position
the ground tap S is moved up or down until the reading of
the ammeter in "Position (2)" is a maximum at the wave-
length upon which it is desired to operate. (The adjustment
of this tap will change the wavelength somewhat.) In the
case of operation at the fundamental, or with high antenna
resistance, it may be necessary to move several turns from
the voltage node to secure sufficient voltage; in these cir-
cumstances it is advisable to neutralize the inductive
reactance between the voltage node and the tap by means
of a condenser of capacity about .005 to .007 mfd., otherwise
the phase difference of the voltages at the two grounds will
be too great for proper compensation. In experimenting
with this circuit it is helpful to remember that the effect of
the inserted e.m.f. is either to repel current from the direct
ground, or to attract it, according to the direction of the
e.m.f. The direct ground is therefore a remedial device
78 RADIO TELEPHONY
for changing the current distribution at the point of its
application. It is obvious that the arrangement may be
used also to secure an optimum distribution of current
between two or more direct grounds.
Instead of the reading of the ammeter, a better method
of adjustment is to use a nearby receiving station which can
note the effect of each change upon the strength of the
received signal. Ammeter readings in cases of this kind are
not altogether trustworthy as an indication of the power
radiated.
27. Losses in Imperfect Dielectrics in the Electric Field
of the Antenna. We come now to the fourth important
source of loss in the antenna system. This is occasioned
by the presence in the electric field of poor dielectrics in
which losses by hysteresis, etc., occur.* The part of the
total antenna resistance by which this loss is represented
increases linearly with the wavelength, or very nearly so.
The antenna is virtually a condenser, in which the plates
are separated by a distance which is large compared to their
dimensions. Its electric field, therefore, extends for a con-
siderable distance beyond its margin; thus not only are the
dielectrics directly beneath it affected, but also those which
might appear at first sight to be innocuously situated. Many
types of poor dielectric are apt to be encountered. All
wooden poles, spreaders, etc., used in the antenna's sup-
port, unglazed porcelain insulators (and indeed most in-
* The importance of this type of loss was first emphasized by Dr. John M.
Miller, at that time Assistant Physicist at the Bureau of Standards, and the
reader is urged to consult his paper, "The Effect of Imperfect Dielectrics in
the Field of a Radiotelegraphic Antenna," Scientific Paper of the Bureau of
Standards, No. 269. Copies of this may be obtained from the Supt. of Public
Documents, Wash., D. C.
ANTENNA CONSTRUCTION 79
sulators), rope, trees, buildings, shacks, etc., should be
looked upon with suspicion. In estimating the effect of any
dielectric in the field account should be taken (1) of its rela-
tive volume compared with the volume of the whole di-
electric, and (2) of the strength (usually determined by its
proximity to the antenna conductors) of the field where it
is situated. Otherwise expressed, the effect will be deter-
mined by the capacity through the dielectric. Without
going into any technical detail, the following general direc-
tions for avoiding losses of this type may be given.
28. Trees in the Antenna Field. Probably the most
prevalent source of dielectric loss in amateur antennae is
the proximity of trees, large bushes, etc. It is common
practice, for instance, to use a tree for the support of one
end of the antenna, usually the end farthest from the trans-
mitter the worst end. This practice, while very excusable
from the point of view of convenience, cannot be too strongly
condemned electrically. If a tree must be used, the active
part of the antenna should terminate some distance from it,
and not in its foliage; this will reduce the loss somewhat.
But whether used for the support of the antenna or not,
and even if fifty feet away from it, trees are deleterious and
to be avoided if possible. This is particularly important
with a type of antenna having very little flat-top capacity.
29. Insulation of the Antenna System. Many amateurs
(and "engineers," too, for that matter) design the insulation
of their antennae apparently as if the only thing to be de-
manded of an insulator is that it shall insulate. But careful
measurements of dielectric loss reveal that there are "in-
sulators and insulators." From this point of view we shall
demand of an insulator that in addition to its being one, it
8o
RADIO TELEPHONY
shall have a low capacity between its terminals and shall
be a good dielectric with low electric absorption. Unglazed
porcelain, hard-rubber, ebonite, fibre, wood, and any syn-
thetic compounds which absorb moisture, are unsuitable.
Either glass or a good grade of glazed porcelain should be
used. After the selection of the material, its disposition and
form are to be carefully attended to. Consider first the
insulation of the antenna itself. (That of the guy-wires is
taken up in Art. 30.)
^-SPR CAPER
*^ (f1tl orWoo*)
24" Roi>
INSULATOR
WOOPEN
SPHlADtBj
ANTENNA WIRE.
Tstr "li Copter
Fig. 56. Spreader construction: (a) Arrangement for wood or metal spreaders;
(b) alternative arrangement for wood spreaders.
If either of the forms of flat- top, (c) or (d), Fig. 46 are used,
an insulator may be inserted in each supporting stay to the
mast. The masts, incidentally, in these as well as all
antenna types should be placed as far as possible from the
active part of the antenna. Space will not permit this in
all cases, but it is very desirable, especially with wooden
poles. In the case of forms (a) and (6), Fig. 46, spreaders
will be used and care should be taken with them. The
spreaders are best made of metal tubing; and a suitable
suspension lay-out is shown at (a), Fig. 56. This same
ANTENNA CONSTRUCTION 81
arrangement may be used also with wood spreaders; and in
this case the wires should be wrapped around the spreader
and joined electrically to the suspension bridles A B C D.
An alternative scheme is shown at (5), Fig. 56, which has
for its object keeping the wooden spreaders out of the in-
tense electric field which exists close to the antenna wires.
The cross jumper tends to reduce the electric force at the
ends of the wires, which is desirable with wooden spreaders.
The strain insulator should be chosen with care. The best
insulator known to the writer for the low powers used by ama-
teurs is a 17 -inch porcelain rod insulator of the type shown
in Fig. 57. They are rather more costly than the common
Fig. 57. Sketch of 17 -inch porcelain strain insulator.
garden variety of insulator, but are well worth the extra
expense if the antenna is otherwise good.
Metal rings are recommended for the construction of the
cage types of antenna (Fig. 46), although many amateurs use
wooden barrel hoops. Brass or copper tubing is suitable
and may be bent without flattening by first filling the tube
with sand or melted sulphur. Metal plugs may be used in
making the joint, and should be securely pinned or soldered
in place. There is little danger of eddy currents hi these
rings. When wooden rings are employed the wires should
be electrically connected to the bridle wires, as in the case
of the wooden spreader. This type of construction is illus-
trated in Fig. 58.
When the antenna is operated at its fundamental wave-
6
82
RADIO TELEPHONY
Fig. 58. Cage antenna construc-
tion using metal rings or wooden
barrel hoops.
length the voltage increases rapidly from the load coil to
the free end; and for that reason the above precautions with
respect to dielectrics are especially important near the free
end. But when any consid-
erable amount of inductive
loading is used, the voltage
is more uniformly distributed
and the entire electric field of
the antenna, down to the load
coil, is strong; consequently
care should be taken with the
whole system, from the load
coil itself, lead-in insulator and lead-in, to the end of the
antenna. A very good and inexpensive lead-in insulator
may be made out of a sheet of window
glass as shown in Fig. 59. The glass may
be drilled with the broken-off end of a tri-
angular file, kept well lubricated with
kerosene. The threaded rod and lead
bushing should be inserted and locked in
place immediately after drilling to prevent
the internal strains from cracking the
plate.
30. Losses by Induced Currents in Masts
and Guy Wires. The reader has already
been advised to keep the masts as far away
from the antenna as possible. This applies
to metallic as well as to wooden masts. In
the latter case the purpose is to reduce the dielectric losses;
in the former to minimize the flow of induced currents and
the losses due to them. Metallic masts made up of sections
59.IIIUS-
inexpensive
insulator
low losses,
made from a sheet
of window glass.
ANTENNA CONSTRUCTION 83
of wrought-iron pipe of decreasing size are very easy to
build, cheap and convenient in every way; in addition, they
present a very business-like appearance. In erecting such a
mast do not dig a hole in the earth and slip the bottom of
the mast into it, and do not bury it in concrete, but secure
a number of good glass or porcelain insulators (telegraph
line insulators will do), and, after bolting a flange and plate
to the bottom of the mast as in Fig. 60, insert three of
these insulators between the plate and the concrete or
other supporting base. This will provide good insulation
FlN TYPE |M4ULAT0*S
(GkASS OR RRCeiAIN)
Fig. 60. Showing method of supporting and insulating metallic masts to reduce
dielectric and heat losses in them.
for the mast; then if it is to be grounded it can be done
properly, with the aid of the direct ground of Art. 23.
There will be induced currents in the metallic guys as
well as in the mast. Each guy should therefore be broken
up into short lengths of 10 to 20 feet by means of porcelain
egg insulators of familiar type. Remember that the mast
with its guys is really a small antenna system, excited by
the main antenna, in which all the losses enumerated for the
main antenna, including radiation, will occur; hence it
should be given its proportionate amount of care, especially
in the matter of insulation.
84 RADIO TELEPHONY
Whether the masts should be grounded or not is still a
matter of controversy. The question is a hard one to answer
directly. If the insulation at the base is good, and there is
no appreciable dielectric loss there, perhaps it had better be
insulated. In any case the only satisfactory way to settle
this question is by appeal to experiment; a distant receiving
station can be requested to note the effect of grounding and
insulating the masts. The reading of the ammeter in the
main antenna circuit will be increased by grounding the
masts; but this has no special significance and does not
indicate that the radiation has increased. In fact, this may
happen if the radiation has been reduced. So the observa-
tions should be made at the distant receiver.
31. Telefunken Method of Reducing Losses in Masts.
A modification of a method which is claimed to reduce the
losses in metallic masts is shown in Fig. 61. This is due to
the Telefunken Company of Berlin and is described hi
German Patent No. 300,782 (September 15, 1919). The idea
is to impress upon the mast an e.m.f . of such magnitude and
phase that will reduce the current flowing in it. This e.m.f.
is regulated by means of the tap S on the load coil in the
antenna circuit, and the inductance and capacity L C of the
mast circuit. The wires AB should be run about 2 or 3
feet above the earth and connected to the masts at their
bases. The masts should be insulated. The adjustment
may be found tedious, since a distant receiving station
should be used, and is very similar to that in the case of the
combination of direct ground and counterpoise (Art. 26).
In the original Telefunken diagram the antenna is connected
to the mast, as shown by the dotted lines in Fig. 61. Theo-
retically there is not much gain to be expected with this
ANTENNA CONSTRUCTION 85
system, but several amateurs have reported favorably upon
it. It is probably worth trying.
32. Other Losses in the Antenna System. The losses
due to direct leakage through insulators and corona are
trivial and hardly worth discussion. The remedy in the first
case is obvious; and the corona condition may be alleviated
by increasing the capacity of the antenna, especially near the
free end. It is not likely to be encountered with the low
powers to which amateurs are restricted unless the antenna
Fig. 61. Modified Telefunken method for reducing flow of currents in metallic
masts and losses due to them.
design is such that an unusually low capacity is obtained,
as, for example, with a single wire of small diameter.
33. Antennae on Houses and Buildings. Up to this point
we have considered the antenna system from the point of
view of the amateur who has plenty of space at his dis-
posal, and is in the fortunate position of being able to sit
down and plan his antenna on a sound theoretical basis and
to carry out his plans. In other words, we have cleared our
conscience by outlining what we think should be done.
86
RADIO TELEPHONY
But there are many of us who, however enthused, do not
feel justified, for instance, in cutting down a particularly
fine tree simply because it is suspected of causing dielectric
loss in our antenna system. And in other things as well, the
Fig. 62. Illustrating a high-capacity house-top antenna supported by one mast
in which the restricted roof area is used to maximum advantage.
theoretical "what should be done" must meet the practical
situation half-way and effect a compromise with it.
Going still farther, consider the less fortunate city brother
who has no space for his antenna except over the roofs of a
row of houses. What shall be said to him? His whole vista
is one of dielectric loss, eddy currents, and poor grounds; yet
ANTENNA CONSTRUCTION 87
he manages to live and to radiate, even if the row of houses
must be heated by his losses in getting it accomplished.
Obviously he needs sympathy and encouragement and not
elaborate descriptions of cylindrical grounds 50 feet in
diameter. Our advice in these circumstances is mainly
non- technical : Do the best you can. But, in addition, the
following remarks concerning house-top antennae may be
ventured:
Figure 62 shows an antenna of this type, in which the
small roof space available is used to maximum advantage.
This antenna has a high capacity and can be supported by
one mast.
The chief disadvantage of these systems is that a great
deal of material, dielectric and conducting, is directly under
the antenna, in the most intense part of its field. The result-
ing dielectric and other losses will therefore generally be
quite high. In the case of a house with a tin roof well bonded
together electrically, there is sometimes an advantage in
grounding the tin at its four corners, or in as many places
as possible, by running a separate lead from each point of
connection directly to the ground. The grounding of these
leads should be well done; otherwise the supposed advantage
may be turned into an increased loss. The ground-lead from
the transmitting apparatus may then be connected to the
tin roof. The effectiveness of this scheme increases with the
amount of load in series with the antenna. This should not,
however, be construed to mean that the antenna should be
operated above its fundamental. Figure 63 illustrates this
type of installation.
Good results have also been reported with counterpoises
erected in the cellar, although this is hard to account for.
88
RADIO TELEPHONY
Other amateurs have used a moderately large cylindrical
ground (Art. 23) in the back-yard with some success.
34. Condenser Antennae. The so-called condenser an-
tenna is an ordinary antenna of small height and very large
flat-top area, in which the ground is usually replaced by a
counterpoise of about the same shape and size as the flat-
ANTCNNA
LEAP
ELECTROPCS
Fig. 63. House-top antenna installed over well-grounded tin roof.
top. This type appears to be particularly suitable for house-
top installation and is recommended for trial in such cir-
cumstances.* The antenna is virtually a large condenser,
* Mr. Malcolm Ferris, Expert Radio Aid of the United States Navy at
Norfolk, Va., reported very good results in 1919 from an arrangement similar
to that shown in Fig. 64. More recently Mr. Melville Eastham, of the General
Radio Co., has experimented extensively with condenser antennae on the
roof of his plant in Cambridge, and by careful design has been able to radiate
about as effectively with it as with the ordinary type of antenna. The
properties of this antenna have been studied experimentally by Fessenden,
and to some theoretical extent by Bennett.
ANTENNA CONSTRUCTION
89
one plate of which is the flat-top of the antenna (which should
be of large area and contain a large number of wires) and the
other is either a well insulated tin roof, or better, an auxiliary
network of wires, or counterpoise, of as large area as space
will permit. The field is largely confined to the region
between the plates and there is consequently less dielectric
ETAU MAST
Fig. 64. Illustrating "condenser type" of antenna especially suitable for house-
top installation.
loss in the material under it and outside of its immediate
neighborhood.
The success of this antenna depends upon securing a very
low resistance in the system, principally in the lead-in, which
should be made of a number of wires in parallel, or a gener-
ously proportioned copper strip. The length of the lead-in
should be small and to secure this the apparatus may be
installed on the top floor of the building as near to the
antenna as possible. The load coil of the transmitter should
90 RADIO TELEPHONY
also have an especially low resistance. The reason for these
precautions is that the radiation resistance of this type of
structure is very low, consequently in order to radiate
effectively with it the other losses should be proportionately
low (Art. 18). A suitable antenna arrangement for house-
top installation is shown in Fig. 64. The leads AB to the
apparatus are two copper strips, about 1 inch wide, TS inch
thick, which are mounted about 1 foot apart with the aid
of as few insulators as possible. Care should be taken with
the insulation of these leads. The antenna is best excited
with a form of oscillating circuit (see Chapter IV) in which
magnetic coupling between the antenna coil and the audion
is employed. The master oscillator system is particularly
suitable, the Meissner circuit being the best substitute for
this.
35. Special Antennae for Receiving. All of the types
described above are good for receiving. We will consider
here the special antennae which may be used for this purpose,
but which are not well adapted for transmitting. As stated
at the beginning of this chapter, almost any kind of elevated
wire may be used. In receiving the antenna losses are not
nearly so important as in transmitting, for their reduction
of the signal intensity is easily compensated by amplifica-
tion, whereas in' transmitting (on account of the large
powers involved) a similar amplification would be ex-
pensive and prohibited beyond a certain point by the law
restricting power input. Hence if the antenna is to be used
only for receiving, the problem of its design has almost dis-
appeared as such.
The special requirement of the receiving antenna not
shared in transmitting is an immunity from the atmospheric
ANTENNA CONSTRUCTION 91
disturbances commonly called "static." The response of
the antenna to static should be as low in proportion to its
response to the signal to be received as it is possible to ob-
tain. This characteristic or property of a receiving system
is often spoken of as its signal-static ratio, and many special
schemes and circuits have been devised for its increase.
Fortunately for the amateur, the static problem decreases
with the wavelength. A number of popular forms of antennae
suitable for receiving will now be described.
36. The Single Wire Receiving Antenna. This consists
merely of a single wire, insulated at both ends, from one end
of which (or from the middle) a lead-in is taken to the re-
ceiving instruments. Its height is generally not important,
at least over ground of not too high conductivity, and its
length (including ground-lead) should be from 100 to 150
feet for 200 meter work. The fundamental wavelength of
this kind of system is from 4 to 4.2 times its total length
in meters. The insulation is not important; ordinary por-
celain cleats will do; but lightning protection of the type
prescribed by the underwriters should be used. (See
Appendix.)
37. The Beverage Wire. This modification of the single
wire antenna, proposed by Mr. H. H. Beverage and described
in United States Patent No. 1,381,089, has for its principal
object the reduction of interference from static and other
stations by means of its sharply directional characteristic.
It consists of a single horizontal wire of length equal to the
wavelength to be received (or an integral multiple thereof).
One end of this is grounded through a resistance approxi-
mately equal to the "surge impedance" of the line (200 to
600 ohms for a line about 10 feet high, No. 16 A.W.G. wire,
92 RADIO TELEPHONY
at radio frequencies) and the other end is connected through
an inductance to the ground in the usual way. The receiving
apparatus may be coupled to this inductance (see Fig. 65).
The system has theoretically a well-defined directional
characteristic and receives best from a direction toward the
end grounded through the resistance, as shown by the arrow,
Fig. 65. The inductance to be used at L may be of the
order of 100 micro-henries for the 200 meter system por-
trayed. The chief merit of this antenna resides in its direc-
tional properties and the immunity 'it provides from static
disturbances; a theoretical examination shows that as an
antenna it has no special virtue, at least over ground of
-A=200m.= 650ft.
34 *
| L ( 1 00>U h.) > | R (200-600.rO
f T T
rr-r;
Fig. 65. Illustrating "Beverage wire" suitable for 200 meter reception.
average conductivity. But the directional property may be
frequently of great use; an example of this was furnished by
the recent Trans- Atlantic Tests conducted by the American
Radio Relay League, in which the antenna was employed
with some absolute success in receiving the signals.
38. Loop, or Coil, Antennae. Another type of antenna,
also with directional characteristics, which is fast becoming
popular on account of the small space required for its in-
stallation, is the loop or coil antenna. It may be acknowl-
edged at the outset that as an antenna the merit of this
device is small, usually about 1 to 10 per cent, of that of the
average elevated type; but since convenient amplifiers are
ANTENNA CONSTRUCTION
93
COIL ANTENNA
Y
easily constructed, this objection is quickly vitiated by its
features of compactness and portability.
The design of a good loop antenna is a complicated and
difficult matter. We will confine ourselves here to a few
general directions by following which a fairly satisfactory
system may be built. The circuit for a simple loop is shown
in Fig. 66. The circuit is tuned by means
of the variable condenser C and the am-
plifier or detector is connected across this
condenser. The e.m.f. induced in such a
coil when acted upon by the wave is nearly
proportional to its area; and the voltage
across the tuning condenser at resonance
will also increase with this area. For best
results the size of the loop should be in-
creased until a single turn is formed which
will resonate at the desired wavelength
with 20 or 30 degrees of the variable con-
denser. This will make a large loop (see
Table I, page 94), even for the short wave-
length of 200 meters; and if it is too large, Fig 66.Show-
its size may be reduced and more turns ing connections of
added at the same time to maintain the coil in use . as a
receiving antenna.
same inductance. Data for square loops
of different sizes suitable for wavelengths from 180 to about
400 meters with a variable condenser of maximum capacity
of .0006 mfd. are given in Table I. In making these com-
putations it has been assumed that No. 18 lamp cord is used,
and that the turns are separated by 2 inches. (There is
some advantage in separating the turns on the coil, since
then a larger number of them may be used for a given in-
To AMPLIFIER
OR DECTECTOR
94
RADIO TELEPHONY
ductance and wavelength.) The wire should be chosen care-
fully. The best method is to wind on about ten No. 18
d.c.c. wires in parallel; this will make a flat stranded con-
ductor of low resistance. Copper strip \ inch by .005 or
.01 inch is also very good. Ordinary lamp cord may be
used if nothing better is at hand.
TABLE I
Data for Coils Suitable for Wavelengths from 180 to 400 m. with Con-
denser .0006 mfd. Capacity.
Side of
Square Loop.
Number of Turns.
Induced E. M. F.
(Relative).
35 ft.
1
19.8
17
2
9.3
10
3
4.8
7
4
3.1
5
5
2.0
3
7
1.0
2^
10
1.0
The loop receives best from directions parallel to its plane,
and very poorly or not at all from directions at right angles
to it. This property has been applied in making of it a
direction finder to determine the directions of distant trans-
mitting stations. It is also frequently of use for avoiding
interference from other transmitters when their directions
differ from that of the station whose signals are being
received.
A loop antenna installation is shown in Fig. 67. Here the
loop is rather elaborately constructed, and is of the type
used by the United States Navy for direction finding pur-
ANTENNA CONSTRUCTION
95
"A" BATTERY
COIL ANTENNA
TEL.
Fig. 67. Receiving set with coil antenna of the type used by the United States
Navy for direction finding.
poses. Very much simpler loops will do for ordinary receiv-
ing purposes.
CHAPTER IV
CONSTRUCTION AND OPERATION OF THE
TRANSMITTER
39. Comparison of Methods for Generating R. F. Power
with the Audion. Self-excitation and Master-oscillator
Systems. In Chapter II the use of the audion as. a gen-
erator of alternating currents of any desired frequency was
explained. Two schemes were considered; one in which the
power for the excitation of the grid circuit was derived
from the plate or output circuit of the same tube, and
another in which a separate audion oscillator furnished the
power for the excitation. The first of these was referred to
as a self -excited system; the second, as a separately excited
system or master-oscillator system (the small extra oscillator,
or exciter, being referred to as the master oscillator).
With the self-excited audion the frequency of the gener-
ated oscillations is determined by the inductance and
capacity of the circuit; in the case of the master-oscillator
circuit it is of course that of the master oscillator, since the
power tube simply acts as an amplifier in these circum-
stances and the inductance and capacity of its output cir-
cuit have little reaction upon the master oscillator. The
power is supplied to the transmitting antenna, and for this
purpose the antenna is coupled in some way to the output
circuit of the tube; thus the constants of the antenna cir-
cuit affect the frequency of the oscillations in the self-
excited system. In continuous wave (c.w.) work the re-
96
CONSTRUCTION OF THE TRANSMITTER 97
ception is accomplished by the "beat" method (see Chapter
VI), consequently it is very important that the frequency of
transmitting current remains constant. This does not apply
to radio telephone work, for in this case ordinary detection
is generally used. It is clear then that the effect of a change
in the antenna constants, such as might be caused by a
swaying antenna, or lead-in, is most pronounced in the self-
excited oscillator and is frequently very troublesome. The
master-oscillator scheme, on the other hand, is almost free
from this objection for the reason just explained.
In addition to this the master-oscillator system is more
convenient to work with, and the adjustment for maximum
output for different wavelengths and antenna resistances is
more easily made. When adjusting for maximum output
in the self-excited system we must adjust for proper feed-
back and stability as well; or in other words, we are juggling
with two oranges instead of one.
But from the amateur's point of view, in which the prob-
lem involved in the design of a transmitter is that of getting
the greatest range for the least money, the master oscillator
is objectionable in that it requires an additional tube. It
is true that this master tube need only be large enough to
supply the losses in its own oscillating circuit and those of
the grid circuit of the main tube, and this is a fortunate
feature; nevertheless the system has not become so popular
with the amateur as the self-excited oscillator. This chap-
ter will be devoted to both systems, but will reflect the
greater importance of the circuits employing but one os-
cillator.
40. Fundamental Circuits for Self-excitation. Such cir-
cuits are briefly referred to as oscillating audion circuits, and
7
98 RADIO TELEPHONY
the various types are usually designated by the names of
their inventors. The fundamental principle is .the same in
all of them, and has already been explained. As we noticed,
the circuit may be thought of as consisting of: (1) a power
source, or plate ("B") battery; (2) audion; (3) load circuit
(transmitting antenna); and (4) means for feeding back
power to the grid circuit for its excitation. In order to
bring out the fundamental form of these circuits all auxiliary
apparatus, such as the filament heating source, biasing bat-
teries, grid condensers, etc., will be omitted and the circuit
will be indicated in three ways: one in which the connec-
tion of the load circuit and its form is emphasized; a second
in which the actual feed-back connections are added (this
is the form in which the circuits are usually drawn); and
finally, the substitution of the actual transmitting antenna
for the load circuit will be shown.
(a) Meissner Circuit. This is the most general and flexible
circuit, and is named after Dr. A. Meissner of the Telefunken
Co. of Berlin. It is extensively employed by the General
Electric Co. in their commercial transmitters, and has
proved to be very convenient, particularly for aircraft in-
stallation where antennae of widely varying characteristics
are encountered. Referring to Fig. 68, it will be seen that
the load circuit is magnetically coupled to the plate circuit
of the tube; and to get the proper phase relation and easy
adjustment for feed-back, or excitation, this circuit is like-
wise magnetically coupled. The chief merit of the circuit
resides in its flexibility, and by means of the transformer the
transfer of maximum power from any tube to any load
(antenna) resistance can be arranged for. The adjustment
of the feed-back is also conveniently made, and does not,
CONSTRUCTION OF THE TRANSMITTER 99
as in the case of the Hartley and Colpitts circuits, depend
upon the voltage drop across a reactance in the load circuit.
(b) "Tickler Coil" Circuit. If either of the transformers of
the Meissner circuit are omitted, there is derived a form of
LOAP CIRCUIT
Fig. 68. Schematic diagram of the Meissner circuit for obtaining self -excited
oscillations with the audion.
circuit which has come to be known as the "tickler" coil cir-
cuit in continuation of a humorous reference made to the
unnamed circuit by Mr. George H. Clark, of the Navy De-
Fig. 69. Schematic diagram of the "tickler coil" circuit for obtaining self-
excited oscillations with the audion.
partment, some years ago. The two forms obtained by
eliminating either coupling are shown in Fig. 69 (a) and (i).
100
RADIO TELEPHONY
At (c) the connection of the antenna to form (b) is shown.
The reader will be able to devise the proper antenna connec-
tion in the case of form (a), so this is not indicated. These
are obviously less flexible than the Meissner circuit from
which they are derived. The first form (a) has been used
extensively for generating oscillations, and for obtaining
regeneration in receiving, an application to which we shall
come in Chapter VI.
(c) Hartley Circuit. In this circuit, invented by Mr. R. V.
L. Hartley of the Western Electric Co., inductive coupling
is entirely eliminated, the connection -of the load circuit and
(b)
Fig. 70. Schematic diagram of the Hartley circuit for obtaining self-excited
oscillations with the audion.
the feed-back being made directly (Fig. 70). This circuit is
not so flexible as either of the preceding, and is particularly
hard to adjust for maximum power output with the low
capacity antennae used by amateurs at wavelengths of the
order of 200 to 300 meters. It has, nevertheless, many
useful applications in both the transmitting and receiving
stations, notably as a master-oscillator circuit and as a
local oscillator hi "heterodyne" reception. In both of these
applications it is possible to use the large capacity at C
which the circuit demands for best operation.
CONSTRUCTION OF THE TRANSMITTER 101
(d) Colpitts' Circuit. By exchanging the reactances in the
Hartley circuit, that is, by exchanging the inductance and
capacity, we derive a form of circuit, invented by Mr. E. H.
Colpitts of the Western Electric Co., which is shown in Fig.
71. This circuit is a very popular one in amateur circles,
and is deservedly so, for it is practically better adapted for
the excitation of the low capacity antennae. It will be seen
from Fig. 7 1 that the antenna is substituted for the capacity
C 2 in constructing the actual transmitting circuit (c).
(e) Armstrong Tuned Plate Circuit. Reversed Feed-back
Circuit. The principle of this circuit is different from any
yet described, in that the feed-back to the grid circuit
(a)
(b)
,r- co
Fig. 71. Schematic diagram of the Colpitts circuit for obtaining self-excited
oscillations with the audion.
takes place through the tube itself by means of the small
condenser formed by the grid and plate electrodes and their
associated connections (batteries, etc.). This is one of the
circuits invented by Mr. E. H. Armstrong, who first studied
the feed-back action in audion circuits experimentally and
gave a qualitative explanation of it. The most general
scheme is shown in Fig. 72 (a). The frequency of the
generated oscillations is determined principally by the con-
stants of the grid circuit, L g C g , although the apparatus in
the plate circuit does have some effect upon it. This is a
102
RADIO TELEPHONY
very valuable feature of the circuit, for by connecting the
antenna to the plate circuit, as shown at (c), the change in
its constants by mechanical swinging, etc., will have a
smaller effect upon the wavelength emitted than in those
circuits where it is directly associated with the circuit con-
stants. Thus we secure to some extent the constant fre-
quency advantage of the master-oscillator system which is
so important in c.w. work. This circuit is sometimes
referred to as the reversed feed-back circuit, although the
Fig. 72. Schematic diagram of the Armstrong tuned-plate, or "reversed feed-
back" circuit, for obtaining self-excited oscillations with the audion.
designation is a little fantastic. The mathematical theory*
of this kind of feed-back action shows that the feed-back
effect increases as the wavelength is shortened, and depends
also upon the grid-plate capacity, so that the action may
*This was first formulated and published by Dr. J. M. Miller, then
Physicist at the Bureau of Standards, and the interested reader is urged
to consult his paper, "The Dependence of the Input Impedance of the Three-
electrode Vacuum Tube Upon the Load in the Plate Circuit"; Scientific
Paper of the Bureau of Standards, No. 351. (Copies may be obtained from
the Supt. of Documents, Govt. Printing Office, Wash., D. C., for 5 cents.)
Without having had knowledge of this publication the author examined the
circuit theoretically (Physical Review (2), Vol. 15, p. 409, May, 1920) and
published some curves which show the action very clearly.
CONSTRUCTION OF THE TRANSMITTER 103
often be improved and controlled by connecting an extra
variable condenser of very small capacity (max. about
.0001 Mfd.) between these electrodes, as shown by the
dotted lines in Fig. 72 (a). This condenser should be
capable of withstanding the high voltages existing between
grid and plate during oscillation.
The Armstrong tuned plate circuit is particularly val-
uable for regenerative reception, which will be considered in
Chapter VI.
41. Electrical and Mechanical Data of Power Tubes
Suitable for Amateur Use. Audions have been built in
this country capable of handling as much as 10 k.w. of
power; and tubes rated at 5, 50 and 250 watts are available
on the market for amateur (experimental) use. Inter-
mediate powers may be obtained by paralleling tubes of
the same type. A general notion of the signalling ranges
attainable with these powers, working by the c.w. method
(not radio telephone), may be obtained from the following
table:
TABLE II
APPROX. C.W. SIGNALLING RANGE WITH VARIOUS POWERS, USING MODERN
RECEIVING APPARATUS
Power.
Antenna
Current.
Range.
5 w.
0.45 amp.
400 miles
10 "
0.64 "
700 "
50 "
1.4 "
1500 "
100 "
2.0 "
2100 "
250 "
3.0 "
3000 "
The antenna currents are computed on the basis of an
oscillator efficiency of 50%, for a 14 ohm antenna resistance
104
RADIO TELEPHONY
PLATE
("T" antenna with flat-top total length equal to its height,
operating 20% above fundamental). These currents are
smaller than those obtained by the usual method of opera-
tion, i. e.j at a wavelength con-
siderably higher than the fun-
damental, but as I have pointed
out, represent more efficient ra-
diation and will cover a greater
distance with a given amount of
power on any wavelength. The
range estimates are very rough,
and may vary several hundred
per cent., but they will furnish the
inexperienced reader with a basis
for calculation. It is to be noted
further that these are not radio
telephone ranges, and require
considerable abridgment to cover
this type of modulation and re-
ception.
A brief specification of the
electrical and mechanical char-
acteristics of the available types
of power tubes may be useful to
the amateur in planning his
transmitter, and is contained in
the following paragraphs :
250-Watt Radiotron; UV-204.
-The UV-204 Radiotron is, with the exception of the
filament, very similar to the "Pliotron," or "P" tube
developed by the General Electric Co. and supplied to
GRID
FILAMENT
Fig. 73. Type "P" power
tube, prototype of the 250-walt
Radiotron; UV-204.
CONSTRUCTION OF THE TRANSMITTER 105
the United States Navy during the war. It is the most
powerful tube marketed for experimental purposes and
is of the very best construction. The plate is of tung-
sten-molybdenum which, as is well known, possesses
the useful property of absorbing gases when heated
("clean-up effect") thus improving the vacuum and not,
like many other metals, spoiling it by the emission of oc-
cluded gases. For this reason the plate may be permitted
to assume a bright red heat during operation without danger.
The general rugged appearance of the tube is illustrated in
Fig. 73, where the connections to the grid, plate and filament
are also designated. A special set of contacts for mounting
the tube are marketed, in which are also incorporated the
safety gaps which will be referred to in Art. 45 on "Pro-
tective Measures." The tube may be mounted vertically;
or horizontally in such a way that the plane of the electrodes
is vertical, with the seal-off tip of the glass bulb pointing
down. If the filament sags this will prevent its coming too
close to the grid, a circumstance which would reduce its
life, and in some cases result in the tube's destruction.
The important electrical and mechanical data of this
tube are as follows:
Electrical and Mechanical Data, 2 50- Watt Radiotron; UV-204
- Overall dimensions 5 x 14J^ in.
Base Special end mountings
Voltage of filament source 12V.
Filament terminal voltage 11 V.
Filament current 14.75 amp.
Plate voltage 2000 V. normal
Amplification factor 25
Plate resistance 4000 ohms min.
Plate current 25 amp.
Watts output 250 normal
io6
RADIO TELEPHONY
50-Watt Radiotron; UV-203. This tube is very convenient
for medium powered sets, and is illus-
trated in Fig. 74. It demands a spe-
cial socket of the 4-prong bayonet type
which can be easily constructed or ob-
tained from the manufacturers, is pref-
erably mounted vertically, but can be
mounted horizontally with the long
dimension of the elliptic plate verti-
cal. The plate may be permitted to
assume a bright red color in opera-
tion, and in special cases overloads
and higher temperatures may be tol-
erated for short periods. For tubes
in which the plate current is high, as
in this case, the filament is best
heated from an a.c. source; this will
prevent one-half the filament from
carrying more, and the other half
less than its share of current. The
important electrical and mechanical
data of this tube are as follows :
Fig. 74. Illustrating
Radiotron; UV-203; a 50-
watt power tube.
Electrical and Mechanical Data, 50-Watt Radiotron; UV-203
Overall dimensions 2 x 7 *A in.
Base Four prong special
Voltage of filament source 12V.
Filament terminal voltage 10 V.
Filament current 6.5 amp.
Plate voltage 1000 V. normal
Plate current 15 amp.
Amplification factor 15
Plate resistance 4000 ohms
Watts output 50 normal
CONSTRUCTION OF THE TRANSMITTER 107
5-Watt Radiotron; UV-202. These tubes are very pop-
ular for low-powered sets, particularly on account of their
cheapness and the ease with which
they may be operated in parallel
to obtain higher outputs. They
are also useful for power amplifi-
cation at the receiving station for
the operation of "loud speakers."
The base is a 4-prong bayonet of
standard type, and the tube may
be mounted in either a vertical or
horizontal position. They are op-
erated with the plate at a bright
red heat and will stand consider-
able overloads. In many cases
power outputs as high as ten watts
may be obtained, but in these cir-
cumstances they should be oper-
ated with extreme caution and
carefully watched for the development of portentous symp-
toms. The general appearance of the tube is shown in Fig. 75
and the chief electrical and mechanical data are as follows :
Electrical and Mechanical Data, 5 -Watt Radiotron; UV-202
Overall dimensions 2J/g x 5 in.
Base Four prong standard
Voltage of filament source 10 v.
Filament terminal voltage 7.5 v.
Filament current 2.35 amp.
Plate voltage 350 v. normal
Plate current 045 amp.
Amplification factor 8
Plate resistance 4000 ohms
Watts output 5 normal
Fig. 75. Illustrating 5-watt
Radiotron; UV-202.
io8
RADIO TELEPHONY
5-Watt Western Electric "E" (V.T.-2) Tube.
tubes were made in accordance with the Western Electric
Co.'s 1918 practice with coated platinum filaments, and on
account of gas effects which have not been completely
eliminated, must be operated with great care. At the rated
heating current the filament will glow with a bright red
color; and the plate should not be allowed to become hot,
or at least to assume more than a barely perceptible red
Fig. 76 Western Electric 5-watt type "E" tube (CW-931 or V.T.-2).
color. A plate ammeter is the best protection (this should
itself be protected from being burnt out by a flash-over in
the tube by a fuse made out of a piece of gold-leaf y 1 ^ inch
wide), and should be watched to see that the plate current
per tube does not exceed 40 milliamperes. In some cases
bright spots will develop on the filament, and this may be
generally regarded as a danger signal, portending flash-over
and destruction of the tube. In starting to operate these
CONSTRUCTION OF THE TRANSMITTER 109
tubes after a period of inactivity it is advisable to first
reduce the filament current and plate voltage about 50%
and to bring the output up to full strength gradually. This
is good practice in the case of any tube that has been idle,
whether gas effects are suspected or not. The tubes are
fitted with a standard 4-prong bayonet base and are mounted
preferably in a vertical position. Their appearance is in-
dicated in Fig. 76 and the chief mechanical and electrical
characteristics are as follows:
Electrical and Mechanical Data, 5-Wati Western Electric "E" Tube
Overall dimensions 2 x 3X in.
Base Four prong standard
Voltage of filament source 10 v.
Filament terminal voltage 7 v.
Filament current 1.35 amp.
Plate voltage 350 v. max.
Plate current 040 amp. max.
Amplification factor 7
Plate resistance 5000 ohms
Watts output 5 max.
Smaller Powers from Overloaded Receiving Tubes. In ad-
dition to the power tubes listed above, the use of overloaded
receiving tubes is often resorted to for c.w. transmission
and radio telephone work over very short distances up to
fifty miles. Some of the tubes described in Chapter VI in
connection with receiving apparatus are suitable for the
purpose. Such a transmitter may be put together very
easily and will appeal to the amateur who wishes to make
his debut and experience the thrills of two-way working with
as little expense as possible. Any of the circuits described
in later sections may be used, with suitable changes, of
course, for the lower power; or the tubes may sometimes be
no RADIO TELEPHONY
used in their native receiving connections by increasing the
plate battery voltage.
42. Collection of Transmitting Circuits with Detailed
Description. We have given in Art. 40 a collection of the
fundamental oscillating circuits suitable for transmitting,
and in order to emphasize in these diagrams the form of the
circuits, the auxiliary apparatus which is necessary in the
practical applications was omitted. In this section a collec-
tion of circuits will be given, following closely the develop-
mental scheme of Art. 40, but in which the actual connec-
tions and all auxiliary apparatus with the exception of the
plate power source and the modulation system will be in-
dicated. The various circuits will be accompanied by such
details of construction and specification of proper circuit
constants, as are considered useful. These data should be
accepted by the reader suggestively, and represent averages
with which he may begin his own experimenting. The wide
variation in antennae, and often in the power tubes them-
selves, render any more rigorous specification of doubtful
value. Besides, the experimenter in many cases will want
to make use of materials and apparatus on hand.
The power supply for the plate circuit and for heating the
filament of the tube is a subject of sufficient importance and
scope to require separate treatment, and this is attempted
in the next chapter; so that in these diagrams the actual
power connections will be omitted, the power sources being
simply represented symbolically. This will simplify the
diagrams a great deal.
TRANSMITTING CIRCUITS
in
112
RADIO TELEPHONY
TRANSMITTING CIRCUIT NO. 1
MEISSNER ARRANGEMENT
LO
=3 i
=>
CD
C?
=>
i 4- c >
.fl
T
ii C|
1 ;&W \
Co
LS ,r tr~
_f t jT-f}!
T- GROUND OR IS -1 L
- j| [4 J 1 jtj 1
/
COUNTERPOISE PLATE
FILAMENT CENTER TAP)
,
J
SOURCE
SOURCE ON TRANSFORMER
(a)
(*)
(b) Showing method of connecting to center-tap of filament transformer when
using a. c. for filament heating.
DETAILED DESCRIPTION
Cir-
Unit.
General
Description.
Power.
cuit
Sym-
bol.
5 w.
50 w.
250 w.
L.
Antenna inductor
see Art
. 44 (a)
Anter
den
ma series con-
ser
f.0003 to X
\.003 mfd./
8000 v.
8000 v.
8000 v.
L K
Grid coupling coil
4" dia.,
No. 22 d.c.c.
20 t.
15 t.
lOt.
L p
Plate coupling coil ....
4" dia.,
No. 22 d.c.c.
30 t.
30 t.
30 t.
C , C
Grid and plate tuning
con
densers . . .
variabl
e .0005
mfd.
max.
800 v.
1500 v.
3500 v.
Ci
Grid condenser
.002
mfd.
1500 v.
1500 v.
1500 v.
Si
Grid
resistance
/5000 to
\ 10,000 ohms
o
Fi
rt m
ent rheostat
2 a
8 a.
16 a.
A
Anter
ma ammeter
0-2 a.
0-4 a.
0-8 a.
CONSTRUCTION OF THE TRANSMITTER 113
CIRCUIT NO. 1
MEISSNER ARRANGEMENT
This is considered the best circuit for reasons which have already been
given. It is adapted as shown for either direct ground or counterpoise.
The antenna series condenser is intended to counteract the effect of the
coupling inductance L in raising the wavelength above the fundamental,
and may be constructed of glass or mica as explained in Art. 44 (6). If a
counterpoise is used and is of the proper capacity, it may be substituted for
this series condenser; in this case its capacity is preferably large enough
to resonate with an inductance in L barely large enough to secure efficient
energy transfer to the antenna, at the fundamental wavelength of the an-
tenna. The filament circuit may be grounded as shown; and when an a.c.
step-down transformer is used to furnish the heating current, the center tap
of the winding may be grounded and the grid circuit and negative terminal
of the plate source may be connected to it. This will diminish the voltage
fluctuation in these circuits. (See diagram at right of circuit.) Remember
also that some part of the antenna coil L (the bottom when the system is
vibrating at the fundamental) is at a high potential with respect to earth and
there will be a capacity current through the small condenser formed by the
windings of L and of L g , L p , to ground. To avoid dielectric loss, the
capacity of this path should be small, that is, the windings should not be too
close together, and all unnecessary insulation and poor dielectrics should
be eliminated. This is especially important if C is small. The condensers
C p and C g are not necessary, but merely a convenience for tuning. Both
L g and L p should be coupled to the antenna coil rather than to each other,
especially when these condensers are used; this will reduce the tendency
for a short wave oscillation in their circuits in which the antenna does not
take part. The purpose of the grid bias resistance Ri and the grid condenser
Ci is to increase the efficiency by regulating the grid voltage during the posi-
tive cycle, and to secure a proper average negative bias during oscillation.
The circuit constants specified in the table are computed for an oscillator
efficiency of 50% and for an antenna resistance of from 5 to 20 ohms at 200
to 300 meters.
RADIO TELEPHONY
ll
8 E
fl 6
o E
"a
W D.
II
S ^ "S
o u c
1 00 8 fe
8 .s c
S a .5 -S
! I1 1
.J "C T3 -O
c 'C "C 'C
< O O O
CONSTRUCTION OF THE TRANSMITTER 115
CIRCUIT NO. 2
"TICKLER COIL" WITH INDUCTIVE GRID COUPLING
Three arrangements are shown, in which: (a) a direct ground connec-
tion is made, and power is supplied by connecting the plate across the series
condenser C ; (b) a counterpoise is substituted for the antenna series condenser,
the filament circuit being grounded to complete the path for the plate alter-
nating current; (c) a direct ground is used, and the plate is connected across
the antenna coil L instead of across the condenser C , or the counterpoise.
(At the fundamental it makes no difference, except that the connections to
L g may need reversal, whether the plate is connected to L or C .) In ar-
rangements (a) and (6) a fuse is inserted in the plate circuit to protect the
antenna and plate circuit ammeters (latter not shown) in case the antenna is
accidentally grounded at some point. In both (a) and (b) the antenna- is at
the d.c. potential of the plate supply, and great care should be exercised not
to touch this part of the circuit without first disconnecting the plate source.
The ground connection of the filament in (b) need not be elaborately made,
as in the case of an antenna ground, for this carries only an alternating plate
current of the order of a fraction of an ampere. Notice that the radio fre-
quency choke X, whose function is to prevent the shorting of the output
circuit by the plate source, is in parallel to the antenna condenser, C , or the
counterpoise. Ordinarily the value assigned (3 milli-henries) will be resonant
with this capacity at wavelengths from 2000-3000 meters, and unless the
counterpoise capacity is unusually low there will be no troublesome resonance
effects. In the third arrangement (c), the antenna circuit is at ground poten-
tial, but the filament circuit will be at the d.c. potential of the plate source;
hence the same caution in touching this circuit should be observed as in the
case of (a) and (6). About 60 micro-henries inductance will be needed be-
tween the plate tap and the ground in the case of (c) ; and a capacity of about
.000185 Mfd. in the condenser C , or counterpoise, in the case of (a) and (b),
for an antenna of 15 ohms resistance at 200 meters with any of the three
types of audions described in Art. 41. This a very small capacity, and would
seem to render the arrangements (a) and (b) inconvenient for short wave-
lengths with the tubes and antenna resistances we have contemplated in the
design. This is, however, best determined by direct experiment.
116
RADIO TELEPHONY
s
Q
W
J
<!
i
it
q
> >
^ > >
er
pt
.s
"a .S
I
CONSTRUCTION OF THE TRANSMITTER 117
CIRCUIT NO. 3
"TICKLER COIL" WITH INDUCTIVE PLATE COUPLING
These circuits are very similar to the preceding, differing in the use of
inductive antenna-plate coupling instead of antenna-grid coupling. Three
arrangements are shown in which: (a) a direct ground is used and the excit-
ing voltage for the grid is obtained from C ; (b) a counterpoise is substituted
for the antenna series condenser C , the filament being grounded to complete
the circuit; (c) a direct ground is used, but the grid voltage is obtained from
the antenna coil L . The purpose of the choke X is to keep the r.f. currents
out of the circuit containing the grid biasing resistance R\. Without this
choke there is a loss of about 20 watts in a 5000 ohm bias resistance using
the 250 watt tube (8%); by its use this loss is reduced to 0.5 watt (0.2%).
In the case of arrangement (c), inductances of the orders of 30, 20 and 10 micro-
henries will be required between the grid tap on the antenna coil and ground,
for the 5, 50 and 250 watt powers respectively (15 ohm antenna at 200
meters) ; while the corresponding capacities in the case of (a) and (b) required
will be .00037, .00056 and .0011 mfds. For this reason the arrangement (b)
will generally be more suitable than the corresponding arrangement (b) of Cir-
cuit No. 2, for use with rather low capacity counterpoises.
n8
RADIO TELEPHONY
TRANSMITTING CIRCUIT NO. 4
HARTLEY ARRANGEMENT
(a). Filament at (-) plate (d.c.) (b). Both L and filament grounded,
potential.
DETAILED DESCRIPTION
Cir-
cuit
General
Power.
Sym-
bol.
Unit.
Description.
5 w.
50 w.
250 w.
c:
Antenna inductor
Antenna series con-
denser
see Art. 44 (a)
f.0003 to \
8000 v
8000 v
8000 v
Grid condenser
\.003 mfds.J
002 mfd
1500 v
1500 v
1500 v
Ri
Grid resistance
(5000 to
Tuning condenser
Filament rheostat *
\1 0,000 ohms
variable, max.
.0005 mfd.
1000 v.
2 a
2000 v.
8 a
5000 v.
16 a
A
Antenna ammeter . . .
2 a.
4 a
8 a
Plate isolation con-
denser
.002 mfd.
5000 v.
5000 v,
5000 v
X
R.F. choke coil
see Art. 44 (d)
CONSTRUCTION OF THE TRANSMITTER 119
CIRCUIT NO. 4
HARTLEY ARRANGEMENT
This circuit is not well adapted for use with a counterpoise when the
filament supply is obtained from 110 volt a.c. mains through a step-down
transformer, on account of the losses which would occur in the condenser
formed by the transformer windings, this capacity being in shunt to that of
the counterpoise. For the same reason . the antenna series condenser is
placed above the antenna coil as shown. In the arrangement (a) the antenna
coil and its connections are at the d.c. potential of the plate source, and
should be shielded from accidental contact. A condenser is required to by-
pass the antenna current around the plate source, and should be made of
glass or mica and have a capacity of at least .002 mfd. Paper condensers
will not do. The tuning condenser 3 serves to reduce the number of turns
required between the plate tap and ground tap on the coil L (which may be
quite large) and is a convenience. The radio frequency choke coil X is
inserted to prevent the short-circuiting of the output circuit by the plate
source. The proper inductance between the plate and ground taps on the
coil L is about 60 micro-henries for all tubes (used singly, not in parallel);
between the grid tap and earth about 40, 25 and 15 micro-henries for the
5, 50 and 250 watt tubes respectively; and just enough inductance should be
included between the antenna and ground taps to secure maximum energy
transfer.
120
RADIO TELEPHONY
TRANSMIT'
cOLPir
T
C4 i
PING CIRCUIT NO. 5
fS ARRANGEMENT
T
<.$
^^+ V
- j jg T x " /^^VonnRN
COUNTERFOISE " A<
^kf^C
'=r ' r^y
-V "i*
Filament grounded. Insertion of isolation condenser at (*) will keep plate
(d.c.) potential off L .
DETAILED DESCRIPTION
Cir-
cuit
Sym-
bol.
Unit.
General
Description.
Power.
5 w. 50 w. 250 w.
1
Ri
A
*
X
Antenna inductor
Grid input condenser . .
Plate output condenser
Grid condenser
see Art. 44 (a)
3000 v.
.001 mfd.
.002 mfd.
/5000 to
\ 10,000 ohms
variable, max.
.0005 mfd.
.003 mfd. .0022 .0015
5000 v. 5000 v. 5000 v.
1500 v. 1500 v. 1500 v.
750 v. 1500 v. 2500 v.
2 a. 8 a. 16 a.
2 a. 4 a. 8 a.
5000 v. 5000 v. 5000 v.
Grid resistance
Plate tuning condenser
Filament rheostat ....
Antenna ammeter ....
Isolation condenser . . .
R.F. choke coil
".002" mfd."
see Art. 44 (d)
CONSTRUCTION OF THE TRANSMITTER 121
CIRCUIT NO. 5
COLPITTS ARRANGEMENT
The antenna is usually inserted in place of the condenser C 2 in the Colpitts
arrangement (see Fig. 71). In this connection it is possible to work only at
wavelengths above the fundamental, for it is only here that the antenna
acts as a condenser. At the fundamental it is a pure resistance, and below
the fundamental, an inductance. So it is evident why many amateurs, using
this circuit in its canonical form (Fig. 71 (c)) have not been able to realize
the advantages of operation at the fundamental. It is possible to operate at
the fundamental by inserting a condenser C 2 in series with the antenna as
shown. The antenna is a pure resistance and Ci provides the capacity effect
essential to the circuit; otherwise the antenna short-circuits the plate output
circuit. The exciting voltage for the grid is derived from C\. Both of these
condensers should be free from losses since they are in series with the antenna
circuit. The tuning condenser C p is a convenience, and is intended to reduce
the number of turns between the plate tap and Ci on the antenna coil re-
quired for efficient energy transfer; with fundamental-operation where the
antenna resistance is quite high, the required inductance is quite high also
(70 micro-henries), so the use of this condenser is recommended. Arrange-
ment (a) is suitable for direct ground; (6) for use with a counterpoise, the
counterpoise capacity taking the place of C\. The capacity of the average
counterpoise will be somewhat lower than the values of C\ prescribed, so the
grid voltage will be too high. It may be reduced by connecting the grid to a
tap on the antenna coil instead of to the bottom of the coil as shown in the
diagram. As the tap is moved from the bottom toward the plate tap, the
voltage is reduced. If the proper capacity is available this will be unneces-
sary. The purpose of the choke X is as usual, to prevent the short-circuiting
of the output circuit by the plate source. The antenna coil and its con-
nections will be at the d.c. potential of the plate source; and to obviate this
an isolation condenser of .002 mfd. capacity (5000 v.) may be inserted at (*).
An inductance of from 15 to 25 micro-h. will be required between the antenna
tap on L and G, with the capacities specified. The data are based upon an
antenna of 15 ohms resistance operating at its fundamental near 200 meters.
122
RADIO TELEPHONY
o
p
5 ,
p
$ % =g -O
i v
_^
II r*,J
ll I 3 I
O ir>
S I
o I
Si'-" *
^^SoS S
'S: rtOO O
1.i
a * S 3
^ ^ o 'o
o o
CONSTRUCTION OF THE TRANSMITTER 123
CIRCUIT NO. 6
ARMSTRONG TUNED-PLATE, OR REVERSED FEED-BACK ARRANGEMENT
The feed-back voltage for the gtid is derived in this arrangement from
the plate circuit through the small capacity between the plate and grid; the
purpose of the small variable condenser C m (max. capacity = .0001 mfd.) is
to control it. This condenser may not be necessary, especially with the
higher powered tubes having a high amplification factor. The wavelength
of the generated oscillations is determined mainly by the constants L g , C g of
the grid circuit. The condenser of this circuit C g is made variable for con-
venience in tuning. There is normally no coupling between the antenna
coil L and the grid coil L g .
Two arrangements are shown: (a) for use with direct ground; and (&) for
counterpoise. In both the function of the condenser C is to give the antenna
a capacity reactance at the fundamental. It will be noted that the second
arrangement applies the method for combining a counterpoise and direct
ground described in Art. 26, in which the direct-earth tap S is set approxi-
mately at the voltage node. The isolation condenser C keeps the d.c. voltage
of the plate source off the antenna coil. The function of the r.f. choke-coil X
is, as usual, to prevent the short-circuiting of the output circuit by the plate
source.
124
RADIO TELEPHONY
I
<
Lo
C =
t.
A)
8
i
I
TI
w
L
|g.
IANSMITTING CIRCUIT NO. 7
^STER-OSCILLATOR ARRANGEMENT
POWER TUBE EXCITER
.s^^ r. - /=^K
c,
n
w
R,
F
I;*,
w\ ii
3 L! 1B
p
i?z
T
L
c =
c '
Ii c
? T
T |
1
|
P L '
L|_J L
*
BETA
r
[LED DESCRIPTION
Cir-
cuit
Sym-
bol.
Unit.
General
Description.
Power.
5 w.
50 w.
250 w.
Lo
C
%
Ci
Ri
X
C
L 2
u
R*
A
An
An
(
Pla
Pla
Gr
Gr
R.]
Os
<
Os
(
O
(
Fil
An
tenna inductor
tenna series con-
ienser
see Art. 44 (a)
f.0003 to \
\.003 mfd.J
4" dia.
variable,. 0005
mfd.
.002 mfd.
/5000 to
\ 10,000 ohms
see Art. 44 (rf)
.001 mfd.
6" dia.
6" dia.
8000 v.
30 t.
500 v.
1500 v.
2000 v.
7
7
2 a.
2 a.
8000 v.
30 t.
1500 v.
1500 v.
4000 v.
9
5
8 a.
4 a.
8000 v.
30 t.
3000 v.
1500 v.
8000 v.
10
4
16 a.
8 a.
ite couplin
ite tuning
id condens
d resistan
F. choke-c
dilating ci
ienser
g coil ....
condenser
er
ce
oil
rcuit con-
:illating circuit plate
oil
dilating circuit grid
oil
ament rhe
tenna ami
ostat . . .
neter ....
Note. The construction of coils LI
antenna inductor L . (See Art. 44 (a).
and L 2 is the same as that of the
)
CONSTRUCTION OF THE TRANSMITTER 125
CIRCUIT NO. 7
MASTER-OSCILLATOR ARRANGEMENT
This circuit is adapted for use with either direct ground or counterpoise,
or with any combination of direct grounds and counterpoises by the method
of Art. 26. The condenser C v is a convenience for tuning and may be omitted
if desired. An inductance of approximately 60 micro-henries will be required
in L p for any of the three types of tubes used singly, with an antenna resist-
ance of 15 ohms at wavelengths from 200 to 300 meters and an oscillator
efficiency of 50 per cent. The choke-coil X serves to reduce the grid loss by
preventing the flow of radio frequency currents through the biasing resistance
Ri. The master oscillator employs the Hartley circuit with two coils L\
and L 2 shunted by a condenser C. This condenser may be made variable
for convenience, or should at least be shunted by a variable condenser if
precise adjustment of the wavelength is to be made. The coils L\ and LZ are
constructed in the same way as the antenna coil as described in Art. 44 (d).
For the excitation of the 5, 50 and 250-watt tubes the master-oscillator tube
should be an overloaded receiving tube with 300 volts or so on its plate
capable of supplying 0.2 watt, a 5-watt tube, and a 50-watt tube in the
respective cases. For a general discussion of this circuit and instructions for
its adjustment see Art. 48.
126 RADIO TELEPHONY
43. Comparison of the Transmitting Circuits. A critical
and comparative examination of the circuits collected in the
preceding pages may be undertaken either from the point
of view of economy or from that of electrical performance
and ease of operation and adjustment. A useful review
would embrace both considerations. But since the various
circuits differ only by a few coils and condensers, which are
cheap or easily constructed, the economical features are of
small importance and this survey may proceed upon an
entirely electrical basis.
The possible exception to the last statement is the sepa-
rately-excited or master-oscillator system, Circuit No. 7.
Here, indeed, we are face to face with the economic matter
of providing an extra audion tube, the value of which is
but partly realized in actual antenna amperes. (The grid
losses of the main tube are seldom over 10% in a properly
operated system.) From an electrical viewpoint the cir-
cuit is superior to any of the self-excited type and a maxi-
mum of flexibility and freedom from frequency changes are
attained with it. The latter feature is most important, of
course, in the case of c.w. signalling. It is suitable for
direct earth connection, counterpoise, and any combina-
tion of direct earth connections or counterpoises. More-
over, the transformation ratio can be' easily adjusted for
maximum power transfer from tube to antenna under con-
ditions of widely varying antenna resistances and antenna
forms, an adjustment which proceeds with almost no atten-
tion to the grid circuit. Every amateur who contemplates
the installation of 50- or 2 50- watt tubes is urged to con-
sider the advantages of this arrangement.
The Meissner self -excited circuit, Circuit No. 1, is in point
CONSTRUCTION OF THE TRANSMITTER 127
of flexibility a good second to the master-oscillator circuit,
and from an economic point of view excels in not requiring
an extra tube. It is adapted for almost any type of antenna,
or antenna resistance, direct ground connection, counter-
poise, etc., and is without doubt the best of the self -excited
circuits for radio telephony. For c.w. work its chief
drawback is that the frequency may be changed by a swing-
ing antenna, and from this point of view only is it inferior
to any other circuit. In these circumstances and for this
method of communication, the Armstrong tuned-plate cir-
cuit is preferable to it.
Next in order of merit, for use with a counterpoise, the
"tickler" coil Circuit No. 3 (ft); Colpitts Circuit No. 5 (ft);
Armstrong tuned-plate Circuit No. 6 (ft); and "tickler"
coil Circuit No. 2 (ft), are recommended in the order given.
For direct earth connection, "tickler" coil Circuit No. 3
(a) and (c) ; "tickler" coil Circuit No. 2 (a) ; Armstrong tuned-
plate Circuit No. 6 (a) ; Colpitts' Circuit No. 5 (a) ; and the
Hartley Circuit No. 4, are recommended in this order. The
arrangements in which the antenna coil and its connections
are at the d.c. potential of the plate, Circuit No. 2 (a) and
(ft), and in which the filament is at this potential, Circuits
No. 4 (a) and No. 2 (c) should be carefully operated. In
many cases they may be made safe by means of an extra
r.f. choke-coil and isolation condenser. The method will
be gleaned from some of the other circuits.
44. Construction of Apparatus. The descriptions of ap-
paratus to follow are merely intended to serve as a guide,
and fairly substantial departures from the specifications
can generally be made with no danger of failure. When such
128
RADIO TELEPHONY
departures are likely to introduce trouble, a notation will be
made to that effect.
(a) Antenna Inductor, L . This inductor is inserted in the
antenna circuit primarily to provide a means for transferring
the power from the plate circuit of the audion to the an-
tenna. It should be constructed so that it will introduce a
minimum resistance into the antenna circuit; and this must
be attained by reducing to a minimum the high frequency
resistance of the conductor, and the
dielectric losses. We speak of dielectric
losses because they are more important
than is generally imagined. When oper-
ating at the fundamental, the part of
the coil nearest ground is at a high
potential (of the order of 1000 volts)
with respect to earth, and when operat-
ing above the fundamental the top of the
coil, as well, oscillates at a potential
higher than that of the earth; hence
there is danger of loss in poor dielectric
paths both between the ends of the coil,
and much more importantly, between
the coil and the earth. To reduce the
latter loss it is advantageous to surround the inductor and
the antenna series condenser C with a metal screen, which
is grounded, and split vertically to prevent the flow of eddy
currents. This will also prevent accidental contact with
these parts of the circuits, which in many of the circuits
are continuously at the d.c. potential of the plate source.
Such a shield is shown in Fig. 77 and may be made of any
piece of ordinary iron fly-screening, for there are no heavy
Fig. 77. Showing
method of reducing di-
electric losses by sur-
rounding antenna coil
and grid condenser .with
a grounded metal screen.
CONSTRUCTION OF THE TRANSMITTER 129
currents in it. The other type of loss, due to the high fre-
quency resistance, may be kept low by the use of copper strip
wound preferably flat in the form of a helix. A minimum
of insulating material should be used in the support of this
helix. Avoid winding the strip on insulating tubes, using
rather two or three longitudinal stiffeners like the upper one
in the helix shown in Fig. 78. An inductance sufficient for
most purposes will be obtained with 30 turns of 6 or 7 inch
diameter. Taps should be provided as indicated in some
of the circuit diagrams. Certain of these circuits, notably
the Meissner and "tickler"
coil arrangements, require one
or more coupling coils. These
coils may be wound with a
smaller wire because they do
not carry heavy currents, and
may be mounted inside the an-
tenna coil, being conveniently
arranged to be rotated so that
the coupling can be adjusted.
The proper way to arrange
this mechanically will occur to the reader. A typical antenna
inductor is shown in Fig. 78.
(b) Antenna Series Condenser, C . The purpose of this con-
denser is generally to neutralize the effect of L in raising the
wavelength above the fundamental, and since it is in the
antenna circuit and carries the antenna current, should be
designed to have low losses. A number of good condensers
are on the market; those in which a good grade of mica is
used are preferable. An inferior condenser in which the
losses are somewhat larger, may be constructed from glass
9
Fig. 78. Typical form of antenna
coil for audion transmitting circuits.
130 RADIO TELEPHONY
plates coated with copper or tin-foil. The proper capacity
will vary a great deal and must be determined experimentally.
In the tables the limits of this variation have been indicated
as .0003 to .003 mfd., these being considered adequate for
most situations, and a condenser adjustable within this
range may easily be designed as
follows :
The capacity in microfarads
Fig. 79.-Illustrating construe- Q & condenser made from
tion of glass-plate condenser.
positive plates, and n + 1 nega-
tive plates (see Fig. 79) is given by the following formula:
~ .000000365 Am .
C = j , (mfds.);
wherein A = area of the conductor in sq. in. ; d = thickness
of the glass plates in inches; n = the number of positive
plates (corresponding to which there are n + 1 negative
plates) and is the dielectric constant of glass, usually about
6. Thus supposing we have a number oi 5 x 7 inch glass
plates at our disposal, of thickness = .12 inch (about
-| inch), which may be coated with tin-foil, 4x6 inches (A
= 24 sq. in.); the capacity per positive plate will be:
C = 00000 36 J 2 X 24 X 6 = .000438 mfd, | '"" ; I
A consenser of 8 positive and 9 negative plates arranged so
that various numbers of positive plates can be included in
the circuit would probably cover the range required.
Other condensers in the antenna circuit, input conden-
sers, etc., as well as the grid and by-pass condensers fre-
quently indicated, can be constructed in the same way.
CONSTRUCTION OF THE TRANSMITTER 131
(c) Variable Transmitting Condensers. In many of the trans-
mitting circuits variable condensers are a great convenience,
especially in the situations indicated by the dotted lines in
the diagrams. Most of the variable condensers on the
market are designed for receiving purposes and will not
stand the high transmitting voltages. There are, however,
a few types which may be adapted for this purpose by in-
creasing the breakdown voltage. These are provided pref-
erably with large plates, half of which may be removed,
thus doubling the distance between plates and quartering
the maximum capacity of the condenser. If the use to which
the condenser is to be put demands a higher capacity than
this and a higher breakdown voltage, the condenser may be
filled with a good grade of transil oil (or any kind of pure
oil, castor coil, Russian white mineral oil, etc.) from which
the moisture has been removed by filtering through blotting
paper. This will increase the breakdown voltage from 2 to
5 times and the capacity in the same ratio. It also in-
creases the losses, so should not be used unless absolutely
necessary.
(d) Radio Frequency Choke-coil, X. This may be wound
with No. 28 enameled, silk or cotton covered wire, in a
single layer coil of about the following dimensions:
Diameter. No. Turns. Length of Winding.
V 500 Sy 8 "
3" 250 4y s "
It is advantageous, if a suitable condenser is at hand, to
shunt this coil with a capacity so chosen that the circuit is
in resonance with the transmitting wavelength; it will then
offer an impedance greater than the coil alone. A variometer
of sufficient range is also useful here and may be adjusted
132 RADIO TELEPHONY
so that with its own distributed capacity the above condi-
tion is satisfied. The operator will be able to tell when the
proper choking action is being obtained.
(e) Grid Resistance, RI. Resistances of from 5000 to
10,000 ohms are suitable for most purposes with single tubes.
When tubes are operated in parallel it will be necessary to
either reduce the resistance and increase its current-carrying
capacity, or to provide a separate resistance for each tube
(see (a) and (6), Fig. 80). Resistance units in very con-
venient form are marketed, and in view of the low prices at
which these are offered it hardly seems worth while to try
to make them. If, however, the reader wishes to attempt
this a resistance suitable for small tubes may be constructed
as follows:
Procure a short length of glass tubing of the type used
for mercurial barometers, of inside diameter about 2 mm.
with a stout wall about -J- inch thick. This is to be closely
packed with lampblack and provided with metal end-caps
for making connections. If a resistance bridge or "megger"
is at hand the proper length of tube and closeness of packing
for any desired resistance can be determined; otherwise
make the tube about 4 inches long for 10,000 ohms re-
sistance.* Carbon, graphite and carborundum rods may
also be found useful. In all these forms and in the lamp-
black resistance, the current-carrying capacity is low and
regular wire-wound resistances may be necessary with the
higher powered tubes. But for 5-watt tubes, in which the
grid current should not exceed 5 milliamperes, they should
be entirely satisfactory.
* This construction is described by Mr. A. J. Funk in the November,
1920 issue of Q S T, p. 20.
CONSTRUCTION OF THE TRANSMITTER 133
45. Protective Measures. Both the audions used for
power generation and the meters used in their circuits are
costly and cannot easily be repaired in case of accident to
them. For this reason proper precautions should be ob-
served in connecting up the circuits to see that every wire
is in its proper place before the power is applied. It is al-
ways advisable in putting a circuit into operation, either
for the first time or after a period of inactivity, to reduce
the voltages at least 50% and to bring the output up to its
full value gradually.
The use of a plate circuit milliammeter during the ad-
justment of the transmitter, so that the mean plate current
may be kept within safe limits, or to indicate the rise of
plate current if for any reason the audion stops oscillating
and is improperly bicfced, or to test the modulation in the
case of a radio telephone installation, is advantageous and
recommended. This meter should be protected from being
burnt out by excess current due to accidental grounding or
flash-over in the tube or its safety gap, by means of a small
fuse. A suitable fuse may be made by pasting some gold-
leaf on a sheet of writing paper, and cutting it up into strips
of the proper size. The width of these strips will vary from
6T to i inch depending upon the current to be interrupted.
There are many times when the voltages in the trans-
mitting circuit will rise momentarily to values considerably
higher than have been contemplated in the design of the
apparatus and tubes. Cases of this arise from: c.w. sig-
nalling with a key in the grid circuit when the voltage across
a d.c. generator in the plate circuit, for example, may rise
to the breakdown point of the insulation; and in ordinary
telephone modulation when the telephone transmitter is
i 3 4 RADIO TELEPHONY
given a sudden jar, or its circuit suddenly opened. In such
cases discharges may take place in the tube, particularly in
the higher power tubes and in parallel operation, which may
be disastrous, as, for example, in the case of an arc from plate
to filament. These surges may be regulated by means of
safety gaps, electrolytic cells or aluminum lightning arrestors.
Of these the safety gap is preferable, because of its low
capacity properties; electrolytic cells are suitable for con-
nection to the motor generator, if one is used, but should not
be connected to the radio circuits. The best position for the
safety gap is between grid and filament, as here it will cir-
cumvent discharge from plate to filament, thus preventing
destruction of the filament. The width of the air gap will
depend upon the type of tube used and will vary from -^
inch in the case of the 50-watt tube, to ^ inch for the 250-
watt type. For lower powers (5 watts) the need for protec-
tion is not so great.
In circuits where an automatic grid bias voltage is ob-
tained by means of a grid resistance and condenser, the bias
vanishes when the tube stops oscillating and the mean plate
current generally increases. This should be carefully
watched, especially if the tube is being operated above its
rated output, and the plate source disconnected at all times
when there is no evidence of oscillation.
46. Operation of Tubes in Parallel. Within reasonable
limits the paralleling of tubes of the same type to secure
greater output is feasible. Beyond a certain point, however,
the plate resistance becomes too low and the use of a higher
powered tube with larger plate resistance and amplification
factor is desirable. In order to secure a good cooperation
between paralleled tubes they should have as closely as
CONSTRUCTION OF THE TRANSMITTER 135
possible the same electrical characteristics. The uniformity
of commercial tubes of the same type will generally be
found adequate.
The first precaution to be observed in parallel operation
concerns the size and design of the plate circuit apparatus
and connecting wires. The ratio of transformation, that is,
the size of the plate coil in the Meissner circuit, and the
inductance between plate and filament taps in the other
arrangements, will have to be decreased; and in this case
the leads or connecting wires in the circuit should also be
shortened so that their inductance does not constitute a
larger percentage of the total inductance of this circuit.
The cross-section of the conductors, including the plate coil
itself, should likewise be increased in proportion to the num-
ber of tubes in parallel. To still further reduce the length
of the radio frequency part of the plate circuit, the by-pass
condensers shunting the plate source should be located as
close to the radio frequency apparatus as possible, and not
placed, for example, at the terminals of feed-wires stretched
across the room.
Sometimes a very high frequency oscillation takes place
between the tubes, which does not contribute to the useful
output and lowers the efficiency, and oftentimes the out-
put, of the set. This tendency may be checked by means
of small choke-coils (20 turns No. 28 wire, 1-J- inches diam-
eter) inserted in the grid leads, close to the grid, as shown
in Fig. 80 (c).
47. Adjustment of the Transmitter. The adjustment, or
"tuning-up," of the transmitter has two objects: (a) secur-
ing maximum power output from the tube, or maximum
antenna current at a given wavelength (preferably the fun-
i 3 6
RADIO TELEPHONY
damental); and (b) obtaining highest efficiency from the
oscillator, that is to say, the highest ratio of power in the
antenna to power supplied to the plate circuit. Both results
can seldom be obtained at the same time, or with the same
adjustment. The amateur will generally want to adjust
Fig. 80. Relating to the operation of power tubes in parallel: (a) with com-
mon grid condenser and resistance; (b) with individual grid resistances and
condensers; (c) showing insertion of small choke-coils to prevent ultra-radio-
frequency oscillation between tubes.
for maximum output regardless of what the efficiency may
be with this adjustment, so the manipulation or tuning-up
will be discussed here from this point of view.
Since the aim is maximum antenna current, an antenna
ammeter should be provided to gauge the current, or in
CONSTRUCTION OF THE TRANSMITTER 137
default of this some form of low resistance, sensitive resonance
indicator may be used, such as a low c.p. automobile lamp
which has been shunted with a copper strip or wire of
proper length for the power used. In addition, it is decidedly
convenient to have a plate circuit ammeter, so that at least
some idea of the variation of the efficiency may be had.
The adjustment of the self -excited system under con-
sideration comprises three principle operations as follows:
(a) Adjustment of the inductance and capacity of
the circuit for the desired wavelength;
(b) adjustment of the ratio of transformation, or
the coupling between the plate circuit and
the antenna for maximum output into the
antenna resistance at this wavelength;
(c) adjustment to secure proper exciting voltage, or
feed-back to the grid circuit.
Before commencing these adjustments reduce the plate
voltage somewhat, so that the tube will not overheat while
experimenting. Carefully check all connections before clos-
ing the power switches. If a modulating tube for telephony,
or apparatus for c.w. or i.c.w. signalling is provided, see
that this is disconnected from the oscillator or otherwise
properly disposed. In case the circuit refuses to oscillate
after the power has been applied, first try reversing the con-
nections to the grid; then make the coupling between the
grid and plate and load circuits as close as possible in those
circuits where inductive coupling is used. The starting of
the oscillation is generally indicated by a movement of the
plate circuit ammeter.
The adjustment for wavelength is first made using a
138 RADIO TELEPHONY
wavemeter or by listening-in on a local calibrated receiving
apparatus. It will be observed that the three adjustments
are not independent, and a constant readjustment of the
wavelength will be necessary in the course of the other ad-
justments, particularly in the less flexible circuits. A maxi-
mum of independence is attained in the Meissner arrange-
ment. In adjusting for maximum power output be careful
that this is not attained at the expense of a heavy plate
current. Remember also that maximum power does not
mean maximum current in the antenna at any wavelength,
but at any particular wavelength, and if operation at the
fundamental is essayed the output current may be very
small, probably one-fifth of that which could be obtained
by moving to a higher wavelength. Do not be deceived by
a large antenna current, and in any case do not go about
boasting of a 2 ampere antenna current with a 5-watt tube,
for this species of self-delusion will never increase your sig-
nalling range, and is besides a pernicious doctrine to spread.
Decide upon the wavelength, keeping as close to the funda-
mental as possible (if the fundamental is too low, build a
larger antenna), and manipulate the clips or couplings until
the largest antenna current is secured at that wavelength.
The grid input will not usually be critical and can be ad-
justed last.
48. Master-oscillator Systems. In this system the power
tube acts as an amplifier of the power supplied by the small
auxiliary exciting, or master, oscillator. The exciter must
develop sufficient power to supply the losses in its own
oscillating circuit and those of the grid circuit of the power
tube. It is difficult to predict exactly what the upper limit
of the losses in any power tube will be, for the grid current is
CONSTRUCTION OF THE TRANSMITTER 139
modified by the load in the plate circuit; an allowance of
from 2 to 10 per cent, will, however, cover most cases. Be-
cause the plate circuit of the tube reacts upon the grid cir-
cuit and changes its effective capacity and resistance, it is
advisable in order to minimize the effect of this upon the
frequency of the oscillations generated by the exciter, to
employ here some form of oscillating circuit in which a large
capacity is employed. The Hartley circuit is well adapted
for this purpose and is applied in Circuit No. 7. Then the
adjustments for maximum transfer of power from the plate
circuit of the power tube to the antenna can be made with
a minimum of embarrassment from frequency changes.
A suitable master oscillator arrangement is illustrated as
Transmitting Cirucit No. 7, and the description of the
apparatus is given in the accompanying table. The circuit
is obviously suitable for direct ground or counterpoise.
When operating above 350 meters, a frequency trap L'C'
tuned to a wavelength equal to half the wavelength of the
master oscillator, should be inserted in the antenna circuit
to suppress the second harmonic of half wavelength which
is emitted and may be disturbing to other stations on this
wavelength. In this circuit the ratio of inductance to
capacity should be small.
The adjustment of the master-oscillator system is ex-
tremely simple and easy, for the three adjustments (a),
(b) and (c) mentioned in Art. 47 are almost independent.
The antenna circuit is first tuned to the proper wavelength;
then the frequency of the exciter is adjusted by varying
either the taps on the coils LI or L 2j or the capacity at C
(this may be shunted by a variable condenser for greater
convenience) until the antenna current is a maximum. The
i 4 o RADIO TELEPHONY
exciting voltage for the grid of the power tube is obtained
by inductive coupling with Z, 2 or by tapping in on this coil
as shown in the diagram. Setting this tap roughly at its
best value, adjust the inductance L PJ capacity C p , and
coupling between the plate coil and antenna for maximum
antenna current. Then return to the grid circuit and re-
adjust the tap 5, and grid resistance and condenser (RiCi)
for best excitation. The last adjustments are not wholly
independent; also some readjustment of the wavelength of
the oscillator may have to be made during these manipula-
tions. But the experimenter will find a great deal of pleasure
in working with this circuit, particularly after experience
with the self-excited systems. It is in all respects a clean-
cut and convenient arrangement.
49. Modulation Methods for C.W. and I.C.W. Telegraphy.
While this book is devoted ostensibly to the radio tele-
phone, yet the methods and circuits are so closely allied to
those used in the continuous wave (c.w.) and interrupted
continuous wave (i.c.w.) methods of telegraphy that it
seems worth while to include a description of the small
modifications necessary for this type of signalling.
In c.w. work the continuous radio frequency output of
the set is broken up into dots and dashes by means of a
telegraph key. These signals are detected at the receiving
station by the heterodyne method, to be described in
Chapter VI. The i.c.w. method differs from this in that
the output is further broken up into small lumps at a rate
chosen to give a pleasing musical note to the signal when
received by the ordinary detection method. The com-
munication range in miles per watt radiated with the first
type of signalling is enormous compared with the range for
CONSTRUCTION OF THE TRANSMITTER 141
the older " spark" transmitters. But this does not apply
to the second method, and except for its selectivity advan-
tages, it is decidedly inferior to the " spark" method. This
contradicts some prevalent notions and dicta, but is never-
theless supported mathematically and by the experimental
results.*
For c.w. and i.c.w. working, the key is usually and
preferably situated in the grid circuit, except when a special
stability of frequency is desired. The exact connections are
indicated in Fig. 81.
The arrangement (a) is simpler, but the throttling action
with tubes having a low grid circuit resistance may be too
Ci
UJU
(a) (b)
Fig. 81. Showing connection of telegraph key to oscillator for c.w. signalling:
(a) arrangement for low power, (b) arrangement necessary with higher power to
prevent voltage surges in the circuits.
rapid and dangerous transients may result. The arrange-
ment (&), in which the key short circuits a condenser large
enough to the give, with the grid circuit resistance and
RiCi, a sufficient time constant, is preferable in such cir-
cumstances. This latter arrangement is especially indi-
cated for paralleled tubes where the effective grid resistance
may be quite low. A capacity of from ^ to 2 mfds. will be
necessary, depending upon the tubes employed.
* In making thrs statement, ordinary detection (with or without regener-
ation, but retaining the note of the signal) is contemplated. This type of
signal may also be heterodyned, and the range considerably increased thereby,
but the note is lost and it is no longer an i.c.w. signal.
i 4 2 RADIO TELEPHONY
Sometimes the key is inserted directly in the antenna cir-
cuit. This may be successful in the case of low powers, but
is bad practice and overheats the tubes. The modulation
may also be accomplished by producing a small change of
wavelength. For this system, two or more turns of the
antenna coil L or a small coil coupled to it, may be short-
circuited by the key. In the master-oscillator system, Cir-
cuit No. 7, this loop may be coupled to Li or L 2 , or one or
two of their turns shorted.
For i.c.w. work, in addition to the signalling key, some
means for interrupting or completely modulating the an-
tenna current at an audible frequency is required. A motor-
(I) M
S~^ ^~^ - *
B
rt*Q) w? ]
1,1.1.
Fig. 82. Two-buzzer scheme for use as modulator of output current for i.c.w.
telegraphy.
driven interrupter or "chopper" will be useful where large
currents are to be interrupted and a generous contact sur-
face should be available. For small currents, the double
buzzer scheme shown in Fig. 82 is frequently useful. The
purpose of the first buzzer (1) is to cause the armature of
buzzer (2) to vibrate, thus interrupting its contacts at the
desired rate.
These contacts (A and B) may be inserted in the antenna
circuit, or across one or two turns of the antenna coil Z, ,
or in the case of a radio telephone installation already pro-
vided with a modulator tube, may take the place of the
speech operated microphone. In the last connection it is
CONSTRUCTION OF THE TRANSMITTER 143
advisable to shunt the contacts with a variable resistance
to be adjusted until the proper modulation is obtained.
50. Methods of Modulation in Radio Telephony. As we
have previously pointed out, the function of the modulating
device in the radio telephone transmitter is to vary the
output current of the r.f. generator in accordance with the
low frequency variations of the sounds to be transmitted.
Many schemes have been devised for the control of the out-
put current, operating principally by three fundamental
methods as follows:
(a) Variable absorption in the load circuit (antenna) ;
(b) variation of the grid voltage of the generator
tube either directly or by modulation of a
master oscillator;
(c) variation of the plate voltage, or power sup-
plied to the plate circuit of the oscillator tube.
Omitting any hybrid schemes, of small technical importance,
we shall indulge in a critical discussion of these methods and
describe circuits for their application. It is not necessary
to remark that the spirit of this discussion will be utilitarian
rather than technical, and will be utilitarian in the special
sense of economy of apparatus and tubes. For a very ex-
cellent technical discussion of these matters the reader is
referred to Mr. E. S. Purington's paper on the "Operation
of the Modulator Tube in Radio Telephone Sets" published
as Scientific Paper of the Bureau of Standards, No. 423
(copies obtainable from Supt. of Documents, Govt. Print-
ing Office, Washington, D. C., for 10 cents).
51. Modulation by Power Absorption. The first of these
methods is mainly of historical importance and at least for
high powers (over 5 watts) is rapidly becoming obsolete.
144 RADIO TELEPHONY
An example of it is furnished by the classical picture of a
speech operated microphone inserted in series with the an-
tenna, or included in a circuit inductively or directly coupled
thereto. For best results the effective " talking resistance"
of the microphone should be equal to the antenna resistance;
and by effective talking resistance I mean the actual resistance
when the device is directly in the antenna circuit, and the
referred resistance (equal to the actual resistance divided by
the ratio of transformation, unity coupling) when it is
coupled to it. The purpose of coupling is in fact to reduce
the effective resistance of the average high-resistance mi-
crophone so that the above simple relation is satisfied.
This method is inherently a poor one, not only from the
viewpoint of distortion, but from that of efficiency as well.
For the amateur its attractiveness lies in its economy, and
for low powers (5 watts or so) fair results can be secured by
its use, in spite of many technical objections. In applying
the garden variety of microphone (borrowed perhaps from
some idle wire 'phone), connect it across a few turns of the
antenna coil and then increase the number of turns until
there is evidence of its overheating or the quality and
intensity of the speech as reported by a distant receiving
station, ceases to improve. Just how far the increase of the
number of turns can be carried out, and the effectiveness of
the microphone thereby improved will depend upon the
power to be modulated. Figure 83 shows the connections
contemplated. Modifications of this idea are frequently em-
ployed in which by amplifying the effect of the microphone,
higher powers can be successfully modulated. Examples of
such systems are furnished by a speech operated audion am-
plifier, and by the magnetic modulator developed by the
CONSTRUCTION OF THE TRANSMITTER 145
Radio Corporation of America. The latter device is of con-
siderable interest to amateurs because it has been put into
a thoroughly practical form, especially designed for short
wave work, and is available on the market.
52. Modulation by Grid Voltage Variation. The second
method of modulation depends upon a variation of the
average (biasing) grid voltage of the oscillator tube in a self-
excited system, or its equivalent, the modulation of the out-
put current of the exciter in the sepa-
rately-excited system. Like the pre-
vious method (modulation by absorp-
tion), this system flourishes mainly
on account of its economy of appara-
tus, for its principle is fundamentally
a poor one. In the self -excited os-
cillator this follows from the experi-
mental fact that the relation between
the grid biasing voltage and the an-
tenna current is not even approxi-
mately linear, as it obviously should
be for good modulation. In fact, the
output current is but slightly affected
by changes in the grid biasing voltage throughout a sub-
stantial range in which the oscillations are stable; and if the
grid voltage is reduced below this range the oscillations
cease altogether and the antenna current falls to zero. The
reader will easily see that such characteristics are unfavor-
able for faithful reproduction of the speech vibrations. The
.situation is, however, not hopeless and by careful adjust-
ment of the circuit and modulating voltage limits a tolerably
satisfactory operation is possible.
10
Fig. 83. Showing con-
nection of microphone to
antenna coil for modulat-
ing output current in
radio telephony (absorption
method).
146
RADIO TELEPHONY
A suitable circuit for grid modulation is shown in Fig. 84.
The parts and apparatus not essential to the modulation have
been omitted. The secondary of the modulation ^trans-
former T takes the place of the grid resistance R^ either
when it bridges the grid condenser as at (a), or is connected
between the grid and filament as at (b). In the latter case
it is necessary to insert the 3 m.h. radio frequency choke-coil
X in series with the secondary to prevent it from short-
circuiting the grid circuit for the radio frequencies. The
design of the transformer will depend upon the electrical
(a) (6)
Fig. 84. Circuit for modulation by variation of the grid voltage of the oscillator
tube.
characteristics of the tube, but since this system is not
suitable for higher powers than 5 watts, or possibly two or
three 5-watt tubes in parallel, the transformer described in
Art. 53 for the Heising modulation and the 5-watt tubes, can
be recommended here. Before taking the trouble to build
a special transformer, however, the reader should try an
ordinary "spark coil." Those of the "Ford" variety have
been successfully employed by many amateurs.
53. Modulation by Plate Voltage Variation. The third
method of modulation, by variation of plate voltage, plate
CONSTRUCTION OF THE TRANSMITTER 147
current, or plate power, excels by reason of the experimental
fact that a substantial proportionality exists between the
output current and the plate voltage over a wide range, and
also because there is no waste of the oscillator's power, as in
the absorption method. In the practical application of this
principle a voice voltage is superposed upon the d.c. voltage
in the plate circuit, causing the plate current and thus the
plate power to fluctuate at speech frequencies. A complete
fluctuation from zero current to double current entails an
amount of power approximately equal to that supplied to
the unvaried oscillator, hence the modulating device must be
capable of either furnishing this power or controlling its
supply from the plate source. For this reason, the micro-
phone in itself, although acting as an amplifier in its normal
connection, is yet incapable of controlling the power sup-
plied by tubes of 5 -watt and high ratings and its effect must
be amplified by means of another audion. This auxiliary
tube is for obvious reasons designated as the modulator tube.
In this role the modulator tube may be regarded in two
ways : as a speech controlled amplifier or generator of speech-
frequency power (the microphone and its battery serving as
an exciter), or as a speech operated resistance, which, in-
serted in series with the plate source and the oscillator tube
causes the voltage of the latter to vary. These points of
view are equivalent, although explanations may be found in
the literature of the subject which make a rather sharp dis-
tinction between them. Reference is made to the classi-
fication of "constant voltage" and "constant current" sys-
tems illustrated in Fig. 85.
The constant voltage scheme is shown at (a). Here the
oscillator tube is represented by the resistance R 2 , and the
148
RADIO TELEPHONY
modulator tube (regarded as a speech controlled resistance)
is represented as Ri. E b is the constant plate voltage.
There are many ways of viewing the action, having the
common result that the voltage across R 2 undergoes varia-
tion with the variation of R^ At (6), instead of the con-
stant voltage a constant current (7 b ) is available, and varia-
tions of Ri will in this case also cause a variation of the
voltage across R 2 . In the practical application the con-
stant current action in the common branch is secured by
inserting a large choke-coil in series with the d.c. plate
source. For maximum effectiveness the simple relation
(a)
Fig. 85. Illustrating schematically the so-called constant voltage (a) and con-
stant current (b) systems of modulation by variation of plate power.
stated for the absorption system that Ri = R 2 , must be
satisfied, which accounts mainly for the familiar prescrip-
tion that the modulator and oscillator tubes shall be elec-
trically identical. Thus if the oscillator consists of a bank
of ten 5 -watt tubes, the modulator should comprise a sim-
ilar bank of tubes.
Now adopting the other viewpoint, the modulator may
be thought of as a generator (excited by the microphone and
battery in its grid circuit) of speech-frequency power. This
power is to be supplied to the oscillator tube in place of the
steady power used when telephonic communication is not
undertaken. The oscillator tube acts therefore as the load
CONSTRUCTION OF THE TRANSMITTER 149
(approximately a resistance), Fig. 86, and, as is well known,
the maximum amount of power will be transferred to it when
its resistance is equal to that of the generator tube. This is
the conclusion already arrived at. The choke-coil X,
serves to prevent the short-circuiting of the output circuit
by the plate source in the same way that the 3 m.h. choke-
coil prevented this in the previous radio diagrams.
The reader will have his own preference of either the first
or second methods of viewing the modulator action. The
MODULATOR
TUBE
SCILLATOR
TUBE
Fig. 86.- Schematic diagram of the Heising system of modulation, in which
the modulator tube is depicted as a speech excited generator supplying speech-
frequency power to the oscillator tube.
latter is perhaps more harmonious with the other explana-
tions of the audion's action as a generator given elsewhere
in this book. Previously we discussed the generation of
r.f. currents; here the currents are of speech frequencies.
A practical system applying this method of modulation
was first described by Mr. R. A. Heising of the Western
Electric Co., a fact which accounts for its being referred to
as the "Heising system" of modulation.
The operation of this system of modulation is nicely illus-
trated in Fig. 87. This is a photograph* of the speech-
* For this very clear and interesting oscillogram I am indebted to Mr.
E. S. Purington, of the Hammond Radio Research Laboratory, Harvard
University.
150 RADIO TELEPHONY
frequency current variations which took place in a circuit
of the type, Fig. 86, when the word "Johnson" was spoken
into the microphone. The upper trace, labeled "Modulated
Load Current," represents the plate current of the oscillator
tube and the r.f. output current will be moulded in this
form; in the middle curve, "Modulator Current," the plate
current of the modulator tube is shown; and the lower trace
represents the current hi the common or choke-coil branch,
which remains practically constant.
Fig. 87. Photograph showing the operation of the Heising system of modu-
lation. Currents flowing in circuit of Fig. 86 when the word "Johnson" was
spoken into the microphone (Purington).
An ideal choke-coil would have an infinite impedance
throughout the entire range of important speech-frequencies
from 100 to 5000 cycles; but since it is impossible to build
coils without resistance, the practical attainment of an ex-
tremely high impedance is checked by the increasing heat
loss in the coil, due to the passage of the normal plate cur-
rent for the two tubes through it. Moreover, the gain in
power transfer from modulator to oscillator is not propor-
tional to the impedance of this coil, but falls off as it is
increased. A small calculation shows that a transfer of
CONSTRUCTION OF THE TRANSMITTER 151
6 HENRY CHOKE-COIL FOR HEISING SYSTEM OF
MODULATION
AIR G-AP
Arrangement of windings and core, showing air-gap for reducing normal
flux-density.
SPECIFICATIONS
Power Rating of Tubes
(Watts).
5 w.
50 w.
250 w.
Average direct current . . .
0.1 a.
03 a
06 a
Conductor (B & S )
No. 28
No. 24
No. 22
No. turns (total)
3000
3000
3000
Amount wire required
Core dimensions
3 Ibs.
3K" x \y 2 "
5^ Ibs.
4" x 1"
Core cross-section ....
V" x "V"
1" x 1"
Width of air gap
_i_" 8
_3_//
Resistance (d c )
77 ohms
64 ohms
44 ohms
Heat loss (watts)
.77
5 8
16
Core. No. 26 B. & S. silicon-steel laminations.
Design Data. Average permeability assumed = 3000; flux-density =
5000 lines/cm. 2 for magnetizing force = 1.75 c.g.s. units; conductor
allowance = 1000 circular mils/ampere.
Fig. 88. Data for the construction of 6-henry choke-coil for modulation by the
Heising method, using storage battery or motor- generator source of plate supply.
152 RADIO TELEPHONY
about 80 per cent, of the power (90 per cent, of the current)
will be obtained when the choke-coil impedance is equal to
twice the resistance of the modulator and oscillator tubes in
parallel. These tubes are to be of the same type so that
their resistances will be equal and the above statement will
mean that this transfer is secured when the impedance is
equal to the resistance of either tube. The impedance of
the available tubes described in Art. 41 averages about 4000
ohms, hence at the lower limit frequency of 100 cycles, an
inductance of 6 henries is indicated. Since the coil carries
normally the steady d.c. plate current of both tubes, it
should be designed carefully so that the iron is not operated
with too high a flux-density. An air gap in the iron circuit
is the most economical solution of this problem. Specifica-
tions for a 6-henry choke-coil suitable for two 5, 50 and 250
watt tubes are given in Fig. 88.
The complete modulation circuit is shown in Fig. 89.
The filament and plate sources are indicated symbolically as
in the previous diagrams, and are usually common to both
oscillator and modulator. The two wires pointing to the
left are connected to the oscillator tube (in both the self-
excited and master-oscillator systems) in place of the "B"
battery or plate source in the other diagrams.
The modulation transformer T should supply, with aver-
age intensity of speech, a secondary voltage sufficient to
completely modulate the oscillator's output. This will
usually be accomplished if the range of modulator plate
current is from zero to twice its normal value. The proper
variation of grid voltage to produce this change in plate
current depends, of course, upon the tube to be employed.
It will not generally be possible to produce this grid variation
CONSTRUCTION OF THE TRANSMITTER 153
in the case of high power tubes because their grid losses
and the dielectric losses in the secondary of the modulation
transformer will exceed the small amount of power available
in the microphone circuit. The 50-watt tube is regarded as
the limit of direct control. A description of a modulation
transformer suitable for the 5-watt tubes, UV-202 Radio-
tron, and Western Electric "E" tube, is contained in Fig. 90.
The grid is kept at a suitable negative potential by
means of the biasing battery "C" which also supplies cur-
rent for the microphone circuit. Regarding the correct
biasing voltage, nothing definite may be said. Ordinarily
Fig. 89. Showing actual connections of modulator tube for Heising modulation.
this would be adjusted so that the modulator plate current
and the normal oscillator plate current (with the tube in
the oscillating state) are about equal. This will be obtained
with a bias of approximately 25 volts in the case of the
5-watt tubes, and 60 volts in that of the 50-watt type.
However, a practice, instituted first I believe, by Mr. L. R.
Damon of the Federal Institute of Radio Telegraphy, of
using an abnormally large negative bias ( 120 volts for the
50-watt tube) has been reported to yield better results and
speech of superior quality. This is hard to account for
theoretically, and is one of the cases where theory must bow
before practice, for these results have been repeatedly
154
RADIO TELEPHONY
MODULATION TRANSFORMER FOR 5-WATT TUBES
(RADIOTRON, UV-202; W. E. CO. "E" OR VT-2)
WINDING SPECIFICATIONS
Primary.
Secondary.
No turns
220
22 000
Conductor
No. 24 B. & S
No. 40 B. & S.
No. feet required
78
9000
No. ounces . .
2 oz.
5 oz.
Resistance
2 ohms
9505 ohms
Core. No. 30 B. & S. gauge silicon-steel.
Notes. Design based upon standard Western Electric microphone with
12 volt battery; average secondary voltage (resistance load = j megohm)
for average talk = 30 v. r.m.s.
Fig. 90. Constructional details of modulation transformer for 5-watt tubes
(see Fig. 89).
verified by a number of experimenters.* It is therefore un-
* Those who have listened to the broadcasting from Mr. Damon's station
"WRP" (Camden, N. ].) will be able to judge the efficacy of this system
of biasing. The quality of speech, especially the sounds "s" "/," "th," etc.,
in the upper range, is unusually good.
CONSTRUCTION OF THE TRANSMITTER 155
hesitatingly recommended for trial where "C" battery
facilities of this size are available. Mr. Damon increases
the bias voltage until the antenna current ceases to increase
when the microphone is spoken into.
The resistance R, of value 1 or 2 megohms (megohm =
one million ohms), controls to some extent the secondary
voltage when there is no grid current and thus reduces dis-
tortion. It also checks the tendency, manifested par-
ticularly with the Western Electric "E" tube, toward
"blocking," that is, sticking at a positive grid voltage.
CHAPTER V
SOURCES OF POWER
54. Enumeration of the Methods of Power Supply for
the Plate Circuit of the Transmitting Tube. As the reader
will know, or have gleaned from the description of the
power tubes in Art. 41, the audion requires for its opera-
tion in radio telephony d.c. plate voltages ranging from
350 to 2000 volts, with which the corresponding plate cur-
rents (per tube) range from 0.040 to 0.250 ampere; and
filament lighting currents of from 2 to 15 amperes. In
radio telephone work the power supply, at least for the
plate circuit, must be substantially steady, otherwise its
fluctuations are heard by the receiver and produce a de-
cidedly disagreeable type of interference with the speech
or other sounds generated by the legitimate variation of
the plate voltage by the modulator. For c.w. work (see
Art. 49), in so far as the actual heterodyne reception is
concerned, it is not important that the plate supply should
be steady; in fact, I shall describe circuits later in which an
alternating voltage is applied directly to the plate, but
in this case the signal will be audible in the ordinary re-
ceiver and the operator should be very careful about caus-
ing interference. This practice, in which a.c. is applied
directly to the plate, is excusable only from the point of
view of economy, and even this excuse is a meager one in
view of the cheapness and ease with which a chemical
rectifier and filter circuit may be constructed. The amateur
156
SOURCES OF POWER 157
is urged to consider the rights of his fellows when con-
templating an installation of this type for c.w. work, or
in maintaining for telephone work a system of supply in
which through defective filtering or for any other reason,
a fluctuating plate voltage is obtained. For nothing more
disagreeably assails the ear than the monotonous drone
of a 60-cycle supply current. This is especially important
in cities and in other places of radio station congestion.
The primary source of power is usually the 60-cycle
a.c. lighting mains, although in some cases it may be pos-
sible through proper arrangements to secure d.c. at 500 to
600 volts from the feeders of a local street-railway system.
But considering only the first supply, the methods of con-
verting it into d.c. of the proper voltage may be divided
into two general classes: (1) by means of a motor-generator
set, and (2) by the use of a step-up transformer, rectifier
and electrical filter. The first method is the more con-
venient, reliable and flexible, but its expense has prevented
any extensive use in amateur stations. The transformer-
rectifier-filter scheme appears to be the most accessible
to amateur operators, and is very cheap and gives good
results. The rectifiers are generally of the thermionic,
chemical (electrolytic), or mechanical types, and used in
conjunction with a proper electrical filter, their output
voltage is reasonably steady and free from the undesirable
"ripples." In the articles to follow the construction of
suitable systems of this type will be taken up.
55. Supply of Power for Heating the Filament. The
life of a power tube is that of its filament, and a considera-
tion of the proper operation of the filament is therefore
of prime economic importance. This is especially true
158 RADIO TELEPHONY
with the higher power and more expensive tubes, and the
amateur should constantly guard against a natural tendency
to give little or no thought to this most vital part of his
equipment, which burning brightly one moment, may be,
dark the next.
For lighting the filament a.c. should be given preference
over a unidirectional current. This is fortunate, for the
proper voltage and current can be easily derived from the
110 v. power mains by means of a small step-down trans-
former. The reason for this recommendation has already
been stated (Art. 41, description of 250-watt Radiotron).
Briefly, the use of a.c. prevents one-half of the filament
from carrying more, and the other half less, than their
shares of current, as would happen if the unidirectional
thermionic plate current were superposed upon a uni-
directional heating current. Like the classical chain,
the filament is no stronger than its weakest link, and the
whole filament may as well be overloaded as one-half of
it in this manner, so far as the reduction of its life is con-
cerned; hence the use of a.c. is to be recommended even
in cases where it is not available and has to be specially
generated for the purpose. The overloading of one side
of the filament with a d.c. supply varies from 1 per cent,
to 1.5 per cent, for normal operation, which involves,
according to life tests made on the filaments of these tubes,
a decreased longevity of from 15 to 20 per cent. The im-
portance of a.c. operation depends therefore upon the cost
of the tube; in the case of a 5-watt tube, the use of d.c.
when a.c. is not obtainable may be justified.
The a.c. for the filament is usually obtained when an
a.c. transformer is used for the plate supply, from an
SOURCES OF POWER
FILAMENT TRANSFORMER
SUITABLE FOR SUPPLY OF TWO 5-, 50- AND 250-WATT TUBES
DETAILED SPECIFICATIONS
Power.
5 w.
50 w.
250 w.
Power rating
40 w.
200 w.
440 w.
Efficiency.
75%
85%
85%
Primary voltage (60 cycles). . . .
Secondary voltage
110 v.
10 v.
llOv.
13 v.
110 v.
13 v
Primary turns
500
500
250
Primary conductor (B. & S.) . . .
Secondary turns . .
No. 22
45
No. 16
60
No. 12
30
Secondary conductor (B. & S.).
Amount of wire required
Core dimensions (a x b)
Core cross-section
No. 12
(P. \X Ibs.)
\S. 1 Ib. /
W x \y 2 "
V x 1"
(6 No. 16
\in parallel
6 Ibs.
3" x 3"
1>" x 1>"
(6 No. 12
\in parallel
7 Ibs.
3" x 3"
2" x 2"
Core Laminations: No. 26 silicon-steel.
Fig. 91. Data for the construction of filament lighting transformers suitable for
two 5-, 50- and 250-watt tubes.
auxiliary winding upon this transformer. When a motor-
generator or storage battery is used for the plate supply
i6o RADIO TELEPHONY
a separate filament transformer will be required. Speci-
fications and full constructional details of suitable trans-
formers for two 5-, 50- and 250-watt tubes are given in
Fig. 91.
A tap should be brought out from the center of the
secondary winding and it is to this that all connections
from the grid and plate circuits should be made (see Trans-
mitting Circuit No. 1, Art. 42). This is done so as not to
introduce into these circuits the voltage fluctuations due
to the drop in the filament which would be included by
connecting to one end of the filament. It will not be neces-
sary to provide a by-pass condenser for this winding be-
cause if the tap is at the exact center of the winding there
will be no effective inductance between it and the ends of
the coil.
The life of the filament depends upon the temperature
at which it is operated, and this temperature will be roughly
proportional to the rate at which heat (= I 2 R) is developed
by the electric current. While the filament is burning,
its material is constantly evaporating, or being thrown off
and deposited upon the electrodes and the walls of the
glass container. The dark appearance of old audions and
electric light bulbs is a familiar manifestation of this action.
As the amount of material diminishes, the resistance of the
filament, of course, increases, with the result that if it is
operated with a constant filament current the operating
temperature goes up continually as the tube gets older.
The abridgment of its life is the result of such practice.
Few amateur stations are, however, equipped with a radia-
tion pyrometer with which to measure the filament tem-
perature, so it is rather hard to keep the temperature con-
SOURCES OF POWER 161
stant during the life of the tube. A practical scheme,
suggested by the manufacturers of the tubes, to restrain
the operation at increasing temperature is to use, rather
than constant current, a constant voltage; the tempera-
ture is roughly proportional to E?/R (voltage squared -*-
resistance) and as the resistance goes up due to evaporation,
the operating temperature decreases, and the life of the
tube is prolonged on two counts. The makers of the tubes
also state that an increase of 3 per cent, of the rated cur-
rent halves the life of the tube, and a decrease of 3 per
cent, doubles it. This will indicate the extreme importance
of proper attention to this detail of operation.
56. Self -rectifying Circuits. The so-called self -rectifying
circuits utilize an a.c. plate supply, the appropriateness
of the terminology being due to the action of the audion
oscillator in permitting the passage of current only during
the positive half of the cycle. If a sinusoidal a.c. is applied
directly to the plate, the current in the plate circuit, like-
wise the output current, will undergo such variation and
the signals will be audible in an ordinary receiver. Tele-
phony is rather difficult with this type of supply; but for
c.w. work its use is perfectly feasible and distances almost
as large as those obtainable with the same amount of d.c.
power may be covered. The heterodyne note is somewhat
peculiar and invested with a 60- or 120-cycle hum, but is
not at all unpleasant. The unpleasant part falls to the lot
of the nearby receiving station.
The self-rectifying systems may be divided into two
classes: those utilizing with one tube but one-half of the
a.c. cycle; and those employing a symmetrical arrange-
ment of two tubes for utilizing both halves of the cycle.
11
162
RADIO TELEPHONY
The latter system may be improved by the addition of a
large choke-coil, which somewhat reduces the current
fluctuation. The canonical form of electrical filter does
not seem to be applicable to this system.
A circuit using one tube and but one-half of the a.c.
cycle is shown in Fig. 92. Here the wires P, G and F con-
nect to the oscillating circuit, suitable forms of which have
been indicated. X is a radio frequency choke-coil. The
application of this scheme to the circuits shown in Art.
42 will occur to the reader. If the plate source is in series
with the radio frequency part of the plate circuit, it will
FiL.Sec.
Fig. 92. Self-rectifying circuit using one-half of the a.c. cycle.
usually be necessary to shunt the secondary winding with
a by-pass condenser of at least 0.002 mfd. capacity. A
center tap on the filament winding is not necessary.
In the symmetrical arrangement with two tubes shown
in Fig. 93, both halves of the cycle are used, one tube operat-
ing during the positive alternation and the other during
the negative alternation. In this case the secondary of
the transformer is provided with a center-tap to which
the common terminal of the filament is connected. The
iron-core choke-coil X 2 is inserted in the common lead to
smooth out the ripple, for which purpose its inductance
SOURCES OF POWER
163
should be as high as possible, 10 to 50 or even 100 henries,
and its design preferably embodies an air gap, as in the
case of the 6-henry choke-coil described in Fig. 88. The
condensers C (capacity at least .002 mfd.) serve to isolate
the oscillating circuit from the high voltage source, and
enable the plates to be connected (from a radio frequency
viewpoint) in parallel without short-circuiting this source.
The wires P, G and F connect to a suitable oscillating cir-
DOV-60'v
Fig. 93. Self-rectifying circuit using both halves of a.c. cycle with symmetrical
arrangement of tubes, and choke-coil to reduce current fluctuation.
cuit. Other arrangements are possible; for example, those
in which the so-called symmetrical oscillating circuits are
employed. These are, however, uneconomical of apparatus
and in general harder to operate, and having no compensat-
ing advantages for amateur use will not be described.
57. System of Plate Supply Employing Step-up Trans-
former, Rectifier and Electrical Filter. This scheme for
securing an approximately steady voltage for the plate
circuit of the power tube is illustrated symbolically in
164 RADIO TELEPHONY
Fig. 94. The method shown at (a) employs but one rectifier
and uses, of course, but one-half of the a.c. cycle; the
method (b) uses a symmetrical arrangement of two rectifiers
Di and D 2 , and operates on both halves of the cycle. This
is vastly superior to the first scheme and will therefore
form the nucleus of this discussion.
The system embraces three essential parts: a step-up
transformer, through which power at the proper voltage
is derived from the 110 volt a.c. line; a set of rectifiers
Di and D 2 'j and an electrical filter. The audions are repre-
sented symbolically as a resistance load. The purpose
POWER
TUBES
(a)
Fig. 94. Schematic diagram of systems of plate supply using a.c. step-up
transformer, rectifier and electrical filter: j (a) using one-half of the a.c. cycle;
(b) using both halves.
of the filter is to suppress variations of the current supplied
to the load, or as commonly expressed, to "smooth out the
ripples." This may take a variety of forms, from a simple
choke-coil to the more elaborate Campbell arrangement
with recurrent similar sections. The rectifiers Di and D 2
are usually of three types : thermionic, chemical and mechan-
ical. Of these the thermionic and chemical types are the
most satisfactory; the latter being especially economical
of power and easy to construct and to operate. In the
following articles the design, construction and operation
of the three elements will be given detailed consideration.
SOURCES OF POWER 165
58. Elementary Theoretical Basis for the Design of
the Transformer-rectifier-filter System. This note is in-
serted for the reader who may be interested in the theo-
retical side of the design and operation of the system under
discussion and who will want to know upon what basis the
constructional data given in the following articles rests;
the average reader may skip this and go on with the next
article. The design here undertaken proceeds on a quasi-
scientific basis as follows:
The rectifier is assumed to be perfect, that is to say,
completely obstructs the passage of current during the
RECTIFIER
RE
ESISTANCE
TIME
(a) (ft)
Fig. 95. Equivalent electrical circuit for system of Fig. 94 (b), and (6) form
of e.m.f. acting therein.
negative half -cycle, and permits its passage in proportion
to the voltage applied, during the positive half -cycle. This
condition is practically fulfilled in the case of the chemical
and mechanical types, if the capacity effect of the film in
the chemical type is small. In these circumstances, and
provided also the transformer is closely coupled, the system
may be replaced electrically by the simple circuit of Fig. 95 (a)
in which an e.m.f., "e" of the form shown ^at the right
(6) acts. This e.m.f. is compounded of a number of alter-
nating constituents, of frequencies 0, 120, 240, 360, 480,
etc., cycles per second, and with amplitudes represented
graphically in Fig. 96. The first constituent (zero fre-
i66
RADIO TELEPHONY
quency, direct current) is all that we desire to flow in the
output circuit, the others represent fluctuations in the
supply current and are to be suppressed by the filter.
The amplitude of the d.c. component is 2/n (=.636)
of the maximum value of the a.c. voltage wave, or with
a sine wave, .9 times the r.m.s. value of the wave. This
is the clue to the transformer design; the suppression of
fj
ll
M n2,4,fc.. . .
a
K
i
VO
8.5
3.6 2Q
IZO Z40 360 480
FREQUENCY
600 7ZO
Fig. 96. Illustrating relative magnitudes of d.c. component (zero frequency)
and harmonies in the voltage wave of (b) Fig. 95.
the higher frequencies will be considered in the article on
filters. Note that the d.c. in the transformer secondary
flows in opposite directions from the center-tap to the two
ends of the coil and produces no steady magnetization,
and that half of the secondary winding works during one-
half-cycle, the other during the other half-cycle. Thus
each coil works, so to speak, on half time and the heating
SOURCES OF POWER
167
TABLE II
SPECIFICATIONS FOR THREE TRANSFORMERS CAPABLE OF SUPPLYING
POWER FOR FOUR 5-WATT, TWO 50-WATT AND TWO 250-WATT
TUBES RESPECTIVELY
Rating of Power Tubes.
5 w.
50 w.
250 w.
Power rating . .
250 w.
700 w
2200 w
Normal primary voltage
Primary current (full load)
Efficiency
Primary turns, for 102 volts
105 "
" " " 108 "
"110"
110 v.
2.5 a.
90%
306
315
324
330
HOv.
6.5 a.
90%
204
210
216
220
220 v.
10.8 a.
90%
204
210
216
220
.. U 1 1 C ((
1 I -
Primary conductor (B. & S.)
336
345
No. 14
224
230
No. 12
224
230
No. 10
Amount of wire required
Filament secondary voltage
4K Ibs.
8 v.
W Ibs.
10 v.
7 Ibs.
11 v.
Filament current .
10 amp.
13 amp.
30 amp.
Filament secondary turns
" " conductor ....
H T. secondary voltage, Ei
24
f3 No. 14
| in parallel
400 v.
20
3 No. 12 1
in parallel/
1200 v.
11
Note 2
2200 v.
" " E 2
500 v.
1500 v.
2600 v.
" " " E 3
700 v.
1700 v.
3000 v.
H.T. secondary turns; Tap No. 1. .
" No. 2. .
" (Center tap) ; No. 3. .
" turns; Tap No. 4. .
" No. 5. .
" No. 6. .
H.T. sec. conductor (B. & S.)
Amount of wire required
450
900
2100
3300
3750
4200
No. 28
2 Ibs.
400
1000
3400
5800
6400
6800
No. 28
3X Ibs.
400
800
3000
5200
5600
6000
No. 24
9 Ibs.
Core dimensions (a x b)
Core cross-section
4" x 2^4"
W x Iff*
4" x3"
iK'x \ys
Note /.In the case of the 2200 watt transformer for two 250-watt tubes,
the 220 volt supply is recommended. For 110 v. operation, in lieu of 220 v.,
wind primary in two sections and connect in parallel, or wind with half the
above number of turns with copper ribbon \" x ^" cross-section.
Note 2. Either 3 strands of the primary wire (No. 10 B. & S.) may be used
in parallel, or copper strip of f " x Y&" cross-section.
effect is halved, so that although the secondary must be
wound for double voltage (the full tube voltage between
1 68
RADIO TELEPHONY
the center tap and each end), the cross-sectional allowance
is halved.
The theory of the thermionic rectifier is, so far as I am
aware, unmanageable by means of the ordinary mathe-
matics since the differential equations are inhomogeneous
and irrational. Practically, however, a system designed to
operate with a perfect rectifier will give the same results
when this type of rectifier is employed.
59. Construction of the Step-up Transformer. Sufficient
data for the construction of transformers for supplying
I**KJl
.T. SECONPAR
ARRAN&EMENTaf WlNPINGS
(Primary and H.T. Secondary)
Fig. 97. Arrangement of windings and plan of step-up transformer.
four 5-watt, and two 50- and 250-watt tubes, including a
secondary for supplying the filament current, will be found
in TABLE II on p. 167. The electrical connections and
arrangement of the windings, and the general plan of the
transformer are illustrated in Fig. 97.
In making the small computations for TABLE II, no
attempt has been made to attain a most economical or
efficient design, neither of which can proceed without
SOURCES OF POWER 169
special data as to the materials to be used, the magnetic
characteristics of the iron, etc.
The primary winding is preferably tapped at six points
as indicated, the purpose of this being to provide an ad-
justment of the filament current without a filament rheo-
stat, which costs money and wastes power. These taps
may be brought out to a suitable rotary switch, as indicated
in the diagram, for convenient adjustment. The filament
secondary is wound upon the same core-leg as the primary,
and except with unusual circuits, an average amount of
low tension insulation between the two and between them
and the core will be sufficient. The filament secondary
is tapped at the center. The high tension (h.t.) secondary
is arranged with a center tap, and subdivided so that three
different voltages between the center tap and taps 0, 1,
2 on one side, and 4, 5, 6 on the other side, may be obtained.
This arrangement may be necessary in changing from self-
rectification circuits to those employing separate rectifiers,
or in changing from the thermionic to the chemical and
mechanical types of rectifier. The h.t. secondary winding
should, of course, be well insulated from the core and ample
clearance should be allowed between this coil and the primary
and filament windings on the opposite core-leg. A 60-
cycle, 110 or 220 volt supply has been contemplated in the
design.
60. Design of the Filter Circuit. Examination of Fig.
96, in which the amplitudes of the various harmonic con-
stituents of the "equivalent voltage," (b) Fig. 95, have been
represented, shows that a filter which will eliminate all
frequencies above 100 cycles is needed. A very elaborate
and effective network for this purpose is the Campbell
170 RADIO TELEPHONY
arrangement of recurrent equal sections of series inductance
and shunt capacity. This network is shown schematically
in Fig. 98 and is referred to in the classification of Prof.
G. W. Pierce* as of Type II, or in engineering parlance
as a "low pass" filter.
This arrangement requires a number of coils and con-
densers and will be rather expensive to construct. For
this reason we shall be content to mention it as the theo-
retically proper network, abandoning its further considera-
tion in favor of simpler arrangements employing a few
coils and condensers. The sophisticated reader who wishes
'WH-'&W^T'^^
*VX_T T
nrC -!-* -T- -T-
.
T T
Fig. 98. Electrical filter circuit of the Campbell type for suppressing currents
of high frequency.
to undertake the design and construction of such a filter
is encouraged to do so and will find plenty of theoretical
guidance in Chapter XVI of Prof. Pierce's book previously
referred to; but the average experimenter will, I believe,
be satisfied with the less perfect but still quite satisfactory
action of a simpler system.
At first sight it might appear that the most desirable
filter would be one whose impedance at the various fre-
quencies would be adjusted to suppress the harmonics in
proportion to their amplitudes. This would indeed give a
* Cf. Pierce: "Electric Oscillations and Electric Waves," p. 299 (New
York, 1920).
SOURCES OF POWER
171
uniform reduction of all frequencies, but this is not particu-
larly desirable for two reasons, as follows:
In the first place, the higher frequencies have a relatively
greater disturbing effect than the lower ones, that is to
say, for a given amplitude the 240-cycle and higher fre-
quencies produce more interference with the voice waves
than the 120-cycle component. So the higher harmonics
should ultimately be smaller than the lower ones. The
second reason why the impedance should not fall off with
increasing frequency is appreciated immediately if the
problem is looked at from the viewpoint of Heising modula-
Fig. 99. Symbolic representation of three simple filter circuits for use with
step-up transformer and rectifier.
tion, Fig. 86. The transformer-rectifier-filter system is in
series with the modulation choke-coil in that circuit, and
obviously no reactances should be introduced in this system
which will with the inductance of the choke-coil cause the
impedance of this path to fall much below the value of
4000 ohms which we have prescribed. Thus the problem
of design must be looked at from both ends of the telescope,
and in this particular arrangement of circuits and rectifiers
will not have the same aspects when so viewed.
Three simple filtering arrangements are shown sym-
bolically in Fig. 99.
172
RADIO TELEPHONY
Arrangement (a). In the first arrangement a single
choke-coil is inserted in series with the circuit; and a single
coil incorporating the modulation choke-coil is recom-
mended. The choking action decreases with the frequency
and the inductance should be large enough to give effective
filtering action with the 120-cycle harmonic. An inductance
of 10 henries will suppress one-half of this harmonic (rep-
resenting a fluctuation of 25 per cent, of the d.c.), 20
henries will suppress three-fourths of it (12 per cent, fluctua-
tion), and so forth. Coils of from 50 to 100 henries are
not too large for this purpose, neither is their construction
(a) (b) (c)
Fig. 100. Showing actual connections in the three simple filter arrangements
illustrated in Fig. 99.
particularly difficult. An air-gap should be provided, as
in the case of the 6-henry modulation choke described in
Fig. 88. Data for the construction of a suitable high-
valued coil will be given in the next article. The actual
connections are shown in Fig. 100 (a). The (+) wire con-
nects to the plates of the tubes (no extra modulation choke
necessary) and the ( ) wire connects to the midpoint of
the transformer filament winding. It is important to note
that this system is not suitable for modulation by the grid
voltage method. In this case the effect of the impedance
in the plate circuit opposes the changes in the plate current
that we are trying to cause by varying the grid voltage.
SOURCES OF POWER 173
Arrangement (b). The next step in improvement and
complication consists in shunting the choke-coil of arrange-
ment (a) with a condenser as shown at (&), Fig. 99. This
is merely a symbolic representation, for the capacity at
C may be the inherent distributed capacity of the choke-
coil itself, or may be due to the combination of this and
an external capacity. In its most desirable form this in-
ductance and capacity would consist merely of a coil upon
which wire had been wound until its natural frequency was
120 cycles, provided, however, that the inherent capacity
does not exceed certain limits to be prescribed. At the
frequency for which LC are resonant the branch circuit
formed by them will offer a very high impedance, and as
the frequency is increased from this point the impedance
will decrease; at 4000 cycles the circuit will act very much
like the condenser C. Thus we have here many of the
features shown to be desirable, viz., a high impedance at
120 cycles falling off not too rapidly and never falling below
4000 ohms throughout a range of speech frequencies from
100 to 4000 cycles. In order for this last statement to be
true a maximum capacity of .01 mfd. at C will be neces-
sary, and with this capacity an impedance of 4000 ohms
at 4000 cycles will be obtained, giving, as explained in
Art. 53, a 90 per cent, modulation effect.
The practical features demand, however, some con-
sideration. For resonance at 120 cycles with a capacity
of .01 mfd. an inductance of 200 henries is required. Such
a large coil would be impractical and the heat losses would
vitiate to a large extent the improvement in the filtering
action. We must, therefore, either sacrifice something in
modulation and increase the capacity C (double it, let us
i 7 4 RADIO TELEPHONY
say); or not preferring this, sacrifice some of the choking
effect at 120 cycles and tune the circuit to a higher fre-
quency, for example, 240 cycles. By doing the first we
reduce the inductance required to 100 henries, by the second
to 50 henries, and by both to 25 henries. Somewhere in
this range a practical set of values may be found which
represent the best compromise between the operator's
constructive energy and his desire for perfection.
Up to this point the design has been governed by the
requirements of the Heising modulation system. For
ordinary c.w. work this problem vanishes, and we have
merely to design an effective filter system. In these cir-
cumstances a larger capacity at C may be used, allowing
some reduction in the amount of inductance required in
the choke-coil. This should not be carried too far, for the
higher frequency quenching action will be impaired.
The actual connections are shown in Fig. 100 (b). The
condenser should be designed to withstand the full voltage
between the midpoint and one end of the h.t. secondary.
The construction of a suitable high-valued choke-coil will
be described in the next article. The wires (+) and ( )
connect respectively to the plates of the tubes and the
midpoint of the filament secondary. It is unnecessary to
provide a separate modulation choke-coil; in fact, the addi-
tion of any inductance in series with the plate supply cir-
cuit will be deleterious. If properly designed this system
of filtering will give excellent results and is easily the best
of the simpler arrangements. The adjustment is not
difficult, for the operator may listen-in on a local receiving
apparatus or with a pair of telephones inductively coupled
to the common lead (+) to the tubes, and vary the capacity
SOURCES OF POWER 175
C or the width of the air-gap in the choke-coil, until the
120-cycle or 240-cycle notes have their minimum intensities.
Like the previous arrangement this system is not well
adapted for use with grid modulation on account of the
high impedance in the plate circuit. The operation with
this method of modulation may be improved by employ-
ing in the filter a large ratio of capacity to inductance.
Arrangement (c). This very popular form of filter is
mentioned here because of its popularity, because it is
fundamentally wrong in principle and because if not properly
designed it will do more harm than good. The condenser
C is in this case shunted across the load, as shown in Fig.
99 (c).
Considering first the c.w. arrangement in which the
modulation choke-coil X does not appear, we may formu-
late the following rule for proportioning L and C:
Starting with the simple choke-coil arrangement, (a)
Fig. 99, the addition of a condenser, as at (c) Fig. 99,
will reduce the choking action on all frequencies lower
than the frequency n unless its capacity is greater than
1/271 VL.
Considering the 120-cycle harmonic as of most im-
portance this means that if the product of L in henries and
C in microfarads is less than 3.5 the addition of C will
magnify the 120-cycle ripple; if the product is equal to
3.5 it will have no effect on it (although the higher har-
monics are more effectively choked) ; and in order to effect
an improvement at 120 cycles the product should be greater
than 3.5. This critical value for C may be determined by
listening-in to the effect of adjustment of C or the air-gap
176 RADIO TELEPHONY
of the choke-coil. I believe that a great majority of the
failures of this filter system can be explained as due to
violations of the above fundamental rule, which is not
empirical, but based on the mathematical theory of the
circuit, and should be rigidly adhered to.
Now from the point of view of telephony, first with
Heising modulation, if the modulation choke-coil X is
omitted, the system will give the same results as arrange-
ment (5) provided the same conditions obtain, viz.: that
C does not exceed .01 mfd. In view of the rule just stated
this will demand an inductance of 350 henries. Thus we
are justified in condemning the system on two counts,
as a defective filtering arrangement whose principle is
entirely wrong, and as eminently impractical from the
Heising modulation viewpoint. From the latter view-
point also the extra choke-coil X should always be omitted.
The arrangement does, however, have a slight appeal
if we consider it from the viewpoint of grid voltage modula-
tion. As already remarked, in order for this system of
modulation to be successful the plate circuit should con-
tain no telephone-frequency reactance. In this respect
it is quite the reverse of the Heising method, which for best
operation demands a maximum reactance. Thus the
arrangements (a) and (b) which were found suitable for
the latter method, will give poor results with grid modula-
tion. The present arrangement is slightly more effective
in this respect. Still keeping the LC product above 3.5
we may increase the ratio of C to L until the reactance
l/2nnC is sufficiently low. With a capacity of 2 mfds.
the effective reactance from the plate circuit point of view
at the lower limit of the speech frequency, 100 cycles,
SOURCES OF POWER 177
will be about 700 ohms; at 4000 cycles it will be about
17.5 ohms. There will be some reduction of the lower
frequencies, but the operation may be considered satis-
factory. Grid modulation is principally an economical
method and will be employed most extensively with the
5-watt tubes. In this case the voltage which the shunt
condenser C must withstand is relatively low (1000 volts
max.) and commercial forms of paper condensers may be
grouped in series and parallel until the required breakdown
voltage and capacity is obtained. With a capacity of
2 mfds. an inductance of at least 3.5 4- 2 = 1.75 henries
should be used; 25 or 50 henries will be better.
The actual connections of this circuit are shown hi Fig.
100 (c). The wires (+) and ( ) connect to the plates of
the tubes and to the midpoint of the filament secondary
respectively.
61. Resume: Comparison of the Filter Circuits. Of
the preceding arrangements, considering their merits as
filters and from the viewpoint of modulation, the arrange-
ment (ft) Fig. 99 is regarded as the best for Heising modu-
lation, with (a) next in order, and (c) as inherently and
practically unsuitable for this method of modulation, but
more suitable than the others for grid voltage modula-
tion.
In applying (a) and (&), the inductance should be at
least 50 henries, and in the case of (6) the total capacity
across the coil, including the distributed capacity of the
coil windings, should not exceed .01 or, possibly, .02
mfds.
In applying (c) for grid voltage modulation the ratio
of C to L should be as large as possible and the product
12
178 RADIO TELEPHONY
of L in henries and C in mfds. should not be less than 3.5.
A minimum capacity of 2 mfds. is suggested for C, and an
inductance of 25 or 50 henries for L.
62. Construction of the Filter Apparatus. The filtering
arrangement (b) utilizes a 50-henry choke-coil and a .01
to .02 mfd. condenser. It will be helpful, I think, to indicate
briefly how these electrical constants may be practically
attained.
(a) 50-Henry Choke-coil, L. This coil carries normally
the steady d.c. for the tubes and should be carefully de-
signed so that the iron is not operated at too high a flux-
density. An air-gap is the most convenient means for
preventing this, and may be made adjustable so that the
inductance of the coil may be varied for experimental pur-
poses. A greater variation of inductance can be secured
by varying the number of turns and taps may be brought
out at intervals for this purpose. In the 12,500 turn coils
to be described, it is suggested that taps be made at 6000,
8000, and 10,000 turns. The arrangement of coils and the
general plan of the windings and core are similar to those
of the 6-henry modulation coil described in Fig. 88. Data
for the construction of coils for two 5-, 50- and 2 50- watt
tubes are given in TABLE III on p. 179.
Instead of the arrangement shown in Fig. 88, the top
core-leg in which the air gap is cut may be made in a solid
piece, hinged at one end and separated from the other end
by an adjustable air gap. The mechanical arrangement of
this will be left to the reader. Since the coil may be sub-
jected during modulation to voltages equal to twice the
normal tube voltage, the insulation of the winding should
be carefully attended to.
SOURCES OF POWER
179
TABLE III
SPECIFICATIONS FOR 50-HENRY CHOKE-COIL FOR FILTER ARRANGEMENTS
(a) AND (b), FIG. 100
Power Rating of Tubes.
5 w.
50 w.
250 w.
Normal direct current
0.1 amp.
0.3 amp
6 amp
Conductor (B. & S. gauge)
No. 30
No. 24
No 22
No turns (total)
12,500
12 500
10 000
Amount of wire required
Core dimensions; (a x b)
2% Ibs.
3" x 2"
16 Ibs.
4" x 3"
25 Ibs.
5" x 3"
Core cross-section
1" x 1"
2" x2"
2M" x iy*"
Width of air gap
1 A"
y
V*"
D c resistance (ohms)
660
500
200
Heat loss (watts) .
6.6
45
75
Note. For arrangement of windings and notes on design, see Fig. 88.
(b) 0.02 Mfd. Condenser, C . The condenser is subjected
to the same voltages as the choke-coil and its dielectric
should be selected with this in mind. The capacity re-
quired is not large, so the use of glass plates is convenient.
The design of a glass plate condenser has already been
considered in Art. 44 (b) ; and with the plate and conductor
dimensions there given a total of 81 plates will be required.
It will be better to use 8" x 10 " photographic plates, which
average about 0.06 " (iVO m thickness. They may be
covered with tin-foil 6^" x 8-J-"; and using the formula of
Art. 44 we find that the capacity per positive plate is 0019
mfds. A total of 11 positive and 12 negative plates will,
therefore, give the required capacity.
63. Construction of a Chemical Rectifier. The chemical
rectifier is cheap and easy to construct and gives quite
satisfactory service. It usually consists of an electrolytic
cell formed by an electrode of aluminum and an electrode
of lead or other metal immersed in a concentrated solution
i8o RADIO TELEPHONY
of borax (NaJ^Oy) and water. It is found experimentally
that at ordinary temperatures such a cell possesses the
valuable property of permitting the flow of current from
the lead to the aluminum electrode, but for limited applied
voltages checks the flow in the opposite direction. If the
voltage, applied in the direction for which the cell is non-
conducting, exceeds a certain critical value, determined
by the electrochemical properties of the solution (480
volts for borax), the thin film of gas on the aluminum elec-
trode which prevents the flow of current, breaks down and
the current is no longer impeded. It is necessary, there-
fore, in using this type of rectifier with high voltages to
use several cells in series, the number being so chosen that
the peak value of the alternating voltage does not exceed
this critical value in any cell, for obviously (see Fig. 100)
the entire voltage between midtap and the end of the trans-
former is impressed across the rectifier during the non-
conducting period. The voltage at which the film breaks
down, and the cell ceases to rectify, is reduced also by a
rise in temperature; hence an ample amount of solution
should be used in order to dissipate the heat, and the elec-
trode surface should not be too small. There is, in fact,
an optimum electrode surface per ampere, which repre-
sents the best compromise between getting a low resistance
during the conducting period and a ready formation of the
barrier gas film for the non-conducting period. If the elec-
trode surface is too large, the formation of the film is impeded.
Unfortunately, because they have no special applica-
tions in the general electric practice and have, we might
say, but lately come into vogue on account of their suit-
ability for the supply of high voltages for audions, there
SOURCES OF POWER
181
is but little engineering data in the literature on the sub-
ject of chemical rectifiers. The following remarks on the
design and construction of the chemical rectifier has there-
fore no sounder basis than a limited number of experiments
and a collation of the published experience of others.
The number of cells will be determined by the maximum
voltage (equal to r.m.s. value x 1.4 for sine wave) between
the midpoint and the end of the h.t. secondary of the trans-
former; and allowing a maximum voltage of 150 volts per
cell the number of cells in the three cases are as given in
TABLE IV. As for the current density, from 5 to 10 sq. in.
per ampere appears to be common practice, and the data
in TABLE IV is based upon an average value of 7.5. There
is no optimum value for the quantity of solution required;
for proper cooling this should be as large as possible.
TABLE IV
DATA FOR THE CONSTRUCTION OF CHEMICAL RECTIFIERS FOR USE WITH
TWO 5-, 50-, AND 250- WATT TUBES
Power.
Normal
d.c.
Electrode
Area of
Immersion.
No. Cells
Each Bank.
Voltage
(r.m.s.).
Quantity
of Solution.
5 w.
.1 a.
2" x y 8 "
6
550
Xpt.
50 w.
.3 a.
1" X %"
12
1100
Ipt.
250 w.
.6 a.
6" x \y
24
2200
Iqt.
The solution is made by dissolving as much borax (the
"20 Mule Team" variety will do) as possible in cold water
and using the clear liquid. After the electrodes are in
place a supernatant layer of oil from J" to \" thick should
be placed on the electrolyte to prevent its evaporation
and loss during operation, and to check the tendency for
arcing from the unimmersed portions of the aluminum
182
RADIO TELEPHONY
ALUMINU
to the electrolyte. The oil should be heavy enough not to
emulsify, but not heavy enough to prevent the free escape
of gas from the electrodes.
The sharp edges of the electrodes may be rounded in
their immersed portions and
the metal should be well
cleaned, first with sandpaper
or by scraping, and then by
boiling in a solution of caus-
tic potash and scrubbing to
remove all traces of grease.
A mechanical arrangement of
cells is shown in Fig. 101 and
the electrical connections are
illustrated in Fig. 102 in rela-
tion to the transformer-rectifier-filter scheme of Fig. 100 (b).
The cells should be placed in trays which have been
carefully insulated from each other and from the earth by
BORAX"
SOLUTION
Fig. 101. Illustrating construc-
tion and arrangement of cells in
chemical rectifier.
ALUMINUM,
LEAP
Fig. 102. Electrical connections of chemical rectifier, step-up transformer, and
filter circuit suitable for Heising modulation.
porcelain "knob" or other available insulators. The trays are
best coated with paraffine and kept dry to prevent leakage.
In putting a new rectifier into operation the plates may
be "formed" or seasoned by operating them in series with
a few electric lamps on the 110- volt mains. The best re-
SOURCES OF POWER 183
suits are obtained if the process is carried out with a sub-
normal current density, say at about 50 per cent, of their
rated current. After about six hours 7 continuous operation
the aluminum plate will be found covered with a light
gray oxide, and the lead plate will have turned a very
dark brown color. They may then be connected in their
normal mode and further seasoned by operation.
64. Thermionic Rectifiers. The fundamental principle
of this type of rectifier has already been explained in Chapter
II. It consists of a plate electrode and a filament, both
enclosed in a glass bulb in vacuo. The filament emits
electrons when heated and the device possesses the im-
portant property of permitting current to flow from plate
to filament, but not in the opposite direction.
Thermionic rectifiers are less efficient and more expen-
sive, both in first cost and maintenance, than the chemical
type. The smaller efficiency results both from a lower
rectification efficiency, that is, a larger loss in the rectifier,
and from the extra power consumed by the filament. This
latter is a pure loss, since the filament contributes nothing
directly to the rectification process. The maximum over-
all efficiency of the usual commercial type of rectifier is
about 40 or 50 per cent. Thermionic rectifiers are, however,
of considerable commercial and military value, and may
also be found convenient by amateurs in situations where
the messier jars and solutions of the chemical type are
unsuitable, as in portable installations.
Several types are available on the market and are sold
under various trade names as "Kenetrons," "A-P" tubes,
etc. These are manufactured in various sizes appropriate
to the 5-, 50- and 250-watt tubes previously described.
184
RADIO TELEPHONY
The connection of a typical thermionic rectifier set is
shown in Fig. 103, hi relation to the filter scheme (fr), Fig.
100. The filaments are supplied by the transformer T 2 ,
or by a separate winding on the power transformer, Z\.
It is to be noted that the filaments of the rectifier tubes
are at the average potential of the plates of the power
tubes, and the apparatus connected to them, filament
rheostats, meters, switches, etc., and the filament winding
IIOV/v
Fig. 103. Illustrating connections for thermionic rectifier, step-up transformer
and filter circuit.
of the transformer should be carefully insulated, and shielded
from accidental contact.
65. Mechanical Rectifiers. Mechanical rectifiers are
simply automatic switches which reverse the direction
of the current in synchronism with the alternating supply
voltage. They are chiefly of two types, classified accord-
ing to the means adopted for actuating the reversing mech-
anism, as of the vibrating type and the synchronous com-
mutator type. The first arrangement employs a vibrating
reed which is acted upon by an a.c. magnet connected to
SOURCES OF POWER 185
the current source; the second type consists of a reversing
commutator driven by a synchronous motor.
Because of the difficulty in controlling the time and
duration of contact, vibrating rectifiers have not been
successfully applied in high voltage low current systems
of the type under consideration. They are chiefly useful
for charging storage batteries, and for other purposes
where low voltages are involved.
The synchronous commutator is more suitable for the
rectification of high voltages because the time of making
contact and its duration is under better control. A syn-
chronous motor is necessary, and if the reader is fortunate
enough to have possessed a synchronous spark gap, its
motor may now be pressed into service. The system of
constant-potential rectification* seems to be most suit-
able and is illustrated schematically in Fig. 104, in relation
to the filter scheme of Fig. 100 (b). The rotating disc is
provided with two sets of segments, connected respectively
to the terminals of the h.t. secondary of the supply trans-
former. The connections are made by means of slip-rings
and brushes of the usual type. The commutator is pref-
erably of large diameter so that the brushes may be suffi-
ciently large for mechanical rigidity without covering too
great an electrical angle. The full secondary voltage will
exist between the consecutive segments and this should
be kept in mind in designing the insulation. This arrange-
ment will be successful for use with 5-watt tubes, and by
careful design may possibly be applied to the higher powered
types.
* Cf. C. P. Steinmetz: "Transient Electric Phenomena and Oscilla-
tions," p. 236 (New York, 1909).
i86
RADIO TELEPHONY
Considering the design of a disc for use with a synchronous
60 cycle 110 volt single phase motor, with synchronous
speed of 1800 R.P.M., four segments will be required,
two for each cycle. Allowing a maximum voltage at inter-
ruption of 15 per cent, of the peak value, a disc diameter
of 8 inches will be indicated. The width of the insulation
segment should not be greater than f inch. The brushes
of each set are, of course, separated by the same distance.
Fig. 104. Illustrating constant-potential mechanical rectifier using synchronous
commutator.
These may be of wire and have a rectangular cross-section,
narrow in the direction tangent to the disc. The function
of the resistances R is to control the change of current and
prevent sparking. For a load of two of the standard types
of tubes, a value of 5000 ohms will be sufficient. Since
the current exists but a short time, and at a time when
the voltage is low, ordinary grid leak resistances, of the
type used with the 2 50- watt tubes will be convenient.
These are usually of 10,000 ohms resistance, with a center-
SOURCES OF POWER 187
tap at 5000 ohms. The mechanical arrangement, includ-
ing the provision of a means for shifting the brushes so
that commutation takes place at the time of minimum volt-
age, is left to the reader. The type of commutator described
by the writer in "Modern Electrics" for January, 1911
may be more convenient than a construction embodying
separate slip-rings. It is to be noticed that in a constant-
potential system of commutation, as here described, the
commutator simply reverses the supply voltage and the
current in the load circuit will be determined mainly by
the characteristics of that circuit. Thus the method of
filter design previously employed will be appropriate in
this case also.
66. Motor-generators. The motor-generator is without
doubt the most reliable source of high voltage d.c. for the
plate circuit of power tubes. This is especially true with
the higher powered tubes such as the 250-watt type, where
suitable chemical rectifiers, involving as specified in TABLE
IV a total of 48 cells, are somewhat unwieldy. The dis-
advantage of this source of supply is, from the amateur's
point of view, its high initial cost. Several types of gen-
erators with d.c. or a.c. motors are available on the market,
capable of supplying enough power for groups of 5-, 50-
and 250-watt tubes.
It is often possible to rewind low voltage motors or gen-
erators to deliver the required voltage. This cannot be
carried too far unless the insulation of the commutator
is designed to withstand high voltages. Old 500-600 volt
d.c. motors, such as are employed for auxiliary purposes,
for air-pumps, etc., by street-railway .companies are very
useful, already having well-insulated commutators, and,
i88 RADIO TELEPHONY
of course, being suitable without rewinding for the excita-
tion of the 5-watt tubes.
While the voltage supplied by a d.c. generator is sub-
stantially constant, a high frequency commutation ripple
or "hum" will often occur, due to faulty commutation
or other more natural causes. This may be removed by
a proper filter; and a proper filter for Heising modulation
does not consist in shunting the generator with a 1 mfd.
condenser, or any reactance which with the reactance of
the modulation choke-coil will give a lower impedance
than 4000 ohms at frequencies in the speech range. This
is a common error, and I know at least three large broad-
casting stations where the modulation in the lower range
suffers from this species of malpractice. When Heising
modulation is used a 50-henry choke-coil in series will
effectively subdue the ripple and improve the quality of
the speech at the same time.
67. Complete Wiring Diagrams for Transmitters. In
concluding this section on the transmitting apparatus,
in which the various parts, oscillating circuits, modulator,
power supply, etc., have been considered separately, it
seems worth while to indicate the method of combining
the various parts into one complete wiring diagram. This
will be illustrated for a model amateur transmitter, using
the Meissner Transmitting Circuit No. 1, the Heising
system of modulation, Fig. 89, the transformer-rectifier-
filter scheme of Fig. 100 (6), and the method of combining
direct earth and counterpoise to reduce earth losses, de-
scribed in Art. 26. In addition, switches are to be pro-
vided so that the modulation tube may be connected in
parallel with the oscillator tube for c.w. operation. The
SOURCES OF POWER
189
gifc
Soo
aq
sill
SgS
a | IB
III
Ml
i5!
vQOOMOafiQOQOi)QMQ90
IQO RADIO TELEPHONY
complete wiring diagram is shown in Fig. 105, and is suit-
able for two 5-, 50- or 250-watt tubes, with proper circuit
constants and apparatus as completely specified in the
diagrams and tables of the last two chapters.
68. Assembly of the Complete Transmitter. It is sug-
gested that the various parts described in the preceding
articles be assembled in a complete transmitting unit, as
in the specimen transmitting set illustrated in Fig. 106.
Fig. 106. Complete low-power radio telephone and c.-w. transmitter.
This shows a 20-watt set, employing four 5 -watt Radio-
tron power tubes. The 60-cycle power transformer is shown
at the left, the radio frequency choke-coil to the right of
this, then the low frequency filter circuit choke-coil, and
to the extreme right the antenna coil, which is in this case
wound with solid round copper wire, spaced to reduce
the resistance. Taps are taken from this coil to the switches
on the front of the panel, for the control of the wavelength
and the adjustment of power output.
CHAPTER VI
RECEIVING APPARATUS
69. Main Processes in the Reception of Radio Telephone
Signals. Reference has already been made (Art. 9) to the
fact that the radio telephone receiver consists essentially
of two parts; (a) an apparatus for tuning the antenna cir-
cuit to the wavelength of the signal to be received, and
(b) a detector or demodulator, whose function it is to
translate the radio frequency waves into the low frequency
vibrations of the speech or other transmitted sounds. In
more elaborate receivers, in addition to these essential
parts, amplifiers may be provided to magnify the radio
frequency waves previous to detection, or to magnify the
audio frequency vibrations after detection, either to secure
greater response in the telephones or for the operation of a
loud speaker. Cascaded audion amplifiers are generally
used for this purpose.
There is some didactic advantage, I think, in parti-
tioning the receiving equipment into parts according to
the distinct functions of these parts, and not only does this
apply to the written description and treatment of the
system, but may also extend to the actual assembly and
construction of the apparatus. As applied in the con-
struction, the above four-part division constitutes a com-
promise between the German method of mounting all parts
of the receiver in one cabinet, and the English method of
keeping every part separate. The treatment of the re-
192 RADIO TELEPHONY
ceiving apparatus given in this chapter will proceed in this
spirit, but at the same time we shall consider the com-
bination of the units into receivers of various degrees of
pretentiousness.
70. The Tuning Apparatus. Single-circuit and Double-
circuit Receivers. The purpose of the apparatus which we
have designated as the " tuning apparatus" is twofold:
first, to reduce by the insertion of proper amounts of induc-
tance and capacity, the impedence of the antenna circuit
to its lowest value at the wavelength of the signal to be
received, so that the current flowing in the circuit will be a
maximum; and second, to provide the proper transformer
action for transferring the greatest amount of energy from
the antenna circuit to the detector. These functions are,
in the main, independent. Concerning the latter, it should
be noted that the audion detector or amplifier is a potential
controlled device and demands a high voltage and small cur-
rent for its operation, whereas the crystal detector requires
a somewhat larger current. According to the two principal
methods used for the transfer of energy from antenna cir-
cuit to detector (or amplifier), receivers are classified as
of the single-circuit, or double-circuit types. These types
are illustrated in Fig. 107. In the single circuit arrange-
ment (a), the inductance L and capactiy C , either of which
or both may be variable, are used to tune the antenna
circuit. The detector or amplifier is connected across the
inductance L and is actuated by the voltage drop across it.
For wavelengths below the fundamental of the antenna,
the detector or amplifier is preferably connected across the
tuning condenser C . In the usual practical application
the inductor L consists of a coil, tapped at certain intervals
RECEIVING APPARATUS
to provide a means for roughly adjusting the inductance.
The finer tuning is accomplished by means of the variable
condenser. The proper transformation ratio is obtained by
varying the proportion of inductance and capacity, keeping
their combined reactance constant, of course, to preserve
resonance with the induced voltage.
In the double-circuit arrangement (6), Fig. 107, the
coupling is inductive and both circuits are to be tuned to
the incoming wave by means of the adjustable inductances
UJ IU
CD
o
<=>
c,
(a) (ft)
Fig. 107. Schematic illustration of the single-circuit (a) and double-circuit (6)
methods of tuning.
and capacities. The detector or amplifier is in this case
connected across the secondary condenser. The trans-
formation ratio is adjustable by means of the coupling as
well as by varying the ratio of inductance to capacity in
the two circuits.
The single circuit receiver is the simpler of the two, in
both construction and adjustment. For this reason it will
be particularly attractive to beginners, and others of
limited experience. But with the usual type of receiving
13
i 9 4 RADIO TELEPHONY
antenna, having a high resistance, it will be considerably
less selective and efficient than the double circuit arrange-
ment. In the first form (a) the voltage acting upon the
detector or amplifier is inversely proportional to the antenna
resistance; in the second (6) it is inversely proportional to
the square root of the product of antenna circuit resistance
and secondary circuit resistance. Thus in the form (b) it
is possible, by building a secondary circuit of low resistance,
to compensate to some extent for a poor antenna. In
these days, when so many radio transmitters are operating
in a narrow band of wavelengths, the increased selectivity
of the coupled circuit is also an extremely important feature.
71. Construction of the Tuning Apparatus. The tuning
unit is preferably mounted in a cabinet by itself. For
best results the coils should be constructed so that their
losses are small. As stated above, if the receiving antenna
is not good, that is to say if it is not carefully constructed
to have low losses by the methods exploited in Chapter III,
its resistance will be relatively high and any improvement
of the tuning apparatus directly in the antenna circuit will
not be particularly effective. In these circumstances it
will be more profitable to devote attention to the improve-
ment of the secondary circuit, using the coupled circuit
arrangement of Fig. 107 (6). But the reader will make no
mistake in minimizing the resistance of both circuits and
the following notes may be of some assistance in connection
with this.
The high-frequency resistance of a coil is higher than its
resistance for steady currents and increases rapidly with
the frequency (decreasing wavelength). This is due mainly
to the crowding of the current toward the outside of the
RECEIVING APPARATUS 195
coil at high frequencies, and to losses in the dielectric mate-
rial used for its insulation and support. At certain wave-
lengths the first kind of loss may be reduced by winding
the coil with a stranded conductor, or cable, made up of a
number of smaller wires in parallel. But for wavelengths
of the order of 100-500 meters with which we are here
concerned, it has been found that solid wire, No. 18-
22 B. & S. gauge copper, yields a coil whose resistance
is lower than one wound with the usual commercial grade
of r.f. cable.* When using the solid wire (No. 18 is to be
preferred) it is desirable to separate the turns by a distance
equal to the diameter of the wire. This optimum spacing
is very well defined for coils wound on cardboard tubes
and with cotton-insulated wire. Variation of the inductance
is usually obtained by tapping the coil at certain intervals
and connecting these taps to a multi-point switch. In
this case the fine tuning is obtained with a variable con-
denser. This is a convenience, and is justified only for
this reason, for electrically it is bad practice. In the first
place it leaves unused portions of the coil more or less
closely coupled to the active portion, and in these unused
turns induced currents will cause a loss and increase of the
resistance of the main circuit (* 'dead-end' ' effect) . Secondly,
the taps being brought out to the multi-point switch on a
bakelite or other panel, produce a very substantial dielec-
tric loss. This can be reduced in several ways: by elimi-
nating the switch entirely and using clips on a coil which
* This fact, as well as other statements of this section, rests upon the
extensive and unpublished experimental study of the losses in radio coils
made by my former colleagues, Messrs. L. B. Dick, W. G. Ellis and W. D.
Loughlin of the U. S. Navy, to whom I am indebted for placing their results
at my disposal.
196
RADIO TELEPHONY
has been tapped as shown in the middle inductor of Fig. 8,
or by using great care in the construction of the switch.
I have found the form of construction shown in Fig. 108 to
yield very low switch losses if the mica plate is carefully
selected. The accompanying figure conveys the important
details of this construction. The panel is cut away and
for it is substituted a sheet of good electrical mica, barely
thick enough to give good mechanical support. Mica
varies a great deal in its electrical properties and unless
^ PAN EL
DIAL a KNOB
Fig. 108. Illustrating construction of mica-insulated switch, designed to reduce
dielectric losses -with tapped coils.
this sheet is selected with great care, the whole purpose of
the special construction will be defeated. The switch is
of the usual rotary type, but in which the contact points
are very small (--" diameter with 3-48 studs). "Spaghetti"
or other insulation is not to be used on the wires leading
from the coil to the switch. These should be heavy, about
No. 16 B. & S. bare copper wires, and self-supporting.
Another scheme for eliminating the dead-end and switch
loss incident to the use of tapped coils, is to use separate
coils of the proper inductance values for the wavelength
RECEIVING APPARATUS 197
ranges to be covered. Two types of such coils have already
been illustrated (Fig. 7). These are marketed in many
forms and in all sizes from coils of a few turns to those
containing 1500 and more. A special form of zig-zag wind-
ing is sometimes employed which renders the coil mechan-
ically firm and self-supporting, and is claimed to reduce the
distributed capacity and coil losses. The reduction of the
distributed capacity increases the wavelength range, but
this is of no importance; the principal result of reducing the
capacity is that it diminishes the dielectric losses due to
the tendency for the currents to forsake the inner turns,
and flow by capacity paths toward the outer ones.
The above instructions are general and pertain to all
radio coils in which a low resistance is to be obtained. As
in the case of the antenna, I have outlined what appears to
be the best practice, but do not want the unsophisticated
reader to suppose that less elegant construction will not
give satisfactory results. He may, if he chooses, omit the
mica-insulated switches, and to space the turns on the
coil, and so forth, and still hear good signals; but the more
serious experimenter who takes pride in a job well done
and who has a proper appreciation of the relative importance
of various improvements, will want to have this information.
(a) Single-circuit Construction. The single-circuit tuner con-
sists of an adjustable coil and a variable condenser. When
regenerative reception is undertaken by the "tickler coil"
method (see Art. 73) an auxiliary coil is coupled to the
antenna coil. In this case the double-circuit construction
may be used, employing the secondary as the "tickler coil"
and omitting the secondary tuning condenser. An example
of this assembly is shown in Fig. 109, where the tapped
198 RADIO TELEPHONY
antenna coil and the antenna tuning condenser are clearly
shown. For the average amateur antenna with a funda-
mental wavelength of 200 meters (including the single
wire receiving antenna, Art. 36), a coil wound with from
60 to 90 turns of No. 18 copper wire on a tube 4" to 5"
in diameter, tapped every 15 turns, will cover the con-
ANTENNA
TUN1NS
CONDENSER
Fig. 109. Illustrating single-circuit tuner provided with tickler coil for regen-
erative amplification.
templated wavelength range (150-600 meters). The turns
are preferably separated by a distance equal to the wire
diameter. A variable condenser having a maximum ca-
pacity of .001 mfd. is suggested for use with this coil. The
connections of the single-circuit tuning unit are indicated
in Fig. 110.
RECEIVING APPARATUS
199
(b) Double-circuit Construction. The primary circuit of the
double-circuit tuner is arranged as described for the single-
circuit scheme. The secondary consists of 50 turns of
No. 18 d.c.c. wire wound on a 3" tube, and is arranged to
rotate within the larger primary coil so that the coupling
can be varied. This arrangement of coils is popularly
referred to as a "vario-coupler." It is suggested that this
coil be wound in a double layer, 25 turns on each side, and
that a tap be taken out at 25 turns. Unless a double layer
1 1 1
ANT.
TUNING
CONDENSER
TAPPED-2
INDUCTANCE
Fig. 110. Electrical connections of single-circuit tuner.
is used the coil will be too long to be rotated within the
5" primary coil. Used in conjunction with a secondary
tuning condenser of maximum capacity, .0005 mfd., the
secondary coil will cover a wavelength range of 150-600
meters. The connections of the complete tuning unit are
given in Fig. 111.
Sometimes the antenna variable condenser is omitted and
the circuit tuned by means of an inductance alone. For
operation above the fundamental of the antenna this
20O
RADIO TELEPHONY
method is more efficient since it involves the introduction
of no more wire (which has resistance) than is absolutely
ANT.
GNO.
~i-- SEC. TUNIN6 CONDENSER
Fig. 111. Electrical connections of double-circuit tuner.
necessary to attain the desired wavelength. The coil may
be tapped every ten turns, and then tapped every single
Fig. 112. Type of vario-coupler for double- circuit tuning, using tapped primary
coil.
turn for ten turns, to secure fine adjustment. An example
of this form of construction is shown in Fig. 112. In this
RECEIVING APPARATUS 201
case the coil is tapped every 7 turns (right hand switch)
and every single turn for 7 turns (left-hand switch).
Another variation of the above tuning arrangement con-
sists in eliminating the secondary variable condenser and
tuning by means of a variable inductance (variometer).
A few writers and manufacturers insist that this is a more
efficient combination. Leaving aside the technical ques-
tions involved, the condenser scheme is considerably more
satisfactory when regenerative reception or radio frequency
amplification is attempted, and gives a quieter and more
easily controlled receiver. This is sufficient to justify its
recommendation in this book.
72. Simple Type of Receiver Employing a Crystal De-
tector. The simplest detector of the modulated radio
frequency currents used in radio telephony consists of a
sensitive mineral or "crystal" upon which a contact is made
with a metallic probe or another mineral. This combina-
tion possesses the important property of allowing more
current to pass through it in one direction than in the other.
The application of this property in the detection of radio
signals will be obvious to the reader, and has been dis-
cussed in Chapter I. The connections of the crystal
detector to the single-circuit tuner described in the pre-
ceding article are shown in Fig. 113.
Here D represents the crystal detector; C is a blocking
condenser whose purpose is twofold: first, to shunt the
radio frequency currents around the telephones, and second,
to give the circuit the proper audio frequency character-
istics for the emission of a smooth note by the telephones.
A paper condenser of approximately .01 mfd. capactiy will
be indicated with the usual type of telephones. This
202
RADIO TELEPHONY
condenser is generally included in the crystal detector sets
supplied commercially (Fig. 114), or can be easily con-
structed by the reader according to the following method:
Cut out two strips of tinfoil 3 in. wide and 2 ft. long, and
three strips of thin paraffined paper, 4 in. wide and 2 ft.
3 in. long. After pasting the tinfoil sheets on each side of
one of the paper strips, sandwich the whole between the
remaining two paper strips and roll it up, binding the roll
UJANT.
SINGLE CIRCUIT TUNER
TEL.
GNP.
Fig. 113. Electrical connections of simple receiving set, employing single-cir-
cuit tuner and crystal detector.
with tape or cord. Connections are made to the two tinfoil
sheets. The telephone receivers require no description.
At the left of the figure is shown a "test buzzer" which is
to be used in adjusting the detector, that is, in probing the
surface of the crystal for a sensitive spot. This consists of
a radio buzzer operated by one dry-cell, from one of whose
contacts a connection is made to the ground and from the
other to a few turns of wire wrapped around, but no con-
nected to, the antenna lead. The buzzer shocks the circuit
RECEIVING APPARATUS 203
into radio frequency vibrations and when the detector is
properly adjusted, the buzzer note will be heard in the
telephones.
Several sensitive minerals are known, and these or com-
binations of them have been employed at various times in
crystal detectors. The most satisfactory of these are:
galena (lead sulphide, PbS), used with a light steel spring
contact ] fused silicon, used with a sharp rather firm metallic
Fig. 114. Commercial type of crystal detector set, providing two detectors with
switch for including either of them in the circuit.
contact, or with a pointed probe made of antimony; the
"perikon" combination of zincite (zinc oxide, ZnO) and
chalcopyrite (copper-iron sulphide, CuFeS 2 ), often used with
a biasing battery; and carborundum, used with a steel point,
firm pressure, and a biasing battery. A typical commercial
detector set is shown in Fig. 114. Here two detectors
are provided, either of which may be thrown into the cir-
cuit by means of the small rotary switch. The blocking-
condenser is in this set mounted in the Lase, the left-hand
204 RADIO TELEPHONY
pair of binding posts being connected to the tuner and the
right-hand pair to the telephones.
The combination of single-circuit tuner and crystal
detector just described constitutes a simple type of receiver
which will give satisfactory response from broadcasting
stations within a radius of five or ten miles, and will appeal
to the beginner who is located within this distance and who
wishes to make his dbut in the radio telephone field with a
minimum of expense. In erecting the antenna, the latter
part of Chapter III, particularly Art. 36, should be con-
sulted; and lightning protection of the type prescribed by
the underwriters (see Appendix A) should be provided.
73. Regenerative Method of Amplification. Reference
has already been made to the amplifying action of the
audion, whereby small changes in the voltage of the grid
are capable of producing substantial changes in the plate
current. This property is utilized, as explained in Chapter
II, in the application of the audion as a self -excited generator
of alternating currents. Here some of the power of the
plate circuit is brought back to the grid circuit for the
purpose of exciting it and sustaining the oscillations. Now
if the back-coupling between plate and grid is less than a
certain critical amount, the audion will not oscillate, but
nevertheless the energy feed-back into the grid circuit is
manifested by an amplification effect in that circuit. One
way of looking at this is to follow the process through
step by step, forgetting for the moment that the whole
action takes place almost instantaneously. For concrete-
ness suppose that the audion is connected across the tuned
secondary circuit of a coupled receiver. The signals im-
pressed upon its grid are amplified, and some of this ampli-
RECEIVING APPARATUS 205
fied, or extra energy, is brought back to the grid circuit and
being in phase, augments the signal voltage. It is then
re-amplified, brought back to the grid, and so forth, the
course of this vicious electrical circle being run until the
steady state is arrived at. This is referred to as "regenera-
tive' ' amplification, and the effect was discovered experi-
mentally by Mr. E. H. Armstrong. In such circumstances
the tuned grid circuit acts very much as if its resistance had
been reduced; in fact, when the back-coupling is sufficient
to cause the system to break into sustained oscillations, the
resistance is sometimes referred to as being zero. But this
view should be applied with caution, and in any case must
not be used as an argument against the importance of
reducing the resistances of the coils and condensers to
which the grid of the audion is connected. For in spite
of the artificial reduction of the circuit resistance by the
regenerative action, the signal response still remains inversely
proportional to the actual circuit resistance and the amount
of amplification which the regeneration yields will be
governed by this resistance* (see Fig. 117).
A number of methods of feed-back have been employed,
and many of these have already been described in Art. 40
in connection with the generation of radio frequency power
by means of the audion. The most satisfactory and popular
forms for receiving purposes are the " tickler" coil circuit
(Fig. 69 (a)) and the Armstrong tuned-plate circuit (Fig.
72). The following discussion will be devoted to these
important circuits.
* This fact was established experimentally by my former colleague, W.
G. Ellis, radio expert of the U. S. Navy, to whom I am indebted for the use of
the curves of Fig. 117.
206
RADIO TELEPHONY
(a) Tickler Coil Circuit. The connections for this method
of feed-back are illustrated in Fig. 115 in relation to the
double-circuit tuner, Fig. 111. If the single-circuit tuner is
used, the audion is connected across L as usual and the
tickler coil is inductively coupled to it. The function of the
condenser in the plate circuit is to pass the r.f. currents
around the telephones and "B" battery. The regenerative
action depends upon the wavelength, tickler coupling, ratio
of L to C in the tuned grid circuit, and the effective resistance
TEL.
""B"BATT.
A" BATT.
Fig. 115. Schematic diagram of the tickler coil method of regenerative ampli-
fication.
of this circuit. It is controlled by the tickler coupling,
which is made variable for this purpose, and is increased by
increasing this coupling, decreasing the ratio of C to L in
the grid circuit, and decreasing the resistance of this circuit.
For a wavelength range of 150-600 meters, about 20 or
30 turns of No. 26 d.c.c. wire on a tube 3" in diameter will
make a proper tickler coil for use in conjunction with the
tuners previously described. The resistance of this coil is
of slight moment and special precautions need not be taken
(as in the case of the tuned-circuit coils) to keep it low.
RECEIVING APPARATUS 207
It is of some importance which terminal of the tickler coil
is connected to the plate and which to the "B" battery; but
the proper connections can be determined experimentally
after the apparatus has been assembled. If oscillations or
regenerative action are not obtained at the first trial,
reverse the connections to this coil, or reverse the coupling
between it and the secondary or antenna coil.
In operating this circuit the reader will note that, starting
with a tickler coupling of zero, as this coupling is increased
the strength of the signal will increase until a certain critical
point is reached beyond which the circuit breaks into oscil-
lation, a state whose existence can be detected by touching
the grid terminal of the audion with the finger. If oscilla-
tions are present this will quench them and a dull thud or
click will be heard in the telephone receivers. If "spark"
signals are being received their tonal characteristics will
probably be obliterated before this critical coupling is
reached, due to the overlapping of their wave trains and
the fortuitous character of the spark discharge as a radio
frequency time event. Their musical sounds will then be
supplanted by a rough note resembling escaping steam, and
generally the strength of this signal will be very much
greater than the strength obtained when the musical note
is preserved. Radio telephone signals will generally be
lost after passing into oscillation, but not necessarily. The
reception of c.w. signals by this method will be considered
in the next article.
I have been accustomed to represent this action graph-
ically for the instruction of Navy radio operators by a
diagram shown in Fig. 116. Here the distance from the
dotted circle, or the depth of the shaded portion, represents
208
RADIO TELEPHONY
crudely the change in the intensity of the signals as the
tickler coupling is increased from zero. The regenerative
region is regarded as extending from zero to the point at
which oscillation starts.
The effect of the tickler coupling upon the intensity of
the signals is also shown graphically in Fig. 117. The
ordinates (vertical distances) represent the telephone (sound)
energy, and the abscissae (horizontal distances) represent
the tickler coupling. The amount of amplification obtainable
POINT OF
OSCILLATION
REGENERATION
OSCILLATION
ARC SIGNALS
SPARK SIGNALS
TICKLER COUPLER
Fig. 116. Representing change in signal intensity with variation of the tickler
coil coupling.
by this method is indicated; in the case of no inserted
resistance (R=O, the normal receiving circuit) this amounts
to a magnification (energy) of about 26 times. Since a
magnification of this amount would be produced by a well
designed one- or two-stage radio frequency cascade ampli-
fier; the substantial economy of the regenerative method is
thus demonstrated. The fact that the tickler coupling, or
rather the feed-back of energy from the plate circuit, does
not compensate for the actual resistance of the grid circuit,
at least so far as the signal response is concerned, and that
RECEIVING APPARATUS
209
this circuit should be well designed, is also shown by the
several curves, representing the effects of various inserted
resistances in the tuned grid circuit.
For reasons which will be brought out in part in the
course of the discussion of the Armstrong tuned-plate cir-
cuit, the tickler coil method is regarded as the most satis-
factory method of regenerative amplification". A receiver
Z5
O ZS 50 75
TICKLER COUPLING.
Fig. 117. Graphical representation of the amplification due to the tickler
coil method of regeneration, showing the effects of inserting resistance in the
tuned grid circuit (Ellis).
compounded of the double-circuit or single-circuit tuner,
and one audion bulb used with a tickler coil as shown hi
Fig. 115 will be very satisfactory for short distance work
and as noted will give the same response as a good two-
stage radio-frequency amplifier with detector, operating in
the usual manner. But it will, of course, be harder to
adjust and to operate than such a receiver.
14
210
RADIO TELEPHONY
(b) Armstrong Tuned-plate Circuit. The connections of this
method of regeneration are illustrated in Fig. 118 in relation
to the double-circuit tuner of Fig. 111. The corresponding
connections for the single-circuit tuner will be obvious to
the reader and need not be indicated. As explained in
Art. 40, the feed-back in this case takes place through the
small condenser formed by the grid and plate electrodes of
the audion and is induced by the insertion in the plate cir-
Fig. 118. Schematic diagram of the Armstrong tuned-plate method of regen-
erative amplification.
cuit, of the inductance L. This inductance is made variable
for the purpose of controlling the regeneration and usually
takes the form of a variometer. A variometer having a
wavelength range (with its inherent capacity) of 150-500
meters will be suitable for this purpose.*
* The inductance required for regeneration is given approximately by the
formula:
L=
1
(C m + C
cj
f*
wherein = 2^n (n = frequency); C total capacity across L including
the plate-filament capacity of the audion and the self-capacity of the vari-
ometer; C m = grid-plate capacity of the audion; // = amplification factor of
the audion. Inductance L and capacity C are expressed in henries and farads
respectively.
RECEIVING APPARATUS
211
As in the case of the tickler coil circuit the regenerative
action is manifested by a decrease in the resistance of the
tuned (LiCi) grid circuit. The audion with its inductive
-80
10
2468
Load Inductance (V..H.)
Fig. 119. Curves showing the relation between the effective input conductance
of an audion and the inductance in its plate circuit, for a typical case <W. .
"/" tube at 627 meters').
plate circuit may therefore be looked upon as supplying a
negative resistance which counteracts to some extent the
actual (positive) resistance of the grid circuit. The
amount of negative resistance so introduced will depend,
212 RADIO TELEPHONY
of course, upon the amount of inductance in the plate
circuit. The relation between the two is clearly illustrated
in Fig. 119, which is reproduced from my Physical Review
paper previously referred to (Art. 40), and is based upon
computations from the theory of a typical case.
The ordinates of these curves are marked "Input Con-
ductance" and are intended to represent the effective
values of the conductance added across the tuning con-
denser Ci by connecting the audion to it. As the reader
will observe, this is negative for a range of inductance
values and represents regeneration throughout this range.
The maximum regenerative effect is obtained in approxi-
mately the middle of the range. The signal strength may
be taken as proportional to the input conductance, and this
will furnish an idea of how the regenerative adjustment
amplifies the signal. The several curves represent the
effects of various inserted resistances in the plate circuit,
and justify a previous statement to the effect that rather
large losses may be tolerated in the plate variometer. The
introduction of 100 ohms, for example, would have an
almost negligible effect. Thus the plate variometer may be
made very much smaller than the average commerical form
and can be wound with fine wire, e. g., No. 30 B. & S., or
contain an iron core, without appreciably affecting the
regenerative action.
Now in addition to its action in reducing the resistance
of the tuned grid circuit, the audion with an inductive
plate circuit also increases the capacity of this circuit, that
is to say, it acts like a condenser as well as a conductance.
Figure 120 shows the effective capacity of the audion as a
function of the plate circuit inductance, in these circum-
RECEIVING APPARATUS
213
stances. The ordinates represent the capacity; the abscissae,
the plate inductance. I would like the reader to notice
that this capacity is in parallel with that of the tuning
condenser (see Fig. 118), and its variations will affect the
tuning of this circuit. It is unfortunate that the audion
4 6
Load Inductance (M.H.)
10
Fig. 120. Curves showing how the effective (input) grid-filament capacity of an
audion is affected by an inductance in its plate circuit (W. E. "J" tube).
capacity changes most rapidly with the inductance values
most favorable to regeneration (compare Fig. 119, curve
for R = O (627 meters), and Fig. 120, curve for 600
meters). This is objectionable in that the adjustments for
tuning and regeneration are not completely independent,
and the inconvenience of this is aggravated by using a
2i 4 RADIO TELEPHONY
tuning system (such as a variometer for tuning the secondary
circuit of the coupled circuit tuner instead of a variable
condenser), in which the tuning capacity is small. This is
one reason for the expressed preference for the tickler coil
arrangement; there the adjustments of tuning and regenera-
tion are independent because the amount of inductance
included in the plate circuit (the tickler coil) remains
constant.
74. Principle of the Heterodyne Method of Reception.
The heterodyne method of reception and detection is used
in connection with the unmodulated or sustained radio
frequency waves employed in c.w. telegraphy. In com-
mercial work these waves are supplied by a high frequency
alternator; in amateur work, by an audion power oscillator.
Its basic principle is easily apprehended and may be
explained as follows:
For concreteness, suppose that a radio frequency wave of
constant amplitude is incident upon the receiving antenna,
and produces hi the tuned circuits currents of the same
nature. The presence of these currents will not be detected
by an audion (or crystal) detector operated in the usual
manner, for the wave simply produces in the plate circuit
an elevation of the average plate current for its duration.
The radio frequency vibrations of the plate current are not
considered, because, as has already been explained, they
are incapable of affecting the telephone receivers, or of
being heard if they did succeed in affecting them. The
presence of this signal is made known to us in the heter-
odyne method by adding to it another radio frequency wave
of constant amplitude but of slightly different frequency.
The combination of the two waves results in a wave of
RECEIVING APPARATUS
215
frequency equal to the mean of the two frequencies, but
with an amplitude which varies at a frequency equal to
one-half the difference of the frequencies. This is known
as the phenomenon of "beats" and is illustrated in Fig.
121.
In the practical application of this principle the extra
radio frequency current is generated at the receiving station
and its frequency is adjusted there until the beat note has
(a)
Fig. 121. Illustrating the formation of "beats" (c) by the addition of waves
(a) and (b) of slightly different frequencies.
the desired frequency within the audible range. Thus the
whole process simply consists in artificially producing a
modulation of the incoming signal. It is then detected
as any other modulated signal, a radio telephone signal for
example, would be detected. According to whether the
local oscillation is produced by a separate audion oscillator
or occurs in the receiver circuit itself, the reception method
is referred to as the separate-heterodyne or self-heterodyne
(autodyne) method. For brevity the former method is
2l6
RADIO TELEPHONY
often called simply the heterodyne method, the latter the
autodyne method.
One of the most valuable features of this method of
detection is the large amplification that is obtainable
simply by increasing the strength of the local oscillation.
It is found experimentally that in both the heterodyne and
autodyne systems the telephone response for signals of
STIMULUS (IMPRESSED E.M.R)
Fig. 122. Showing the relation between the telephone response and the signal
voltage, for various methods of signalling and reception.
moderate strength, is roughly proportional to the product
of the signal strength and the strength of the local oscilla-
tion. This is in contrast to the detection of "spark" or
other modulated signals (radio telephone signals) where the
response is proportional to the square of the signal strength.
The relation as determined experimentally is shown in
Fig. 122, the ordinates representing the telephone response
RECEIVING APPARATUS
217
and the abscissas the signal voltage (stimulus).* The
linear relation between signal voltage and response, char-
acteristic of the heterodyne and autodyne methods, com-
pared with the second power curve of the ordinary detec-
tion, indicates that by this means the feeble signals are
amplified just as much as the strong ones, whereas in the
latter method the weak ones lose in detection while the
strong ones gain. Thus
in c.w. signalling the sig-
nal strength goes down
inversely as the distance
from the transmitting
station, and in spark or
i.c.w. telegraphy or radio
telephony the response
decreases inversely as the
square of the distance.
The inherent superiority
of the c.w. method of
telegraphic communica-
TEL.
LOCAL
OSCILLATOR
Fig. 123. Connections of a simple re-
ceiver for c.w. signals using the heterodyne
method with separate oscillator.
tion is therefore manifest.
75. Receiving Circuits Applying the Heterodyne Method.
The term heterodyne is used here to mean the separate-
heterodyne method in which the local oscillation is furnished
by an extra oscillator. This oscillator consists simply of a
small receiving audion, connected to a suitable oscillating
circuit, and so arranged that the frequency of the gen-
erated oscillations can be readily controlled by the operator.
A very reliable circuit for this oscillator is the Hartley
* This illustration is reproduced from my article, "The Radio Compass,"
published in the Year Book of Wireless Telegraphy and Telephony, London, 1921.
2i8 RADIO TELEPHONY
circuit illustrated in Fig. 70, and Fig. 123 shows the con-
nections for the complete heterodyne receiver, in which
an oscillator of this type is used. Here the receiver con-
sists of a coupled-circuit tuner and ordinary audion detector.
It will be understood that several stages of audio frequency
amplification may be added. If radio frequency amplifica-
tion is to be used the local oscillator should be coupled to
the grid of the detector audion or to the plate circuit of the
last radio frequency amplifier audion, and not to the sec-
ondary of the tuner as shown above. The coil L 2 (100
microhenries) is an ordinary receiving circuit coil (for
example, 30 turns of No. 18 d.c.c. wire on a tube 4" diameter),
tapped approximately in the middle. The frequency of the
oscillations is adjusted by means of the variable condenser
C 2 (max. cap. = .001 mfd.). The range of the oscillator is
from 150 to 600 meters. If it should be found that the
local oscillations are too strong, the oscillator may be
set at a wavelength equal to 2, 3, 4, etc., times the wave-
length of the received signal instead of approximately at
this wavelength; then the feebler 1st, 2d, 3d, etc., har-
monics can be used to produce the beat note. The adjust-
ments of both the strength of the local oscillations and the
frequency of the beat note are more easily made when the
harmonics are used.
The heterodyne receiver is rather hard to operate, for in
picking-up signals the tuner and oscillator must be manipu-
lated together, keeping their wavelengths approximately
equal while exploring the wavelength scale. Its main
advantage is that once the adjustment is made the fre-
quency of the beat note is not affected by small capacity
variations due to a swinging antenna or to the movements
RECEIVING APPARATUS 219
of the operator's hand or body. It is assumed, of course,
that the local oscillator employs a large tuning condenser
and is either placed out of the operator's way or well
shielded from his electrical influence.
76. Receiving Circuits Applying the Autodyne Method.
The autodyne or self -heterodyne method requires no special
apparatus, a regenerative circuit being used and the feed-
back coupling adjusted until oscillations are obtained. The
frequency of the oscillations corresponds to the wavelength
adjustment of the circuit, so that by slightly detuning the
receiver the wavelengths of the signal and local oscillation
may be made sufficiently different to give the desired beat
note. The amount of detuning required is not large at
short wavelengths and the signal strength is but slightly
diminished by the maladjustment.
The theory of the autodyne audion receiving c.w. signals
is very complicated and not generally well understood.
The merit of the system, or the strength of the telephone
response for a given signal strength, depends upon the
instability of the oscillations in the circuit. Specifically,
with the tickler coil arrangement, the sensitivity is closely
proportional to the rate at which the average plate current
changes with a slight change in the tickler coupling.* This
will depend to a large extent upon the stability of the oscil-
lating state, and as a general rule we may say that anything
that can be done to make the oscillations less stable will
increase the intensity of the beat note. For this reason the
tickler coupling should not be increased much above the
* This and other statements of this section rest upon an experimental
study of the heterodyne and autodyne methods of reception which the writer
made for the Navy in 1918.
22O
RADIO TELEPHONY
point at which oscillations start. I have also found the use
of a grid biasing voltage preferable to a grid condenser and
resistance for promoting instability, and the bias should be
made adjustable so that the greatest change in plate current
for a given change in tickler coupling can be obtained. The
tickler coil circuit is more suitable for this work than the
tuned-plate circuit, and connections for a complete receiver
i B
Fig. 124. Connections of a simple receiver for c.w. signals using the heterodyne
method with self-oscillation.
embodying this circuit and including an adjustable grid
bias are given in Fig. 124.
The operation of the autodyne circuit is much simpler
than that of the heterodyne; in picking up signals it is
simply necessary to run over the gamut of wavelengths,
adjusting the regeneration at the same time to maintain
the circuit in the oscillating state. In the next article a
special receiver will be described in which even this latter
adjustment is eliminated, the oscillating state being auto-
matically maintained during the adjustment of the tuning.
RECEIVING APPARATUS 221
The main disadvantage of the method resides in the effect
of the circuit constants upon the wavelength of the local
oscillation, whereby small capacity changes produced by a
swaying antenna or by the hand of the operator cause
troublesome changes in the frequency of the beat note.
It is desirable for this reason (and also to reduce the radi-
ation of undamped waves from the receiving antenna which
may interfere with other receiving stations) to use a double-
circuit tuner and to properly shield all parts of the circuit
likely to be within the electrical influence of the operator.
The use of the double-circuit tuner, of course, considerably
complicates the operation and makes it very difficult to
pick-up signals; nevertheless in the hands of a patient and
expert operator it has many advantages.
77. The Reinartz Tuner. This receiver, devised by Mr.
J. L. Reinartz, is, according to the testimonials of many
amateurs who have used it, the most satisfactory apparatus
for the reception of c.w. signals. It is simple, easily con-
structed and has the special advantages for c.w. work of
having but one adjustment for the wavelength, and oscil-
lating freely and with proper restraint over the entire
wavelength range (130-400 meters). The latter feature is
especially convenient in picking up c.w. stations, and is
attained by a modification of the Chambers feed-back
circuit and the combination of inductive and capacitative
feed-back first exploited and put into practical apparatus
by Messrs. Jones and Priess, formerly engineers of the
U. S. Navy. The following description of the Reinartz
receiver is abstracted from articles appearing in the June,
1921 and March, 1922 issues of "QST," which the reader is
urged to consult for a more detailed account.
222
RADIO TELEPHONY
A schematic diagram of the connections is shown in
Fig. 125. The coils Zi and L 2 are wound with No. 26 s.c.c.
copper wire in spider-web fashion on nine " spokes" around
a 2^-inch center, the completed coil being about 5 inches in
diameter (see Fig. 126). The plate coil comprises the first
45 turns; the wire is then cut and another coil of 40 turns is
wound on the same form and in the same direction. Taps
OUTSIDE END
/-SWITCH FOR
WAVELENGTH ADJUSTMENT
INSIDE END
OF COIL.
Fig. 125. Connections of the Reinartz receiver for c.w. signals.
are brought out to the switches as follows: for the plate
winding, at 15, 30 and 45 turns; and upon starting the
second coil at 2, 4, 5, 6, 7, 8, 9 turns, with a ground tap at
the 10th turn; and continuing the winding, at the 25th,
30th and 40th (outside) turns. The connections of the
taps to the switches are clearly shown in the circuit diagram;
also the functions of the switches are marked.
RECEIVING APPARATUS
223
The condenser C 2 (23 plates; max. cap. about .0005 mfd.)
controls the feed-back and the condenser Ci (13 plates;
max. cap. about .0003 mfd.) is used for tuning. For min-
imization of the capacity effects of the operator's body the
rotary plates of the condenser C 2 should be connected to
the antenna, those of Ci to the earth. The function of the
r.f. choke-coil X is to pre-
vent the short-circuiting of
the radio frequency output
circuit of the tube by the
telephones (or primary of
an audio frequency ampli-
fier transformer) and the
"B" battery. This may be
of the 3 m.h. type de-
scribed in Art. 44 (d).
A good mechanical ar-
rangement of the parts is
shown in the exterior and
A CLOSE-UP Or THE INDUCTANCE
Fig. 126. Illustrating "spider-web'
of the Reinartz c.w. receiver.
coil
interior views of the re-
ceiver, Figs. 127 and 128.
78. The Armstrong Super-heterodyne Method for Re-
ception at Short Wavelengths. The super-heterodyne
method of reception is another product of the ingenuity of
Mr. E. H. Armstrong and represents his solution of the
important and difficult problem of amplifying signals of
short wavelength. The performance of cascaded radio
frequency amplifiers at the high frequencies represented
by wavelengths of the order of 100-600 meters is consider-
ably poorer than their performance at lower frequencies,
due to actions which will be more fully considered in the
224
RADIO TELEPHONY
PLATE CIRCUIT GRID CIRCUIT
FEEDBACK CONDENSER
TUNING CONDENSER
ANTENNA CIRCUIT
Fig. 127. External view of the complete receiver.
CONNECT COILS FOR LONGER WAVELENGTHS HERg;
PLATE FILAMENT GRID.
TAPPED
SPtDERWES
INDUCTANCE
Fig. 128. Showing arrangement of coils and condensers in Reinartz receiver.
RECEIVING APPARATUS 225
section on amplifiers. The method of the super-heterodyne
consists in changing the natural high frequency of the
signal to a lower frequency more suitable for amplification
by the cascade process. The frequency change is produced
without appreciable distortion and all the characteristics
of the original signal are preserved. This is accomplished
by an application of the heterodyne principle; by adding
to the signal another oscillation whose frequency differs
from that of the signal by an amount necessary to give a
beat note of the lower radio frequency. This combination
of currents is then rectified and yields an alternating current
whose frequency is equal to the difference of the two
frequencies.
This process has already been explained, and is nicely
illustrated in Fig. 42. Suppose that the middle curve ( g )
in this photograph represents the beats which have been
formed as a result of the combination of the signal and local
oscillation. The lower curve then represents the plate
current of the detector; and if the plate circuit contains an
LC branch tuned to the beat frequency, the higher frequency
currents will be suppressed, leaving the beat frequency
current shown in the upper trace. Thus the system acts
simply as a frequency transformer; and the output is of the
proper wavelength to be efficiently handled by a cascaded
radio frequency amplifier of the usual type.
The electrical scheme is illustrated in Fig. 129. This
shows a coupled circuit tuner and detector connected to-
gether in the normal fashion. The detector may be ar-
ranged for regenerative amplification; this is not shown, but
a better response will be obtained by its use. The local
oscillator is of the type previously described in connection
15
226
RADIO TELEPHONY
with the separate-heterodyne method. Autodyne, or self-
heterodyne methods, are not convenient because to produce
a beat note of say 50,000 cycles (6000 meters wavelength)
with a signal of 300-meter length, the oscillator's wavelength
must differ by 15 meters, and if the receiver furnished its
own oscillation a detuning of this amount would appre-
ciably diminish the signal voltage. The transformer T is
tuned to the wavelength to which the signal has been
changed. It is customary, I believe, to change to a wave-
length of 6000 meters. This could be conveniently reduced
OSCILLATOR
|p
Fig. 129. Schematic diagram of the Armstrong super-heterodyne receiver.
to 2000 or 3000 meters, thus permitting the efficient use of
the choke-coil coupled amplifier to be described later.
After amplification the radio frequency signal is detected
in the usual way. The use of two detectors may confuse
the reader, but it should be remembered that the first
detector is used in the transformation of the frequency; the
second detector is necessary, for the new frequency is still
above audibility.
A complete system is illustrated in Fig. 130 and a suf-
ficiently detailed description of the principal constants and
parts is given to enable the reader to construct it. The
RECEIVING APPARATUS
227
73 ^
.
JH JcJ
228 RADIO TELEPHONY
separate heterodyne oscillator employs an ordinary receiving
audion, and while enough power is available, rather close
coupling between it and the secondary coil LI is necessary
to force the proper current through the high impedance
which the tuned secondary circuit offers at the oscillator's
frequency. For example, to change a signal of 300 meters
wavelength to 3000 meters, the oscillator would be adjusted
to 330 meters; a sharply tuned secondary circuit would
offer a substantial impedance to a 330 meter oscillation.
The oscillator employs the Hartley circuit and has already
been described (Art. 75). The 3000 meter transformer L 2 C 2
may contain an iron core The double winding is simply
a convenience to eliminate the isolation condenser in the
grid lead of the first amplifier audion, which would be
indicated with a single winding in order to keep the plate
voltage off the grid of this tube. These coils may be wound
together in a multi-layer form of rectangular cross-section,
with 250 turns apiece, 3^-" mean diameter; or they may
consist of two 3 00- turn honeycomb coils, bound closely
together with tape. The writer prefers the use of iron-
cored coils at these wavelengths; a core of "radio-frequency-
iron," built up of laminations of .002 " (2 mils) thickness,
approximately 1-J-" x 3" inside, and f " cross-section, wound
with 300 turns of No. 36 s.s.c. wire in a coil of diameter
If ", will be suitable. The same construction may be used
for the output coil, but in this case but a single winding is
necessary, for a grid condenser is used (for detection) in any
case, and will serve to keep the plate voltage off the detector
grid. If the coils are wound close together so that their
coupling is close, but one tuning condenser will be neces-
sary as shown in the diagram. The r.f. amplifier will be
RECEIVING APPARATUS 229
described later. The detector arrangement is the usual
one; and the signal may be amplified after detection by
means of audio frequency amplifiers of the usual type. This
amplifier is connected in place of the telephones.
In putting the circuit into operation the input transformer
C 2 I/ 2 and the output transformer L 3 C 3 are first tuned to a
convenient wavelength, let us say 3000 meters. This may
be done with a wavemeter, or may be approximately
adjusted in the ordinary course of operation by manipulating
C 2 and C 3 until the signal has its maximum intensity. Sub-
sequent adjustment of these circuits should not be made.
There are three tuning adjustments: the antenna circuit,
secondary circuit, and local oscillator; in addition, there is
the tickler adjustment for regeneration. This multiplicity
of adjustments makes the operation of the system rather
tedious and difficult, and for that reason and also to avoid
interference, many amateurs prefer to eliminate one adjust-
ment by using a coil antenna. The coil is connected in
place of the secondary coil LI (Art. 38) and a small series
inductor is included between it and the tuning condenser Ci,
to which the tickler coil and local oscillator may be coupled.
For the reception of c.w. signals the use of a second local
oscillator is convenient. This need not have a wide range
of wavelengths, in fact, once adjusted it requires no further
attention since all signals may be transformed to the same
wavelength and will then give the same beat note with the
local oscillator. The wavelength of the local oscillator
should, of course, be in the neighborhood of that to which
the r.f . amplifier system is adjusted, and to which the signals
are changed by the heterodyne process. One or two stages
of r.f. amplification are usually necessary to compensate
230
RADIO TELEPHONY
RECEIVING APPARATUS
231
-
|
c
<y . .
W to >
U So
ri
4-1 3
O -M
w
w
V-
IS
^ -S^^JS^^I
1 l|llil^*J
03 JS^S
r<3
o v>
O 'g
^x? ^
^ s
^ fc fe i
232 RADIO TELEPHONY
for the unavoidable losses occurring in the wavelength
changing process; stages in. excess of this are pure gain.
This is a very remarkable system and one which yields
amplification which cannot be attained with ordinary r.f.
amplifiers operating at the natural signal wavelength (using
the high-capacity tubes available for amateur use), with
less than two or three times the number of stages. It was
mainly through the use of this receiver that Mr. P. F.
Godley was able to hear in Scotland many of the American
amateur stations some of them using but a few watts of
transmitting power during the recent trans-Atlantic tests
conducted by the A.R.R.L. In making this statement
there is no intention to discount the very important factor
of expert operation.
79. Electrical Characteristics of Audions Used as Detec-
tors and Amplifiers. In the section treating of the applica-
tion of the audion as a power generator, a compilation of
the electrical and mechanical data and characteristics of
the various commercial types of tubes was given. It is
proposed here to furnish a similar collection of data for the
tubes used for receiving purposes, as amplifiers and detectors.
A rather wide variation exists in the mechanical construction
and the electrical properties of commercial types of tubes,
expressing the ideas and preferences of their designers, and
adapting them for different applications.
Figure 131 illustrates some of the popular and familiar
types of audions, of which many are available for amateur
(experimental) use. The accompanying table gives the
important electrical data of these, and another of which a
photograph was not available. This data rests almost
entirely upon tests made by the writer, or under his direc-
RECEIVING APPARATUS 233
tion, so that full responsibility for their correctness will
have to be assumed. While an effort has been made to
record good average values, but a limited number of tubes
were available for test and wide variations from these
values may be expected with most types. The inclusion
of the manufacturers' names is slightly irregular, but is
justified by the common practice of referring to the audions
by these names. The common designations of the praxis,
and the type numbers assigned by the Navy and Army, are
also given.
With regard to "amplification factor," "plate resistance"
and "mutual conductance," these are technical terms
defining certain electrical properties as follows:
The amplification factor is the ratio between the plate
voltage and grid voltage for constant plate current.
The plate resistance is the rate at which the plate
voltage changes with respect to the plate current.
The mutual conductance is the rate at which the plate
current changes with the grid voltage.
These definitions are given in most text-books on the
subject, to which the non-technical reader is referred for
further information. The plate-grid and plate-filament
capacities of the tube (not including the socket) have also
been given; these are important in the design of radio
frequency amplifiers.
All the tubes illustrated are of the so-called "hard" or
high vacuum type. The Radiotron, UV-200 and certain
other types of which the data are not available, notably the
so-called Audiotron, contain small quantities of gas, which
gives to the tube a characteristic more suitable for detection
234 RADIO TELEPHONY
than that of the hard tubes. The use of gas, however,
makes the device critical in adjustment and a careful
regulation of the grid and plate voltages and the filament
current is necessary in order to realize the substantial
improvement thus afforded. These adjustments are con-
veniently made if the proper "potentiometer" connections
are made to the batteries. It has not been sufficiently
noticed that such tubes are also superior to hard tubes as
amplifiers, although the range of operation is considerably
smaller and sometimes hard to find. For this reason the
gas-filled tubes have found no commercial application; this
should not, however, deter the amateur from using them.
One of the best audion detectors ever made was the old
DeForest-McCandless bulb which flourished about 1914,
and has unfortunately now become obsolete. These were
made back in the days when an audion was called an audion,
before the skill of operators had been dulled by two-stages
of audio frequency amplification. Comparative tests made
with 75 of these detectors, and several approved commercial
types, showed an average increased current response of 10
times. This record is eclipsed, so far as I know, only
by the gas-filled tubes developed by Prof. ChafTee of
Harvard; two of which, according to tests made last win-
ter, gave improvement factors of 12 and 16 respectively.
From this experience I feel justified making an unqualified
plea for the more general use, and on the part of the manu-
facturers for a better production, of gas-filled tubes for
detection and amplification as well. The noises developed
in these tubes will, however, restrict the amplification use
to radio frequency systems.
The best detector action of the hard tubes is usually
RECEIVING APPARATUS 235
obtained at a lower plate voltage than would be used for
amplification. In most cases this voltage is about 20.
Coated filament tubes should not be burned at a higher
temperature than that corresponding to a bright red or
orange color (1000 C.); the tungsten filaments may be
operated at a bright white color (2300 C.).
The electrical characteristics of an audion depend upon
the size of the electrodes and upon the plate and grid volt-
ages of operation. They are not affected, for a filament of
constant length and structure, by varying the size of the
electrodes provided the geometrical configuration remains
similar to itself and all voltages are changed to correspond.
Thus it is possible to build very small tubes having the
same characteristics as the larger ones. This fact has been
utilized in the design of the so-called "peanut" or liliputian
tubes recently placed on the market by several manufac-
turers, and which are very economical in their current con-
sumption from batteries. The "Aeriotron" (see 3, Fig. 131),
the "N" tube of the W. E. Co., and the "Lfflputfonnat" tube
of the Germans, are examples of this practice.
80. Amplification. Radio Frequency and Audio Fre-
quency Methods. So far we have considered merely what
may be regarded as the essential processes of reception,
although certain simple methods of amplification have
been described, notably the regenerative method for modu-
lated signals (radio telephone, i.c.w., and "spark"), and
the heterodyne method for c.w. signals. The Armstrong
super-heterodyne cannot be regarded, as a method of ampli-
fication; it is rather a method for changing the wavelength
of the signal so that it can be more effectively amplified.
The response with all these methods of reception may be
236
RADIO TELEPHONY
further improved by straightforward amplification, and the
amplification may take place before the signal has been
detected, after it has been detected, or both. Amplification
of the radio frequency currents previous to detection is
commonly referred to as radio frequency amplification,
and amplification of the low frequency currents produced
by the detector is called, for obvious reasons, audio fre-
quency amplification. The scheme of these methods is
shown in Fig. 132. An indicated, the r.f. amplifier is in-
serted between the tuning apparatus and the detector,
and the audio frequency amplifier is inserted between the
TUNING-
APPARATUS
RADIO
FREQUENCY
AMPLIFIER
Fig. 132. Illustrating symbolically the positions of the radio- and audio-fre-
quency amplifiers in the receiving system.
detector and the telephones, loud-speaker or other electro-
phonetic device.
Concerning the relative effectiveness of the two methods
there is this to be said: Remembering that when radio
telephone or spark signals are being received by the ordinary
detection method, as Fig. 122 shows, the telephone current
produced by the detector is closely proportional to the
square of the voltage of the r.f. signal which acts upon it,
it is clear that amplification previous to detection will
be more effective than a similar amplification after it. Thus
supposing that we have an amplifier capable of amplifying
either the radio or audio frequency currents 10 times;
RECEIVING APPARATUS 237
using it as a r.f. amplifier will give an increase of 10 2 = 100
times, whereas if it is used after detection the increase
will be only 10 times. Hence if r.f. and audio frequency
amplifiers can be built with equal effectiveness the first
method is inherently superior to the second solely on
account of the peculiarities of the detector. With hetero-
dyne reception of c.w. signals this does not apply, for in
this case the response varies as the first power of the signal
voltage and the two methods are, at least from this point
of view, equally effective. But there are other considera-
tions which modify these conclusions somewhat.
In the first place r.f. amplifiers cannot be built, par-
ticularly for the short wavelengths used by amateurs, to
function as well as those designed for audio frequencies.
This would seem to favor the audio frequency system.
But the audio frequency system also has its drawbacks.
Because it does amplify the low frequencies, it amplifies
not only the signal but all those extraneous and parasitic
variations of current which are the unavoidable accom-
paniment of mechanical vibrations of the audion bulbs,
and of irregularities in the electrochemical action in the
batteries. The magnification of these tube and battery
noises, and other disturbances of this nature, limits the
number of audio frequency stages that can be comfort-
ably employed to two, or possibly three.
Taking these practical matters into consideration, and
looking at the problem from an economic angle, the best
way to improve the response by amplification would be to
first provide a two-stage audio frequency amplifier, then
to add radio frequency stages according to the availability
of the tubes and the constructive ability of the operator.
238 RADIO TELEPHONY
Of course a greater number of low frequency stages may
be used if a loud-speaker or other sound amplifying device
is to be operated, but this will generally increase the volume
of sound without increasing the range of reception. This
is not due, as many people believe, to the detector having
a threshold stimulus characteristic whereby any signal of
deficient strength is incapable of affecting it, but to the
simultaneous magnification of the undesirable noises. So
far as I am aware, there is no quantum theory of detector
action to support the former idea. When the addition of
r.f. stages is contemplated, it should be remembered that
the regenerative method of amplification described in
Art. 73 accomplishes as much with one audion as one or
two stages of this type; hence this is the first step to be
taken.
81. Concerning Cascaded Audion Amplifiers in General.
Both the radio frequency and audio frequency amplifiers
usually employed are of the cascaded type, so that a con-
sideration of this general method will furnish a foundation
for the subsequent separate treatment of the two applica-
tions. The cascaded system consists of a series of audion
amplifiers arranged electrically so that the amplified out-
put of each tube is passed on to the next, to be again am-
plified, passed on, and so forth. Each tube with its passing-
on mechanism, or coupling, is referred to as a stage or step
of the amplifier. Several methods of linking the tubes are
employed; and the reference to the amplifier as of the
resistance-coupled, transformer-coupled or inductance-coupled
type is based upon the electrical nature of this link. The
three important methods of coupling may be explained as
follows:
RECEIVING APPARATUS
239
(a) Resistance Coupling. The method of employing a
resistance for the purpose of passing the amplified output
of the plate curcuit of one tube on to the grid of the next
is illustrated in Fig. 133. This shows a two-stage amplifier,
in which a resistance R has been inserted hi the plate cir-
cuit of the first tube. The variation of the plate current
flowing through this resistance causes a variable voltage
drop across it which is impressed upon the grid of the next
tube. On account of the fact that the point to which the
grid is connected is normally at the d.c. plate voltage, a
special means of connecting the grid is employed so as to
INPUT *
B
Fig. 133. Scheme of resistance coupling in audion amplifiers.
avoid operating the grid at this unsuitable potential. As
was pointed out in Art. 13, if the audion is to give its maxi-
mum amplification, the normal potential of the grid must
be adjusted so that the proper part of the characteristic
curve (see Fig. 34) is used. The grid is, therefore, isolated
from the d.c. plate voltage by the condenser C, and the
proper negative bias is obtained by means of the resistance
R f which is either connected to a "C" (biasing battery)
as shown, or to a suitable part of the circuit. Thus C and
R r are not part of the coupling means, although they do
affect the impedance of this means to some extent, especially
at high frequencies.
Theoretically the resistance R should be as large as pos-
2 4 o RADIO TELEPHONY
sible, for the amplification per stage is proportional to the
ratio of this resistance to the total circuit resistance (R
+ plate resistance of the audion). There are practical
reasons, however, why this resistance cannot be inimitably
increased. When this resistance is used the normal plate
voltage is less than the "B" battery voltage by the IR
drop across it caused by the passage of the normal plate
current; hence in order to operate the audion at the rated
plate voltage the voltage of the "B" battery must be in-
creased by this amount. For example, suppose that we
wish to use the "D" tube (1, Fig. 131) having an amplifica-
tion factor of 40, a plate resistance of 70,000 ohms and
requiring 150 v. for normal operation. A resistance of
100,000 ohms would give an amplification per stage of 23.5
(at low frequency), but allowing a normal plate current of
1 m.a. the "B" battery voltage would have to be increased
to 250 volts to get 150 volts on the plate. The extra "B"
battery requirements of this system constitutes one of its
chief drawbacks.
Resistance-coupled amplifiers possess the important
property of amplifying currents of all moderate frequencies
to the same extent. This is of especial value for the audio
frequency amplification of radio telephone signals, for in
order that these may be amplified without distortion it is
important that the entire range of speech frequencies from
100 to 5000 cycles undergoes uniform magnification.
(b) Inductance Coupling. This method is illustrated sche-
matically in Fig. 134. The inductance, or choke-coil L,
replaces the resistance of the preceding method, and the
grid of the next tube receives its stimulus from the voltage
drop across this impedance. The extent to which signals
RECEIVING APPARATUS
241
are amplified by this method depends upon the proportion
of the choke-coil reactance (= 2 n L X frequency) to the
total circuit impedance; consequently the amplification
decreases with the frequency, very slow current variations
producing practically no e.m.f. across L. As applied in
radio frequency amplifiers, the impedance of the choke-
OUTPUT
INPUT
Fig. 134. Scheme of inductance (choke-coil} coupling in audion amplifiers.
coil branch is often increased by connecting a condenser
across it and tuning the branch circuit so formed to the
wavelength of the signals to be amplified.
INPUTS
Fig. 135. Scheme of transformer coupling in audion amplifiers.
In both the resistance and choke-coil methods of coupling
the highest amplification per stage that can be attained
is numerically equal to the amplification factor of the audion.
(c) Transformer Coupling. In this case the grid of the
second tube is inductively coupled to the plate circuit
of the first tube as shown schematically in Fig. 135. The
transformer PS is usually designed to operate at resonance,
16
242 RADIO TELEPHONY
and at low frequencies it is possible to use more turns on
the secondary than on the primary, thus giving a stepping-
up of the voltage. This aids the amplification and in these
circumstances the amplification per stage may exceed the
amplification factor of the tube. Were it not for the reso-
nance effects due to the capacity of the windings and audions
and dielectric losses in the secondary circuit, the step-up
ratio could be increased indefinitely and enormous ampli-
fication might be obtained. This follows from the fact
that when the grid of the audion is biased negatively no
grid current flows and the tube consumes no power in its
operation.*
The grid is not conductively connected to the plate circuit
of the first tube, hence the use of the isolation condenser,
and the grid resistance for securing the proper bias, is avoided.
The correct bias is imparted by including the "C" battery
in the secondary circuit as shown hi the diagram, or con-
necting to a point of the circuit giving an equivalent voltage.
82. Combination Regenerative and One-stage Radio
Frequency Amplifier. The simplest, most economical and
effective scheme of radio frequency amplification for the
average amateur is one embodying both the regenerative
and a single stage of straightforward amplification. This
accomplishes with a single tuning adjustment and with
one tube what would ordinarily require two tuned stages of
the usual type to do. An experimental test of this system
made by the writer in 1919 showed that the regenerative
contribution is of about the same amount as the straight-
* Provided that the load in its plate circuit is of such a nature to give
no effective grid input conductance (see pp. 211-213). This condition
practically obtains at audio frequencies with any plate load.
RECEIVING APPARATUS 243
forward amplification. The fact that the adjustment of
the plate inductance for maximum regeneration in the
Armstrong tuned-plate circuit corresponds approximately
under usual conditions to the adjustment for maximum
impedance of the LC branch circuit formed by the plate
inductance, and its distributed capacity and the capacity
of the leads and audions, permits us to take advantage of
both methods of amplification simultaneously with one
adjustment. The complete connections of a receiver with
this method of amplification are shown in Fig. 136, to which
the following description applies.
The construction of the double-circuit tuner has already
been described (Art. 71 (5)). A d.p.d.t. switch is added,
by means of which the amplifier may be connected across
the secondary condenser, thus using a double-circuit tuner,
or across the antenna coil, using the single-circuit scheme.
The latter arrangement will be preferable for pick-up work
because there are then but two adjustments to be made.
The tuned-plate circuit increases the selectivity of the single-
circuit arrangement to a degree which for ordinary work
will be found ample; but if the operator prefers, the addi-
tional selectivity of the double-circuit connection may be
realized after the signals have been picked-up by simply
throwing the switch and tuning the secondary condenser.
The mechanical arrangement of this switch needs no
description. It will be noted that this also opens the
secondary circuit when in the single-circuit position to pre-
vent its abstracting energy. It is recommended that the
amplifier and tuning units be built in separate cabinets.
The use of a two-stage audio frequency amplifier is pre-
sumed and is indicated in the diagram. This will be
244
RADIO TELEPHONY
RECEIVING APPARPTUS 245
described later (Art. 87). The radio frequency stage should
be carefully wired. All filament leads are to be kept to-
gether and away from the plate and grid leads, and the leads
marked "A" and "B," connected respectively to the grid
and plate electrodes of the first tube, should be separated
as far as possible* The lead "A" should be run direct to
the grid and kept as short as possible. The object of this
is to minimize the grid-plate capacity (upon which the
regeneration depends) so that the regeneration and oscilla-
tion may be properly controlled. The stabilizer will assist
in this control, and may consist of a 300-ohm "potenti-
ometer" of the usual type connected across the "A" battery.
The extra battery in this circuit may be necessary with
certain types of tubes and wiring to get sufficient range of
control. The plate circuit inductance is most conveniently
a variometer having a natural wavelength range of 150-
500 meters. Many suitable types are available on the
market. There is no necessity for the variometer having
low losses; resistances of 200 or 300 ohms can be tolerated
in this circuit. It is sometimes helpful to connect between
the variometer and filament lead a paper condenser C' of
large capacity (1 mfd.) to diminish the feed-back from the
detector and give more stable action. The grid condenser
C is of the usual type with capacity .0002 mfd. The grid
leak resistance R f (J to 2 megohms) may not be necessary
with gas-filled detector tubes. The low voltage (20 v.)
necessary for the detector is obtained by tapping in on the
"B" battery as shown. The audio frequency amplifier will
be described later.
In the assembly of the parts avoid close magnetic associ-
ation of the variometer and the primary and secondary
246 RADIO TELEPHONY
coils of the tuner. If there is too much interaction between
the units a proper shielding of both cabinets will be helpful
(see Art. 86).
Regarding the selection of the tubes, preference should be
given in the radio frequency stage to audions having a high
amplification factor; the Meiers (both types) and Moorhead
type have proved by general experience to be best suited
for this purpose. Audions containing gas, as previously
recommended, are preferable for the detector stage. For
the audio frequency amplifier, the selection of the tube will
be governed by the transformers available. Most com-
mercial transformers will operate successfully with the
type "J" Western Electric, Meiers' low "mu," or Radio tron,
UV-201 tubes.
The writer considers this receiver to be most suitable for
the average amateur, for the reception of radio telephone,
i.c.w. and "spark" signals. It is not very convenient and
less desirable than the Reinartz receiver (Art. 77) for c.w.
work.
83. Resistance-coupled Amplifiers for Audio and Radio
Frequencies. Since the construction of resistance-coupled
amplifiers for radio frequencies is substantially the same as
that for audio frequencies, both systems may be treated
under the same caption. As already pointed out, the most
valuable property of this type of coupling is its ability to
amplify all moderate frequencies to the same extent. This
follows, of course, from the fact that the impedance of the
coupling link through which the plate current flows and
the voltage drop across which actuates the grid of the
succeeding tube, is simply its resistance and is not affected
by the frequency. For high frequencies corresponding to
RECEIVING APPARATUS 247
the short wavelengths employed by amateurs, however,
this statement demands modification for two reasons:
First, across the resistance coupling two audions are con-
nected so that it is shunted effectively by a capacity equal
to the plate-filament capacity of the first tube plus the
grid-filament capacity of the next tube, and at high fre-
quencies the impedance of this capacity branch may be
low enough to eclipse the effect of the resistance link. (For
two tubes with plate-filament and grid-filament capacities
of 9 mmfds., the impedance of the capacity path at 250
meters is only 7400 ohms.). Secondly, due to the grid-
plate capacity of the second audion the resistance load in
its plate circuit further increases the apparent capacity of
this audion so that it adds not only its grid-filament capacity
but this extra input capacity as well. In the case of the
"J" tube with a 60,000 ohm resistance link hi its plate
circuit, at 200 meters, the effective impedance considering
only the plate-filament and gird-filament capacities is
4800 ohms; assuming that this capacity could be made
zero and considering only the effects of the feed-back
action, the effective impedance of the coupling link is equal
to 6000 ohms; but since both are present the actual imped-
ance will be 4150 ohms. Thus an amplifier, compounded
of "J" tubes, which at long wavelengths or for audio fre-
quencies gives a voltage amplification of 4.5 per stage,
would give at 200 meters an amplification of 1.02 per stage!
In other words, the 2 per cent, amplification obtained at
this wavelength would be completely negligible.
I have carried through this calculation to emphasize the
effect of the audion capacities in diminishing the impedance
of a resistance coupling link at short wavelengths, and to
248 RADIO TELEPHONY
emphasize particularly that the failure of this type of ampli-
fier is not due solely to the plate-filament and grid-filament
capacities, but is conditioned to about an equal extent by
the feed-back action through the grid-plate capacity of the
second tube, which we assume also has a resistance in its
plate circuit. This contributary effect of the feed-back
action has not been sufficiently noticed. This example will
indicate the extreme importance of keeping these capacities
small, using short leads, and small apparatus, grid con-
densers, leaks, etc. In selecting the tube preference should
be given to those tubes having a high amplification factor,
such as the W. E. Co.'s "D" type (ji = 40) and the Meiers'
high "mu" tube (ji = 20). The latter is preferable for
other reasons and is available for amateur use.
In view of the difficulty in obtaining audions with low
inter-electrodic capacities and the failure of the usual
tubes over the range of wavelengths with which this book
is mainly concerned (150-600 meters), I do not feel justified
in devoting further space to this application or recom-
mending it. Nor does the action improve to an expected
extent as the wavelength increases, for although the im-
pedance of the plate-filament and grid-filament capacities
is no longer important, yet the capacity due to the feed-
back action increases very rapidly with the wavelength.
Consequently it may be concluded that the transformer
and choke-coil methods of coupling are inherently more
attractive, since in these cases the parasitic capacities may
be turned to good use. This statement should not be
construed to mean that resistance-coupled amplifiers will
not work at wavelengths of 3000 meters, such as are utilized
in the Armstrong super-heterodyne receiver; the point is
RECEIVING APPARATUS
249
that choke-coil-coupled amplifiers will work better. If,
however, the reader wishes to follow the general practice
and use a resistance-coupled amplifier with the Armstrong
system described in Art. 78 (Fig. 130) he may find the
following description of a suitable amplifier of some interest,
(a) 3000-Meter Resistance-coupled Amplifier for Armstrong
Super-heterodyne. The connections of a 4-stage amplifier are
given in Fig. 137. Four stages are indicated because this
is the minimum number that should be used, and the
maximum number that can be cascaded without the special
(TABIUZER
Fig. 137. Electrical connections of 4-stage 3000-meter resistance-coupled am-
plifier for use with Armstrong super-heterodyne receiver, Fig. 130.
shielding and stabilizing precautions outlined in Art. 86.
This discussion should be consulted if the addition of more
tubes is contemplated. Because they do amplify all fre-
quencies, resistance-coupled systems also magnify the low
frequency tube and other extraneous noises to a consider-
able extent, hence the interposition of the tuned radio
frequency LC circuit between the amplifier and detector is
desirable to block these disturbances and to allow only the
radio frequency currents to reach the detector. I am
aware of the possible detecting action that may occur at
250 RADIO TELEPHONY
any stage of the amplification, especially in circumstances
of incorrect bias and plate voltages and that by the use of
an audio frequency filter this additional rectification is
lost, but the first seems to be the greater evil. Anyone who
has connected a 10-stage resistance-coupled amplifier di-
rectly to the detector will, I think, concur with this recom-
mendation. I have found the use of a Type I ("high-
pass") Campbell filter of value here, or what amounts to
the same thing, the use of a choke-coil coupled stage (radio
frequency) proximate to the detector. The L 3 C 3 circuit
of Fig. 130 will be adequate, and a small ratio of L to C
(using a low resistance coil) will aid the elimination of the
low frequencies.
The coupling resistances R shown in Fig. 137 are of
value 100,000 ohms and may be constructed in various
ways. Sputtered film grid-leaks which have been reduced
by reheating their filaments, carbonized cellulose films, non-
polarizing electrolytic cells (these are sometimes very
noisy), and various forms of carbon and graphite rods are
useful, or a satisfactory resistance unit can be easily con-
structed by the amateur as follows:* Coat a piece of bristol
board 2 in. wide with a mixture consisting of 6 parts of
Higgins' "American" India drawing ink and one part pow-
dered graphite shaved from a grade H pencil. The mixture
is applied to both sides of the bristol board strip with a
camel's hair brush, stroking across the strip not from end
to end. After the strip has been thoroughly dried in a
warm oven, cut a piece 1^- in. long from it and carefully
wrap it around a glass tube \ in. in diameter and 2 in.
* This construction was described by Mr. P. F. Godley and is here repro-
duced together with Fig. 138 from Wireless Age, November, 1920, p. 13.
RECEIVING APPARATUS
251
long, on the ends of which two strips of soft brass (cov-
ered with tinfoil) have been placed (see Fig. 138). The
strip is now clamped in position with a second strip of soft
brass (inserting a sheet of tinfoil between the brass and
coated bristol board). The space between the clamps is
then covered with two layers of cotton or silk tape, var-
nished with a clear, light, insulating varnish and again
baked until dry. The inner and outer connection strips
should be connected together. The resistances R' are
Soft t>rvssc/o/np/ng
termrn0/.
632 screw-
Section A- A
&&sj tot*
Connect/on strip
Clomping sfr/p fa
Tree fed 6r/sfo/bo0nJ fowl
-String
Soffbrvss connec
t/ng
On smog or
co f Jon fape rornshtd
offer in place
- - 6/oss rubing
Fig. 138. Illustrating construction of 100,000 ohm coupling resistance (Godley).
ordinary grid leak resistances of the order of 1 -megohm.
The condensers C may be paraffined paper condensers of
capacity not less than .01 mfd. (The effective grid-filament
capacity of the "D" and Meiers' high "mu" tubes at 3000
meters with a resistance in their plate circuits will be about
.0001 mfd.). A suitable condenser of this capacity has
already been described on p. 202 in connection with the
single-circuit, crystal detector receiver. The "stabilizer" is
useful for controlling the amplification and the tendency
toward oscillation and may consist of the usual 300-ohm
252
RADIO TELEPHONY
"potentiometer." The 1-mfd. condensers shown across the
stabilizer and the "B" battery are necessary to reduce the
radio frequency impedance common to the grid and plate
circuits of the tubes. Other stabilizing precautions may be
necessary (see Art. 86). In connecting this amplifier to the
super-heterodyne receiver, Fig. 130, omit the direct connec-
tion from the coil L 2 to "A." This coil is to be con-
nected across the input terminals of the amplifier. The
"B" battery voltages and amplification per stage corre-
sponding to the use of several of the audions described in
Art. 79 are given in the following table:
Type of Audion.
Amplification
per Stage.
"B" Battery
Voltage.
Meiers (high "mu")
4 6
110 v
W. E. Co. "D"
3
250 v.
Moorhead
2.8
150 v.
Radiotrorr UV-201
2 5
150 v
W. E. Co. "J"
1 5
110 v
The low amplifications and high "B" battery voltages
required by this system are rather discouraging.
(b) 2-Stage Audio Frequency Resistance-coupled Amplifier.
On account of its ability to amplify currents of all moderate
frequencies to the same extent, the resistance-coupled ampli-
fier is very valuable for the amplification of the low fre-
quency detected currents, especially in radio telephony
where the entire range of from 100 to 5000 cycles is used
and all frequencies within this range must receive the same
amplification if there is to be no distortion. A suitable
2-stage amplifier for this purpose is shown in Fig. 139.
RECEIVING APPARATUS
253
The coupling resistances R (100,000 ohms), the grid
resistances R' (1 megohm) and the condensers Ci (.01 mfd.)
are of the same construction as described for the 3000-
meter amplifier. The condenser Ci connected across the
resistance in the detector audion serves to shunt the radio
frequency currents around this resistance and to keep
them out of the amplifier. Because the detector requires
a low plate voltage, the second audion contains a resistance
in its plate circuit, and the third does not, three different
2-STAfrt AF AMPLIFIER
Fig. 139. Electrical connections of 2-stage audio frequency resistance-coupled
amplifier.
"B" battery voltages, obtained by the three taps (1), (2),
(3) on the "B" battery as shown, will be required to give
the proper plate voltages in each case. Presuming the use
of a detector of the ordinary type operating on 20 volts,
the "B" battery voltages of the three taps, and the ampli-
fication per stage that may be expected with several tubes
used in the amplifier positions (2) and (3), are given in the
table on page 254.
The great differences between the amplifications obtainable
at these frequencies, and with the same arrangement at
254
RADIO TELEPHONY
Type of Audion.
Amplifica-
tion per
Stage.
"B" Battery Voltages.
Tap (1)
Tap (2)
Tap (3)
W. E. Co. "D"
19.0
10.0
5.5
5.1
4.3
50-70 v.
14
250 v.
110 v.
150 v.
150 v.
110 v.
150 v.
60 v.
ti
ii
Meiers' (high "mu")
Moorhead
Radiotrorr UV-201
W E Co "J"
3000 meters (cf. the preceding table, page 252) are no-
table.
84. Transformer- and Choke-coil-coupled Radio Fre-
quency Amplifiers. Transformer- and choke-coil-coupled
amplifiers offer some advantage over those of the resistance-
coupled type at radio frequencies, in that the parasitic
audion capacities which are so deleterious in the latter type
can be associated with a proper inductance to give the
coupling link whether transformer or choke-coil a very
high and favorable impedance. Unfortunately this favor-
able adjustment obtains only at one frequency, or for a
narrow band of them. Concerning the transformer-coupled
type for radio frequencies and with the tubes available, a
one-to-one turn ratio between primary and secondary
turns is generally all that can be obtained, hence the trans-
former does not step-up the voltage and simply acts like a
choke-coil of the same number of turns. The only advan-
tage of the transformer is that by conductively separating
the primary and secondary, the necessity for providing an
isolation condenser to keep the plate voltage off the grid
of the next tube, and a biasing resistance, is avoided. In
view of this the transformer- and choke-coil-coupled systems
may be considered together.
RECEIVING APPARATUS 255
The simplest form of choke-coil amplifier is the regenera-
tive and one-stage amplifier described in Art. 82. Here
the choke-coil consists of a variometer, so that the adjust-
ment to make the impedance of the branch circuit formed
by the inductance of the variometer and its inherent ca-
pacity plus the capacities of the audions, can be made for any
wavelength from 150 to 500 meters. Besides the straight-
forward amplification due to this arrangement, a consider-
able regenerative effect is caused by the variometer, this
action being manifested in the tuned secondary circuit. In
the same way, if we attempt to add another stage of this
type, regenerative action would be felt in the first variometer
circuit. Unrestrained cascading of such stages would yield
an amplifier which would be quite unstable and unmanage-
able. This is due largely, as I have previously remarked, to
the feed-back through the grid-plate capacity.
The solution of this difficulty is contained in the curves
shown in Fig. 119, which clearly exhibit the damping effect
of various inserted resistances in the plate circuit for a
typical case. In this case a resistance of 3000 ohms would
give an appreciable diminution of the regenerative action,
would promote stability, and at the same time would not
interfere with the kind of amplification we are interested in;
for inserted in series with the choke-coil, or being incor-
porated in the choke-coil itself by winding it with high-
resistance wire, its effect in decreasing the impedance of
the coupling link would be almost negligible. Since the
feed-back decreases with the wavelength, the insertion
of such a stabilizing resistance will not be necessary at the
higher wavelengths. This idea might be practically applied,
and has been practically applied by English and French
256 RADIO TELEPHONY
engineers (although their descriptions show that the im-
portance of the grid-plate regenerative action was not
clearly appreciated), by winding the choke-coil or trans-
former with resistance wire or by using an iron-core.
Transformers and choke-coils of fixed value are often
employed. In this case the maximum amplification will
be obtained, of course, only at one wavelength and the
response will fall off as the signals differ from this wave-
length. Some very extravagant claims have been made
for amplifiers of this type, and the reader is cautioned
against being mislead into purchasing transformers which
are alleged to cover a wavelength range of from say 400
to 4000 meters. Remember that an "amplifier" with but
a single turn of wire in the coupling link would cover a range
of from 50 to 5,000,000 meters, but it would not be an am-
plifier, but an absorber! The legitimate extension of the
wavelength range of an amplifier with fixed coils can be
accomplished in two ways: by regenerative action, and by
taking advantage of the variable reluctance due to the
skin effect in iron cores. Other extensions are obtained
only by decreasing the response at the most favorable
wavelength, or tuning point, by the insertion of resistance
or its equivalent. Such extension does not seem to the
writer to be bonafide, and hardly merits the use of the term.
The response for wavelengths differing from the tuning
wavelength may be unproved by careful design, and by
employing a choke-coil with a high ratio of L to C. The
capacity due to the audions and feed-back cannot be con-
veniently reduced, but the distributed capacity of the vari-
ometer may be reduced by decreasing the physical dimen-
sions of the coils, using at the same time a greater number
RECEIVING APPARATUS 257
of turns of smaller wire to get the proper inductance. The
writer recommends a construction in which the variometer
both rotor and stator is wound with two paralleled
wires. This gives virtually a transformer with a 1 : 1 ratio,
in which the primary and secondary inductances are varied
together to preserve this ratio and at the same time to pro-
vide a means of adjusting the inductance for each wave-
length. One winding is connected to the plate circuit,
the other to the grid of the succeeding tube (see Fig. 141).
This arrangement is electrically equivalent to a variometer
of the same number of turns, but is more convenient, in
that the isolation condenser and grid biasing resistance
required with the variometer are not necessary. The name
"vario-transformer" might be applied to it. The mechanical
details, number of turns, etc., will have to be worked out
experimentally. Use equal numbers of turns on the rotor
and stator. Iron-cored variometers of small size may be
designed, incorporating all the desirable features: high
ratio: of L to C, variable L, some extension of the operating
range by changing flux penetration in the iron, and a loss
which will reduce the instability due to regeneration. That
a variable transformer (1:1 ratio) is preferable to a fixed
one is obvious; that the present commercial variometers
are too large (giving a large distributed capacity), and
possess proportionately intense stray fields, is equally
obvious.
A suggestion for the design of a liliputian vario-trans-
former of the type recommended, suitable for amplification
over a wavelength range of 150-500 meters, is contained in
Fig. 140. The rotor (moving coil) is wound on two slots
cut in a bakelite or hard-rubber disc, with 100 double turns
17
2 S 8
RADIO TELEPHONY
STATOR --- c
TOTAL -,-
100 D.T. f
M0.36-S.&C
f ROTOR WINDING
I EACH SLOT
\ 50 O.T No.36-S.SC.
(50 on each side of the shaft) of No. 36 s.s.c. or enameled
wire. The stator (fixed coil) is wound with a similar number
of turns, in a haphazard fashion if the wire is found too
small to wind regularly. In winding the coils if two spools
of wire are available the paralleled wires can be fed out
more conveniently. The coil sizes indicated are suitable
also for a plate circuit variometer of the type used in the
regenerative and one-stage re-
ceiver, Fig. 136.
The application and the
electrical arrangement of this
type of coupling may be most
usefully discussed in relation
to the receiver, Fig. 136.
Supposing that this receiver is
to be further improved by
the addition of radio frequency
stages, how shall we go about
it? The resistance-coupled
amplifier has already been condemned for operation at
these wavelengths; the present type of coupling is con-
sidered to be most suitable. One or two stages of this kind
may be inserted between the regenerative first-stage and
the detector, the scheme of connections being indicated
in Fig. 141. The proximate old apparatus (Fig. 136) is
shown dotted in order to show more clearly the method of
addition. The insertion of stabilizing resistances in the
plate circuit of the audions (2) and (3), if found to be neces-
sary with the small number of stages likely to be added
by the average amateur, will be left to the reader. If such
resistances are indicated the vario-transformers are pref-
Fig. 140. Suggested design for lili-
putian vario-transformer.
RECEIVING APPARATUS
259
erably wound with high-resistance wire. It is to be noted
that the stabilizer has the same effect, when adjusted so
that the grids are positive. The first variometer has been
replaced in Fig. 141 by a vario- transformer. This variom-
eter may either be moved up to the last stage, where the
grid condenser of the detector audion will serve also as an
isolation condenser, or eliminated altogether. It will be
found that the settings of the vario- transformers (1) and
(2) will be smaller than that of (3), due to the extra capacity
VARIO -TRANSFORMERS
V-T. OR
VARIOMETER
\ PET.
\ /Sv
T+^swi }
^
I
Fiff. 141. Illustrating addition of vario-transformer-coupled amplifier to the
receiver of Fig. 136.
across (1) and (2) which is the concomitant of plate-grid
feed-back in the audions (2) and (3). It may, therefore,
be desirable to wind the last variometer to a larger induc-
tance, or if the range of this is already correct, to reduce
the numbers of turns on vario-transformers (1) and (2).
In general the adjustments of all three bear a close relation
over the useful wavelength range, so that it is quite feasible
to simplify the adjustment by connecting them together
mechanically in such a way that the proper rotations are
made by turning one knob. This should be carefully ar-
260
RADIO TELEPHONY
ranged in order that no magnetic couplings between the
vario-transformers may be
introduced. If (2) and (3)
are not adjusted, the equiv-
alent of the usual fixed 1 : 1
transformers is obtained; the
beauty of this arrangement
being that if the operator
wishes to realize the extra
amplification conditioned by
proper adjustment, he may.
Separate stabilizers for the
first tube, and for the second
and third tubes are recom-
mended in place of the single
stabilizer shown, and should be
shunted by 1 mfd. condensers
as indicated in some of the
other diagrams (e. g., Fig. 142).
The "B " battery should also be
shunted by such a condenser.
85. Description of a 3000-
meter Amplifier for the Arm-
strong Super-heterodyne Re-
ceiver. The method of coup-
ling discussed above may be
advantageously applied in the
construction of a 3000-meter
Fig. 142. Electrical connections amplifier for USC with the
of 4-stage 3000-meter transformer- Armstrong sup er-heterodyne
coupled amplifier for super-hetero- m *
dyne receiver, Fig. 130. receiver described in Art. 78
f
RECEIVING APPARATUS
261
-N-
(Fig. 130). The connections of a 4-stage amplifier are
shown in Fig. 142. The coil L 2 of the original circuit (Fig.
130) is to be connected across the input terminals of the
amplifier, omitting the connection from L 2 to "A" shown
in that figure. The tuned L 3 C 3 circuit, used with the re-
sistance-coupled amplifier to block the low frequency dis-
turbances, while giving slightly more selectivity, is not
necessary and in its place one of the
transformers may be used as indicated
in Fig. 141 by the dotted lines. On
account of the exclusive use of one
wavelength the coils T can be ad-
justed for this wavelength and fixed.
Suitable dimensions for these coils
are indicated in Fig. 143. As in the
case of the vario-transformer, a dou-
ble winding is provided to obviate
the use of an isolation condenser and
grid leak which are necessary when
one winding (choke-coil) is used. On account of the wide
variations in the tubes and wiring of the circuits, the num-
ber of turns listed in the table on page 262 are to be ac-
cepted suggestively, and the exact arrangements to suit the
reader's special conditions may best be determined by direct
experiment. The other details will either have been described
or will be obvious.
For comparison purposes I have indicated in the follow-
ing table the amplification per stage that may be expected
with this amplifier when several familiar types of tubes are
employed. The figures for the "D", "J" and Moorhead
tubes were experimentally determined in 1919.
Fig. 143. Illustrating
construction of coupling
coils.
262
RADIO TELEPHONY
Type of Audion.
Amplification
per Stage.
Number
of Double
Turns.
*W. E. Co. "D"
13 4
700
Meiers' (high mu)
6 7
1600
*Moorhead
4 1
Radiotron, UV-201
4
M
*W. E. Co "J"
3 6
* Experimental values.
86. Stability of Cascaded Amplifiers. The regenerative
influence of the reintroduction of energy from the plate
circuit into the grid circuit of the amplifying audion by
means of the -various types of back-coupling, electrostatic,
electromagnetic and conductive, has already been con-
sidered. It has also been noticed that if this energy is rein-
troduced in sufficient amount a state of self-sustained oscil-
lation is set up in the audion circuits. For a given degree
of back-coupling the feed-back is proportional to the ampli-
fication experienced by the current in passing through the
tube; hence in cascade amplifier systems, where enormous
amplification is produced as compared with that produced
by a single tube, the tendency of the circuits toward oscilla-
tion is proportionately pronounced. The consideration
of the natural back-coupling in a cascaded system and its
reduction to the point where the system is tolerably stable
and devoid of regenerative tendencies is, therefore, a matter
of prune importance. This result is to be obtained if pos-
sible by actually eliminating the feed-back action and not
by reducing the amplification.
The parasitic couplings are of three types: electrostatic,
electromagnetic and conductive. We have already had an
RECEIVING APPARATUS 263
example of the electrostatic (capacity) coupling in the
regenerative action through the grid-plate capacity of the
auction, which is used in the Armstrong tuned-plate circuit.
When several audions are used, not only is there feed-
back from the plate of an audion to its own grid, but also
from the plate to every other grid. In a resistance-coupled
amplifier, and in the transformer- and choke-coil-coupled
types at certain wavelengths, when the grid of the first
tube goes positive, its plate goes negative, and the feed-
back through this tube is anti-regenerative; but the plate
of the second audion goes positive, giving through its ca-
pacity with the first audion, a regenerative effect. Thus
the feed-back contributions in the first tube are alternately
positive and negative and the stability of the system will
depend upon the sum of these effects.* In these circum-
stances evidently combinations of 1, 3, 5, . . . stages
will be more stable than those of 2, 4, 6, etc. Stability
may be provoked in several ways, the most effective of
which consists in preventing the electric flux from each
tube from reaching any other tube by "grounding" it through
a shield which completely incloses the stage except for the
small holes required for the connecting wires. Then there
will remain only the feed-back action through the tubes
themselves, which is localized and not so prolific of in-
stability. An example of the application of this method is
* In technical discussions of the stability of amplifiers I have found the
employment of the terms conditional- and absolute stability convenient, apply-
ing to the resistance- and inductance-coupled types respectively; these t rms
being in analogy with the mathematical series upon the convergence of which
the ultimate stability of such systems depends. These terms were defined in
papers read at meetings of the Institute of Radio Engineers, Philadelphia
Section, June, 1920, and Boston Section, February, 1921.
264
RADIO TELEPHONY
furnished by the amplifier illustrated in Fig. 144. Here
every tube with its apparatus occupies a small metallic
cell by itself. Small holes are drilled to pass the necessary
connecting wires to the other cells, and usually the cell is
STABILIZER
BY- PASS CONDENSER
Fig. 144. Illustrating assembly of multi-stage amplifier in which the separate
stages are inclosed by individual metallic cells.
connected to the A battery terminal and grounded.
The top and sides of the cell are formed by the sheeting,
which covers the inside of the wooden case, and when in
position makes a firm electrical contact with the rest of the
cell. It is very important that no cracks or holes appear
RECEIVING APPARATUS
265
in the final assembly; the shielding should be complete or
left out altogether. Tin-foil will be adequate for electro-
static shielding; sheet copper at least 0.010" thick should
be used for electromagnetic shielding at radio frequencies,
and sheet electrical-iron -%" thick, at audio frequencies.*
The necessity for shielding at low frequencies is practically
eliminated by using transformers which are completely
iron-clad; the induction from the wires themselves at these
frequencies is almost negligible.
^-MOVABLE PLATE
STABILITY*
Fig. 145. Device used by the French for promoting stability or regeneration in
a cascade amplifier.
Returning to the capacity coupling at radio frequencies,
an interesting connection used extensively by the French
for utilizing at will the regenerative effect of a capacity
path to alternate tubes, and the damping effect of capacity
paths to the others, is illustrated in Fig. 145. This may
be used to secure stability, or to increase the signal strength
* These figures are based upon a shielding effectiveness of 90 per cent.
If copper is used for low frequency shielding (1000 cycles) it will have to be
.3" thick to give the same effect as .01" copper at 300 meters. Consequently
it is better to use iron at these frequencies, for which a thickness of .073" is
indicated.
266
RADIO TELEPHONY
by regenerative action. The latter is not very satisfactory.
The condenser C is, of course, of very small capacity a
few micro-microfarads, or cms. and the damping and
regenerative effects are obtained by moving the middle
plate to the left and right respectively.
The second kind of coupling provocative of instability is
the inductive coupling which exists between the wires of
the various stages, and especially between the coils if they
are used, as in the transformer and choke-coil coupling
schemes. This may be reduced by shielding, according to
the specifications given above. The leads should be kept
B' BATTERY--'
Fig, 146. Illustrating direct coupling between the stages of a cascade amplifier
due to the use of a common "B" battery.
as short as possible, and in view of the reaction between
coils, the desirability of reducing their physical dimensions
and stray fields as much as possible, is patent. This is
one reason why the liliputian variometers and vario-trans-
formers and the small choke-coil illustrated hi Fig. 143
have been recommended.
Another type of coupling is provided by the combination
of conductive and direct inductive coupling due to the use
of a common "B" battery and stabilizer. The electrical
situation is indicated symbolically in Fig. 146, for the case
of a common "B" battery. Obviously the internal resistance
of the battery is common to all four tubes and they are there-
' I
RECEIVING APPARATUS 267
fore conductively coupled together. While this resistance
may not amount (in the case of new batteries) to more than
15-20 ohms, yet with the enormous amplification obtained,
for example, with a 5 -stage amplifier, the feed-back from
the last tube to the first tube may be quite appreciable.
The "D" section of the circuit is likewise common to all
tubes and since the reactance (for example) of two No. 18
wires separated 2 in. is 5.8 ohms per foot at 200 meters,
the importance of this inductive coupling is evident. The
actual resistance of the wires is of small importance. Other
portions of the circuit, "A," "B," etc., are common to other
stages of the amplifier. The above situation may be reme-
died in several ways as follows: The "B" battery is first of
all shunted by a large condenser, of 1 or 2 mfds. capacity
(1 mfd. gives a reactance of -^ ohm at 200 m.), this con-
denser being located as close to the apparatus as possible.
The* other mutual reactances may be reduced by closely
associating the two lead wires, by twisting or using two
flat strips with their broad faces close together. The most
effective scheme is to complete each plate circuit right at
the tube through a large condenser, as shown in Fig. 147.
If each tube is inclosed by a metallic cell as recommended
above, this condenser should also be included in the cell;
and. in any case its connection should be as close to the
position shown in the diagram as possible. If properly
designed and proportioned the choke-coils L will still further
reduce the circulating common r.f. currents. They should
be connected as shown in Fig. 147 and likewise included in
the cell of the tube to which they belong. If the grid
leaks are connected together to a stabilizer or common
biasing means, similar precautions with respect to their
268
RADIO TELEPHONY
circuits should be taken. The stabilizer resistance, which
will be common to all grids, should be eliminated from the
r.f. circuit by shunting with a 1 mfd. condenser, as shown
in many of the diagrams (e. g., Fig. 142). If transformers
are used, the connection of a condenser between the lower
secondary terminal and the filament will be helpful; but not
so effective in the case of the 1-megohm leaks.
The free oscillations which are encouraged by the various
feed-back effects enumerated will take place at the most
favorable frequency, as determined by the LC constants
Fig. 147. Method of reducing instability in cascade amplifier due to inter-
stage coupling illustrated in Fig. 146.
of the system and the loss of energy with which the various
possible modes of vibration are attended. The short leads
and small capacities of the cascaded amplifier cause the
oscillation to select a high frequency, usually an ultra-
radio frequency, and at such frequencies the reactances of
even short leads assume surprising magnitudes. The
reactance of the two No. 18 "B" battery leads, for example,
which was previously calculated to be 5.8 ohms at 200
meters, will be 58 ohms at 20 meters. And the other modes
of energy transfer as well are more effective at the higher
frequencies. It is not surprising, therefore, that multi-
RECEIVING APPARATUS 269
stage amplifiers, especially those employing resistance
coupling, often persist in developing very high frequency
oscillations, which may be above audibility, or above the
frequency of the r.f. signals, yet are still troublesome
because they aggravate tube noises and prevent the simul-
taneous efficient use of the tube for the legitimate signal
frequencies. This tendency may be checked by the method
adopted in the case of power tubes operated in parallel
(see Fig. 80 (c)), i. e., by inserting in each grid lead, close
to the grid, a small r.f. choke-coil. The coils should be
very compact, and may consist of 30 or 40 turns of No. 28
s.c.c. wire wound haphazardly around the thumb or ringer.
Many other stabilizing measures are employed, as indi-
cated in special circumstances. In the case of transformer
coupling, at both radio and audio frequencies, an increased
stability may be often obtained (generally with a sacrifice
of amplification) by reversing the terminals of the secondary
winding in every other stage. The cores of transformers,
tube sockets, and other non-circuital metallic elements are
best bonded together and connected to earth. The leakage
of radio frequencies into telephone receivers and its reintro-
duction into the initial stages of the amplifier through the
medium of the operator's hand, often invokes oscillations.
This may be reduced by inserting a radio frequency trap or
filter circuit between the apparatus and the telephones, or
by shielding the telephone cords with flexible belden-braid
which is connected to the metal receiver cases, and to the
ground or A battery terminal at the other end. The in-
sertion of a telephone transformer between apparatus and
telephones is also helpful, but not usually necessary if the
telephones have been shielded as above described.
270 RADIO TELEPHONY
87. Transformer-coupled Audio Frequency Amplifiers.
Amplification with audions at the low frequencies of the
detected currents is fortunately not attended with the
many and divers parasitic actions to which the radio fre-
quency systems are subject. As a result the method is re-
markably successful, giving amplifications per stage as high
as 40 or 50. It avoids perfection, however, by means of the
high efficiency with which extraneous disturbances, tube and
battery noises, etc., are amplified. This has already been
discussed in Art. 80 and requires no further comment. The
practical result is that the number of stages to which ampli-
fication by this method can be usefully pushed, is limited
by the simultaneous magnification of noises, to two or three.
When the operation of a super-telephone, or loud-speaker,
is essayed, additional stages may be added to form what is
commonly referred to as a "power amplifier." In general
such amplification is resorted to only for the production of
a greater intensity of sound, and will not result in an in-
creased receiving range.
Presuming that the average amateur will not be interested
in "power amplifiers," this discussion will be restricted to
the ordinary amplification of the detected currents previous
to their passage through the telephones. The use of a 2-stage
amplifier is common practice and a resistance-coupled am-
plifier for this purpose has already been described. The dis-
advantages of the transformer-coupled amplifier, not in-
herent but present in 90 per cent, of the commercial types,
is the variation of their amplifying power with the variation
of the frequency within the speech range. Obviously some
distortion of the sound accrues, which increases with the
sharpness of tuning and the amount of amplification. It is
RECEIVING APPARATUS
271
to be noted that such effects are not merely additive, that is
to say, a transformer whose resonance characteristics give
an amplification of 1 unit at 1000 cycles and 4 units at 2000
cycles will not give in two stages an amplification change
from 1 to 8 units, but will give 1 unit at 1000 cycles and 16
units at 2000 cycles. The desired flat characteristic is easily
obtained by sacrificing amplification, and may be also aided
by taking advantage of the variable flux penetration in the
iron core, or in a multi-stage amplifier by tuning the trans-
TAP FOR DETECTOR J
Fig. 148. Electrical connections of 2-stage transformer-coupled audio fre-
quency amplifier.
formers to different frequencies, or using a special connecting
network such as the Campbell filter.
The electrical connections of a 2-stage transformer-
coupled amplifier are shown in Fig. 148. The detector audion
is indicated by the dotted lines in order to bring out the
essential connections of the amplifier. Switches or jacks are
usually provided for connecting the telephones either in the
plate circuit of the second audion (as shown) for full ampli-
fication, or in the plate circuit of the first audion for the use
of a single stage. The transformers T are of the usual type
and many forms are available on the market. Two com-
272
RADIO TELEPHONY
mercial types are illustrated in Fig. 149. Preference should
be given in selecting a transformer to those types which are
iron-clad; and by iron-clad I do not mean inclosed in an iron
case with five sides and a fibre top, but one which is com-
pletely surrounded and shielded magnetically. If the trans-
formers are not shielded in this way, be careful with their
mounting and polarity of the secondary connections. In
view of the low cost of these transformers it hardly seems
Fig. 149. Illustrating two commercial types of transformers for audio frequency
amplifiers.
worth while to try to make them, but if the reader wishes
to attempt this, the specifications of Fig. 150 may be help-
ful. This design is not theoretically correct nor above re-
proach, and will not give the flat characteristic mentioned
above, but represents the average commercial construc-
tion.
The completed transformer may be properly shielded
by placing it in a box made of soft iron, -fa" thick. Fila-
ment rheostats and other auxiliary apparatus may be pur-
RECEIVING APPARATUS
273
AUDIO FREQUENCY AMPLIFIER TRANSFORMER
SUITABLE FOR W. E. "J", MEIERS' (LOW MU), MOORHEAD, RADIOTRON;
UV-201 AUDIONS
WINDING SPECIFICATIONS
Primary.
Secondary.
No. turns
4500
11 000
Conductor
No 40 B &
S enameled
No. feet required
1050
3500
No ounces required
^ oz
\ 1 A oz
Resistance (ohms)
1100
3680
Approx. inductance (1000 c.)
8 h
50 h
. Silicon-steel ("transformer sheet") laminations, 10 mil thick;
or electrical-steel ("dynamo sheet") 5 mil thick.
Turn Ratio. 2 A5 : 1.
Normal flux-density (% m.a. plate current) = 5000/cm 2 .
Fig. 150. Constructional details of audio frequency interstage amplifier trans-
former.
chased at any radio supply house, and the other construc-
tional details of the amplifier require no comment.
88. Desirable Electrical Characteristics for Audions Used
as Amplifiers. At the present time the patent situation
respecting the audion is such that the amateur is forced
18
274 RADIO TELEPHONY
to accept what a few licensed manufacturers are pleased to
market, yet the life of these patents is rapidly nearing its
end and it is soon to be expected that the keener competi-
tion among a greatly increased number of producers will
stimulate the development and sale of better tubes and a
greater variety of them. At this date there is little oppor-
tunity for the exercise of any choice in selecting tubes for
their electrical characteristics. The average characteristics
of a number of commercial types have been tabulated on
page 231, and the reader will notice that with the exception
of the W. E. Co. "D" and the Meiers' types, the electrical
constants are about the same in all types, centering about
an amplification factor of 6 and an internal plate resistance
of 20,000 ohms.
The amplification which a tube is capable of producing
is ultimately proportional to the thermionic emission from
the filament. Within certain limits the variation of the
amplification factor and internal plate resistance with the
variation of the grid dimensions and position, take place
together in such a way that their quotient remains invariant.
Hence, since all the coupling devices which have been de-
scribed in discussing the cascade amplifier offer a definite
impedance in series with the plate resistance of the tube,
the drop across this impedance increases with any change
in the grid design which increases the amplification factor.
It follows that tubes used for amplification should have a
high amplification factor ("nm"); not too high, for the di-
electric losses in apparatus and distributed capacity and
feed-back effects conspire to vitiate the improvement,
but appreciably higher than the factors of the present
tubes. A "mu" of 20 appears to be about right for most
RECEIVING APPARATUS
275
purposes. The W. E. Co. "D" has a "mil" of 30-60,
but requires a plate voltage of 150 for proper operation.
I tested the correctness of the above idea by having the
"D" tube modified by moving the plate electrode nearer
to the grid, thus giving a lower operating voltage (80-100 v.)
and a lower "mu" (20). The experiments were remarkably
successful and established, as I think, the desirability of a
change from, the present practice of using "mu's" of 5-10.
Another important characteristic
to be secured in the audion is a low
grid-plate capacity. Indeed, for
radio frequency amplification this is
even more important than the increase
in "mu," as evidenced by the ampli-
fication factors tabulated on page
252 in the case of a resistance coupled
amplifier. It is there shown that the
"D" tube, even though possessed of
a "mu" 3-5 times as great as those
of most of the other types, does not
produce a proportionately greater
amplification on account of the feed-
back effects through its large grid-plate capacity. The
other capacities, plate-filament and grid-filament are of
small importance compared to this one. A few progressive
manufacturers have placed on the market tubes which
have been especially designed to have a small grid-plate
capacity. The W. E. Go's. "N" type, the "Liliputformat"
tube of the Germans, and the special tubes used by the
English and French (Fig. 151) in which the plate and grid
leads are taken out of opposite sides of the glass bulb hi-
\
Fig. 151. Showing
method of reducing grid-
plate capacity of audion by
bringing leads out opposite
sides of the bulb.
276
RADIO TELEPHONY
stead of through the mash in the usual way, are examples
of this. Few of these tubes have the desirable high "mu"
properties, however. Some reduction of the grid-plate
capacity may be had by bringing out the grid and plate
leads from opposite ends of the tube. This is done in the
Meiers' tube and is largely responsible for its low g-p capac-
ity. In the Moorhead and some of the other types the grid
Fig. .152. Illustrating two commercial sound distributing devices for use with
radio telephone-receivers.
and plate leads have been separated in going through the
mash. This is obviously favorable to a low capacity be-
tween these leads, especially in view of the high dielectric
constant of glass.
89. Loud-speakers.^-In its present usage the term "loud-
speaker" is applied generically to all electro-phonetic de-
vices employing a means for distributing the sound, such as
a large horn. We may, however, distinguish between two
RECEIVING APPARATUS 277
principal classes; those which are merely telephone re-
ceivers equiped with a sound distributing device, and those
designed to produce, usually with the aid of an audion
amplifier, a greater initial sound intensity, which is then
distributed by the customary means. Of the first class
there are many commercial types, two of which are illus-
trated in Fig. 152. In each of these instruments an ordinary
telephone receiver is applied, this application being clearly
shown in the right-hand type. Attachments are also mar-
keted, by means of which the sound-box and horn of the
ordinary phonograph may be pressed into service, the re-
ceiver being in this case substituted for the reproducer.
All these arrangements are of such technical simplicity as
to merit no further discussion.
The second class of loud-speaker, which in a strict sense
may be regarded as the real loud-speaker, produces sound
in great volume and generally finds its chief use in dis-
tributing the radio telephone speech and music over large
areas, such as in auditoria. The design of a good loud-
speaker is a matter of considerable technical difficulty, for
as in the case of the amplifier we are face to face with the
problem of devising a system which functions indifferently
over the entire range of speech frequencies, from 100 to
5000 cycles. As is well known, the mechanical system com-
prising the vibrating diaphragm and its actuating connec-
tions, when acted upon by a periodic force acts very much
like an electrical circuit consisting of an inductance, capacity
and resistance in series when acted upon by a periodic e.m.f .
The analogy is not flawless but will suffice for this discussion.
As the electric current depends upon the frequency of the
e.m.f. and reaches its maximum value at the frequency for
278 RADIO TELEPHONY
which the circuit is in resonance, so also in the mechanical
case the vibration of the diaphragm depends upon the fre-
quency of the vibro-motive-force and reaches its maximum
when this frequency corresponds to the mechanical res-
onance frequency of the diaphragm. The mechanical res-
onance frequency is often spoken of as the natural frequency
of the diaphragm, and in the usual telephone receiver with
diaphragm 2.1" diameter and 0.015 " thick, averages about
1000 cycles. It obviously depends upon the physical dimen-
sions of the diaphragm, and upon its elastic properties, as
well as upon any extra stiffness due to electromagnetic
reactions, or pressure reactions from air films, that may be
present.
Thus the design of a good loud-speaker involves not only
the production of a great volume of sound, but the impor-
tant matter of accomplishing this without acoustically mis-
representing the electric currents with which it is stimulated.
And in order that there shall be no distortion, the vibration
of the diaphragm must remain practically constant over a
wide range of frequencies. At the date of this writing, so
far as I know, there is not a loud-speaker on the market
which possesses the last-named characteristic essential for
the production of high quality speech*. The object can
hardly be hoped to be attained by means of the ordinary
diaphragm system unless by a great sacrifice of sensitivity
the tuning of the diaphragm system to a very high frequency
is resorted to. The most promising line of attack, and one
which any amateur with a little mechanical ingenuity may
successfully undertake for himself, is the application of air
* As the book goes to press a very excellent loud-speaker appears on the
market, so that a less pessimistic view of the situation may now be taken.
RECEIVING APPARATUS 279
damping to automatically increase the stiffness of the sys-
tem with increasing frequency, as in the Wente-Crandall
condenser transmitter. This should be reinforced by an
electrical network for maintaining constant the current
through the windings as the frequency is changed. There
are a number of physical " tricks" that await application
to this problem, and it is certainly to be hoped that they will
speedily receive this application. For there is room for
enormous improvement in existing apparatus, and the
popular demand for reproduction more appropriate to the
high quality of the speech and music being broadcasted by
many stations, instead of the present indescribable electro-
acoustic abortions ranging from those of the nasal and
"phonographic" type to the deep-throated disturbances re-
sembling a distant artillery engagement, will soon stimulate
the practical evolution of the more nearly perfect equip-
ment that exists in theory.
APPENDIX A
UNDERWRITERS' SPECIFICATIONS GOVERNING
INSTALLATION OF RECEIVING ANTENNA
EVERY amateur who plans to erect a receiving antenna
of the usual elevated type, out-of-doors, whether his
house is insured or not, should be guided by the rules and
methods of lightning protection prescribed by the National
Board of Fire Underwriters. Of course if the building in
which the radio receiving station is located is insured, the
provision of such lightning protection is usually insisted
upon by the insurance company. In view of this the
following reproduction of the most recent specifications of
the National Board may be of interest:
Rule 86. National Electrical Code
Specifications (For Receiving Stations Only):
ANTENNA
(a) Antennae outside of building shall not cross over or under electric light
or power wires of any circuit carrying current of more than six hundred volts,
or railway trolley or feeder wires, nor shall it be so located that a failure of
either antenna or of the above mentioned electric light or power wires can
result in a contact between the antenna and such electric light or power wires.
Antennas shall be constructed and installed in a strong and durable manner
and shall be so located as to prevent accidental contact with light and power
wires by sagging or swinging.
Splices and joints in the antenna span, unless made with approved clamps
or splicing devices, shall be soldered.
Antennae installed inside of buildings are not covered by the above speci-
fications.
LEAD-IN WIRES
(b) Lead-in wires shall be of copper, approved copper-clad steel or other
approved metal which will not corrode excessively, and in no case shall they
be smaller than No. 14 B. & S. gauge except that approved copper-clad steel
not less than No. 17 B. & S. gauge may be used.
281
282 APPENDIX A
Lead-in wires on the outside of buildings shall not come nearer than four
(4) inches to electric light and power wires unless separated therefrom by a
continuous and firmly fixed non-conductor that will maintain permanent
separation. The non-conductor shall be in addition to any insulation on the
wire.
Lead-in wires shall enter building through a non-combustible, non-ab-
sorptive insulating bushing.
PROTECTIVE DEVICE
(c) Each lead-in wire shall be provided with an approved protective
device properly connected and located (inside or outside the building) as near
as practicable to the point where the wire enters the building. The protector
shall not be placed in the immediate vicinity of easily ignitible stuff, or where
exposed to inflammable gases, or dust, or flying of combustible materials.
The protective device shall be an approved lightning arrester which will
operate at a potential of five hundred (500) volts or less.
The use of an antenna grounding switch is desirable, but does not obviate
the necessity for the approved protective device required in this section.
The antenna grounding switch if installed shall, in its closed position, form a
shunt around the protective device.
PROTECTIVE GROUND WIRE
(d) The ground wire may be bare or insulated and shall be of copper or
approved copper-clad steel. If of copper the ground wire shall be not smaller
than No. 14 B. & S. gauge, and if approved copper-clad steel it shall be not
smaller than No. 17 B. & S. gauge. The ground wire shall be run in as straight
a line as possible to a good permanent ground. Preference shall be given
to water piping. Gas piping shall not be used for grounding protective de-
vices. Other permissible grounds are grounded steel frames of buildings or
other grounded metallic work in the building and artificial grounds such as
driven pipes, plates, cones, etc.
The ground wire shall be protected against mechanical injury. An ap-
proved ground clamp shall be used wherever the ground wire is connected
to pipes or piping.
WIRES INSIDE BUILDINGS
(e) Wires inside buildings shall be securely fastened in a workmanlike man-
ner and shall not come nearer than two (2) inches to any electric light or
power wire unless separated therefrom by some continuous and firmly fixed
non-conductor making a permanent separation. This non-conductor shall
be in addition to any regular insulation on the wire. Porcelain tubing or
approved flexible tubing may be used for encasing wires to comply with this
rule.
APPENDIX A 283
RECEIVING EQUIPMENT GROUND WIRE
(/) The ground conductor may be run inside or outside of building. When
receiving equipment ground wire is run in full compliance with rules for Pro-
tective Ground Wire, in Section d t it may be used as the ground conductor
for the protective device.
On account of its connection across the receiving appa-
ratus, the usual electrolytic type of protective device, having
a high electrostatic capacity, is unsuitable. Accordingly the
Fig. 153. Two commercial types of lightning protective devices for installation
with receiving antenna.
present practice is to provide a very short spark gap which
will break down at 500 volts. Several forms of spark gaps
are marketed and may be purchased at almost any radio
supply house. Two types are illustrated in Fig. 153.
Some amateurs further protect their receiving instruments
by providing the familiar and previously prescribed light-
ning switch in addition to this protective device.
APPENDIX B
RADIO CLUBS
THE AMERICAN RADIO RELAY LEAGUE
WHEN one has a hobby it is very pleasant and natural to
seek intercourse with others of similar propensities. It is
largely to this impulse that clubs and associations of all
kinds owe their existence. So in the delightful field of radio,
particularly non-professional radio; from the early days
amateurs have been wont to band themselves together into
radio clubs- and associations. Not only is this beneficial
for the ordinary reasons, but is of especial value for the
proper protection of the rights of the private citizen pur-
suing radio for amusement or instruction, and in defending
it from the onslaughts of the military and of mercenary
professionals.
I feel that many of my readers will be novices in this
radio business and wish therefore to address to them the
appeal that after getting their radio house in order, one of
their first moves be to seek out and become affiliated with
their local radio club. Here you will come in contact with
many kindred spirits, with the radio beau esprit of your
community, and the ideas to be there gathered, the free
instruction, exchanges of experience and so forth, are of
inestimable value. The familiar prospect of a radio meet-
ing at which 60-year-old presidents of large institutions and
influential men will be found enthusiastically and deferen-
tially discussing the merits of this or that "hook-up" with
284
I
APPENDIX B 285
14-year-old school-boys is a curious one to contemplate and
to think about.
The domain of influence of a local organization is, how-
ever, very restricted, and from the point of view of pro-
tecting the amateur's rights when radio legislation is con-
templated by the Government, is quite impotent. This
indicates the need for an organization of national scope;
one great organization embracing the grand hierarchy of
radio amateurs, and not two or three. Fortunately such
an organization, the American Radio Relay League with
headquarters in Hartford, Connecticut, exists in this coun-
try and is probably the most powerful amateur radio club
in the world, having a present membership of ten thousand.
In view of the importance of this body in amateur radio
affairs, and the plea which is here made that every amateur
make it his immediate business to become a member of it,
a few remarks on its history and aims will perhaps be
appropriate. For this information I am indebted to Mr.
K. B. Warner, Secretary of the League and editor of its
admirable little journal, "QST."
"The American Radio Relay League is the only associa-
tion of its kind in the country, being of national scope,
entirely non-commerical in its nature, and truly of, by and
for. the amateur. It is a corporation without capital stock,
with a charter under the laws of Connecticut. Its gov-
erning body is a board of seventeen directors, elected by
popular vote every two years, and no man is eligible to
membership of the Board who is in any way financially
interested in the manufacture or sale of radio apparatus.
The officers of the League are elected by the Board members
and serve for two years.
286 APPENDIX B
"The purpose of the League is the advancement of
private radio, especially as exemplified by the American
amateur. We are bonded together for the more effective
relaying of friendly messages between our stations, for
legislative protection, orderly operating and scientific
growth. We have seventeen divisions in our Operating
Department, embracing the United States, Canada and
Alaska, and each division is in charge of a manager who
is a well-known and qualified amateur. In turn he has
district superintendents and city managers as assistants,
forming a field organization of about 400 men, who keep
closely in touch with the individual station owners all over
the country. A. R. R. L. is a hobby with these men and
all serve in their spare time without financial remuneration,
as do all of our officers with the exception of the Traffic
Manager and Secretary, who, devoting their entire time to
the work at the headquarter's office, must necessarily make
their living thereby.
"The League owns and operates "QST" as its official
organ, chronicling the activities of the amateurs all over
the country. QST is devoted solely to the interest of the
amateur and that interest is principally the practical im-
provement of short-wave communication. The A. R. R. L.
has represented amateur radio in legislative hearings ever
since its formation, and it may be safely said that there
have been several occasions when if no League had existed,
there would be no amateur radio today. Our substantial
prestige at Washington is due largely to our being bonded
together in a non-professional organization into which the
taint of commercialism cannot enter. We have made our-
t
APPENDIX B 287
selves into that kind of an association which the United
States itself can recognize and deal with.
"Thus whenever any matter affecting the amateur is
under consideration in Washington the view of the A. R.
R. L. is sought. When that expression is secured it repre-
sents the best opinion of seventeen men from all over the
country who in turn represent the general amateur in their
communities. To help in this business of being truly repre-
sentative of the amateur, there are some 400 clubs scattered
throughout the land which are affiliated with the League.
Affiliation costs a club nothing and nothing tangible is
given in return except a charter, but it bonds all together
with hoops of steel in a common brotherhood that of the
American 'ham.'
"From time to time our Operating Department stages
special stunts just to get some fun out of radio. We regularly
handle some thousands of messages every night over relay
routes, but occasionally knock off and try for a record.
The result is that we have handled a message from the
Atlantic Coast to the Pacific Coast and got the message
back to the east coast again in a total elapsed time of six
and a half minutes. Recently we handled messages from the
governors of the various states to the President, and forty
of the forty-eight messages were delivered, five not starting
and three only being lost in the process of transmission.
The A. R. R. L. recently conducted experiments in con-
nection with the 'fading' of radiotelegraphic signals for the
Bureau of Standards, and thousands of curves and data
sheets were obtained which are still being analyzed at the
Bureau. It was the A. R. R. L. that sent Mr. Paul F.
Godley to Scotland in the recent amateur trans-Atlantic
288 APPENDIX B
tests, in the course of which about three dozen American
amateur stations were heard across the Atlantic.
"It costs nothing to belong to the League except the
annual dues of two dollars. One does not even have to be
an amateur station owner, the only requirement being that
the applicant possess a bona fide interest in amateur radio.
The dues include, of course, a year's subscription to QST"
INDEX
A battery in audion circuit, 46
A-P tubes, 183; see also Thermionic rectifier
Absolute stability of cascade audion ampli-
fiers, 263
Absorption method of modulation, 143-145
Aerial, see Antenna
Aeriotron (audion), 235; electrical character-
istics of, 231
Alternating current circuit, containing re-
sistance, 28; containing resistance and
inductance, 29; containing resistance,
inductance and capacity, 29; resonance in,
29, 30
Alternating currents, 26-30
American Radio Relay League, amateur
trans-Atlantic tests, 92; history and pur-
pose (Appendix B), 285
Ammeter, 14
Ampere, def., 15
Ampere's rule concerning directions of cur-
rent and magnetic force, 16.
Amplification, audio frequency, 236; per
stage (3000-meter transformer-coupled
amplifier), 262; (3000-meter resistance-
coupled amplifier), 252; (2-stage audio fre-
quency resistance-coupled amplifier), 254
Amplification factor (of audion), def., 233;
comments on, 274
Amplifier, audion, 49; fas power amplifier),
51; regenerative audion, 204-214; cascade
audion, 238; radio- and audio frequency,
236; resistance, inductance and trans-
former coupling methods, 239-242; com-
bination regenerative and 1-sta.pe radio
frequency, 242; description of 3000-meter
resistance-coupled, 249; of 3000-meter
transformer-coupled, 260; of 2-stage audio
frequency resistance-coupled, 253; of 2-
stage audio-frequency transformer-cou-
pled, 271; of vario-transformer-coupled
for short wavelengths, 258; stability of
cascade, 262
Amplifier transformer, construction of radio
frequency (3000-meters), 261; construc-
tion of audio frequency, 273
Angular velocity, def., 29
Antenna, action of receiving, 36, 37; Alex-
anderson's multiplex, 64; cage construc-
tion, 62; circular flat-top, 60; condenser,
88-90; construction of spreaders, 80, 81;
construction of cage, 62, 82; description of
various types, 59-66; ellipsoidal (vertical),
60; fan, 65; hemispherical, 59, 60; house-
top, 85-90; insulation of, 79-82; inverted
"L," 61, 62; loaded, 36: loop, 64; require-
ments of, for transmitting and receiving,
58, 66; symmetrical multiplex, 63; "T"
type, 61, 62; triangular flat-top, 61; um-
brella, 65; vertical, 60; (radiation of elec-
tric wave from), 33
Antenna coil (trans.), see Antenna Inductor
Antenna efficiency (trans.), 66, 67
Antenna inductor, construction of (trans.),
128, 129; losses in, 128
Antenna losses, 66; conductors, 67, 68; earth
resistance, 70-78; imperfect dielectrics, 78,
79; leakage and corona, 85; in trees, 79
Antenna series condenser, construction of
(trans.), 129, 130
Antennae for receiving, Beverage wire, 91,
92; coil (loop), 93-95; (directional proper-
ties), 94; single wire, 91; Underwriters'
specifications governing installation of, 281
Antimony, use of, with silicon in crystal
'detector, 203
Armstrong, E. H., 101, 205, 223
Armstrong super-heterodyne receiver, 223-
232; 4-stage resistance-coupled amplifier
for, 249; transformer-coupled amplifier
for, 260
Armstrong tuned-plate circuit, comparison of,
for transmitting, 127; fundamental form
for oscillating audion, 101, 102; for regen-
erative amplification (rec.), 210; for trans-
mitting, 122, 123
Audio frequency amplifier, 236; construction
of 2-stage resistance-coupled, 252; con-
struction of 2-stage transformer-coupled,
270
Audion, amplifier, 49, 50; detector, 53-57;
invention of 12, 44; modulator, 144, 147;
oscillator, 50-53; uses for in radio tele-
phone system, 44; electrical and mechan-
ical data of transmitting types, 103-110;
electrical data of receiving types, 231
Audion detector, action of (with grid bias),
53-56; (with grid condenser), 56, 57; use
of gas in, 233, 234
289
2 go
INDEX
Audion oscillator, flow of power in, 52; self-
excited, 52; separately excited, 51; as
speech frequency generator, 146
Audiotron, detector audion, 233
Autodyne detection of c.w. signals, 216, 219
B battery in audion plate circuit, 46
Back-coupling, see Feed-back
Base, standard bayonet type for audion s, 46
Beat frequency, 215
Beats, formation of, in heterodyne reception,
215
Beverage, H. H., 91
Beverage wire (receiving antenna), 91, 92;
design of for 200-meters, 92; reduction of
static and interference by, 93
Bias, grid (for amplification), 49; (for de-
tection 53; (for modulation), 153; use of
with autodyne receiver, 220
Blocking action of modulator audion, 155
Borax solution for chemical rectifiers, 180,
181
Bridle wires (antenna construction), 80-82
Brush discharge loss in antenna, 66
Buzzer, double, for i.c.w. transmitter, 142;
test- for crystal detector (rec.), 202
C battery in audion grid circuit, 48
Cage antenna, 62; construction for reducing
conductor losses, 68; construction of, 81, 82
Capacity of conductors or condenser, 23, 24;
effect of in a.c. circuit, 29; farad, unit of,
24; grid -plate of audion (regenerative
effect of), 101, 210; see also Grid-plate
capacity; input of audion, 212; mechanical
analogy of in electric circuit, 25; of parallel
glass plate condenser, 130; of condensers
in parallel and series, 24; specific induc-
tive, 25
Capacity ground, 72; see Counterpoise
Carborundum for crystal detector, 203
Carrier-wave, 42
Cascaded audion amplifier, resistance coup-
ling in, 239, 240; inductance coupling in.
240, 241 ; transformer coupling in, 241. 242;
stability of, 262
Chaffee, B. L., 234
Chalcopyrite for crystal detector, 203
Chambers, F. B., 221
Chemical rectifier, construction and opera-
tion of, 179-183
Choke-coil, action of, in Heising modulation,
148-150; construction of 3 m.h. for r.f.,
131; construction of 6-henry for Heising
modulation, 151; construction of 50-
henry for filter circuit, 178, 179; grid for
paralleled tubes (trans.), 135; grid for
cascaded amplifier, 267, 269; coupling in
audion amplifiers, 240; use of in filter
circuit (trans.), 172-177
Chopper, use of, in i.c.w. transmitter, 142
Circuit, simple electrical, 14; Armstrong
super-heterodyne, 226, 227; Armstrong
tuned-plate (trans.), 101, 122, 123; (for
regenerative amplification), 210; 4-stage
resistance-coupled amplifier for 3000
meters, 249; 4-stage transformer coupled
amplifier for radio frequencies, 260; com-
bination regenerative and 1 -stage r.f.
amplifier, 244; vario-transformer-coupled
amplifier, 259; 2-stage resistance-coupled
amplifier for audio frequencies, 253;
2 -stage tiansformer-coupled amplifier for
audio frequencies, 271; Colpitts (trans.),
101, 120, 121; comparison of transmitting,
126; comparison of filter, 177, 178; filter
(trans.), 171, 172; fundamental oscillating
audion, 97-102; Hartley (trans.), 100, 118,
119; Meissner (trans.), 98, 112, 113; re-
versed feed-back (trans.), 107, 122, 123;
self-rectifying (trans.), 161, 163; tickler
coil (trans.), 99, 114-117; (for regenerative
amplification), 206
Clark, G. H., 99
Coefficient of self induction, 18, 19; of
mutual induction, 18
Coil antenna for receiving (data for con-
struction of), 93, 94; reduction of inter-
ference by, 94; for direction-finding, 94, 95
Coils for receiving, losses in, 193-197; con-
struction of, 195, 196
Colpitts, E. H., 101
Colpitts' circuit, fundamental form for oscil-
lating audion, 101; for transmitting, 120,
121; comparisons of, 127
Commutator, synchronous (mechanical recti-
fier), 184, 185
Condenser, blocking (rec.), construction of,
201; construction of .01 mfd. for receiving,
202; construction of .02 mfd. for filter
(trans.), 179; construction of antenna
series (trans.), 130; capacity of in series
and parallel, 24; electric, 23; energy
stored in, 24; forms for radio telephone
circuits, 25, 26; mechanical analogy of
charged, 24; use of oil in, 131; variable,
26, 131
Condenser antenna, description of, 88; for
installation on house-tops, 89, 90
Condenser transmitter, Wente-Crandall, 278
Conductance, clef., 14; input of audion,
211
Conduction current, def., 14
Conductive back-coupling in cascaded audion
amplifiers, 262
Constant- current modulation system, 147,
148
Constant-potential modulation system, 148
Constant-potential rectification (mechanical),
185
Continuous wave (c.w.) telegraphy, modula-
tion methods for, 140; Reinartz receiver
INDEX
291
for, 221-223; situation of telegraph key in,
141 ; signalling ranges with various powers,
103
Convection current, def., 14
Converter, audion as, 50, 51
Counterpoise, action of, 71, 72; design and
construction of, 73-76; combination with
direct ground, 76-78; distribution of earth
currents with, 72; in cellars, 87; trans-
mitting circuits for use with, 127
Coupling, magnetic, of two circuits, 18; para-
sitic in cascade audion amplifier, 262-268;
retroactive in audion, 52, (see also Feed-
back); resistance-, inductance-, and trans-
former- in audion amplifiers, 239-242
Coupling coils, construction of, for 3000-
meter amplifier, 261
Coupling resistance, construction of, for re-
sistance-coupled amplifiers, 250
Crystal detector, 203, 204
Current, alternating, 26-30; Ampere's rule
for relation with H, 16; conduction, 14;
convection, 14; (in audion), 46; electric, 13:
electron in audion, 47; flow of, in antenna
system, 6972; heating effect of, 15; mag-
netic effect of, 16; losses due to earth- in
antenna systems, 66, 69, 70; eddy- in
masts, 82-85.
Current distribution in antenna top, 62; in
earth, 69; (with direct ground), 69; (with
counterpoise), 71, 72; (with direct ground
and counterpoise), 76, 77.
D type audion (W. E. Co.), 230, 240, 248, 274;
electrical characteristics of, 231
Damon, L. R., 10, 153, 154
Damon's method of grid bias for modulator
tube, 153-155
Dead-end effect in receiving coils, 195
DeForest, Lee, invention of audion by, 12,
44; -McCandless audion, 234
Demodulator, see Detector
Detection of radio telephone signal, 55;
audion (with grid bias), 56; (with grid
condenser), 57
Detector, function of, in radio telephone sys-
tem, 43; audion, 53-57; crystal, 201-
203
Dick, L. B., 195
Dielectric, def., 25; losses in antenna, 66, 78,
79; losses in masts and guy wires, 82-85;
losses in house-top antennae, 87; losses in
receiving coils, 195
Dielectric constant, def., 25
Direct ground, combination of, with counter-
poise, 76-78; construction of cylindrical,
70, 71; distribution of earth currents in
case of, 69; transmitting circuits for use
with, 127
Direction-finding, 94
Directive antenna (loop), 65, 94
Distortion due to transformer-coupled am-
plifier, 270; in telephones and loud-speak-
ers, 276
Double-circuit tuner, 192, 193; construction
of, 199-201
E type audion (W. E. Co.), description and
electrical characteristics of, 108
Earth, direct, 70; counterpoise, 71
Earth currents, distribution of (direct
ground), 69, 70; (counterpoise), 72, 73;
losses due to, in antenna system, 66, 69,
70
Earth resistance, losses due to, 66, 69, 70;
reduction by combination of direct ground
and counterpoise, 7678
Eastham, Melville, 88
Eddy-currents, losses due to, in antenna
masts, 66, 82-85
Edge-effect in antenna, 62
Electric condenser, 23, see also Condenser
Electric current, see Current
Electric field of charged conductor, 22;
energy of, 24; of antenna, 33-35, 78,
79
Electric filter, see Filter
Electric force, def., 22; lines of, 23; (in wave
from vertical antenna), 35
Electric waves, 30-33; electric and magnetic
forces in, 34 (see also Radio waves)
Electricity, what is it? 13
Electrolytic rectifier, sec Chemical rectifier
Electromagnetic induction, 17; prevention of,
by shielding, 265, 272
Electromotive force (e.m.f.\ def., 14
Electron current in audion, 47
Electron emission by filament of audion,
Electron theory of matter, 13
Electrostatic capacity, see Capacity
Electrostatic shielding, 264, 265
Ellipsoidal antenna, 60
Ellis, W. G., 10, 195, 205
Emission, thermionic, 46
Ether, 31
Ether waves, see Electric waves
Farad, unit of capacity, def., 24, 25; micro-,
25; micro-micro-, 25
Faraday's first law of induction, 17
Feed-back in audions, 52; inductive in
Meissner circuit, 98, 99; in tickler coil cir-
cuit, 99; in Hartley circuit, 100; in Col-
pitts' circuit, 101; in Armstrong tuned-
plate circuit, 101; through grid-plate
capacity of audion, 102, 103; regenerative
effect of (rec.), 204, 205; effectsof, in cas-
cade amplifiers, 263
Ferris, Malcolm, 88
Fessenden, R. A., 71
2Q2
INDEX
Filament electrode of audion, 45; coated
platinum, 108, 235; tungsten, 235; life of,
in power tubes, 158, 160, 161; construction
of transformer for lighting, evaporation of,
160; operation at constant voltage, 160,
161; supply of power for heating (trans.),
157, 161; use of a.c. for heating, 158
Filament transformer, construction of
(trans.), 159
Filter, electrical, for plate voltage supply,
157, 164; Campbell type, 164, 170, 250;
comparison of, 177, 178; construction of
50-henry choke-coil for, 178, 179; design
of, 169-177; "low-pass," 170; rule for
proportioning I. and C in simple type, 175;
simple types of, 171-177; Type II, 170;
use of Campbell in resistance-coupled am-
plifier, 250
Flat-top antenna, triangular, 60; circular, 60;
rectangular, 61, 62; construction of, 80-82
Fogg, W. S., 10
Ford coil, use as modulation transformer, 146
Frequency, def. of a.c., 27; relation to wave-
length, 22
Frequency trap for eliminating harmonics
(trans.), 139
Fundamental wavelength of antenna, 36; re-
lation to dimensions, 60; voltage distribu-
tion at, 82; of single-wire receiving an-
tenna, 91; operation of transmitter at, 60,
138
Galena for crystal detector, 203
Gap, safety, for transmitting tubes, 105, 134;
for lightning protection, 282, 283
General Electric Co., 12, 98, 104, 231
Godley, P. F., 232, 250, 251, 287
Goldsmith, A. N., 12
Grid electrode of audion, 45; controlling ac-
tion of, in audion, 48, 49
Grid leak resistance, construction of, for
transmitter, 132; action of, with audion
detector, 57
Grid voltage modulation, 145, 146
Ground, counterpoise or capacity, 72; com-
bination of direct and counterpoise, 76-78;
direct, 70; transmitting circuits for use
with direct or counterpoise, 127
Ground lead, 68
Grounding of masts, 84; of tin-roof under
house-top antenna, 87
Harmonics, elimination of, in audion gener-
ator, 139; in transformer-rectifier-filter
system of plate supply, 166; use of in
heterodyne reception, 218
Hartley, R. V. L., 100
Hartley circuit fundamental for oscillating
audion, 100; comparisons of, for transmit-
ting, 127; for heterodyne oscillator, 100,
218; for master oscillator, 100, 125, 139;
for transmitting, 118, 119
Hazeltine, L. A., 9
Heat, generation of, by electric current, 15;
wavelengths of radiation, 33
Heising, R. A., 149
Heising modulation system, 146-153; com-
plete wiring diagram of transmitter using,
189; filter for use with, 172-174, 177
Hemispherical antenna, 59, 60
Henry, unit of inductance, def., 19; subdi-
visions of, 10; micro-, 19; milli-, 19
Hertzian waves, wavelength of, 33; see also
Electric waves
Heterodyne, detection of c.w. signals, 214;
separate, 215, 217; Armstrong super-,
223-232
Heterodyne oscillator, construction of, 218;
Hartley circuit for, 100, 217
Honeycomb coils for receiving circuits, 195,
197
House-top antennae, 85-90
Howling of amplifiers, sec Stability of ampli-
fiers
Hull, L. M., 9
Hysteresis losses in imperfect dielectrics, 78
Impedance of a.c. circuit, 29, 30; grid- in
audion detector, 57; input of audion, 102
Inductance, effect of, in a.c. circuit, 29;
coupling in audion amplifiers, 240; def.,
19; effect of, in electric circuit, 18; mechan-
ical analogy of, 19
Induction, electromagnetic, 17; (first law
of), 18
Inductor, construction of antenna-, for trans-
mitter, 128, 129; form of, for radio tele-
phone circuits, 20; fixed and variable types,
20, 21; see also Choke-coil and Variom-
eter
Input impedance of audion, 102, 211; con-
ductance due to inductive plate circuit,
211; capacity due to inductive plate cir-
cuit, 121; effect of, in radio frequency am-
plifier, 247
Insulation of antenna system, 79-82; of
counterpoise, 74; of guy -wires, 83; of
single- wire receiving antenna, 91; of
antenna masts, 83; of lead-in, 82
Insulators, 14; 17-in. for antenna and coun-
terpoise, 74; construction of lead-in, 82;
porcelain egg-, 83
Interrupted continuous wave (i.c.w.) teleg-
raphy, def., 140; compared with "spark"
telegraphy, 141; modulation for, 142;
position of telegraph key in, 141
Interstage transformer, construction of, for
audio frequency amplifier, 273
Inverted "L" antenna, description of, 61, 62
J type audion (W. E. Co.), 230, 246, 247;
electrical characteristics of, 231
Jones, Lester, 221
Joule an effect (loss), 15
INDEX
2 93
Kenetrons, 183; see also Thermionic rectifiers
Lead-in of antenna, 62; cage construction
for, 62, 68; insulation of, 82
Leakage, losses due to, in insulators, 85
Lee, J. W., 71
Light, velocity of, 31; wavelengths of, 33
Lightning protection, 91; underwriters' speci-
fications for, 281
Liliputian audions, 235, 275
Liliputian vario-transf ormer, construction of,
258
Lines of force, electric, 23; magnetic, 16
Loop antenna, 64, 93; data for construction
for receiving, 94; design of, for receiving,
94, 95; for direction-finding, 94, 95
Losses in antennae, 66, 85; in antenna con-
ductors, 67, 68; in antenna coil, 128; in
antenna condenser, 129-131; due to earth
currents, 69, 70; in masts, 82, 85; in re-
ceiving coils, 194-197; in switches for
receiving coils, 195, 196; in trees, 79;
Telefunken method for reducing in
masts, 84, 85
Loud-speaker, description, 276; distortion
due to, 277; design, 278, 279
Loughlin, W. D., 10, 195
Low-pass filter, 1 70
Magnetic field of antenna, 33-35; energy
stored in, 20; of solenoid, 16; due to
straight current, 16
Magnetic force, def., 16; Ampere's rule for
direction of, 16; in wave from simple
antenna, 35
Master oscillator, Hartley circuit for, 100,
125, 139
Master-oscillator system compared with
self-excited generator, 96, 126; circuit for
transmitting, 124, 125; adjustment of, 139;
modulation of, 143 (see also Grid voltage
modulation)
Masts, construction of metal, 82, 83; insula-
tion of, 83; losses in antenna, 66, 84, 85;
Telefunken method of reducing losses in,
84, 85
Mechanical analogy of inductance, 19; of
charged condenser, 24
Mechanical data of power audions, 103-110
Mechanical rectifiers, 184; constant-poten-
tial, 185; construction of, 187
Meiers' audion, 230, 246, 248; electrical
characteristics of, 231
Meissner, A., 98
Meissner circuit compared with other trans-
mitting circuits, 127; fundamental form
for oscillating audion, 98, 99; for trans-
mitting, 112, 113; complete wiring dia-
gram of transmitter using, 188, 189
Micro-henry, def., 19
Micro-farad, def., 25
Microphone, function of, in wire telephone,
38; use of simple, for modulation, 144
Miller, J. M., 9, 71, 78, 102
Milli-henry, def., 19
Mineral detector, see Crystal detector
Modulation in wire telephony, 39; in radio
telephony, 42; in c.w. and i.c.w. telegraphy,
140; methods in radio telephony, 143; by
grid voltage variation, 145, 146; by power
absorption, 143-145; by plate voltage
variation, 147-155; constant-current and
constant-potential systems, 147, 148;
Heising system of, 149; filters for use with
grid voltage and Heising, 171178
Modulation transformer for grid voltage
modulation, 146; for plate voltage modu-
lation, 152; construction of, for Heising
system, 154
Modulator tube, operation of, 147, 148
Moor head audion, 230, 246, 276; electrical
characteristics of, 231
Motor-generator for plate voltage supply,
157, 187, 188
Motors, rewinding old, for plate voltage
supply, 187
Multiple-tuned antenna, Alexanderson's, 64;
symmetrical, 63
Mutual conductance, def. (audion), 233
Mutual inductance, def., 18
Mutual induction, coefficient of, see Mutual
inductance
N type audion (W. E. Co.), 235, 275
Natural frequency of diaphragm, 278
Negative resistance due to regenerative
audion, 205, 211, 212
Ohm, unit of resistance, 15
Ohm's law for steady currents, 14
Oil, use of, in transmitting condensers, 131;
dehydration of, 131
Oscillating audion circuits, Armstrong tuned-
plate, 101, 102; Colpitts, 101; Hartley, 100;
Meissner, 98; reversed feed-back, 101;
symmetrical, 163; tickler coil, 99
Oscillations, generation of, by audion, 50-53;
ultra-radio frequency (in paralleled trans-
mitting tubes), 135; (in cascade amplifiers),
269
Oscillator, audion, flow of power in, 52 ; con-
struction of heterodyne-, 218
Overloaded receiving tubes for transmitting,
110; for master oscillator, 125
P type audion (G. E. Co.), 104
Parallel operation of transmitting tubes,
134-136
Peanut audions, 235
Perikon, crystal detector, 203
Period of a.c.. 27
Periodic time, 27
Phase, lead and lag, in a.c. circuit, 28-30
294
INDEX
Pierce, G. W., 9, 170
Plate electrode of audion, 45
Plate resistance of audion, def., 233
Plate voltage modulation, 146-155
Potentiometer, see Stabilizer
Power, flow of, in audion oscillator, 52;
transfer from audion to antenna, 98;
supply of, for plate circuit (trans.), 156,
157; supply of, for filament, 157161;
method of, supply using step-up trans-
former, filter and rectifiers, 163, 164
Power amplifier, audion as, 51; use of, for
operation of loud-speaker, 270
Power audions, electrical and mechanical
data of, 103-110; operation of, in parallel,
134
Priess, W. H., 221
Protective device (lightning), 182, 283
Protective ground wire, 282
Protective measures for transmitting audi-
ons, 133, 134; for lightning, 281
Radiation resistance of antenna, 66, 67; of
condenser antenna, 90
Radio compass, see Direction-finding
Radio Corporation of America, magnetic
modulator, 145
Radio frequency amplifier, 236; combination
regenerative and 1-stage, 242; construc-
tion of 3000-meter resistance-coupled for
Armstrong super-heterodyne receiver, 249;
construction of 3000-meter transformer-
coupled for Armstrong super-het. receiver,
260; performance of resistance-coupled,
247
Radio telephony, development of, 12; com-
pared with radio telegraphy, 43; carrier-
wave system of, 42; (principles of), 41-
43
Radio waves, production of, 33; transmission
of, 36; reception of, 36
Radiotron (trans, audions), UV-204, UV-203,
UV-202, description and electrical data of,
104-107; (receiving audions), UV-201,
UV-200, electrical characteristics of, 231
Ratio of transformation in audion trans.,
135, 137
Receiver, single circuit, see Single-circuit
tuner; double circuit, see Double-circuit
tuner; Armstrong super-heterodyne, 223-
232; autodyne for c.w. signals, 221-223;
construction of employing combination
regenerative and 1-stage radio frequency
amplifier, 244; heterodyne for c.w. sig-
nals, 217; Reinartz for c.w. signals, 221-
223; simple type employing crystal detec-
tor, 201-204
Receiving antenna, Beverage wire, 91, 92;
loop or coil, 92-95; requirements for, 58,
59, 90, 91; single wire, 91; Underwriters'
specifications governing installation of,
281
Rectification, action due to asymmetrical
audion characteristic, 55; constant-poten-
tial (mechanical), 185; in grid circuit of
audion, 57; in plate circuit of audion, 54
Rectifier, thermionic, 47, 183; constant-
potential mechanical, 185; construction of
chemical, 181-183; construction of me-
chanical, 186, 187; connections of thermi-
onic, 184; description of chemical, 179,
180; mechanical, 184, 185; use of, for sup-
plying d.c. plate voltage, 157, 164
Regeneration, action in audion circuit, 204;
due to plate inductance, 211
Regenerative amplification with audion,
204-214
Reinartz, J. L., 221
Reinartz tuner for c.w. signals, 221-223
Resistance, def., 14; unit of, 15; effect of in
a.c. circuit, 28-30; of conductors in series
and parallel, 15; losses due to, in antenna,
66; speech controlled, in Heising modula-
tion, 148; high frequency of coil, 194, 195;
negative due to regenerative audion, 205,
211, 212; def. of plate, of audion, 233;
coupling in audion amplifier, 239
Resistance, coupling, construction of, for
cascade amplifiers, 250
Resistance-coupled amplifier, def., 239; de-
scription of 3000-meter for Armstrong
super-het. receiver, 249; description of
2-stage audio frequency, 252; effect of
audion capacities in, at radio frequencies,
247
Resistance, grid, construction of (trans.),
132
Resonance in a.c. circuit, 30; mechanical, in
telephone receiver, 277, 278
Retro-active coupling in audion, 52; see also
Feed-back
Reversed feed-back circuit, comparisons
with others for trans., 127; fundamental
form for oscillating audion, 101; for
transmitting, 122, 123
Right-hand rule for electric current and
magnetic force, 16
Round, H. J., 71
Round's round ground, 71
Safety gap for power tubes, 105, 134
Self-excitation, fundamental circuits for
(audion), 97-103
Self-excited audion oscillator compared with
separately excited audion oscillator, 96;
adjustment of, 137, 138
Self-heterodyne, see Autodyne
Self-induction, coefficient of, 18; see also
Inductance
Self-rectifying circuits (trans.), using one-
half a.c. cycle, 162; using both halves of
a.c. cycle," 162, 163
Separate-heterodyne, detection of c.w. sig-
nals, 215, 217
INDEX
295
Separately excited audion oscillator, 51;
compared with self-excited a.o., 96
Shielding cascade audion amplifier, 263
265; telephone cords, 269
Signal-static ratio of receiving antenna, 91
Signalling range, c.w. with various powers,
103; i.c.w. and spark, 141
Silicon, fused, for crystal detector, 203
Simple-harmonic alternating current, 27
Single-circuit tuner, 192, 193; construction
of, 197, 198
Single-wire receiving antenna, 91
Sinusoidal alternating current, 27
Spark coil, use of, as modulation transformer,
146
Specific inductive capacity, 25
Speech, nature of, 39; frequencies present in,
40
Speech-signal in radio telephony, 42; de-
tection of, 54, 55
Spreader, construction of, 80, 81; wood and
metal, 80, 81
Squirrel-cage antenna construction, 62
Stability of cascade audion amplifiers, 262;
conditional- and absolute-, 263
Stabilizer for i.f. amplifier, 245, 251; use of
resistance in plate circuit, 255
Static, 90, 91; reduction by Beverage wire,
92
Super-heterodyne, Armstrong, for short
wave amplification, 223; 3000-meter re-
sistance-coupled amplifier for use with,
249; 3000-meter transformer-coupled am-
plifier for use with, 260
Switch for receiving coils, construction of
mica-insulated, 196, 197
Symmetrical multiplex antenna, 63
Symmetrical oscillating circuits, 163
Synchronous commutator, 184, 185
Synchronous motor, use of, with mechanical
rectifier, 186
T type antenna description of, 61, 62
Telefunken Company, use of umbrella
antenna, 65; method of reducing mast
loss, 84, 85; Meissner circuit, 98
Telegraph key, connection of, in c.w. trans-
mitter, 114
Telegraphy, c.w., see Continuous wave
telegraphy
Telephone receiver, function of, in wire
telephony, 38
Telephony, principles of wire, 38-11 ; develop-
ment of radio, 12; comparison of wire and
radio, 43
Test buzzer for crystal detector, 202
Thermionic emission, 46; current in audion,
46
Thermionic rectifier, 47, 183
Tickler coil, association of, with single-circuit
tuner, 197, 198; construction of (rec.),
206
Tickler coil circuit, comparisons with other
transmitting circuits, 127; effect of tickler
coupling in receiving, 208; fundamental
form for oscillating audion, 99; for regen-
erative amplification (rec.), 206; operation
of (rec.), 207: for transmitting, 113-117
Transformer, construction of (for Heising
modulation system \ 151; (for heating
filaments of power tubes), 159; (of step-up
for power supply for plate circuit), 167-169;
coupling in audion amplifier, 241 ; construc-
tion of radio frequency amplifier, 268;
construction of audio frequency amplifier,
273; input in audion amplifier, 49; output
in audion amplifier, 49; step-up, for plate
supply, 157; vario-, 257
Transformer-coupled audion amplifier for
radio frequencies, 254; description of
3000-meter for Armstrong super-hetero-
dyne, 260; description of 2-stage audio
frequency, 270; amplification per stage, 262
Transformer-rectifier-filter system, use of,
for plate supply, 157, 164; complete wiring
diagram of transmitter using, 188, 189;
design of, 165-168
Transmission of radio waves, 36
Transmitter, adjustment of self-excited, 135
138; adjustment of master-oscillator, 139;
assembly of, 190
Transmitter, telephone, see Microphone
Transmitting antenna, requirements for, 58,
59, 66, 67
Transmitting circuits, audion, collection of,
110-125; comparison of, 126; complete,
188, 189; Armstrong tuned-plate, 122, 123;
Colpitts, 120, 121; Hartley, 118, 119;
master-oscillator, 124, 125; Meisner, 112,
113; reversed feed-back, 122, 123; tickler
coil with inductive plate coupling, 116,
117; tickler coil with inductive grid coup-
ling, 114, 115
Trees, losses due to, in antenna field, 79
Triangular ftat-top antenna, 60
Tubes, vacuum, 44; see Audion; power-, see
Power audions
Tuned-plate circuit, comparison with other
transmitting circuits, 126, 127; funda-
mental form for oscillating audion, 101;
for regenerative amplification (rec.), 210;
for transmitting, 122, 123
Tuner, double-circuit, def., 192, 193; con-
struction of, 199-201
Tuner, single-circuit, def., 192, 193; con-
struction of, 197, 198
Tuning of a.c. circuit, 30; of transmitter,
135-138; receiving antenna, 36, 37; single-
circuit and double-circuit methods of, in
receiving, 192
Tuning apparatus, construction of (rec.), 194;
losses in (trans.), 66, 68. 128, 132; receiv-
ing, 102-201
Type II filter, 170
296
INDEX
Umbrella antenna, 65
Underwriters' specifications for lightning
protection and installation of receiving
antennae, 281
Variable condenser, 26
Vario-coupler, def., 199
Vario-transformer, def., 257; design and
construction of liliputian, 257; use of, in
radio frequency amplifier, 257-251
Vario-transformer-coupled amplifier, de-
scription of, 259
Variometer, principle of, 21; use of, as choke-
coil in transmitter, 121; use for tuning
receiver, 201, 214; use- of, in radio fre-
quency amplifier, 245; use of, in regenera-
tive amplification, 210; description of
liliputian, 257
Velocity, angular, 29; of light, 31
Vertical antenna, 60
Volt, unit of e.m.f., 15
Voltage node in loaded antenna, 77
Voltmeter, 14
Warner, K. B., 10, 285
Wavelength, def., 32; relation to freqiu icy,
32; of X-ray, heat, light and radio waves,
33; fundamental of antenna, 35, 36; best
operating in transmitting, 67; adjustment
of, in transmitter, 135-137
Waves, electric, 30-33; double-cored, 40;
guided, 40
Western Electric Company, 12, 100, 101,
108, 149
Wire telephony, principles of, 38-41; current
variations due to vowel sounds in,
39
Wireless telephony, using voice currents
directly, 40
Wiring diagrams, complete for transmitter,
188, 189 9
Wood spreaders, 80-82
X-radiation, 31; wavelengths of, 33
X-ray, see X-radiation
Zincite for crystal detector, 203
THIS BOOK IS DUE ON THE LAST DATE
STAMPED BELOW
AN INITIAL FINE OF 25 CENTS
WILL BE ASSESSED FOR FAILURE TO RETURN
THIS BOOK ON THE DATE DUE. THE PENALTY
WILL INCREASE TO SO CENTS ON THE FOURTH
DAY AND TO $1.OO ON THE SEVENTH DAY
OVERDUE.
JAN 11 1933
NOV 9 133
MAR M 1946
rv
UNIVERSITY OF CALIFORNIA UBRARY
