(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Proceedings"

n 



Digitized by the Internet Archive 

in 2010 with funding from 

University of Toronto 



http://www.archive.org/details/p3p4proceedin1905inst 



2 F 



THE INSTITUTION 



•f 



OF 



MECHANICAL ENGINEERS. 



ESTABLISHED 1847. 



PROCEEDINGS. 



1905. 

Parts 3-4. 




PUBLISHED BY THE INSTITUTION, 
Storey's Gate, St. James's Park, Westminster, S.W. 



The right of Puhlication and of Translation is re^^erved. 



TJ 

I 

I ^ 



Ill 



CONTENTS. 



1905. 
Parts 3-4. 



List of Past-Presidents .... 

List of Officers ...... 

Pboceedings, Belgian Meeting. — Keception 

Election of New jVIembers 

Transferences ..... 

Votes of Thanks ..... 

"Superheaters in Locomotives" ; by J. B. Flamme (Plates 15-17) 

" Electric Winding-Machines " ; by P. Habets . 

" Ferro-Concrete " ; by E. Noaillon (Plates 18-20) . 

" Steam-Jacketing " ; by A. L. Mellanby (PJate 21) . 

" Large Gas-Engines" ; by K. E. Mathot (Plates 22-37) 

" Strength of Columns " ; by W. E. Lilly 

Excursions, &c. ....... 

Notices of Works visited ..... 

Memoirs ......... 

Proceedings, October Meeting — 

Deceases of Sir E. H. Carbutt, Bart., Mr. W. Dean, and Mr. J. Mansergh 

Business ...... 

Election of New Members 

Transferences ..... 

" Cartridge-Case Manufacture " ; by Col. L. Cubillo 

(Plates 38-49) 

Proceedings, November Meeting. — Business 

Transferences ...... 

"Alloys Kesearch"; by H. C. H. Carpenter, R 
P. Longmuir (Plates 50-61 j 
Pboceedings, December Meeting. — Business 

Election of New Members .... 

Trnnsferences ...... 

Memoirs ........ 

Index to Proceedings 1905, Parts 3 -4 . 
Plates, 15-61. 



and A. P. Head 



A. Hadfield, and 



PAGE 

iv 

V 

401 
402 
404 
406 
409 
429 
485 
519 
619 
697 
723 
729 
777 

785 
786 
786 
790 

791 
855 
855 

857 
1043 
1043 
1046 
1047 
1059 



IV 



®^c Institutian of ^let^anital Engineers. 



PAST-PRESIDENTS. 

George Stephenson, 1847-48. {Deceased 1848.) 

Robert Stephenson, F.R.S., 1849-53. {Deceased 1859.) 

Sir William Fairbairn, Bart., LL.D., F.R.S., 1854-55. {Deceased 1874.) 

Sir Joseph Whitworth, Bart., D.C.L., LL.D., F.R.S., 1856-57, 1866. 

{Deceased 1887.) 
John Penn, F.R.S., 1858-59, 1867-68. {Deceased 1878.) 

James Kennedy, 1860. {Deceased 1886.) 

The Right Hon. Lord Armstrong, C.B., D.C.L., LL.D., F.R.S., 1861-62, 1869. 

{Deceased 1900.) 
Robert Napier, 1863-65. {Deceased 1876.) 

John Ramsbottom, 1870-71. {Deceased 1897.) 

Sir William Siemens, D.C.L., LL.D., F.R.S., 1872-73. {Deceased 1883.) 

Sir Frederick J. Bramwell, Bart., D.C.L., LL.D., F.R.S., 1874-75. 

{Deceased 1903.) 
Thomas Hawksley, F.R.S., 1876-77. {Deceased 1893.) 

John Robinson, 1878-79. {Deceased 1902.) 

Edward A. Cowper, 1880-81. {Deceased 1893.) 

Percy G. B. Westmacott, 1882-83. 

Sir Lowthian Bell, Bart., LL.D., F.R.S., 1884. {Deceased 1904.) 

Jeremiah Head, 1885-86. {Deceased 1899.) 

Sir Edward H. Carbutt, Bart., 1887-88. {Deceased 1905.) 

Charles Cochrane, 1889. {Deceased 1898.) 

Joseph Tomlinson, 1890-91. {Deceased 1894.) 

Sir William Anderson, K.C.B., D.C.L., F.R.S., 1892-93. {Deceased 1898.) 

Sir Alexander B. W. Kennedy, LL.D., F.R.S., 1894-95. 

E. Windsor Richards, 1896-97. 

Samuel Waite Johnson, 1898. 

Sir William H. White, K.C.B., LL.D., D.Sc, F.R.S., 1899-1900. 

William H. Maw, 1901-02. 

J. Hartley Wioksteed, 1908-04. 



C^e Institution of iilet^anital (gngtneers. 



OFFICERS. 



1905. 



PRESIDENT. 
Edward P. Martin, Abergavenny. 

PAST-PRESIDENTS. 

Samuel Waite Johnson, Nottingham. 

Sir Alexander B. W. Kennedy, LL.D., F.R.S., London. 

William H. Maw, London. 

E. Windsor Richards, Caerleon. 

Percy G. B. Westmacott, Ascot. 

Sir William H. VVhite, K.C.B., LL.D., D.Sc, F.R.S., .. London. 

J. Hartley Wicksteed, Leeds. 

VICE-PRESIDENTS. 

John A. F. Aspinall, Manchester. 

Edward B. Ellington, London. 

Arthur Keen, Birmingham. 

Sir William T. Lewis, Bart., Aberdare. 

T. Hurry Riches, Cardiff. 

A. Tannett- Walker, Leede. 

MEMBERS OF COUNCIL. 

SirBenjamin Baker,K.C.B.,K.C.M.G.,LL.D.,D.Sc.,F.R.S., London. 

Sir J. Wolfe Barry, K.C.B., LL.D., F.R.S., London. 

Henry Chapman, London. 

George J. Churchward, Swindon. 

Henry Davey, London. 

H F. Donaldson, Wo(»lwich. 

H. Graham Harris, London. 

Edward Hopkinson, D.Sc, Manchester. 

J. Rossiter Hoyle, Sheffield. 

Henry A. Ivatt, Doncaster. 

Henry Lea, Birmingham. 

Michael Longridge, Manchester. 

The Right Hon. William J. Pirrie, LL.D., Belfast. 

Sir Thomas Richardson, Hartlepool. 

John F. Robinson, London. 

Mark H. Robinson, Rugby. 

James Rowan, Glasgow 

John W. Spencer, Newcastle-on-Tyne. 

Sir John I. Thornycroft, LL.D., F.R.S., London. 

John Tweedy, Newcastle-on-Tyne. 

Henry H. West, Liverpool. 

HON. TREASURER. AUDITOR. 

Harry Lee Millar. Robert A. McLean, F.C.A. 

SECRETARY. 

Edgar Worthington, 

Tlie Institution of Mechanical Engineers, 

Storey* s Gate, St. James's Park, Westminster, S. W. 

Telegraphic address: — Mech, London. Telephone: — Westminster^ 264. 



June 1905. 401 



€\t |nstit«tioit of Pedjaiucal Engineers. 



PROCEEDINGS. 

SUMMER MEETING IN BELGIUM. 



June 1905. 



The Summer Meeting of the Institution, held in Liege, Belgium^, 
commenced on Monday, 19th June 1905, at Nine o'clock p.m., by 
a reception of the President, Edward P. Martin, Esq., the Council, 
Members, and Ladies in the Town Hall, Liege, by the" College 
consisting of Bourgmestre Kleter and the Echevins. Professor 
Alfred Habets, President of the Liege Association of Engineers^ 
introduced the President to the Bourgmestre, who offered a hearty 
welcome to the Institution on behalf of the City, andlrecalled the 
great progress made by the engineering profession since the 
previous Meeting of the Institution in 1883. 

The President, speaking in French, thanked the Bourgmestre 
and College for their kind and hospitable reception. 

The Eeception was followed by a Conversazione in the rooms of 
the Town Hall, attended by a large number of Belgian guests as 
well as by the English visitors. 



2 G 



402 



BUSINESS. 



June 1905. 



The General Meeting was held in the Academic Hall of the 
University, Liege, on Tuesday, 20th June 1905, at Ten o'clock a.m. ; 
Edward P. Martin, Esq., President, in the chair. 

The Minutes of the previous Meeting were read and confirmed. 



The President announced that the Ballot Lists for the election 
of New Members had been opened by a committee of the Council, 
and that the following seventy-eight candidates were found to be 
duly elected : — 



MEMBERS. 

Alexander, Walter, . 
Austin, Edmund George, 
Baechtold, Charles Albert, 
Brodie, George Wallace, 
Carr, Andrew Custance, 
Cook, William Hall, 
Grew, Frederick William, 
kuroda, tsunema, 
Litton, Francis Henry, 
Maxwell, David William Francis, 
Paget, Cecil Walter, 
Rhodes, William Harrison, 
Ross, James Russell, 
Stanfield, Richard, . 
Stewart, James, 
Tandy, John O'Brien, 
Utting, John, . 
Watson, Frank Leslie, 



Brighton. 

Calcutta. 

New York. 

Woolwich. 

Asansol, India. 

Stalybridge. 

Sudbury, Suffolk. 

Chikugo, Japan. 

Tientsin. 

Weybridge. 

Derby. 

Wakefield. 

Glasgow. 

Edinburgh. 

Glasgow. 

Crewe. 

Bombay. 

Leeds. 



associate members. 

Baker, Arthur, .... London. 

Bazin, John Ralph, .... London. 
Berry, Arthur Osborne, . . . London. 



June 1905. 



ELECTION OF NEW MEMBERS. 



403 



BowEN, Charles William, 
Brewer, Albert Ernest, 
Brooke, Harry, 
Carter, Henry Charles, 
Charnock, John Aiton, 
Clarkson, Sydney Samuel, 
Cox, Lionel Maidstone Russell, 
Davidson, William Warburton, 
DowsoN, Ernest Alfred, . 
Ellison, William Thomas, . 
Gamble, George Martin, . 
Hadwbn, Frederick Walter, 
Hamilton, James Alexander, 
Hartness, John Anton, 
HiBBERD, Frederick Charles, 
Hill, Thomas, . 
Holland, Louis Carl, 
HowARTH, Edward, . 
Hughes, Arthur Mumford, 
Jackson, Robert Montresor, 
JopLiNG, Hugh Lanzi Woodwell, 
Kennedy, William, . 

LiDBETTER, ChARLES FREDERICK, 

List, John Forster, . 
Longley, Roland George, . 
Millar, William Pettigrew, 
MuNBY, Ernest John, 
Newman, Kenneth Charles Horton, 
Nichols, Frederick Albert, 
Parrott, Arthur George, 
Pashby, Arthur Harold, 
Pope, Ashley Philip, 
QuELCH, Arthur Temple, 
RoscoE, Edwin Borton, 
Rosenberg, Herbert Morris, 
Scott, Woolby Lockwood, 



Eastbourne. 
London. 

Rochdale. 
Reading. 

Birmingham. 

London. 

Rugby. 

Dublin. 

Birmingham. 

Manchester. 

Birmingham. 

Halifax. 

Singapore. 

London. 

Manchester. 

Runcorn. 

Manchester. 

Glasgow. 

London. 

Monte Caseros, Arg. Rep. 

Leeds. 

Singapore. 
London. 
London. 
Mysore. 

Glasgow. 

London. 

Hong Kong. 

Calcutta. 

Manchester. 

Aliwal North. 

London. 

London. 

London. 

New York. 

Ipswich. 

2 g 2 



404 



ELECTION OF NEW MEMBERS. 



June 1905. 



Sharpley, Eeginald, 
Tasker, Edward Ernest, 
Wood, Wilfred Pimm, 



Madras. 
London. 
Birmingliain. 



graduates. 
Bruckmann, Otto James, . 
Cook, William Stanley, 
Cranwell, Harry, 
Davies, John Ernest, 
Davies, Victor Charles, . 
Gamble, Stuart Arthur, . 
Huntley, Horace Frederick, 
KiEFERT, Charles, 
Lalor, Fintan John, . 
MoNAGHAN, Thomas Joseph, 
MuNYARD, Matthew Henry, 
Needham, Cyril ArmitagEj 
Simpson, Frederick Dudley, 
Ward, Charles William Dawson, 
Whittington, William Ewart, 
Willcox, Benjamin Bruce, 
Wilson, Percy Hutchinson, 
Worth, Henry Norman, 



London. 

Manchester. 

London, 

Camborne. 

London. 

London. 

Peterborough. 

London. 

London. 

London. 

London, 

Birkenhead. 

Birmingham. 

Nottingham. 

Dublin. 

London. 

Alfreton. 

Nottingham. 



The President announced that the following nine Transferences 
had been made by the Council since the last Meeting : — 

Associate Members to Members. 



Berrington, Ernest Evans Willoughby, 

Brotherhood, Stanley, 

Cleave, Arthur Harold Wyld, 

Graham, Walter, 

Haste, Frederick Charles, 

Herschmann, Arthur Julius, 



Wolverhampton. 

London. 

London. 

Greenock. 

London. 

New York. 



JuNEJlOOo. TRANSFERENCES. 40i 

Orr, John, Johannesburg. 

Page, James Handford, . . . Liverpool. 

ToTTLE, Edward George, . . . London. 



The following Papers were then read^and discussed : — 

" Superheaters applied to Locomotives on the Belgian State 
Kailways"; by M. J. B. Flamme, Inspecteur-General de 
I'Administration des Chemins de fer de I'Etat Beige, 
Brussels. 

" Electric Winding-Machines " ; by Professor Paul Habets, of 
Brussels. 

" Ferro- Concrete, and some of its most characteristic 
applications in Belgium " ; by M. Ed. Noaillon, of 
Chenee, near Liege. 

" An Investigation to determine the effects of Steam- Jacketing 
upon the Efficiency of a Horizontal Compound Steam- 
Engine " ; by Mr. A. L. Mellanby, M.Sc, of the Municipal 
School of Technology, Manchester, 

At Half -past Twelve o'clock p.m. the Meeting was adjourned to 
the following morning. 



406 BUSINESS. June 1905. 

The Adjourned Meeting was held in the Academic Hall of the 
University, Liege, on Wednesday, 21st June 1905, at Ten o'clock 
a.m. ; Edward P. Martin, Esq., President, in the chair. 

The Discussion on Mr. Mellanby's Paper on " Steam- Jacketing '* 
was resumed and concluded. 

The following Papers were then read and discussed : — 

*' The growth of Large Gas-Engines on the Continent " ; by 
M. EoDOLPHE E. Mathot, Member, of Brussels. 

"The Strength of Columns"; by Professor W. E. Lilly, 
Member, of Trinity College, Dublin. 



The President said it was his pleasure and duty to propose the 
following Votes of Thanks : — 



*o 



To the College, consisting of Bourgmestre Kleter and the 
EcHEViNS of Liege, for their Welcome of the President, 
Council, and Members of the Institution to the Town of Liege. 

To the LiEGE Association of Engineers for their invitation to the 
Members of the Institution to hold their Summer Meeting in 
Liege, and especially to Professor Alfred Habets, President, 
M. Louis Canon-Legrand and M. Charles Thonkt, Vice- 
Presidents, and the other Members of the Reception Committee 
organised by the Liege Association of Engineers, for the 
arrangements which they have made for enabling the Members 
to visit Works ; and to the Members of the Association who 
acted as Guides during the Visits to Works, especially to 
M. F. Leclercq and M. E. Detienne for kindly accompanying 
the Members on their Visit from Spa to La Gileppe ; also to the 
Association for their great hospitality to the Members at 
the Renommee Hall. 

To the University of Liege, for the loan of the Academic Hall in 
which the Meetings have been held. 



June 1905. VOTES OF THANKS. 407 

To M. Emile Digneffe, President, and the Executive Committee of 
THE Liege Exhibition, for kindly granting free admission to 
the Exhibition to Members during their stay in Liege, and for 
the " Fete de Nuit " arranged for Members and Ladies. 

To M. L. Garnir (I'Administrateur de la Direction de I'Exploitation) 
and the other Officials op the Belgian State Railways, 
for travelling facilities afforded to the Members. 

To the John Cockerill Company, for the opportunity they have 
afiforded the Members of visiting their works at Seraing ; to 
the Fabrique Nationals d'Armes de Guerre and the 
Collieries de l'Esperance and du Hasard ; and to the 
Owners and Officials of Works in Antwerp, Brussels, 
Charleroi, Ghent, Liege, Malines, Mons, and Seraing, for 
their kindness in throwing open their Works to the Members. 

To Professor Herman Hubert, for the trouble he has taken in 
kindly arranging for the various visits to the Works in the 
neighbourhood of Liege and Seraing, and other towns, and 
for conducting the Visits to the Liege University Laboratories 
and the Ecole Professionelle de Mecanique ; to M. le Baron 
DE Laveleye, for his kind assistance ; to M. Rene d'Andrimont, 
for the admirable arrangements which his forethought and 
energy have provided during the whole Meeting ; and also to 
M. LE Baron Edgar Forgeur. 

To M. G. A. Rovers, Chief Engineer of the Municipality of 
Antwerp, and his Assistants, for kindly showing the Members 
over the Docks, the Quays, and the new Municipal Works ; and 
to M. Carlo Spruyt and M. H. Cruysmans, for their kind 
assistance in making arrangements for Members in Antwerp. 

Also especially to The Committee of the Liege Ladies, for so 
kindly arranging for the entertainment of the English Ladies 
during their stay in Belgium. 

Mr. William H. Maw, Past-President, had great pleasure in 
seconding the resolutions, and assured the Meeting that from the 
time the visit to Liege was first mooted the negotiations had been 
of the most pleasant kind. Everywhere the Institution had been 



408 VOTES OF THANKS. JuNE 1905. 

(Mr. William H. Maw.) 

met with the greatest possible assistance by all the Belgian 
engineers with whom it had been in communication, and since the 
Members had come into the city they well knew that all the 
engineers of Liege had done their utmost to show that they were 
really welcome. He was convinced that the Members had made 
numerous personal friends and that they would all look back on 
the meeting with the most pleasant memories. 

The resolutions were carried with acclamation. 



The Business Meeting terminated at Twelve o'clock noon; the 
attendance was 185 Members and 35 Visitors. During the 
meetings which followed several other Belgian engineers and a 
considerable number of Belgian and English ladies attended. 



June 190.-). 409 



SUPEKHEATEKS APPLIED TO LOCOMOTIVES 
ON THE BELGIAN STATE KAILWAYS. 



By M. J. B. FLAMME, Inspecteur-Geneeal de l'Alministeation 

DES ChEWINS de fee DE l'EtAT BeLGE, BRUSSELS. 



(Translated from the French.^ 

The Belgian State Bail ways have recently put in service a series of 
simple expansion locomotives, the boilers of which carry a pressure 
of 14 atm. (205 • 8 lbs. per sq. in.), with an inside diameter of 
1'600 m. (5 ft. 3 ins.) while that of the cylinders is 520 mm. 
(20J ins.). This class of engine givee the maximum power 
obtainable by the simple expansion of steam. In fact, every new 
enlargement of the cylinders would demand larger dimensions for 
the crank-axle and moving parts ; on the other hand, the necessity 
for clearing the loading-gauge limits the diameter of the boiler. 

Under these conditions and in view of further increasing the 
power of the engines, it becomes necessary to have recourse to some 
other system for increasing the useful work of the steam without 
enlarging the existing boilers. 

The two solutions under consideration are compound working 
and superheating of the steam. The first of these does not strictly 
come within the limits of this Paper. Arrangements for producing 
superheated steam and the results obtained with a system that has 
been in service for more than a year will now be considered. 



410 SUPERHEATERS IN LOCOMOTIVES. June 1905. 

Schmidt Superheater for Simple Expansion Locomotives. 

For some time the Locomotive Department had their attention 
drawn to the favourable results obtained by using superheated steam 
in industrial stationary engines. By superheating, the theoretical 
cycle is improved, and the pressure is maintained. The volume of 
steam is augmented proportionately to the rise of temperature 
diminishing, however, its density. In other words, when the degree 
of superheat is sufficient to prevent the loss due to condensation 
in the cylinders, then the surplus heat contained in superheated 
steam is sufficient to reheat the walls of the cylinders, maintaining 
the temperature necessary to get rid of the condensation and the 
loss of work during expansion. These trials have brought to light 
a valuable property of superheated steam. It was recognised as a 
bad conductor of heat, contrary to that which obtains when steam 
is in the saturated state. 

These numerous advantages, tested by many trials undertaken by 
most competent engineers, are specially valuable to the locomotive 
engine. The employment of a practical superheater augments the 
power of the boiler, and the utilization of superheated steam is 
most economical. This is noticeable in hauling heavy goods 
trains on sections of the Hue having steep gradients ; for it is 
then indispensable to reduce to the minimum the consumption of 
water and steam. For the suburban trains having frequent stoppages 
superheat is again highly recommended, because it reduces the 
condensation necessitated by the frequent stops. High speed is also 
favourable to the employment of higher superheated steam, the great 
fluidity of which, as well as its dryness, permits running with early 
cut-offs, thus helping the boiler just at the time when it is most 
hard pressed. 

On the other hand, the passage of saturated steam through 
the pipes (and steam ports) is more difficult, and entails inevitably 
an increase of condensation. Having in mind these various 
theoretical and practical considerations, the Administration of 
the Belgian States recognized the great utility of pushing on their 
investigations in this direction. 



June 1905. SUPERHEATERS IN LOCOMOTIVES. 411 

It was in 1900 that the Administration of the State Eailways 
opened negotiations with Herr Schmidt, the German expert, who at 
that period had already introduced some locomotives with steam 
superheaters formed principally of a series of [rings placed in the 
smoke-box. 

This last plan, described in most of the technical newspapers, 
and applied to a Prussian State locomotive shown in Paris in 
1900, adapted itself without difficulties to outside-cylinder engines. 

It is not quite the same for inside-cylinder engines which, 
as in England, are generally used in Belgium. In this case it 
becomes impossible to clear from the bottom of the smoke-box the 
cinders brought by the large flame-tube placed at the base of the barrel. 

On the other hand, a superheater, placed in the barrel of 
the boiler and described later. Fig. 5 (page 417), offers some real 
advantages. It is lighter, less cumbersome, easier to clean and 
maintain, and its introduction does not necessitate any important 
modifications in the smoke-box. Consequently it was this kind of 
apparatus that the Locomotive Department adopted in a new type 
of powerful locomotive then being built in the Cockerill Works at 
Seraing. 

At the same time another important question presented itself. 
Was it absolutely necessary to superheat the steam to a temperature 
reaching 300° to 350° C. (572° to 662° F.)? It is evident that the 
more the steam is superheated, the more necessary it becomes to give 
attention to the oiling of the piston-valves and cylinders and to the 
construction of the stuffing-box. With a view to getting a clear idea 
of the actual amount of superheat, some trials were made with 
a superheater of small surface installed in the barrel of one of the 
locomotives, type 35, which will be described later. After several 
months of experiments it has been recognised that the utilization 
of steam slightly superheated does not offer any appreciable economy 
of fuel or increase of power. 

On the other hand, with the Schmidt apparatus placed on a 
locomotive, type 35, Fig. 1 (page 412), and Plates 15 to 17, and 
provided with steam with a temperature varying between 300° and 
350° 0.(572° to 662° F.), some favourable results have been obtained. 



412 



SUPERHEATERS IN LOCOMOTIVES. 



June 1905. 



The locomotives compared, one using saturated steam and tlie 
other superheated steam, are both of type 35, with six coupled- 
wheels of 1*600 m. (5 feet 3 inches) with bogie in front. The 
boiler has a round-topped fire-box, the roof of the furnace being 
connected to the arch by vertical stays. The fire-box, of a medium 
depth, burns coal with briquettes varying in quantity with the weight 
of the train. The inside cylinders are made with piston slide- 
valves placed above, steam being admitted in the middle of 

Fig. 1. — Six- Wheels- Coupled Locomotive. Belgian State Railways. 




the valve. This arrangement, with the Stephenson valve -gear, 
involves the employment of a rocking-shaft, which reverses the 
position of the valves compared with those having the exhaust-port 
in the middle of the piston-valves. 

The six coupled- wheels and the bogie are fitted with compressed- 
air brakes. The engine is illustrated in Fig. 1 and Plates 15 to 17. 

The principal dimensions are given in the following Table : — 

Cylinders : — 

. 520 mm. (20J ins.) 

. 660 mm. (26 ins.) 

. 14 atm. (205-8 lbs. per sq. in.) 

. l-600m. (5 ft. 3 ins.) 

. 2-6o0m. (8 ft. Sfgins.) 



Diameter 

Stroke .... 
Working Pressure . ;, 
Diameter of driving wheels . 
Height of centre of boiler above rail 
Tubes:— 

Length ..... 

Exterior Diuinotrr 

Number ..... 



4-130 m. (13 ft. 6^ ins.) 
50 mm. (1^1 ins.) 
271. 



JrxE 1005. 



SUPERHEATERS IN LOCOMOTIVES. 



413 



Heating surface : — 

Interior of tubes 

Fire-box . 

Total 
Grate area . 

Weight in running order 

1st Axle 

2nd „ 

3rd „ 

4tli „ 

5tli „ 
Total weight 
Adhesion weight 



158-25 m.2 (1703 sq. ft.) 

14-90 m.2(i6osq. ft.; 
173-15 m.2 (1863 sq. ft.) 
2-81 m.2 (30J sq. ft.) 



Tractive eflfort 



D 



9710 kg. (9-5 
9740 kg. (9-5 
18215 kg. (17 
17850 kg. (17 
17500 kg. (17 
72965 kg. (71 
53565 kg. (52 



tons), 
tons). 
9 tons). 
'6 tons). 
2 tons). 
'8 tons). 
7 tons). 



16128 kg. (15-8 tons). 



The engine provided with the Schmidt superheater has less 
heating surface than the above, owing to the substitution of 21 tubes of 
118 mm. (4|- inches) diameter for 103 tubes of 50 mm. (ii|^ inches). 
For this locomotive the internal heating surface in the tubes is 
98-10 m.'^ (15O56 square feet), and the total heating surface is 
130*056 m? (1,400 square feet). 

The exterior superheating surface is equal to 27* 15 m.^ (292 sq. ft.). 

The superheater proper is illustrated in Plates 15 to 17, and 
consists essentially of two parts : — 

(1) A series of iron tubes of 118 mm. (4! inches) external 
diameter, occupying the upper part of the nest of tubes and offering 
like them a passage for flame and hot gases ; and 

(2) Some U shaped tubes grouped in pairs among the flame-tubes 
and used for the circulation of the superheated steam. 

A steam collector in several divisions is placed on the top of the 
smoke-box. Some supplementary parts complete the system. 

There must also be a diaphragm to close the flame tubes when 
steam does not circulate in the superheating tubes. This diaphragm 
is handled by the aid of a lever near the engine driver. 

A mercury thermometer shows the temperature of the superheated 
steam at the entrance of the steam-pipe. The degree of superheat is 
read on a graduated quadrant placed in the cab. 



414 SUPERHEATERS IN LOCOMOTIVES. June 1905. 

The large flame-tubes, which are of solid-drawn iron, are screwed 
into the fire-box tube-plate and expanded in the smoke-box tube- 
plate. The superheating tubes, also of solid-drawn iron, are 
protected against the action of the flame at the fire end by cast-steel 
caps. 

In the smoke-box these tubes are expanded into flanged bushes 
fixed by bolts. The tightness is assured by means of asbestos 
joints. 

Copper, bronze, and brass are usually excluded from all parts 
that come in contact with the superheated steam. For this reason 
the steam-pipes are of iron, and the joints between these pipes and 
the cylinders are formed with cast-iron flanges. 

The metallic packings of the piston-rods and valve-spindles are 
composed of cast rings and white metal, the contact of which on 
the rod is obtained by a spring permitting small side-movements of 
the rod. 

The slide-valves are cylindrical with steam admission in the 
middle of the valve, which reduces the packing to simple bronze 
rings with lubricating grooves. The slack between each valve and 
the cylindrical chamber against which it rubs is closed by means of 
three cast-iron rings of suitable section, the steam pressing on the 
interior of the principal segment. 

The oiling of the cylinders and valves is done by a lubricator in 
six sections. The lubricant used is a mineral oil with a high 
flash-point. 

The trials of these two locomotives took place with goods 
trains of accelerated speed and local passenger trains running on 
the Luxemburg line, the extremely undulating profile of which 
contains many gradients of 1 in 62. 

Each locomotive worked twenty-four goods trains weighing 250 t. 
(246 tons) and twelve passenger trains weighing an average of 150 t. 
(147*6 tons). The total journey made by each engine amounted 
to 11,500 kilometres (7,146 miles). The saving of coal per 
train-kilometre in favour of the superheated-steam engine was found 
to be 13*33 per cent., and the water consumption was reduced 



June 1905. SUPERHEATERS IN LOCOMOTIVES. 415 

18 per cent. On the other hand, the expenses of lubricating 
increased in a fixed proportion. 

After four months of trials on the Luxemburg line, more precise 
experiments were organised with the through passenger trains on the 
Brussels and Charleroi line, which has a series of gradients of 
1 in 77. For ten days, during which the climatic conditions 
remained invariable, these two locomotives hauled alternately the same 
train of 250 t. (246 tons). The saving in favour of the superheated- 
steam locomotive amounted to 12*5 per cent, for fuel and 16*5 per 
cent, for water. Moreover the speed raised at the top of the incline 
showed an average increase of 9 • 5 per cent., all the conditions being 
exactly the same. 

As regards maintenance the superheated-steam locomotive, type 
35, has not required special attention during its 1 J years' service. 

These early favourable results have led to the Belgian State 
Railways venturing on the application of superheat to locomotives 
on a larger scale. With this in view twenty-five locomotives, 
comprising five difi'erent types, all provided with the Schmidt 
superheater described above, are actually in course of construction 
or are about to be put to work. 

Amongst these last are a certain number of locomotives of 
type 35, which have fully confirmed the favourable results obtained 
by the first engine of this kind. 

Among the number of services actually and successfully run by 
these engines is to be particularly noted the hauling, from Brussels 
to frontier, of express trains going to Paris. These trains, whose 
tare weight of vehicles exceeds 340 tonnes (334I tons), surmount 
the 17 kilometres (10-56 miles) between Mons and the frontier 
in 17 minutes, against a continuous up-grade with inclines varyincr 
from 1 in 125 to 1 in 55. 

CocKERiLL Superheater for Compound Locomotives. 

It. is seen from the preceding that it is now known that 

superheated steam, as applied to locomotives, is susceptible of giving 

remarkable results which come within the range of practice. The 

State Railways have decided to persevere with their experiments in 



416 



SUPERHEATERS IN LOCOMOTIVES. 



June 1905. 



combining superlieat with compounding, because they perceive that 
there is a most interesting question to elucidate. 

Is it more economical to divide the superheater into two parts in 
such a manner as to raise the temperature at the entrance to both 
the H.P. and the L.P. cylinders, or, on the other hand, to devote the 
whole power of the apparatus to superheating the steam before it 
enters the L.P. cylinders ? The Cockerill Company, after numerous 
investigations, have just completed a superheater which will enable 
them to answer this question. 

This entirely new system is being continually tested on a series 
of compound engines, with four cylinders, and six coupled-wheels 
of 1*80 m. (5 feet 10 inches) diameter with a bogie. This 
locomotive, called 19&^s, has a boiler having an interior diameter 
of 1*65 m. (5 feet 5 inches) diameter, and carries a pressure of 
15*5 atm. (227 lbs. per square inch). The H.P. cylinders are inside 
and connected to the leading coupled-axle ; the L.P. cylinders are 
outside and drive the second axle. The four cylinders are placed on 
the transverse axis of the bogie. The two valve motions of the 
Walschaerts type are outside. They present several peculiarities 
due to the employment of cylindrical valves, with the steam 
introduced in the middle. The leading dimensions of the engine, 
type l%isy are shown in the Table below. 



Diameter H.P. cylinders . 

„ L.P. „ 
Stroke of piston 
Initial pressure 
Diameter of driving wheels 
Height, rail to centre of boiler 
Tubes :— 
Length .... 

Number and exterior diameter < 

Heating surface : — 
Interior of tubes . 
Fire-box .... 

Total 

Grate area .... 



360 mm. (i4j^6 i^^^O 

620 mm, (24^2 ins.) 

680 mm. (26|f ins.) 

15*5 atm. (227 lbs. per sq. in.) 

1-80 m. (5 ft. II ins.) 

2-80 m. (9 ft. 2 1 ins.) 

4-0 m. (13 ft. I J in.) 
30 of 107 mm. (4-/2 i°s-) 
219 of 50 mm. (i§| in.) 

157-62 m.2 (1697 sq.ft.) 
18-35 m.2 (198 sq.ft.) 

175-97 m.2 (1894-1 sq. ft.) 
3-01 m.2 (32 sq.ft.) 



June 1905. 



SUPERHEATERS IN LOCOMOTIVES. 



417 




2 H 



418 SUPERHEATERS IN LOCOMOTIVES. JuNE 1905. 

The apparatus for superheating the steam may be used in two 
ways. The steam may be heated near the entrances to the H.P. 
cylinders, and afterwards near to the L.P. cylinders, or at the 
entrances of the L.P. alone. The superheater shown in Fig. 5 
(page 417) indicates the general arrangement, comprising two series 
of large flame-tubes containing the circulating pipes intended to 
superheat the steam. 

The role of the compartments C and H, placed inside the 
barrel, and of the collectors J and D, installed in the smoke-box, will 
be dealt with later on in connection with the explanation of the 
working of the apparatus. 

In B there is a valve with three pistons intended to divert the 
steam coming from the regulator towards the compartment C, or into 
the tube L, according as it is required to operate the superheat to 
H.P. and L.P. or to L.P. only. The movements of the valve B are 
automatically repeated, due to the presence in the tube L of an 
identically similar valve located within B^. The arrangements of the 
different pipes are made clear by following the course of the steam as 
explained below. 

First Case. — Superheat at the entrance to H.P. and L.P. Cylinders. 
— The steam on leaving the regulator A makes its way, after passing 
B, towards the compartment C ; from there it traverses the left set of 
superheater tubes and enters the collector D, whence it goes to the 
H.P. cylinders by passing through the valve B^ and pipes E. 

The superheated steam, after doing work in the H.P. cylinders, goes 
out by the exhaust pipe, traverses the valve B^, after that the pipe G, 
which is lodged in the interior of the barrel to enable it to enter the 
compartment H. From there the steam goes into the superheating 
tubes (the right set), and is conducted by the pipes K leading to 
the L.P. cylinders. 

Second Case. — Superheat at the entrance of the L.P. Cylinder. — 
The valve B is placed by the driver in a position that diverts the 
direction of the steam, directly from the regulator into the pipe L ; 
from there it goes to the H,P. cylinders after having passed through 



June 1905. SUPERHEATERS IN LOCOMOTIVES. 41 9 

the valve B^ and the delivery pipes E. On leaving the H.P. cylinders 
the steam traverses the pipes F, the valve B^, and enters into the 
collector D. From the front it passes back through the left set of 
superheater tubes and arrives at the compartment C. From this it 
passes through the valve B into the compartment H, and traverses 
through the right group of superheater tubes, whence it goes 
into the collector J, and from there by the delivery pipes K into 
the L.P. cylinder. 

A locomotive of type Idhis, showing this pattern of superheater, 
is exhibited in the Liege Exhibition. Trials are about to be 
continued with a second identically similar engine, to determine 
which is the more advantageous mode of working to adopt for the 
new superheater. 

It is manifest that if the superheat is required at the entrance of 
the L.P. cylinders only, it will be possible to dispense with a certain 
number of parts of the superheater, and by that means remedy the 
obstruction in the smoke-box. 

The Paper is illustrated by Plates 15 to 17 and 2 Figs, in the 
letterpress, and is accompanied by an Appendix. 



APPENDIX. 

A diagram (Fig. 6, pages 420 and 421) shows the working of 
engine 3003, fitted with Schmidt superheater, on 31st May 
1905, hauling a train of 327J tonnes (322*3 tons), consisting of 
11 carriages with three axles, one bogie carriage and two brake- 
vans of 3 axles, from Brussels to Ans, near Liege. The dotted 
line shows the varying temperature of the superheated steam. A 
full line shows the speed of the train, the grades and curves being 
given above these curves. 

(^Continued on page 422.) 2 n 2 



420 



SUPERHEATERS IN LOCOMOTIVES. 



June 1905. 



(Continued on opposite page.) 






I 



o 
o 

V. . 

2 <» 

"■«* fie 



e 






CD 






o 



8 



*<s> 




BRUSSELS v ^§ 

1^^ 



b<ooooopoo 
^ Speed 



< 

t 



June 1905. 



SUPERHEATERS IN LOCOMOTIVES. 



421 



{Concluded from opposite page.) 



Teiiiperaliire Cen/igrcufe 



LIEGE 

Haut-Pre 
1148 Ft. 




u2 



o!^ 



Tlrlemont 





- - 














— 


"\ 


--- 




--- 


■ 


1 
1 
1 




■ 


1 

V 
\ 














\ 
















\ 


1 








■ 




■ \ 


i 








■ 






\ 












\ 
\ 
\ 

1 
1 


\ 


\ 






. 




1 






* 










1 












- 


1 
1 


1 
1 

I - - 
/ 

/ 


- 


\ 


— 


— 


■ 


- 




-7 


\ 





^^ 








/ 




% 








\^ 


■ 


- 


- 


^ 


7 


*rrz 


rr. 


\ 


1 

1 


1 
1 


■( 


/- 






■ 



?5 






•^ 



«^.'5« 



^'^ 
^ 



a:i 



422 



SUPERHEATERS IN LOCOMOTIVES. 



Junk 1905. 



The principal dimensions of the locomotive are given in the 
following Table : — 



Engine : 

Diameter of cylinders 

Stroke of piston . 

Number of coupled- wheels . 

Diameter of ,, ,, 

Number of carrying- wheels . 

Diameter of ,, ,, 

Boiler : 

Length of barrel 

Diameter (greatest) . 

Thickness of iron plates 
Fire-box shell-plate thickness . 

,, copper plate „ 
Tubes : 

Length .... 

Number and exterior diameter 

Thickness .... 

Heating surface : 

Tubes (inside) . 

Fire-box .... 

Total .... 

Superheating surface (outside) 
Grate area .... 
Working pressure 

Tractive effort . . .2x0' 



435 mm. (17^ ins.) 

610 mm. (24 ins.) 

6 

1-98 m. (6 ft. 6 ins.) 

4 

0-900 m. (2 ft. iijlins.) 

4-000 m. (13 ft. i^ins.) 
1-650 m. (5 ft. 5 ins.) 
18 mm. (§1 in.) 
18 to 30 mm. (|| to i^^ in.) 
18 to 27 mm. (|| to i^V ^^') 



4-]02m. (13 ft. sJins.) 
25 of 127 mm. (5 ins.) 
180 of 50 mm. (ijf ins.) 
3-5 mm. (^^5 in.) 
2 - 5 mm. (/j in.) 

138-87 m.2 (1495 sq.ft.) 
16-88 m.2 181 -7 sq.ft.) 

155-75 m.2 (1677 sq.ft.) 

38-95 m.2 (419-26 sq.ft.) 
3-01 m.2 (32-4 sq.ft.) 
14 atmospheres (205 • 8 lbs. per sq. in.) 

65 X ^-^ kg. = 10930 kg. (10-78 tons) 



Junk 1905. SUPERHEATERS IN LOCOMOTIVES. 423 

Discussion. 

Mr. John F. Eobinson, Member of Council, was sorry the author 
was unable to be present at the meeting, because he would have 
been able to add some further information to that contained in the 
Paper, which would have made his communication even more 
interesting than in its printed form. Nevertheless the description of 
the Schmidt superheater and the forecast of what was going to be 
in the future were very interesting. The Schmidt superheater 
had been tried on a great many railways in different parts of the 
world, with somewhat varying success. He believed the Canadian 
Pacific Railway had had a great number of their engines fitted with 
it, and the officials were very much pleased with the result. On 
the other hand, the Cape railways had had two engines fitted 
with a somewhat similar contrivance to the Schmidt apparatus which 
had not proved satisfactory, the reason being, he thought, that the 
superheat arrangement was placed in tubes in the lower part of the 
valve. He believed there was only one large tube about 8 or 10 
inches in diameter, containing a series of small U-shaped tubes, 
through which the steam was carried backward and forward inside 
the tube. For various reasons, partly due probably to the 
construction of the fire-box of the engine running on a narrow- 
gauge railway, he fancied the amount of gases which came through 
the lower tube was not very great, and therefore the effect of 
superheating was not obtained with that arrangement. On the 
Canadian Pacific Railway the tubes had been adopted arranged as 
shown in Plates 15 to 17. There were two rows of 4-inch tubes, or 
thereabouts, on the upper part of the boiler — three rows were shown 
in the diagram — and in that part of the boiler the superheating 
tubes were located. With that arrangement he believed they 
obtained a far better result, because the gases had a tendency to go 
along the higher rows of tubes. 

With regard to the experiments carried out on the two different 
engines, the loads did not seem to be very great, and it would be 
very interesting to know what the actual consumption of coal was 
with the engines. The figures only gave the percentage in favour 



424 SUPERHEATERS IN LOCOMOTIVES. June 1905. 

(Mr. John F. Robinson.) 

of the superheated steam-engine. Of course, in comparing the 
consumption of coal in different countries, the variation in the 
quality of the coal had to be borne in mind, but he thought a figure 
or two would make the subject more interesting and valuable. 
That remark applied not only to the first experiment mentioned, 
that on the Luxemburg line, but also to those on the Brussels and 
Charleroi line, where there was only 13 per cent, in favour of the 
superheated steam. With such a large engine, 246 tons could not 
be said to be a very heavy load. 

It was interesting to note that the superheated steam locomotive 
had done exceedingly well in not requiring attention during a service 
of a year and a half — a fact which reflected very great credit on 
the builders. He noticed in the case of the engines referred to 
(page 415) that among the number of services actually and successfully 
run by the engines in hauling express trains from Brussels to the 
frontier and on to Paris, was one of 334 tons haulage over continuous 
gradients varying from 1 in 125 to 1 in 55. That was a very much 
better performance than those which had been referred to in the 
previous paragraphs of the Paper. He thought the experiment it 
was proposed to make with the superheating, either at the entrance 
of the low-pressure cylinder or divided between the low-pressure 
and the high-pressure, would be very interesting. In all probability 
it would be found desirable to do some superheating before admitting 
the steam into the high-pressure cylinders. 

He noticed (page 416) that a new type of engine was spoken of, 
an engine with four cylinders and coupled six-wheels with a bogie 
and with the high-pressure cylinders inside and the low-pressure 
cylinders outside. That was exactly opposite to the De Glehn 
system, of which the members had heard so much from M. Sauvage 
a year or two ago.* The results in France with the De Glehn 
engines were exceedingly good, and it would be interesting to see 
how the two types of engine compared after working. The drawback 
seemed to be in having the large cylinders outside away from the 
centre line of the engine and the small cylinders on the inside 

* Proceedings, 1904, Part 2, page 327. 



June 1905. SUPERHEATERS IN LOCOMOTIVES. 425 

nearer the centre line. He did not tbink the balancing of this 
new type of engine could be quite as good as that of the De Glehn 
engine. Superheating, speaking generally, had not been much 
utilized in England, partly because it was of more benefit to the 
compound engine than to the simple engine. Compound engines, 
with few exceptions, were not much in favour in Great Britain, 
but if superheating were introduced, he thought it would be probably 
found that the compound engine would very largely result from it. 

Mr. Mark Robinson, Member of Council, suggested it would add 
to the value of the experiments in prospect if they were arranged so 
that the effect of superheating the steam before entering the high- 
pressure cylinder, without reheating between the cylinders, could 
be tried for the sake of direct comparison with the methods of 
superheating before entering both cylinders, or before entering 
the low-pressure cylinder only. 

The President regretted very much the absence of the author, 
who was detained in Brussels in connection with his duties as Chief 
Engineer of the Belgian State Railways. Probably some of the 
members, after carefully studying the Paper, might have views 
which they wished to express, and the Secretary would be pleased to 
receive from them in writing any remarks they had to. make. He 
was sure the members would accord the author a very hearty vote of 
thanks for his most able Paper. 



Communications, 

Mr. John Barr wrote asking for further information on the 
following points : — 

(1) How were the boiler tubes which contained the superheating 
ones cleaned; also how were the superheater tubes themselves 
cleaned ? 

(2) Was the regulation of the superheat left entirely to the 
driver? Would any damage be done if this regulation were 
neglected ? 



426 SUPERHEATERS IN LOCOMOTIVES. June 1905. 

(Mr. John Barr.) 

(3) Were the bushes shown in the glands and necks of the 
stuffing-boxes of gun-metal or cast-iron, as bronze was said to be 
" usually excluded " ? 

(4) Metallic packing. Were the cast rings mentioned on page 414 
of cast iron ? 

(5) Why were bronze rings put in the piston-valves when this 
metal was said to be excluded ? 

(6) Was the extra cost of lubrication not such as would reduce 
the economy of fuel appreciably ? 

(7) Were cylindrical or piston valves a sine qua non for 
superheated steam ? 

M. Flamme wrote, in reply to Mr. John F. Robinson (page 423), 
that at the end of 1904 the Canadian Pacific Railway possessed 
22 locomotives fitted with the Schmidt superheater placed in the 
boiler tubes, and on the Belgian State Railways there were now 31 
locomotives with this type of superheater. The consumption of 
coal for each locomotive was not given, because it was a matter 
of making comparative experiments. It depended chiefly on the 
quality of the fuel used — a fact which was of no importance in the 
present enquiry, which dealt exclusively with the comparison of 
two locomotives of the same type, one with saturated and the other 
with superheated steam. 

On the Brussels-Charleroi line (which has gradients of 1 in 
77), the train-load during the trials was, for running purposes, 
fixed at 246 tonnes (246 tons). The gradients had to be taken with 
the greatest possible speed, and the time-table of the train reduced 
accordingly. The load of this train, drawn by the same locomotive, 
type 35, with superheater, has often exceeded 250 tonnes and at 
times reached 350 tonnes, on which occasions the whole power of 
the locomotive was employed in keeping the train to its scheduled 
time. 

The new compound locomotives of the Belgian State Railways 
differed from those of the De Glehn system in the position of the 
cylinders. The low-pressure cylinders were placed outside instead 
of inside, and the four cylinders were also situated in line on the 



Junk 1905. SUPERHEATERS IN LOCOMOTIVES. 427 

transverse axis of the bogie. In this way the connecting of the 
outside cylinders by a cross-stay (which would have made the 
examination of the top of the connecting-rod of the inside cylinders 
inconvenient) was obviated. 

In reply to Mr. Barr's questions (page 425), the boiler tubes 
containing the superheater tubes were cleaned by the stoker by means 
of a jet of compressed air. The superheater tubes did not become 
greasy, because the steam was continually passing through them at a 
high velocity. The engine driver could regulate the admission of 
the superheated steam by means of a valve in the smoke-box closing 
the large tubes. If he neglected to work this valve, the regulator 
being closed, no immediate damage would ensue. If this omission 
were repeated very often, the superheater tubes would get out of 
order prematurely. 

The metal used for the bushes in the glands and stuffing-boxes 
was cast-iron and not bronze, and the packing-rings of the piston- 
rods were also cast-iron. The piston slide-valves were provided 
with cast-iron rings. Bronze, however, could be used for the stuffing- 
boxes of the cylindrical slide-valves, because these, as already stated 
in the Paper, admitted the steam in the middle while the exhaust 
was at the ends ; it followed, therefore, that the steam in contact 
with the stuffing-boxes was very little superheated. The extra 
lubrication expenses accompanying the introduction of the 
superheater were very small compared with the economy in fuel. 

"With regard to the type of valve used, it had been found that 
cylindrical slide-valves were especially suitable for use with 
superheated steam. It might be mentioned that the Belgian 
State Eailways had applied these valves to numerous locomotives in 
which the steam was not superheated, and the results obtained were 
very satisfactory. It was absolutely necessary that the slide-valves 
should be of cast-iron ; and that recourse should be had to a process 
of balancing, to be sure of having sufficient lubrication on the flat 
surfaces. The author was not aware, however, of an application of 
superheat to a locomotive with slide-valves. 



June 1905. 429 



ELECTRIC WINDING-MACHINES. 



By Professor PAUL HABETS, op Brussels. 



(Translated from the French.) 

The ever-increasing application of electricity to mining 
macliinery, and the economy resulting from centralising the 
generation of energy required for working large undertakiags, 
has not been kept back for want of finding means for working 
winding-machinery electrically. When it is a question of 
underground mechanical winding, the electric transmission offers 
enormous advantages over all other systems (steam, compressed- 
air) with regard to economy and facility of installation. But the 
exaggerated fear of introducing electric plant into explosive 
mines, the cheapness of machines by compressed air and also 
the possible utilization of existing compressed-air plants, have 
retarded the great development of electric winding in underground 
workings. Even in the case of the main winding-machines, which 
can be placed close to the steam-engines, electric driving has been 
introduced during the last years with the object of reducing the 
consumption of steam. This economy is, however, doubtful. 
The losses due to the transformation of electric energy may, 
in fact, equal the unfavourable conditions under which a steam 



430 ELECTRIC WINDING-MACHINES. JUNB 1905. 

winding-engine is working if the latter be an economical 
condensing engine.* The economy is certain, however, when the 
winding-machine must be placed further away from the steam- 
engine or when the use of electric transmission permits of doing 
away with boilers. 

In designing a steam winding-engine, it was generally 
considered sufficient to determine the resultant statical moment 
of the rope-roll. The use of electric driving has necessitated a 
study of the dynamic conditions of the problem, which is not 
without importance also in designing steam winding-engines, if 
it be desired to work the engine under the most economical 
conditions. The author brought to notice, at the Exhibition 
of electric winding-engines at Diisseldorf 1902, f the study of 
the dynamic conditions of winding for machines with ropes and 
perfectly balanced ; winding-machines with cylindrical drums, and 
with counterbalancing ropes ; and machines with Koepe's pulley. 
The latter was at that time the only machine to which direct 
driving by electric motors had been applied. 

At present the application has been extended to machines with 
variable rope-roll, notably to the machine installed at the coal- 
mines of Grand Hornu.J It seems therefore desirable to 
investigate the question more closely. 

Dynamic Investigation of Winding. 

Statical Moments. — The law governing the relation between 
the statical moments and the number of revolutions made by the 
shaft which carries the winding gear of the ropes may be 



* See R. A. Henry, "Eltude theorique et experimentale de la Machine 
d'Extraction " ; ' Revue Universelle des Mines,' 4th series, 1903, vol. ii, and 1904, 
vol. vii. 

t See Paul Habets, "Exposition de Diisseldorf, 1902, des Machines 
d'extraction electriques " ; 'Revue Universelle des Mines,' 4th series, 1903, 
vol. i, and 1904, vol. vi. 

X See E. Troussart, " L'Installation de Transport d'Energie electrique aux 
Usines et Mines de Houilles de Grand Hornu " ; ' Revue Universelle des Mines 
4th series, 1904, vol. vii. 



Junk 1905. 



ELECTRIC WINDING-MACHINES. 



431 



determined by calculation or by a graphical method shown by 
M. H. Dechamps.* As an example, the author has selected the 
electric winding-engine installed at the coal mines of Grand 
Hornu. The following data have been furnished by M. E. 
Troussart: — f Depth, 1,000 m. (3,300 ft.); useful load, 6 wagons, 
2,600 kgs. (2J tons); dead load, 6 wagons, 1,260 kgs. (ij tons); 
dead load, one cage, 2,000 kgs. (2 tons). The flat aloes rope is 
taper and of the following dimensions : — 



TABLE 1. 



i Length. 


Thickness. 


Weight per 


M. 


Feet. 


Mm. 


Inches. 


Metre. 


Foot. 


1 










Kgs. 


Lbs. 


to 120 


to 394 


50 to 48 


1*97 to 1-89 


14-65 


9-84 


120 „ 220 


394 „ 722 


48 , 


, 46 


1-89 „ i-8i 


13-75 


9-24 


220 „ 320 


722 „ 1050 


46 , 


, 44 


i-8i „ 1-73 


12-50 


8-40 


320 „ 420 


1050 „ 1378 


44 , 


, 42 


1*73 >, 1*65 


10-90 


7-32 


420 „ 520 


1378 „ 1706 


42 , 


, 39-5 


1-65 „ 1-56 


9-25 


6-21 


520 „ 620 


1706 ,, 2034 


39-5 , 


, 36-75 


1-56 „ r-45 


8-90 


5:97 


620 „ 720 


2034 „ 2362 


36-75 , 


, 35-25 


1*45 „ 1-39 


8-40 


5-64 


720 „ 820 


2362 „ 269c 


35-25 , 


, 33 


1-39 „ 1-30 


8-25 


5'54 


820 „ 920 


2690 „ 3018 


33 , 


, 33 


1-30 „ 1-30 


8-12 


5*45 


920 „ 1020 


3018 „ 3347 


33 , 


, 32-9 


1-30 „ 1-29 


7-68 


5-15 


1020 „ 1120 


3347 " 3675 


32-9 , 


, 32-7 


1*29 „ 1-28 


7-10 


4-77 



The diameter of the bare drum is 1-40 m. (4 ft. 7I ins.), the 
initial radius is 1-20 m. (3 ft. 11^ ins.), and the final radius 



* See H. Dechamps, "Applicatioa de la Me'thode graphique a I'e'tude de 
I'Equilibre des Cables d'Extraction " ; * Revue Universelle des Mines,' 3rd 
series, 1902, vol. Iviii. 

t See E. TrouBsart, loc. cit. 



432 ELECTRIC WINDING-MACHINES. June 1906. 

3*73 m. (i2 ft. 2f ins.). The number of revolutions for one 
haul is 64. To draw 56 tonnes (55 tons) per hour it will be 
required to make 25 hauls per hour ; it takes 144 seconds to 
make one ascension, and reckoning 38 seconds for manipulating 
the machine, there remain 106 seconds for the actual ascension. 
To prevent the speed from exceeding 15 m. (49 ft. 2 J ins.) 
per second in the pits, a constant angular acceleration is allowed 
for 15 seconds to reach the rated speed of 42 revolutions per 
minute in 77 seconds. The stopping is effected in 14 seconds by 
a constant retardation. The moment of inertia of the two drums, 
including the brake-drum, is 5,606*5 kg.-m.^ (133,000 Ib.-feet^). 
The moment of inertia of the motor is 2,548*4 kg.-m.^ (60,460 
Ib.-feet^), and that of a head-gear 407*7 kg.-m.^ (9?^75 Ib.-feet^). 
By the aid of these data the conditions for manipulating the 
winding-machine can be found. 

The curves (Figs. 1 to 5, pages 433 to 437) may now be drawn 
according to M. Dechamp's method for flat tapered ropes winding 
on drums : — 

(1.) Curve AqAj^q representing the weight of the rope per 
metre run as a function of the depth. 

(2.) BqB^Biq representing the weight of the unrolled part of 
the rope as a function of the depth. 

(3.) BqB'^B^q representing the weight of the unrolled part of 
the rope as a function of the number of revolutions of the 
machine. 

(4.) The curve of statical moments iCoX^Q of the ascending 
loads referred to axis CO. 

(5.) The curve of statical moments x'qx\q of the descending 
loads referred to axis HH'. 

(6.) The invert x"qx"iq of x'qx\q referred to axis CC. 

(7.) The curve XqX-^q of the resultant statical moments. 

Besistances (Figs. 1 and 4, pages 433 and 436). — Besides the 
statical moments, the motor has also to overcome the resistance 
of the air to the motion of the cages, the frictions in the pit 
and at the bearings of the head-gears and of the shaft of the 

(^Continued on page 188.) 



June 1905. 



ELECTRIC WINDING-MACHINES. 



433 



Fig. I. 




35 



^0 g 

o 



25 



(M 



M a 



-10 ^ 



a 

o 

d 

.2 
m 



AbycissaB Scale. 
1 Div. = 100 m. (328 ft.) or 6-4 Revs. 



2 I 



434 



ELECTRIC WINDING-MACHINES. 



June 1905. 



Fig. 2. 




1 



June 1905. 



ELECTRIC WINDING-MACHINES. 



435 



Fig. 3. 




2 I ? 



436 



ELECTRIC WINDING-MACHINES. 



June 1905. 



Fig. 4. 







CO 



20 a 



a 



a 



e8 



Abscissae Scale. 
1 Div. = 100 m (328 n.) or 6-6 Revs. 



June 1905. 



ELECTRIC WINDING-MACHINES. 



437 



Fig. 5. 




438 ELECTRIC WINDING-MACHINES. June 1905. 

machine, and finally tlie resistance of the rope to bending. 
These various factors may be assimilated to a load suspended 
at the ascending rope and whose moment may be added to the 
statical moment. This load has been found to be 15 per cent, 
of the useful load in the trials made on the electric winding- 
machine at Preussen II.* In the absence of exact data, and also 
for the sake of simplicity, the moment of resistances may be taken 
as constant during the whole winding, and a value given equal to 
15 per cent, of the maximum statical moment of the loads. The 
value of the statical moment, including resistances, is measured by 
the ordinate of the curve ^o-^io ^7 lowering the axis of 
abscisses an amount CKq equal to 15 per cent, of the maximum 
moment C^Xp 

To the statical moments must further be added the moments 
of the accelerating forces required for the moving masses. These 
masses are fixed to the shaft of the motor and participate in its 
rotation. The moment of the accelerating forces may be deduced 
from the moment of inertia I of the moving masses referred to the 
shaft of the motor, and from the law of rotation of the shaft, as 
will be shown further on. 

For the purpose of determining the moments of inertia of the 
masses, it will be convenient to consider the following : — 

(1.) The masses of the loads suspended on the rope. 

(2.) The masses of the rope-rolls. 

(3.) The masses of the head-gears. 

(4.) The mass of the drums and pulleys. 

Moments of Inertia of the Suspended Loads (Figs. 1 and 4, 
pages 433 and 436). — The suspended loads P may be considered as 
concentrated at the end of the instantaneous radius p of the 
rope-roll, which is, therefore, also their radius of gyration. The 
moment of inertia is then — 

9 ^g' 

* See P. Habets, " Les Machines d'Extraction electrique " (2* article) ; ' Kevue 
Universelle des Mines,' -ith scries, 1904, voJ. vi. 



June 1905. ELECTRIC WINDING-MACHINES. 430 

Pp is the statical moment of the suspended load ; it will, 
therefore, be sufficient to multiply the statical moment by ^. The 

moments of inertia may be found graphically from the values 
of the statical moments cf the ascending and descending loads. 
It may be noticed that in determining the statical moments, the 
lengths of rope between the winding-gear and the pit's mouth 
were neglected, which are in equilibrium. But as these lengths 
are 40 to 50 metres (131 to 164 feet) they cannot be neglected in 
making determination of the moment of inertia. New curves of 
moments Pp must be determined, by adding to the curve BqE^B^q 
the weight of 50 metres (164 feet) of rope ; one thus obtains B"qB'\q 
which serves to determine the curves of moments 2/o2/52/io *^^ y'oy'bV'ic' 
The latter should be inverted and referred to axis CC This 
operation may be done on the curve of moments of inertia. 

Proceeding as follows : — On the ordinate through 0, the 
ordinates of the curves 2/o2/52/io ^^^ y'oy'bV'io ^^^ projected, and 
through the projections thus obtained lines are drawn to pole D 

in case of curve 2/02/52/10 ^^^ ^^ P^^® -^ ^^ "^'^^^ ^^^® ^^ y'oy'^y'io' 
One of the moments of inertia that is easy to obtain is next 
calculated; for instance, the moment corresponding to the small 
radius of the rope-roll. Taking - r^ for the ascending load, its 

value is Ciq; a horizontal line drawn through ^o intersects the 
pole-ray Dy^ in a^ ; a vertical through a^ intersects the pole-rays 
in points a-a^ott'oa's and a\Q ; these points are then projected on 
the ordinate through C, and through the projections thus obtained 
pole-rays are drawn emanating from poles J) or J^ as the case 
may be. To obtain the moments of inertia «o*5ho ^^^ ^Vs^'io ^^ 
the ascending and descending loads the pole-rays are produced 
until they intersect the corresponding radius of the rope-roll. 
In fact, Aa, = Ci, = Cy, ' ; Aa, = Vy,^; Vi, = ^«5 | = ^2/5-^, 
and so on. 

Curve ^'o^'lo of the moments of inertia of the descending loads 
referred to axis HE' must be inverted and referred to the same 
axis, CC\ as the moments of inertia of the ascending loads. Curve 
*"o*'"io ^^ *^^s obtained. 



440 ELECTRIC WINDING-MACHINES. JuNE 1905. 

The moments of inertia of the ascending and descending loads 
are now added, by adding the ordinates of their respective curves. 
Curve IqI^Iiq is thus produced. 

Moment of Inertia of the Mope-JRoll (Figs. 1 and 4, pages 433 
and 436). — The weight F of the rope-roll may be determined by 
taking the differences between the weight of rope OB"q and the 
weight of the unrolled rope represented by curve B"qB"^B'\q, 
The moment of inertia depends on the radius of gyration. 

In the case of winding on constant radius, the radius of 
gyration is the same as that of the rope-roll, and it is not necessary 
to consider the different parts of the rope separately. The useful 

load of the cages and the ropes is simply multiplied by - . 

In the case of conical and spiral drums, the variation of the 
radius of gyration must be taken into account. When the rope is 
not taper, the radius of gyration of the rope-roll (radius p) is 



r- + p2 



,* where r is the smallest radius of the rope-roll. If the rope 



2 

is taper, the error committed by not taking into account the 

decreasing thickness of the rope is negligible. The moment of 
inertia is determined by multiplying *^^^ by the mass — of the 

rope-roll. The curves CqC'^X and OG\C'iq referred to axis OX 
represent the moments of inertia of the two ropes. 

The two curves are identical, but inverted with regard to each 
other ; the sum of their ordinates will give the total moment of 
inertia (CqC^C\q) of the rope-rolls. 

The latter moment should be added to the moments of inertia 
{IqI^Iiq) of the loads, thus determining the total moments FqI^Fiq, 
Fig. 1 (page 433). 

Moment of Inertia of the Head-Gears. — The moment of inertia 
(i„) of the head-gears may be determined from the drawings. 

2 

i"„ must, however, be multiplied by ^— , p and Bm being the radius 

* The moment of inertia of that part of the rope which is never unrolled is 
constant, and may be added to the moment of inertia of the winding-gear. 



June 1905. ELECTRIC WINDING-MACHINES. 441 

of the rope-roll and head-wheel-rim respectively ; the two sets of 
reduced moments of inertia are represented by the ordinates to 
curves mQmr^m^Q and m'Qm^m'^Q (referred to OX). These curves are 
identical but inverted with regard to each other ; by adding their 
ordinates, the total moments of inertia (^MqMiq) of the head-gears 
are found. The ordinates to curve I"qI"^T\q, Figs. 1 to 5 (pages 
433 to 437) are equal to the sum of the ordinates to curves MqM^q 
and TJ'.r,,. 

Moment of Inertia of the Winding-Gears and of the Motors (Figs. 1 
and 4, pages 433 and 436). — The moment of inertia of the drums, 
brake-drum and gearing may be determined by referring to the 
drawings. In the case of electric driving, the moments of the 
motors must be added. The moment of inertia of the winding- 
machinery is constant, and must be added to the moment 
determined above. This may be done by setting its value CU off 
from GO' downwards ; the ordinates of curve I"qT\T\q, referred 
to UV as axis of abscissae, will then be a measure of the total 
moment of inertia. The curve I"qT\T\q may be called the 
" Characteristic of Inertia " of the winding-engine. 

Sjpeed of Botation of Machine-Shaft. — The moments of the 
accelerating forces may be determined by the total moment of 
inertia and by the speed of rotation of the motor-shaft. 

For this purpose, curve I"qI'\I'\q must be referred to abscissae 
representing time. The abscissae of I"qI'\T\q are proportional to 
the number of revolutions made by the motor-shaft, and the total 
number of revolutions of this shaft during each draw is 64, so that 
each division of the axis of abscissae will represent 6 * 4 revolutions. 
The time required for each revolution must also be known, if the 
relation between speed and time is to be determined. This 
depends, however, on the manner in which the machine is 
manipulated by the driver. 

There are two simple cases : — 

(1) Starting with a uniformly accelerated angular velocity, 
which implies a constant angular acceleration. 



442 ELECTRIC WINDING-MACHINES. June 1905. 

(2) Starting with a uniformly decreasing angular acceleration. 
Analogous laws may be adopted during the period of 
retardation. 

The first case has been adopted in connection with the data 
supplied by M. Troussart, by allowing a time of fifteen seconds to 
get up a speed of 42 revolutions per minute, which is an angular 

velocity w = — — — = 4-39 radians per second; the constant 

angular acceleration is thus -~ = 0*292; the rate velocity is 

maintained 77 seconds, and the stopping is effected in 14 seconds 

with a constant negative acceleration — - = 0*313. Curve 

OMNX (Figs. 2, 3 and 5, pages 434, 435 and 437) represents the 
variation of the angular velocity. The integral of MNX for a 
given abscissa t represents the total angle fl through which the 
motor-shaft has turned since starting, and by dividing O by 27r, 
the total number of revolutions made by the shaft during time t 
is obtained. 

The curves which represent the relation between the statical 
moments (XqXio, Figs. 1 to 5) and time, and also between the 
moments of inertia (I"qI'\q) and time, can now be drawn, by 
taking the ordinates of these curves, Fig. 1 (page 433), which 
correspond to each revolution. 

Moments of the Accelerating Forces. — The energy exerted by the 

2 

accelerating forces during time t is '^ J, where w is the angular 

velocity at the end of time t and 7 the moment of inertia (FqI'^q), 
If M be the moments of the accelerating forces, then the 
energy exerted by these forces during time t may also be 

expressed by \ M aydt ; hence 

J 

lMoidt = ^L 

•' 



June 1905. ELECTRIC WINDING-MACHINES. 443 

As M, (0 and I are functions of <, by differentiation the following 
is obtained : 

Mco = Ico ^ + 2- ^^, or 

W— T^^ A.'^ ^^ 
^~ dt "^2 dt 

^ is constant during the period of acceleration and retardation, and 

is zero during the period of constant speed. Multiplying the 

ordinates of curve T'qF'iq by the corresponding values of -y, and 

then adding the results (algebraically) to the ordinates of XqX^q, 
the portions AB and CD, Fig. 2 (page 434), are obtained. 

For the purpose of calculating w ~^y the derived curve of I"qV\q 
is graphically constructed. The ordinates of this new curve 
(^Vsho)' estimated from the line through Iq as axis of abscissas, are 
now multiplied by the corresponding values of ^ taken from 

OilifiVX, and the results are added (algebraically) to the ordinates 
of ABX^X^CB. In this manner the total torque (AB'B"C"C'D, 
Fig. 2) which must be exerted by the motor at any instant is 
obtained. 

It will be seen that the torque varies considerably ; it attains a 
maximum of 16,500 kg.-m. (119,345 ft.-lbs.) at the end of the 
period of acceleration, and then suddenly drops to 10,000 kg.-m. 
(72,330 ft.-lbs.) at the beginning of the constant speed period; 
during this period the maximum value is 10,600 kg.-m. (76,670 
ft.-lbs.), and the minimum 5,200 kg.-m. (37,612 ft.-lbs.). The 
torque is negative during the period of retardation. 

Power and Expenditure of Energy. — The torque of the motor 
multiplied by the corresponding value of w is the power which 
must be developed at each moment. This power is represented 
by curve OEFGHX (Figs. 2, 3 and 5), the integral of the latter 
is the energy spent during each draw. 

At the end of the acceleration period, the power attains a value 
of 950 H.P., whilst during the constant speed period, it varies 
between 620 and 370 H.P. ; the latter figures differ from those 



444 ELECTRIC WINDING-MACHINES. June 1905. 

found by M. Troussart, whicli have undoubtedly been found by a 
less exact method than the one adopted by the author. 

The case of uniformly decreasing acceleration has also been 
treated; the results are indicated by the curves, Fig. 3, which 
are marked with the same letters as the corresponding curves in 
Fig. 2 ; it may be noticed that the curve OM, which represents 
the variation of the angular velocity o> during the acceleration 
period, is the integral curve of the straight line which represents 

the variation of the acceleration ~^. It will be seen that althous-h 

at ° 

a rather greater torque is required during the acceleration period, 
which is not difficult to obtain with an electric motor, the power 
required varies much less. It would, therefore, be advantageous 
to start the machine with a decreasing acceleration. 

It might be asked whether the power would not be less 
variable if the radii of the rope-roll were chosen so as to make 
the statical moments less regular. The author has treated the 
machine at Grand Hornu with regard to this question for r = 
1 • 1 m. (43 • 4 ins.) instead of 1 • 20 m. (47 • 24 ins.). The results are 
shown in Figs. 4 and 5 (pages 436 and 437), the same letters as on 
Figs. 1 and 2 (pages 433 and 434) being used all through. It will 
be seen that the power diagram is much more regular. The 
ordinates of the power diagrams which have been determined 
should be divided by the efficiency of the motor so as to obtain 
the total power spent; the integral of the latter curve would 
give the total energy required. The useful work divided by 
this energy is the efficiency of the winding-machinery. The 
efficiency of electrical motors differs with the kind of motors and 
their starting devices. 

Electric Mo-^ors for Winding : their Starting Properties. 

The machines used for electric winding are either three-phase 
asynchronous motors or continuous-current motors. The latter 
may be shunt-wound, or separately excited. Both types of 
machines possess a remarkable degree of self-regulation, that is, 
their speed only varies slightly with a considerable variation of 



June 1905. ELECTRIC WIN DING -MACHINES. 445 

load.* This property makes the machine very easy to manipulate. 
The driver need not trouble about the speed, as it never exceeds 
the normal speed much, even when the resisting torque is negative, 
in which case the motors turn into generators sending energy 
back to the station. Continuous-current series machines, although 
used for electric traction, are not suitable for winding, as they run 
away when the load diminishes. The advantage of a better 
starting torque, which these machines can furnish, is, however, 
only relative, as motors of the other types can easily produce 
torques three or four times the normal one. Such torques are 
necessary, as the resisting moment at the end of the acceleration 
period is often double that at other velocities. 

The starting of electric motors requires special contrivances, on 
the design of which the consumption of energy during acceleration 
depends. It has been shown that during the acceleration period 
the motor must develop a torque given by curve AB\ Figs. 2, 3, 
and 5, in the cases already treated. 

As the author wishes to discuss the starting properties of these 
motors more thoroughly, he thinks it would not be out of place to 
give a short account of their mode of working. 

TJiree-pliase AsyncJironous Motors. — The torque produced by a 
polyphase asynchronous motor is (1) : — 



1 + ^^^ ^_^co, 



K^ is a coefficient; H the rotary magnetic field due to the 
inducing currents ; Wj is the slip which is equal to the difference 
of the angular velocities (o>) of the rotary field and (w^) of the 
induced field or rotor ; r the electric resistance of the rotor 
circuits ; L their coefficient of self-induction. 

2 TT f 

The angular velocity w is equal to '^'^ where / is the 
frequency of the current and 2 p the number of inductor-poles. 

* On condition that the difference of potential between the terminals of 
shunt or separately excited machines is constant. 



446 



ELECTRIC WINDING-MACHINES. 



June 1905. 



Under normal conditions, the velocity w^ does not differ more 
than 2 or 3 per cent, from w, hence (o^ = 0*02 to 0*03 w. 

When the intensity of the field is constant, the torque will be 
a function of the slip only. In curve I, Fig. 6, the ordinates 
represent the torque, and the abscissae, which are reckoned 
positive to the left of the origin 0, represent the slip. When the 
slip is zero the torque is also zero ; and if the rotor was driven at 
a speed w^ larger than w, due to a negative resisting moment, the 



Fig. 6. 



V ^ m II I 

' •-£wrref)f ^ , 



i*j' 



-^t tJ 




Fig. 7. 

Liquid Rheostat at Preussen II. 
and at Grand Hornu. 





motor would become a generator and would oppose the resisting 
moment. Curve I also shows that in the neighbourhood of 
synchronism, the torque of the motor may vary between wide 
limits without causing much variation in speed. 

When (Oi increases progressively from zero, the torque will also 



increase progressively, and becomes a maximum when h 

j 



June 1905. ELECTRIC WINDING-MACHINES. 447 

is a minimum, that is, when = * or w, = ^ ; the maximum 

E} H} . ... 

torque will be G^^^, = ^£ - The torque will now diminish as 

coi increases progressively from ^ , and when w^ is equal to w, then 
(0^ will be zero and the motor will come to a standstill ; then — 

0=—- X 



r fo^ 1?' 
1 + 

It will be noticed that the torque changes sign with ooi, that is, 
it becomes a resisting moment when w^ is greater than w. The 
maximum torque may attain a value three to four times greater 
than the normal torque corresponding to coi = * 02 to • 03 w. 

Curve 1, Fig. 6, shows how the current varies with wi ; it will 
be seen that at starting (standstill of rotor), the current is notably 
stronger than at the normal speed. In the case of powerful 
motors (above 5 or 7 H.P.), a current of such strength cannot be 
taken from the mains without causing an inadmissible fall of 
voltage, nor are the motors designed to carry such a current ; the 
small torque which corresponds to 0)1 = 0) is not sufficient to 
start the rotor, especially if it has to drive a winding-machine. 

For this reason, some contrivance must be adapted by which 
the torque can be increased and the current reduced. 

Starting Bheostat. — The only method which has been used with 
three-phase motors for driving winding machinery consists in 
inserting variable resistances r^ in series with the three circuits of 
the rotor. Substitution must therefore be made of r -f »'^ for r in 
the formulas, if r^ is non-inductive. The maximum torque may 

thus be produced when the slip is wj = — j — , and by giving r^ a 

value which will make w^ = oi the maximum torque at starting 
will be arrived at. These conditions are represented by curves 
VI and 6. The resistance r^ inserted in the rotor circuits 
prevents the currents from rising too much, notwithstanding the 
large slip o)i, and hence the currents in the inductors will not 
exceed the normal strength. By reducing the resistance r^ 



448 ELECTRIC WINDING-MACHINES. June 1905. 

successively, the torque and current curves change respectively to 
V, 5; IV, 4; III, 3; II, 2; and finally to I, 1, when the last 
resistance has been cut out. 

The use of a starting rheostat causes considerable loss of 
power during the period of acceleration. The current will be 
proportional to the resisting moment given by curve AB\ and as 
the voltage is constant, the electrical energy absorbed will 
be equal to this moment multiplied by the angular velocity 
and the efficiency of the motor at such a charge. This loss 
will be less felt when the startings are less frequent ; when 
the pit is very deep, the period of acceleration will be small 
compared with the period of constant speed. With regard to 
simplicity, this method is certainly superior to any other, and 
for this reason it is always adopted where three-phase motors 
are used for driving winding machinery. 

The Allgemeine Elektricitats Gesellschaft of Berlin have 
designed a rheostat with liquid resistance, consisting of two 
vessels, the one above the other. The liquid is kept in motion 
by a small centrifugal pump driven by an electric motor of 2 to 
3 H.P. The pump takes the liquid from the lower vessel and 
forces it back into the upper vessel in which three metallic 
plates (one for each rotor circuit), electrically insulated, are 
suspended, Fig. 7 (page 446). The upper vessel has an overflow, and 
the plates are only immersed to a depth corresponding to the 
required resistance. By manipulating the starting lever, which 
works the reversing switch and closes the inductor circuit, a stop- 
valve Y2 will be closed ; the level of the liquid in the upper vessel 
will therefore rise, and the plates will be more and more immersed. 
The rate at which the water rises can be regulated by a cock on 
the inlet pipe. The maximum speed of the machine is reached 
when the liquid has risen up to a second overflow, which, is always 
open and which causes the liquid to overflow continually from 
the upper vessel into the lower one. The circulation of the liquid 
prevents it from getting heated. When the lever of the reversing 
switch has been pushed home, to the right or left, according to 
the direction in which the machine is to turn, it may be moved 



June 1905. ELECTRIC WINDING-MACHINES. 449 

back through a certain angle without breaking the circuit. In doing 
this the stop-valve is gradually opened so that the water can run out 
again and the speed will be reduced as desired. The circuit will 
not be broken until the lever is turned back into its vertical 
position ; as the level of the liquid is falling steadily, the current 
as well as the speed of the motor w^ll be gradually reduced; 
the interruption therefore will not be sudden. 

The immersion of the plates may be further limited by an 
overflow V^ so as to reduce the speed when persons are carried 
up and down. When the pit or the ropes are to be examined, 
the speed may be still further reduced by regulating V^. The 
conditions, when using starting resistances, are very unfavourable 
to the electric brake; one cannot count on recovering the 
energy, but one must try to utilize the energy stored in the moving 
masses as much as possible, and to let the cage arrive at the 
mouth of the pit with a much reduced speed which will only 
require the brake for holding the cage. 

The intermittent working of winding machinery and the 
increased energy wanted for starting a three-phase motor with the 
rheostat causes sudden and irregular demands on the generators 
at the power-station. The latter machines are to be designed 
accordingly large enough and supplied with sufficiently heavy 
fly-wheels. If this is not the case the rosult will be a fall of 
speed and voltage which will be felt all over the area of 
distribution ; this is the case at Preussen * and even at the coal 
mines at Grand Hornu, where the generating plant comprises 
engines, one of 2,100 H.P. with 75 tons fly-wheel, and one of 
4,200 H.P. with 100 tons fly-wheel. This engine is to be worked 
with three winding-machines. Only one is at work at present with 
the 2,100 H.P. engine. The electrical energy delivered varies from 
300 kw. to 900 kw., and the fall of voltage is about 10 per cent. 

Direct driving of winding-machinery by asynchronous motors 
requires generators capable of standing a considerable overload, 



* See Paul Habets, "Les machines d'extraction electriques." 'ReYue 
UniverBelle des Mines,' 4th series, 1904, vol. vi. 

2 K 



450 



ELECTRIC WINDING-MACHINES. 



June 1905. 



and the installation of very powerful sets at the power-station. 
This means an increased expenditure for extra boilers, steam 
engines and alternators, but, on the other hand, the use of three- 
phase machines is much simpler and much less costly than 
transforming alternating currents into continuous currents. 



Continuous-Current Motors. — The fundamental equations for 
continuous-current motors are (1) : — 
E - e 



I = 

r 

p = i e = i n N ^ 



J — 



a — 



27r 

27r 



60 60 

E — ir 



[ 



n^E-2TrrC 



n"^' 



J 



X 



(1)* 

(2) 
(3) 

(4) 

(5) 



60 ' ^ nm. 

i is the current passing through the armature. 
E the difference of potential between the terminals. 
e the counter electromotive force. 
r the resistance of the armature. 
p the mechanical power given off. 
n the number of inductors on the armature. 
01 the total number of magnetic lines through the armature. 
N number of revolutions per minute. 
0)^ the angular velocity of the armature. 

Shunt motors and separately excited motors possess somewhat 
similar properties, at least when they are supplied at a constant 
difference of potential. In separately excited motors the flux 01 
is constant, save for the slight reduction due to armature reaction. 
The torque (3) in this case is proportional to the current i. 
When the motor is supplied with a constant current, its torque 
is also constant, and its speed (4) is proportional to the difference 
of potential between the brushes. By varying the difference of 
potential, the motor could be worked with a constant torque at 



* See " Principes d'Electrotechnie," par Em, Pierard, page 138 and following. 



June 1905. 



ELECTRIC WINDING-MACHINES. 



451 



different speeds, provided that the current through the armature 
is kept constant. Under these circumstances the mechanical 
characteristic of the motor, that is, the curve which represents the 
torque as a function of the speed, will be a straight line parallel 
to the axis of abscissa?, Fig. 8. The machine may then be 
started with any desired torque by progressively increasing the 
difference of potential between the brushes. 



Fig. S. 

C 

Conslani Torque for i Constant 



Fig. 9. 
Auxiliary Starting Dynamo. 




If the constant difference of potential be maintained, the current 
in the armature will always be strong enough to produce a torque 
which can overcome the resisting moment (3) ; the speed varies 
with the torque ; when the torque is zero, the speed will be 

2 TT E 



J — 



60 ^ w3fl 



This is the maximum theoretical speed which the motor can 
attain when running with no load ; the current in the armature is 
thien zero, and the counter electromotive force is equal to the 
difference of potential between the brushes. 

2 K 2 



462 ELECTRIC WINDING-MACHINES. JuNE 1905. 

On account of tlie losses of energy due to hysteresis, eddy 
currents and mechanical friction, the torque of the motor will 
become zero before the current and at a lower speed. When the 
resisting moment becomes negative, tending to drive the motor 

at a speed greater than w^ = -— x -^, then the counter 

electromotive force e will be greater than E ; the current in the 
armature will be reversed and will flow from the motor back 
nto the mains ; and the motor becomes a generator and acts 
as a brake by transforming the mechanical energy which 
it receives into electric energy. As soon as there is a 

n jp 

resisting moment, the speed will fall below -— X —5^, and 

the current absorbed by the motor will be just sufficient to 
produce a torque equal to the resisting moment. The relation 
between w' and C is represented by a straight line (4), 

passing through the point corresponding to w^ = ~— X ~^, and 
inclined slightly towards the axis of ordinates (Fig. 8, page 451) ; 
the ohmic resistance of the armature r is small, and hence zr~ai2 

is also very small. The speed is therefore practically constant for 
all loads ; it varies between the values B and B' for loads 
between + C and — C (Fig. 8). Separately excited motors 
supplied at constant diflference of potential possess therefore the 
same self-regulating properties as three-phase motors. The same 
is the case with shunt motors, which behave, when supplied at a 
constant difference of potential, like separately excited motors. 

When a shunt motor is supplied at a constant voltage, the 
field current as well as the flux will be constant, and the motor 
will therefore behave like a separately excited machine. It is 
however not always possible to keep the flux constant ; for 
instance, when starting the motor, the mains will be almost 
short-circuited through the armature, whose resistance is very 
small, and this causes a momentary fall of potential across the 
terminals, resulting in the field-current, and therefore also the 
flux, being reduced. 



June 1906. ELECTRIC WINDING-MACHINES. 453 

For this reason, a resistance must be inserted in the armature- 
circuit to prevent the armature-current from exceeding three or 
four times its normal strength. The field-current, and therefore 
also the flux, will remain constant, as they are not affected by the 
resistance. The resistance is now gradually reduced, as the speed 
of the motor and, therefore, also the counter electromotive force 
increase, and in this manner a constant armature-current as well 
as a constant torque may be maintained. Under these conditions, 
a shunt motor will behave like a separately excited machine 
supplied with a constant current. Continuous-current motors 
may be started without much loss of energy, and the different 
methods which have been used for this purpose will now be 
examined. 

1. Starting Rheostat. — This method is used when a shunt 
motor is connected across a pair of mains, between which a 
constant difference of potential is maintained. A buffer-battery 
may also be inserted across the same mains. The field current is 
thus constant. 

A variable resistance It is inserted in series with the armature. 
If the difference of potential between the brushes be E^, the power 
absorbed in the resistance will he (^E — E^) i = i"^ It. 

During the period of acceleration, the resistance It is reduced 
progressively, and the current ^, and therefore also the torque G, 
may be kept constant ; this requires that E^ — e must also be kept 
constant. 

The speed cannot be kept quite constant independent of 
the load, with the resistance It inserted. This resistance 
increases the ohmic resistance of the armature. The relation 
between the torque 0, and the speed (4) will be found by 
substituting r -{- B for r. This relation is represented by a 

27r E 

straight line through the point corresponding to w^ = g^- X -^' 

and the inclination of the line towards the axis of ordinates will 
be the greater, the greater J^ is. The straight lines AB, AB\ AB", 
&c., Fig. 8 (page 451), which correspond to increasing values of B, 



454 ELECTRIC WINDING-MACHINES. June 1905. 

show that the greater the resistance is, the more will the speed 
increase in proportion to the torque. 

This method of starting a shunt motor causes the same loss of 
energy as is found with three-phase motors. 

2. Series-Parallel Method. — When using two motors for driving 
the winding-machinery, the starting may be done by the series- 
parallel method, the same as used for regulating the speed of 
electric tram cars. It consists in starting the two motors by 
connecting them in series ; the difference of jjotential across each 

armature, if no resistance is inserted, will be ^ • 

The normal speed will thus be reduced to half; when later on 
the motors are connected in parallel and the total difference of 
potential E is applied across the armature, the full speed will be 
attained. For the purpose of preventing jerks, while switching- 
over from series to parallel, a kind of rheostat (controller) must 
be used. 

This method of starting gives a more satisfactory efficiency 
than the preceding methods, but the two motors together cost 
more than one motor of the same total power, and the circuit 
controller is complicated. 

This principle has been adopted by Messrs. Siemens and 
Halske at the Thiederhall Mines, at Thiede, near Brunswick. 

By using electric sources with variable pressure, the voltage 
between the brushes may be varied. For this purpose, a battery 
of accumulators split up into sections may be used, or a special 
generator, or an auxiliary dynamo. 

3. Batteries in Sections. — A battery of accumulators, which is 
inserted across the mains, is split up into sections ; the latter are, 
however, connected in series so as to make up a total voltage 
equal to the pressure between the mains. During the period of 
acceleration, the brushes of the motor are connected across one 
or more sections of the battery, the number of sections being- 
increased as the speed of the motor rises ; the normal speed will be 
attained when all the sections are switched on. 



June 1905. ELECTRIC WINDING-MACHINES. 455 

The motor may be reversed by reversing the connections 
between the brushes and the poles of the battery (the current in 
the armature will thus be reversed), and the different sections of 
the battery will then be equally discharged, provided that their 
number is even. 

The motor will have as many normal speeds as there are 
sections ; for instance, four sections will give normal speeds equal 
to ^, J, j, and full speed. For the purpose of avoiding jerks 
when switching a new section into the motor circuit, a rheostat 
must be used. The total resistance B required for each battery 
section must be equal to the resistance of the armature; B is 
however subdivided into smaller resistances, which are cut out 
progressively as the speed of the motor increases. 

The economy of this system as compared with that of the 
armature-rheostat is only obtained by great complications, which 
was shown, only too clearly, in the case of the winding-machine 
at pit No. II. Zollern, installed by Messrs. Siemens and Halske. 

4. Special Generator. — In this system the winding-motor is 
separately excited by a constant current, and the difference of 
potential is varied, and therefore also the speed, by regulating 
the electromotive force of the generator, which supplies the 
motor, by means of a field rheostat; the loss of energy is 
almost nil. But as one cannot use a buffer-battery, the generator 
must be able to stand all the variation of the load, and the 
engine which drives the generator will be running on no load 
during the stoppages. The work of the engine might be 
equalised by providing it with a large fly-wheel, which could 
store up the energy during the stoppages of the winding 
machines; or the engine might drive another generator, which, 
in connection with a buffer-battery, could supply a circuit at 
constant pressure. It is, however, always necessary to have 
reserve sets at the power-station, and this makes the undertaking 
costly. The use of a special generator is not practical unless 
the winding machinery is close to the supply -station. This 
system has been adopted by the AUgemeine Elektricitats- 



456 ELECTRIC WINDING-MACHINES. JuKK 1906. 

Gesellschaft in 1896 at the Hollertszug miues, and recently for 
two machines at the Alexandre pits of the Arnim Collieries at 
Planitz, near Zwickau, in Saxony. 

The system of a special generator may be worked more 
economically by adopting high-tension three-phase currents. In 
this case, the special generator is driven by an asynchronous 
motor, and the two machines are installed close to the winding 
machinery ; by using high-tension currents, the distance from the 
power-station is not of much importance. 

For the purpose of storing the energy during the stoppages 
of the winding-machinery, M. Ilgner uses a large flywheel 
mounted on the shaft of a motor-generator, the latter consisting of 
an asynchronous motor and a continuous-current generator. As 
the motor- generator is running continuously, the current from 
the supply-station will not vary much. The mean power is also 
less than in the case of an asynchronous motor driving the 
winding-machinery direct. The continuous-current generator, as 
well as the winding motor, is excited by a small continuous- 
current machine driven by an asynchronous motor. This system 
has been adopted by the Allgemeine Elektricitats-Gesellschaft at 
the pits of St. Nicolas at the collieries of Esperance and Bonne- 
Fortune, Montegnee, near Liege. The installation, together with 
the results of some trials, will be described further on in the 
Paper. Messrs. Siemens and Schuckert have adopted the Ilgner 
fly-wheel system to the machine at pit No. II. Zollern.* In this 
case the motor is a continuous-current machine. 

5. Auxiliary Starting Dynamo. — In this system, the armature of 
the winding motor M, Fig. 9 (page 451), is connected in series with 
the armature of an auxiliary dynamo Da, across the mains of a 
constant-pressure circuit and a buffer battery can also be inserted ; 
the fields of both machines are connected direct to the mains. The 
auxiliary dynamo is driven, at a constant speed, by the steam- 

* See Paul Habets, "Les machines d'extraction electriques." 'Revue 
Universelle des Mines,' 4th series, 1904, vol. vi. 



JlTME 1905. ELECTRIC WINDING-MACHINES. 457 

engine Q G working a sot of dynamos. The field is regulated by 
a rheostat, and the electromotive force generated in the armature 
opposes the pressure between the mains. When the difference of 
potential E" between the brushes is equal to the pressure between 
the mains, no difference of. potential will be available at the 
brushes of the winding motor M, which will therefore be at a 
standstill. By reducing the resistance JR, the difference of 
potential E' between the brushes of M will he E' = E - E" ; a 
torque will now be exerted by M which will be increased by 
reducing E". When the motor M has attained its full speed, the 
counter electromotive force in the armature of Da should be as 
near zero as possible, and E" should only be sufficient to overcome 
the resistance of the armature. The dynamo Da works therefore 
as a motor, aiding the steam-engine ; but when the resisting 
moment becomes negative, then the auxiliary dynamo will work as 
a generator, taking energy from the engine. The current will 
now flow against the pressure of the mains, and the energy 
generated by M, which has also become a generator, will increase, 
and therefore cause M to go slower. In this manner the auxiliary 
dynamo will keep the speed of M almost constant. The speeds 
as well as the startings are obtained without other losses than 
those due to driving the auxiliary dynamo. It is, however, 
inconvenient to have to regulate the field of the dynamo if this 
machine cannot be placed near the haulage motor. 

The auxiliary dynamo may be driven by a motor. The two 
machines mounted on the same shaft constitute a booster, which 
may be connected with the mains and the winding motor in 
different ways, but can be placed close to the winding-machine. 
This arrangement makes the winding-machine more independent 
of the power-station, which may be at a considerable distance 
from the pit. 

6. Starting Boosters. — The booster, which has just been 
described, consists of an auxiliary dynamo with a field rheostat 
driven by a shunt motor which receives current from the 
mains. With regard to economy, the booster is inferior to an 



458 ELECTRIC WINDING-MACHINES. JuNE 1905.- 

auxiliary dynamo driven by an engine, because the efficiency of 
the booster is less, on account of the double transformation of 
energy. The cost of installation is higher, as the motor and the- 
auxiliary dynamo of the booster must both be of the same power 
as that of the winding motor. The speed of the booster may 
however be high, whilst the winding motor, being direct coupled 
to the winding-machine, cannot have a speed exceeding 50 to 
60 revolutions per minute. 

By reversing the field-current of the auxiliary dynamo, the 
booster will Avork as a booster proper, that is, it will raise the 
voltage of the mains. Matters might therefore be arranged so that 
the pressure of the circuit will be doubled when the winding motor 
is running at full speed. Under these conditions, the power of each 
of the booster-machines need not be more than half of that of 
the winding motor, thus making the booster less costly. For 
this purpose, the voltage of the winding motor must be twice that 
of the mains, and one must therefore either reduce the pressure 
of the distributing system or design a high-tension motor, which 
with continuous-current machines is difficult when the tension 
exceeds 500 volts. The problem may also be solved by driving 
the winding-machine by two motors connected in series, each of 
which will therefore only receive half of the total voltage. 

The latter system, which was brought forward by the Union 
Elektricitats Gesellschaft, has been adopted by Messrs. Schuckert 
and Co. in the case of the winding-machine at the Friederich- 
Franz mine at Liibtheen (Mecklenburg). Fig. 10 is a diagram of 
connections. An Ilgner fly-wheel may be mounted on the booster 
shaft. This solution has been applied by W. Lahmeyer and Co. 
to the winding-machine of the Tiremonde pit of the Mining Co. 
of Liguy-les-Airs. For the purpose of bringing the fly-wheel into 
proper play, a second booster is connected in series with the 
armature of the starting booster, in such a manner as to vary the 
speed of the latter automatically, and also inversely as the 
strength of the armature current of the winding motor. 

M. Creplet reduces this group to a single dynamo, the shaft of 
which carries a special fly-wheel. As it is possible to allow a great 



J 



June 1905. 



ELECTRIC WINDING-MACHINES. 



459 



variation of speed for this dynamo, the fly-wheel can be much 
reduced. A machine on this system has been installed by the 
Compagnie Internationale d'Electricite de Liege for the Fleron pit 
of the Hasard Collieries at Micherot, near Liege. 

Very few trials on electrically-driven winding-machines have 
been published up to the present time ; the trials made on 
the machine at pit No. II. Preussen afford a unique account of 
the working of a machine driven by a three-phase motor.* A 
great number of trials have been made on the machine installed 

Fig. 10. 



I fa 1_>^ neversiiiff 
^AAA/wv 1 ^T Field ~Svn 




-Switch 
Field SwiZch 



Tteversi/iny'^ 
for Moton 



Release 



Eieclro -magnet 
for the Brake 
With counlerpoise 



oMaxinuun 
Currenl Release 




AmmeZer 



at St. Nicolas, at the collieries of Esperance and Bonne-Fortune, 
Montegnee, and a detailed description will now be given of this 
installation, together with the results obtained. 



IV. Electrically-driven Winding-Machines at the jpit of St. Nicolas 
at the Collieries of Esperance and Bonne-Fortune, Montegnee. 

1. Description of the Installation (Fig. 11, page 461). — The 
machine was supplied by the Societe Beige d'Electricite A.E.G. 



* See Schulte, " Gluckauf," No. 13, 26 March 1904, and P. Habets, " Les 
machines d'extraction electriques " ; ' Revue Universelle des Mines,* 4th series, 
1904, vol. vi. 



460 ELECTRIC WINDING-MACHINES. JuNE 1905. 

It is designed to draw 500 tonnes (492 tons) from a depth of 
800 m. (2,625 feet) in eight hours. The loads are : four tubs 
holding 535 kgs. (1,102-3 lbs.) of coal or 750 kgs. (1,653-4 lbs.) 
of stone; each tub weighs 250 kgs. (551*1 lbs.) and the cage 
weighs 1,800 kgs. (if tons). The electric energy is supplied 
from the power-station, which was constructed in 1899, at the pit 
of Esperance, and which contains three sets of three-phase 
generators working at 1,000 volts and with a frequency of 44; the 
output of each machine is 200 kw. Two of the sets also supply 
current for pumping, ventilation, picking, etc., at the pits of 
Esperance and St. Nicolas ; the third is a reserve set. 

The coal is drawn from four underground landing-stages at 
depths of 368 metres (1,207 feet), 342 metres (1,122 feet), 288 
metres (945 feet), and 185 metres (607 feet). On account of the 
nature of the deposition of the beds and the extent of the field of 
exploitation, it may eventually be necessary to draw simultaneously 
from several seams, which will require one or two stoppages 
during each lift. The time spent in unloading the cage at the 
bank of the pit, and simultaneously loading the cage at the 
bottom landing-place, is 30 seconds, and loading the cage at 
intermediate landing-stages takes 20 seconds. Under these 
conditions, the maximum winding speed need not exceed 12 metres 
(39 feet 4J inches) in order to draw a total of 500 tonnes (492 
tons) from a depth of 800 metres (2,625 feet). When the depth 
is less or the tonnage is smaller, the speed may be slower. Local 
circumstances have led to the adoption of a Koepe-puUey with a 
central groove for a flat-wire rope, and two brake-rims, one on 
each side of the groove. The motor and the Koepe-pulley are 
mounted on the same shaft ; the diameter of the pulley is 3 metres 
(9 feet io|^ inches) and requires therefore only a medium size 
motor. The winding speed is 10 metres (32 feet 9I inches) when 
the motor makes 64 revolutions per minute. 

Notwithstanding the high frequency, a three-phase motor 
might have been designed for this slow speed. Preference has 
however been given to a continuous-current motor ; and a motor- 
generator of the Ilgner type, consisting of a three-i^hase motor 



Junk 1905. 



ELECTRIC WINDING-MACHINES. 



461 



Fig. 11. — Winding-Engine Installation. 
(See also Fig. 12, page 463.) 




462 ELECTRIC WINDING-MACHINES. June 1905. 

and a continuous-current generator, has been adopted for the 
following reasons : — 

During the period of acceleration, the three-phase motor would 
have absorbed 615 H.P., a power which the central station would 
not have been able to supply with the present plant. Besides, a 
three-phase motor with a starting rheostat is very wasteful, and 
the loss of power in this case would have been still higher on 
account of the intermediate stoppages. 

Notwithstanding the conversion of the three-phase current, the 
motor- generator is in this case more economical, as it almost entirely 
prevents any losses during acceleration, and during the slow speed, 
which must be adopted when the winding is reduced, and also 
during the night-shift. A motor-generator with an Ilgner 
fly-wheel will also equalise the demand on the power-station, no 
matter how the load varies. The power-station will then be 
able to supply the other motors, if the pumping is done during 
the night. 

The winding motor M, Fig. 12 (page 463), is a separately excited 
continuous-current machine, its rated power is 320 H.P. at 64 
revolutions per minute, and at 600 volts. It will carry an overload 
of 40 per cent, for half-an-hour, and of 100 per cent, for 5 minutes. 
It will reverse, without moving the carbon brushes, and gives no 
sparking even when the voltage is considerably reduced. 

The terminals of the winding motor are electrically connected 
with the generator Gj, Fig. 12, of the Ilgner motor-generator 
Mi-Gi; and a maximum current release, which acts when 
the current reaches 1,500 amperes, is inserted in one of the 
leads. The output of the continuous-current generator, which is 
separately excited, is 650 kilowatts at 500 volts, when making 
285 revolutions per minute ; the rise of temperature, measured by 
the increased resistance, does not exceed 45°. The carbon brushes 
have no lead, even when the machine is generating the maximum 
current required by the motor. The machine will also generate 
the current without sparking, at voltages varying between zero 
and ± 500. These results are due to the large proportions of the 
generator, whose power is more than double that of the winding 
motor. 



June 1905. 



ELECTRIC WINDING-MACHINES. 



463 




464 ELECTRIC WINDING-MACHINES. JUNE 1905. 

The generator and the winding motor are of similar design. 
The cast carcass carries the pole-pieces of laminated iron, which 
are fixed by bolts in such a manner that they can be removed 
laterally without dismounting the armature. The field-coils are 
protected by the carcass, and the bobbins are divided into sections 
for the purpose of ventilation. The armature is a drum, the 
conductors being tooled and placed in the grooves of the sheet-iron 
core ; the openings between the conductors allow of an effective 
cooling by circulation of air. 

The winding motor M and the generator Gi are both excited 
by a small motor-generator M2-G2» which supplies a continuous 
current at 110 volts. This motor-generator consists of a 27-H.P. 
asynchronous motor M2 — supplied with current from the power- 
station and making 820 revolutions per minute — and an 18 kilowatt 
continuous-current shunt-generator G2. 

The armature of the main generator Gi is mounted on the end 
of the shaft, which also carries, between its two bearings, a fly- 
wheel of cast steel, 4 metres (13 feet ij inches) in diameter and 
weighing 40 tonnes (39*3 tons). The rotor of the asynchronous 
motor Ml, which drives the Ilgner-group, is mounted on the other 
end of the shaft ; it will exert 250 H.P. at a continuous run of 
285 revolutions per minute. For the purpose of inserting a 
liquid rheostat in the rotor circuits, on starting the machine, the 
rotor is provided with slip-rings. An electro-magnetic brake F 
when excited will generate eddy currents in the fly-wheel rim, and 
will stop the motor-generator in about a quarter-of-an-hour. 
"Without this contrivance, it would take several hours to bring the 
machines to a standstill on account of the large amount of energy 
stored in the fly-wheel. 

The bearings of the motor-generator are lubricated by means of 
rings ; the steps are cooled by circulation of water from a vessel at 
a higher level. The water flows into the grooves of the bearings, 
and into the reservoir of the starting rheostat, and falls into 
a cistern. A small centrifugal pump, worked by a 2 H.P. 
asynchronous motor M4, takes the water back to the high-level 
vessel. The pump can lift 5 litres ( i • i gallon) of water per 



June 1905. ELECTRIC WINDING-MACHINES. 465 

second through a height of 15 metres (49 feet 2 J inches), and 
forces the water as a spray against a cooling screen. 

The reversing of the winding-motor is brought about by 
manipulating the lever L^ of a controller. According to the 
position of the lever, the following will take place: (1) reversing of 
the field-current of the generator G^, whereby the rotation of the 
winding-motor M will be reversed ; (2) reducing the field-current 
of the winding-motor and interrupting the field-circuit of the 
generator, when the motor is stopped ; (3) increasing the field- 
current of the generator progressively, by cutting resistances out 
and thus causing the voltage of the generator as well as the speed 
of the motor M to increase progressively. While the lever L^ is 
turned back towards the position " stop," the counter electromotive 
force generated in the winding-motor (driven by the inertia of 
the moving masses) will become higher than the electromotive 
force generated in G^. The motor will become a generator and 
will act as a brake, thereby reducing the period of retardation. 

The Koepe-pulley K is provided with two brake-rims. The 
brake, which is worked by compressed air, is only used by the 
driver for holding the pulley during the standstill of the machine, 
and is manipulated by means of a lever for admitting the 
compressed air into the cylinder. It also acts as a safety-brake 
when the three-phase circuit is interrupted. An electro-magnet E, 
excited by the field-current, will then be demagnetised ; a 
counterpoise will be released and the brake will be brought into 
action. The brake will also act when the maximum release 
breaks the circuit, as the field circuit will also be interrupted. 
When the cage overshoots the landing-stage, the indicator, 
showing the travelling of the cages, will cause the maximum 
release to break the circuit, and the brake will consequently be 
brought into action. The brake may also be applied by working 
the pedal P, which will release a counterpoise. 

The retardation is also brought about by the indicator, which 
causes the lever L^ to be moved towards the position " stop," as the 
cage approaches the pit-bank. This automatic movement is often 
limited, so as to allow the driver to finish the "lift," and to 

2 L 



466 ELECTRIC WINDING-MACHINES. June 1905. 

manipulate the machine. The compressed air required for the 
brake is supplied by a small compressor driven by a 5-H.P. three- 
phase motor at 120 volts. 

The different machines and apparatus, as well as the electrical 
connections, are shown in Fig. 11 (page 461), and Fig. 12 (page 463). 

2. Description of the Trials and the Besults. — The winding- 
machine just described has been submitted to a series of trials 
with the object of giving an account of the performance of the 
Ilgner system ; also of testing the consumption guaranteed by the 
contractor, and finally of determining the consumption of energy 
under actual working conditions. The trials were made on the 
8th, 10th, 29th and 30th of January, 1905. 

(a) No-Load Trials. — The difficulty of starting the heavy fly- 
wheel, which squeezes the oil out from the bearing-surfaces, 
makes it necessary to keep the Ilgner machine-group in permanent 
motion, as the winding-machine does not stop for very long. The 
consumption of energy, while the Ilgner group is working on no 
load, has an important bearing on the economy of the system, 
and it was therefore decided to determine this consumption. A 
Siemens wattmeter and an electricity-meter were inserted in one 
of the motor-leads (M^), and readings were taken every 6 minutes 
during a period of 35 minutes, the machine making 313 revolutions 
per minute. According to the readings of the wattmeter, which 
had been specially standardized, the consumption was 31*94 kw.- 
hours ; the electricity-meter showed a consumption of 30 • 85 kw.- 
hours. The temperature of the oil on the bearings was 30 • 8° C. 
(87-4° F.). By stopping the circulation of water for 4J hours, the 
temperature of the oil on the bearings rose to 45 • 5° C. (11 3 • 9° F.) ; 
under these circumstances, and the shaft making 315*5 revolutions 
per minute, the consumption according to the wattmeter and the 
electricity-meter was 30*55 kw.-hours and 29*82 kw.-hours 
respectively. According to another trial, repeated on the 29th of 
January, the consumption was found to be 28 kw.-hours at a speed 
of 314 revolutions per minute ; the temperature of the oil was 



I 



Junk 1905. ELECTRIC WINDING-MACHINES. 467 

35* 7"^ C. (96*3° F.). The latter consumption corresponds to that 
guaranteed by the contractor. The efficiency of the motor at this 
reduced load was 0*6; the windage of the Ilgner group and the 
friction between the journals and the bearings absorbed, therefore, 
about 17 kw.-hours. 

(6) Passive Resistances in tlie Pits and in the Winding-Motor, — 
These resistances were determined by giving the masses a certain 
velocity and measuring the height through which the useful load 
was lifted. The energy consumed by the passive resistances would 
then be equal to the kinetic energy stored in the moving masses 
minus the work done on the useful load. 

The suspended loads were : — 

460 + 400 = 860 metres (2,820 feet) of rope at 

5*6 kgs. per metre (3 • 76 lbs. per foot) . 4,800 kgs. about (10,600 lbs.) 
2 cages with attachments at 2,100 kgs. (4,630 

lbs.) 

8 empty tubs at 250 kgs. (551 lbs.) . 
4 loads of coal at 535 kgs. (1,180 lbs.) 

Total . 13,140 kgs. „ (28,988 lbs.) 



4,200 


J) 


„ (9,260 lbs.) 


2,000 


J5 


„ (4,408 lbs.) 


2,140 


J» 


„ (4,720 lbs.) 



\ 



The diameter of the Koepe-puUey (reduced by wear) was 
2 • 88 metres (9 ft. 5 1 ins.). The moment of inertia of the suspended 

loads was = ^^'^^^ ^ ^'^^' = 2,780 kgs.-metres^ (65,970 Ibs.-feet^). 

The moments of inertia of the rotating masses were : — 

1. Moment of inertia of the Koepe-pulley 630 kgs.-metres- (14,950 Ibs.'feet^) 

2. „ „ „ motor armature 250 „ „ ( 5,932 „ „ ) 

3. „ „ „ head pulleys 675 „ „ (i6,oi8 „ „ ) 

1,555 „ „ (36,900 „ „ ) 



The moment of inertia of the masses was thus : I = 2,780 + 

1,555 = 4,335 kgs.-metres^ (65,970 + 36,900 = 102,870 Ibs.-feet^). 

The kinetic energy corresponding to a speed of 8 metres 

7" 2 Q 

(26 ft. 3 in.) per second equals — o- w, = jrn = 5 • 55, and therefore 

2 T ^ 

2" = 15 '4, hence the kinetic energy = —^ = 66,759 kgs.-metres 
(482,900 foot-lbs.). 2 L 2 



468 ELECTRIC WINDING-MACHINES. June 1906. 

The loads travelled through a height of 29 to 30 metres (95 to 
98 feet) after the current had been cut off from the motor. The 
work done by the non-balanced load was thus 2,140 x 29 = 61,060 
kgs.-metres (441,647 foot-lbs.) to 64,200 kgs.-metres (464,400 

66 759 
foot-lbs.). The kinetic energy of the masses is between g/oQn = 

1 '04 and gyWo ~ 1 *095 times the work done in raising the non- 
balanced loads, that is, the useful work. The resistances in the 
pits and in the machine are therefore only 4 to 9J per cent, of 
the useful work. It will be seen that these figures, which may at 
first seem rather small, agree with other tests. 

(c) Guarantee Trials. — The object of these trials is to determine 
whether the consumption in kw. per useful H.P. developed by the 
machine comes within that guaranteed by the contractor. The 
guarantee had been fixed by assuming a winding of 30 up-and-down 
journeys per hour ; each journey included a loading at the bottom 
landing-stage at a depth of 368 metres (1,207 ^^et), and a stoppage 
on the way at one of the intermediate loading stages, at depths 
of 342, 288 and 185 metres (1,122, 945 and 607 feet). 

Three series of trials were made on the 10th of January and 
repeated on the 30th of January ; each trial included 10 up-and- 
down journeys made as quickly as possible. 

Series I comprised journeys with loadings at 368 and 342 metres (1,207 f^et and 

1,122 feet). 
Series II comprised journeys with loadings at 368 and 288 metres (1,207 feet and 

945 feet). 
Series III comprised journeys with loadings at 368 and 185 metres (1,207 feet 

and 607 feet). 

At each of these trials, the following observations were taken 
every ten seconds : — 

(1) Readings of the Siemens wattmeter (the same as used at 
the no-load trials) which was inserted in the joined circuit of the 
Ilgner group and the motor-generator MgGg. These readings 
were plotted with time as abscissae, and the integration of the 
curve gave the total electric energy spent during the test. This 



June 1905. ELECTRIC WINDING-MACHINES. 469 

energy was also determined by means of the electricity-meter by 
taking readings at the beginning and at the end of the tests. 

The results obtained by these two instruments differ only 
1 to 3 per cent. 

(2) Eeadings of a Weston ammeter and voltmeter which were 
connected to the same circuits. From these data may be plotted the 
curve of apparent alternating watts ; and by dividing the watts 
found by the wattmeter by the apparent watts, a curve is obtained 
representing the variations of cos ^ of the installation. 

(3) Readings of the ammeter in the field-circuit. The latter 
was maintained at 110 volts. 

Eecording instruments give an account of the volts and amperes 
of the continuous current which drives the winding-motor. By 
multiplying the simultaneous volts and amperes, the watt-curve 
of continuous current is obtained, which represents the power 
spent at any moment during the winding ; and by integrating the 
corresponding parts of the curve, the energy spent during the 
different parts of the journey is determined. 

The variation of speed of revolution of the Ilgner group was 
shown by a recording speed-indicator. The variation of kinetic 
energy of the group, from the beginning to the end of each trial, 
had to be added (algebraically) to the consumption of the electric 
energy, found by means of the electricity- meter or the Siemens 
wattmeter, in order to determine the total expenditure of energy. 

The moment of inertia of the rotor of motor 

Ml 65 kg.-m^tres^ ( 1,543 Ibs.-ft.^) 

The moment of inertia of fly-wheel . . 1,000 ,, „ (23,730 „ „ ) 
The moment of inertia of the armature of 

generator Gi 160 „„ ( 3,796 „ „ ) 

The moment of inertia of shaft . . . 5 „ „ ( ir8 „ „ ) 



Total I = 1,230 „ „ (29,187 „ „ ) 



The variation of kinetic energy =(<_-_<) ^i^ ^""^^ " !£ ^ ^ ""' 

°'' 2 ^ 3,600 

Xl, 

where n represents the number of revolutions of the Ilgner group 
as indicated by the recording speed-indicator. 



470 



ELECTRIC WINDING-MACHINES. 



Junk 1905. 



Fig. 13. — Results from Three Up-and-Down Journeys. 
(Trial UK 30 January 1905.) 
(continued on next page.) 

I 1 



30 



25 



20 



<1> "^ 




u M 




Is 




i ". 




o > 




S '•B 


1-5 


II -^ 


f\ 15 




> 


rH ft 


'3 


m S 


l-H 


£ f 


-©^ 


O -M 


02 


B* o 


O 


IS 

p! <1 


°10 

GO 
O 


11 


-0) 

p. 


a 


-ta 


a 


d ^ 


o 


O r-J 


r-H 


"^ 


II 5 


oa -f= 


(> 


3^ 


S 


> w 


1-1 


o t^ 


43 


'" 1e 


s 


II ^ 


s 


> ^• 


pi 


I— 1 


d 


CQ ■*■' 




-»a -t= 


-+^ 


II 


1^ 


CO tn 




d pi 




O O 




Pl d 




fl PI 


5 


.r-t -rj 


-♦^ -i^ 




d d 




o o 




Q O 





A_ 



n 






ll / ll^ 

/ ! 






i ■"■^-^a.X'-T'i--. I 



■f-i 



^\ M \ 




AbscissfiB Bcale, 1 div. = 10 sees. 



Junk 1906, 



ELECTRIC WINDING-MACHINES. 



471 



Fig. 18. — Results from Three Up-and-Doivn Journeys. 
(Trial 11^ 30 January 1905.) 
(concluded from opposite page.) 




References. 
Continuous Volts. 

„ Amperes. 

Watts. 

Number of Revolutions. 

Alt. Watts. 

Alt. Volt-amp6res. 

Cos<j>. 

Exciting Current. 

Abscissae scale, 1 div. = 10 sees. 



472 ELECTRIC WINDING-MACHINES. June 1905. 

The author does not consider it necessary to publish all the 
figures and curves obtained by these trials, as their number is very 
considerable. Fig. 13 (pages 470 and 471) represents the results 
from three up-and-down journeys of trial 11^ made on the 
30th January. The trials comprised ten complete journeys from 
landing-stages at depths of 368 and 288 metres (1,207 ^^^ 945 f^^t). 
The performance of the Ilgner system is indicated by this diagram ; 
it will be noticed that the power absorbed by the winding-motor 
(continuous watts) varies considerably, namely, from zero to about 
300 kw., then falls to 200 kw. and returns to zero ; the power 
supplied by the power-station (alternating watts) varies only 
between 110 and 160 kw. 

The results which may be deduced from these trials will be 
found below, and from these have been determined the efficiency 
Rj = ratio of useful power to power supplied to the winding- 
motor (continuous watts), and efficiency Eg ~ ratio of useful 
work to the sum of the energy spent by the alternating current 
and the kinetic energy stored in the Ilgner group. The watts 
consumed per useful H.P. spent in drawing coal is determined by 
dividing 736 watts by Eg. 

The consumptions given above are slightly smaller than those 
guaranteed by the contractor. 

The consumption at the power-station of Esperance is 11 kgs. 
(24! lbs.) of steam per kw.-hour ; the consumption at the trials 
would therefore be between 17*75 and 20*3 kgs. (39*1 and 
44*75 lbs.) of steam per useful H.P.-hour, which is very small for 
a winding-machine. It is possible to produce a kw.-hour by a 
much smaller consumption of steam ; the consumption guaranteed 
for a steam turbo-alternator working at a pressure of 6*5 kgs. 
(14*3 lbs.) and superheated 300° C. (572° F.) is 8*95 kgs. 
(19*73 lbs.). Adopting this figure, one would require 14*40 to 
16*5 kgs. (31*74 to 36*37 lbs.) of steam per useful H.P.-hour, 
and one may even go below these figures. The electric winding- 
machine can therefore compete with the very best steam winding- 
machines. M. Henry, engineer to the Hasard Collieries, has not 
succeeded in reducing the consumption below 18*94 kgs. (41*7 



June 1905. 



ELECTRIC WINDING-MACHINES. 



473 



lbs.) of steam per useful H.P. -hour, notwithstanding the best 
known improvements.* 

It will be noticed that the efficiencies E^ are very high. If the 
mean value of R^, 0*855, (from the trials on 30 Jan.) be divided by 
the efficiency of the winding-motor, the reciprocal of this quotient 
will be the factor by which the useful power must be multiplied, 



! Trial. 

i 


Number of 
up-and- 
down 
journeys 
per hour. 


Ri 


R2 


Mean 

useful 

work per 

hour in 

H.P. 


Consumption 

per useful H.P. 

in kw. 


I 10 Jan. 


Windi 


ng from stages at depths of 368 and 342 metres 
(1,207 ^^^ 1,122 feet). 


28^ 


0-80-0 -89 


0-4513 


81-25 


1-63 


I^ 30 Jan. 


28i 


0-855 


0-453 


80-00 


1-61 


II 10 Jan. 


Winding from stages at depths of 368 and 288 metres 
(1,207 and 945 feet). 


25 


0-73-0-80 


0-413 


67-50 


1-78 


IP 30 Jan. 


29 


0-85 


0-429 


76-75 


1-715 


Ill 10 Jan. 


Winding from stages at depths of 368 and 185 metres 
(1,207 and 607 feet). 


28f 


0-68-0-88 


0-399 


64-25 


1-845 


lir 30 Jan. 


28| 


0-86 


0-406 


63-00 


1-81 



in order to obtain the power required by the winding-motor to 

overcome the useful loads and the passive resistances. If the 

efficiency of the winding-motor (neglecting the field-current) be 

0-855 r, a^^ A ^ 
^= 0-90 and ^,^0 



taken as 0-95, then we have qtS? =0*90 and ^"^ = 1*11 ; the 



* See R. A. Henry : " Etude th^orique et experimentale de la Machine 
d'Extraction." 'Revue Universelle des Mines.' 4th series, 1904, vol. vii. 



474 ELECTRIC WINDING-MACHINES. June 1905. 

passive resistances therefore do not exceed 11 per cent., which agrees 
fairly well with the figure already found above. 

(^d) Consumption of Energy under Working Conditions. — For the 
purpose of determining the consumption in kw. per useful H.P., 
the winding-machine at the pits of St. Nicolas has been submitted 
to a permanent set of trials, lasting from 17th January to 1st March 
1904. The consumption was measured by the electricity meter, 
inserted as in the guarantee trials. The readings were taken 
every day at 6 and 8 a.m., and at 3 and 6 p.m. The number of 
revolutions made by the Ilgner group were read off simultaneously 
with the readings of the electricity-meter, and a record was made 
of the coal and stones drawn during the different intervals ot 
time. By means of these data, the useful work done by the 
machine may be determined, as well as the mean work per 
hour. The consumption of electric energy, to which is added 
or subtracted the kinetic energy of the Ilgner group, admit o± 
determining the consumption in kw. per useful H.P. The large 
number of results from the trials, which lasted over a period of one 
month and a half, have been plotted. Fig. 14 (page 475), by taking 
as ordinates the consumption in kw. per useful H.P., and as abscissae 
the mean useful H.P. given off by the machine. The points lie 
evidently on an hyperbolic curve. During the night, when the 
winding is less, the ordinates are longer on account of the 
permanent consumption of the Ilgner group. The ordinates will 
be shorter as the winding becomes more active, and will be about 
2 kw. per H.P. between 8 a.m. and 3 p.m. The results from the 
guarantee-trials have also been plotted ; these points lie on the 
continuation of the curve. It may be remarked that the 
hyperbolic curve is similar to that representing the consumption 
of a steam-driven winding-machine.* 

The high consumption during the night may be attributed 
to the Ilgner system, and would therefore not take place if the 
winding-machine were driven by an asynchronous motor; but 

* See R. A. Henry, loc. cit. 



June 1905. 



ELECTRIC WINDING-MACHINES. 



475 







03 



o in 

Kilowatts per useful H.P. 



476 ELECTRIC WINDING-MACHINES. June 1905. 

then one would have to take into account the consumption of 
energy required to keep a dynamo at work at the power-station, 
which should be large enough to generate the maximum 
power required for starting the winding -machinery. As the result 
of the trials of the machine at pit No. II Preiissen, the author 
has shown that the variation of consumption of a winding-machine 
driven direct by an asynchronous motor follows also an hyperbolic 
law.* 

The electric winding-machine at St. Nicolas has replaced a 
steam-driven machine, which was put down in 1874. According 
to trials made on the 17th February 1903, the steam-driven 
machine consumed 53*1 kgs. (117 lbs.) of steam per useful H.P.- 
hour during the period of maximum winding from 8 a.m. to 3 p.m. 
By installing the electric machine, a battery of six boilers, 
working at a pressure of 5 atmospheres, has been abolished. 
These boilers were in such a bad state that they would soon have 
had to be replaced by new ones. Besides being economical in 
consumption, the electric machine has also caused a reduction in 
wages, as the stokers, who were dismissed from St. Nicolas, have 
not been taken into service at the power-station. As the electrically- 
driven machine is much simpler to manipulate than the steam- 
driven machine, the drivers (two in number) needed very little 
instruction before taking over the new machine. The consumption 
of the accessories is very small. 

The rope is a flat steel rope. The two ropes, which were used 
at first, consisted of six hawsers of four strands each, on a core 
of hemp ; each strand contained seven wires of 2 mm. (j^^ inch) 
diameter. The ropes were 92 mm. (3! inches) wide and 20 mm. 
(fi inches) thick, weighing about 5" 7 kgs. per metre (3 '8 lbs. 
per foot). The wires were of steel with a breaking-strength of 
145 kgs. to 150 kgs. (320 to 330 lbs.), and would bear fourteen 
bendings over a mandrel of 5 mm. (^ inch), and 30 torsions on a 
length of 200 mm. (7^ inches) before rupture. The total breaking 



* See Paul Habets : " Les machines d'extraction electriques." * Revue 
Universelle des Mines.' 4th series, 1904, yol. vi. 



June 1906. ELECTRIC WINDING-MACHINES. 477 

strength of the rope was 77,750 kgs. (76*5 tons). The first 
rope was removed after only 3 months of service, on account of 
unnecessary precaution. It was found by a tensile test that the 
total breaking strength of the weakest part of the rope was still 
66,000 kgs. (145,504 lbs.) ; the second rope lasted 4J months. 
It was replaced by a rope consisting of seven hawsers of four 
strands each, on a core of three wires of 1*2 mm. (^^2^ inch) 
diameter, each strand containing seven wires of 1 • 8 mm. (/^ inch) 
diameter. The rope, being 110 mm. (4IJ inches) wide and 18 mm. 
(IJ inch) thick, weighs about 5*75 kgs. per metre (3*8 lbs. per 
foot). The wires are of steel with a breaking strength of 130 kgs. 
to 135 kgs. (286*6 to 297*6 lbs.); the 1*8 mm. (-^^^ inch) wire 
would bear eighteen bendings over a mandrel of 5 mm. (|^ inch) 
and 30 torsions on a length of 200 mm. (7I inches). The 1 • 2 mm. 
(^^^ inch) wire would bear 36 bendings, and 30 torsions on a 
length of 150 mm. (5|| or 5*9 inches). The total breaking 
strength of the rope is 77,000 kgs. (75*8 tons). On account of 
the small diameter of the Koepe-pulley, it was desirable to give 
the new ropes more flexibility by making them thinner and of 
finer and more ductile wire. 

It is interesting to compare the cost, with regard to ropes, 
of the Koepe-machine with that of the old steam-driven 
machine with bobbins and flat aloes ropes. The latter ropes 
weighed about 4,600 kgs. (10,141 lbs.), and cost during later 
years from 1*65 francs to 2 francs per kg. (7 to 8"jid. per lb.). 
Adopting the lowest figure, a rope will then cost 7,600 francs 
(£304), and two ropes will therefore be 15,200 francs (£608). 
These ropes last about 26-32 months, say 30 months on an average. 
The rope expenses per month would therefore be 500 francs (£20). 
The ropes which were rejected cost 2,250 francs (£90), which, for 
a period of 4J months, is 500 francs (£20) per month, the same as 
was found for the aloes ropes. One may however expect to find 
the steel ropes more economical, as their construction is more 
suitable for the purpose, and may therefore last more than 4J 
months. A rejected winding-rope may be used as counter-rope. 



478 ELECTRIC WINDING-MACHINES. June 1905. 

It may safely be concluded from these trials that the electric 
winding-machine, even if it is not more economical than the best 
steam-driven machine, is certainly not more expensive. The 
greater facility and safety with which it can be manipulated, 
the smoothness with which it works, and its much greater 
flexibility, will often make it preferable to the steam-driven 
machine, even in the case where transmission of energy is not 
required. There can be no hesitation in the choice between the 
two systems, when the power has to be transmitted from a 
distance, or, as in the case of the pits of St. Nicolas, where the 
production of energy can be centralised at one power-station. 

The Paper is illustrated by 14 Figs, in the letterpress. 



Discussion. 



Professor Paul Habets said that in the Paper, which the 
Institution had been good enough to accept, he had not given all 
the data obtained from experience and during the tests that had 
been made with the electric winding-machines employed at the 
colliery of St. Nicolas. To do so, it would be necessary to publish 
innumerable tables and exceedingly voluminous drawings. He 
however proposed to put up in the engine-room, during the visit of 
the Members of the Institution to the colliery, all the graphical 
representations of results obtained. The Members would thus be 
able to realise the procedure, and would be better able to study 
the behaviour of these machines. 

M. Creplet described the electrical winding-plant laid down at 
the Hasard collieries at Micheroux by La Compagnie Internationale 
d'Electricite of Liege. This electrical winding - plant was 
calculated to extract 2 tons of coal from a depth of 1,313 feet 
at a maximum speed of about 16*3 feet a second. The feeding 
current disposed of was three-phase, 2,000 volts, 50 periods, produced 
by two units of 250 H.P. each. 



June 1905. ELECTRIC WINDING-MACHINES. 479 

Three suggestions, A, B and C, were successively considered, 
namely : — 

(A) Direct Tliree-Phase System.^The winding-machine being 
driven by a three-phase motor of 2,000 volts, and the alterations of 
speed being obtained by means of a starting-switch, either metallic or 
liquid. 

With such a driving-system, the power necessary for starting 
would have been 400 H.P. coming down to 200 H.P. during running, 
Fig. 15 (page 480). This system, however, was abandoned on 
account of the enormous variations in the power required, and 
subsequent losses in the starter. 

(B) Ugner System, Fig. 16 (page 480). — The winding-machine 
being driven by a continuous-current motor, which takes its current at 
a rotary transformer, regulation of speed being obtained by acting on 
the excitation of the transformer's generator. A fly-wheel is fixed on 
the transformer to render the demand for power at the central 
station uniform. According to the calculations established it was 
shown that the weight of this fly-wheel would have amounted to at 
least 15 tons. This second solution was rejected for the following 
reasons, viz. : — 

(1) — The excessive weight of the fly-wheel might have given 

rise to mishaps of dijBferent natures. 
(2)— If at any time something had happened to the fly-wheel — 

such as might have been occasioned by overheating of the 

bearings — it would have caused the machine to stop forthwith. 
(3) — On account of the rotary transformer turning incessantly, 

the power expended for maintaining the speed of the fly-wheel 

would have been considerable. 

(C) Continuous Current System (Creplet). Figs. 17 and 18 
(pages 480 and 481). 

Through the medium of La Compagnie Internationale 
d'Electricite of Liege, a continuous-current system was suggested, 
bearing many points of resemblance with the previous one, at the 
same time there being a great many differences. 



480 

(M. Creplet.) 



ELECTRIC WINDING-MACHINES. 



June 1905. 



Fig. 15. — Direct Three-Phase System. 





^^ 4-00 H.P 




r\ 




\ 200 H.P. 


. Vj 


\ 


5^ 


\ \ 


\_ 


^ 


1 

v 

J 



rvrrve 



F1G.U6. — llgner System. 



Fig. 17. — Creplet System. 






MmUng Motor 




rvih 




Wvndvrv^ Motor 



JUNB 1905. 



ELECTRIC WINDING-MACHINES. 



481 



(1) — The fly-wheel, in the present instance, was entirely 
independent of the rotary transformer ; it was coupled to a 
dynamo connected in series between the transformer's generator 
and the winding motor. Fig. 17 (page 480). 

(2) — The generator incessantly produced current, which caused 
its power to be invariable. Whilst stopping, this power was 
absorbed by the fly-wheel dynamo, which was accelerated 
accordingly ; whereas, whilst winding, the energy absorbed 

Fig. 18. — Creplet System. 



^Power of the Windlrv^-up Motor 



-Energy absorbed^ by the Fly-wheel 




Power prochccedy 
by the Gerurator 



Speed of the Ffy- wheel 



Ttrrve 



by the fly-wheel was conducted into the winding-motor, so 
as to be added to the energy produced by the generator. 
Fig. 18. 

(3) — With this system the variation of speed of the fly-wheel 
was enormous, inasmuch as a fall of speed of 50 per cent, 
was obtained, which, according to established calculations, was 
equal to a restoration of 75 per cent, of the kinetic energy of 
the fly-wheel. 

(4) — With the Ilgner system it has been observed that the 
variation of speed of the fly-wheel hardly ever surpassed 8 per 
cent. Hence the energy produced by the fly-wheel was only 

2 M 



482 ELECTRIC WINDING-MACHINES. JuNE 1905. 

(M. Creplet.) 

16 or 16 per cent. From this it may be deduced that with 
this system the weight of the fly-wheel required was reduced 
to one-fifth. 
The working of the machine necessitated the employment of : 

(1) — A bar or lever, reversing the direction of the current 
in the armature of the motor, also regulating the excitation 
process. 

(2) — A bar or lever, acting upon the excitation of the fly-wheel 
dynamo, so as to reverse it, by making it pass through nought. 

The engineer moved the first bar with the right hand, and the 
second with the left hand, hence it would be observed that the 
working was exactly the same as with an ordinary steam-winding 
engine. 

The Creplet system contained the following advantages over other 
systems : 

(1) — The fly-wheel being independent of the rotary transformer, 
it was possible to do the work without the aid of the latter. 
The working of the winding-machine was thus arrived at by 
regulating the excitation of the transformer's generator. At 
night-time, or when not winding, the power required was 
small. It being useless to render the discharge uniform, the 
fly-wheel was stopped. 

(^2) — On account of the great variation of speed of the fly-wheel, 
its weight might be reduced to one-fifth. 

(^3) — On account of the rotary transformer being able to turn 
with a constant speed, a synchronous motor could be used in 
order to improve the power-factor at the central station. 

(4^ — Owing to the slight weight of the fly-wheel, to the facility 
of doing the work without its assistance, and also to the 
complete absence of starting-switch, the current consumed was 
far less than with any other system. At a medium discharge 
of 20 E.H.P. in lifted coal, the trials have shown that the 
consumption did not exceed one 8-kilowatt-hour for 1 E.H.P. 



June 1905. ELECTRIC WINDING-MACHINES. 483 

The whole installation had been calculated for the power of 
80 E.H.P. For such an installation the consumption would go 
down to one 3-kilowatt-hour for 1 E.H.P. 

The winding-machine was started in February last, and had 
since been at work without interruption. The fact that since the 
laying down of the electric plant the boilers have not had to be 
re-lit, went far towards proving the excellence and absolute reliability 
of this plant, and its steadiness of working. The whole plant 
complete cost £3,000, and the economy, in fuel only, amounted to 
£440 a year. 

Mr. Walter Dixon said he had had an opportunity some months 
ago of seeing the plant which had been described as the Ilgner 
system, and he thought it certainly was of mechanical interest, and, 
under certain conditions, would probably be an advantage in Great 
Britain. The last remark in the Paper was somewhat disquieting, 
namely, that it might safely be concluded from the trials that the 
electric winding-machine, even if it was not more economical than 
the best steam-driven machine, was certainly not more expensive. 
It was also pointed out that where the transmission was over long 
distances the system might certainly work both economically and 
well, and perhaps that was the one condition where its use in Great 
Britain might be recommended. The system that had been explained 
by M. Creplet appeared to have advantages, and it would be well 
for those interested in the subject to investigate the matter while 
they were on the spot. 

The President thought that few Papers required more careful 
study than the present one, and he was sure they all would agree 
that Professor Habets deserved the best thanks of the Institution for 
the able Paper he had communicated. 



2 M '^ 



*4 



June 1905. 486 



FERRO-CONCRETE, AND SOME OF ITS MOST 
CHARACTERISTIC APPLICATIONS IN BELGIUM. 



By M. ED. NOAILLON, op Ch^inee, near Liege. 



(^Translated from the French.) 

Properties and advantages of Ferro-Concrete. — Ferro-concrete 
is a material which was unknown to the general public a few 
years ago, but has now entered with phenomenal rapidity into 
all branches of constructional work. This result is due to its 
remarkable properties, which may be stated as follows : — 

(1) The economy rendered possible by its use as compared 
with other competitive systems. 

(2) Its resistance to fire, which is now put beyond doubt by 
numerous tests, some made for the purpose and others the result of 
accident. It is moreover the only flexible material which possesses 
this quality of fire resistance ; and after the results of the disastrous 
fire at Baltimore it is clear that very little confidence can be felt in 
the use of metallic frameworks covered with thin coatings of 
refractory materials. 

(3) It is unaffected by atmospheric action. Concrete is from 
this point of view comparable with stone of the best quality and it 
improves with age. As to the metal built into the concrete, it has 
been proved that it is perfectly preserved without loss of weight, 



486 FERRO-CONCRETE. JuNE 1905. 

and that even if used in a rusty state it will recover after some time 
the bluish tint which it possessed when leaving the rolling mill. This 
almost incredible result is due to a chemical action of the cement 
and probably to the formation of a protective coating of silicate of 
iron. Concrete also resists equally well the effects of corrosive 
fumes and liquids which are feebly acid. It may be used for 
marine works if the proportion of cement employed be high. 

(4) The ease with which the material may be made to take 
any desired form. While preserving the architectural appearance 
of stone a boldness in construction may be attained which is 
impossible with the latter material. It is merely necessary to 
measure the materials precisely ; an error can be corrected during 
construction and unforeseen details can be improvised. This 
adaptability is specially valuable in dealing with existing structures. 

(5) Its homogeneity, and the mutual support which 
neighbouring parts give in resisting concentrated loads. Joints are 
no longer weak places. Girders which cross, pass through each 
other without a break. Monolithic constructions are rendered 
possible which are far more resisting than others to secondary 
stresses. 

(6) Its rapidity of execution. The constituent parts are 
merely raw materials requiring no previous preparation, and 
therefore procurable without delay. The individual importance of 
the single parts is negligible, thus rendering them easy to obtain, 
transport, and erect. Night work also does not occasion the noise 
caused by riveting. 

(7) Its impermeability, if it has been " floated," immediately 
after construction in a careful manner. Under such conditions it 
may be used in the construction of flat roofs, reservoirs, sewers, &c. 
The monolithic structure which is also watertight may be produced 
without crack or re-entering angle, so that it can be freely washed 
down with the hose. Such a structure is essentially hygienic. 

(8) Its great rigidity and the localisation of the effects of 
shocks. 



June 1905. FERRO-CONCRETE. 487 



General Statements. 

Concrete. — The material is an agglomerate of hard stones bound 
together with cement. As cement is expensive and contracts 
considerably, it is of advantage to use the least possible quantity by 
reducing the volume of the voids between the pieces of stone. This 
result is obtained by using a mixture of materials of different sizes 
such as gravel and sand. Moreover, in order to disperse the cement 
with great certainty equally through the mass it should be mixed 
with sand only. Instead of gravel, granite chips — the refuse of the 
quarries — may be used with advantage. Although these chips cause 
more voids, they nevertheless give a tougher product owing to their 
angular form, which increases the adherence. Granite dust may 
also be used instead of sand. The choice of cement is of the utmost 
importance, and in order to be quite certain it is desirable to use 
only Portland cement of a well-known brand. 

According to the nature of the work, the concrete contains from 
350 lbs. to 700 lbs. per cubic yard. 

In systems of construction where the metal framework can carry 
by itself the whole load and where the concrete is merely intended 
to protect the metal, to fix it, and to hold its different parts together, 
then the quantity of cement may be reduced to the minimum, and 
the gravel may be replaced by clinker or by coke breeze. 

The concrete is far weaker than the other, but it is less expensive, 
lighter, more refractory and more sound proof, and nails can be 
driven into it. 

Metal. — At the present time the metal almost exclusively 
employed is mild steel, with an ultimate tensile strength of 27 tons 
per square inch. This costs no more than iron, and has the 
advantage of possessing a greater tensile strength and a higher 
coefficient of elasticity than the latter metal. 

Round bars are generally used, as they facilitate the escape of air 
and the proper ramming of the concrete ; they also possess no sharp 
angles which would cut the concrete, but, on the other hand, the 



488 FERRO-CONCRETE. JuNE 1905. 

round section gives the lowest coefficient of adhesion for a given 
cross-section of metal. 

Centering. — The construction of the centering is the most 
important part in the employment of structures in ferro-concrete. 
It takes up the most time, and seriously enhances the cost of the 
work. In the design of the centering the contractor has an 
opportunity to exercise all his ingenuity ; to use wood which can be 
again employed, and to avoid cutting the wood into short lengths 
and so causing waste of material. If vaulted forms have to be 
constructed, the cost of the centering may be greater than that of the 
ferro-concrete itself. 

Certain systems reduce or even obviate altogether the use of 
centering by the employment of metal work of sufficient strength or 
pieces of concrete specially made for the purpose. Mouldings as a 
rule are roughly formed in the concrete by centering, and then 
finished in gauged work ; but the latter is a difficult process, for the 
neat cement takes some time to set and is not sufficiently plastic. 

Deflection. — It is not possible by a simple examination to ascertain 
the strength of a finished structure in ferro-concrete, for the metallic 
members are no longer visible and their precise size and position 
cannot be gauged. 

The only method is to measure the deflection of the structure 
under given loads. The results obtained are, however, not precise, 
and useful information can only be gained by comparing similar 
structures. The deflections of structures in ferro-concrete are much 
less than those which would be given by a structure of equal strength 
built of wrought-iron ; for when concrete is stressed up to its 
elastic limit its deflection is less than that of iron under similar 
conditions. 

Principles of Construction. — Professor Eabut has summed up in 
the following six rules the principles which experience and theory 
recommend should be followed in the construction of ferro-concrete 
buildings. 



June 1905. FERRO-CONCRETE. 489 

(1) No connections should be made of iron to iron, as the 
concrete itself holds the parts together in the most economical 
manner. 

(2) At least two distinct systems of reinforcement should be 
used, the one system to take up the tensile stress, and the other to 
take up the shearing stresses in the concrete ; when it is necessary, a 
third system should be used to take up the compressive stresses. 

(3) To so arrange the reinforcement that the separate members 
may be stressed in the direction of their length, so that the stresses 
produced between the iron and the concrete shall be tangential and 
not normal to the axis of the members of the reinforcement. 

(4) To profit by all means of increasing the homogeneity of the 
various parts of the structure. This may be done by prolonging the 
iron parts of one portion of the structure into the thickness of the 
concrete of the adjoining portions, at a negligible cost ; while the 
construction of rigid joints in a metallic structure is very expensive. 

(5) On the other hand, advantage should be taken to the utmost 
extent of the homogeneity thus obtained to economise materials. 

(6) In view of this homogeneity sudden alterations in the cross- 
section of the parts should be avoided, as the parts tend to assist one 
another and to distribute the stresses, the constitution of ferro- 
concrete being, so to speak, democratic. 

The component parts of a ferro-concrete structure may be classed 
under three headings, which the author will examine in turn. 
They are (1) the parts which resist tensile stresses ; (2) those which 
resist compressive stresses ; and (3) those which resist more complex 
stresses. 

Parts Kesisting Bending Stresses. 

These comprise chiefly the beams and the platform beams. In 
the majority of cases the beam supports a platform which is solid 
with it, and can therefore be used as a framework in compression ; 
and this is one of the most characteristic properties of ferro- 
concrete. The beam is therefore really composed of the rib and the 
part of the platform on each side, and this has a cross-section in the 



490 FERRO-CONCRETE. June 1905. 

form of a T. The tension member consists of one or of several 
metai bars embedded in tbe lower part of the rib. 

The materials are therefore used in the most rational manner, the 
concrete of the platform is subject to compression and the metal 
resists tension ; but this specialization of work is only possible 
owing to the adhesion of the concrete to the metal. In reality there 
are forces tending to produce sliding, and these are proportional to 
the shearing forces, and therefore attain their maximum value near 
the supports. These forces tend to shear the concrete of the rib, and 
are concentrated at the contact surface of the concrete and metal. 
The mutual adhesion of these two materials has a very great 
importance, and it is well to bear this in mind. 

Numerous tests have shown that, in a carefully built structure, 
this adhesion is not of lower value than the shearing coefficient of 
the concrete itself. If however the concrete be very poor in cement, 
or if it has been gauged too dry and insufficiently rammed, then the 
adhesion may be low and the use of bars of special section has 
advantages, these bars having projections which prevent all slipping 
of the metal in its concrete sheath. Such bars are very commonly 
used in American practice. 

It has been stated that the adhesion was illusory, and that in 
reality the effect was merely due to a high coefficient of friction 
between the iron and the concrete which compressed it in shrinking. 
But the fact that beams subject for lengthened periods to incessant 
vibration, such as those in the floors of flour mills, have remained 
sound tends to prove that the adhesion is real and lasting. Special 
tests have always given reassuring results, except when they have 
been made upon flat contact surfaces ; but this is a condition which 
does not occur in practice, and it is probable that during setting the 
contraction of the cement produces tangential stresses which destroy 
adhesion as it is produced. In the case of cylindrical surfaces, on 
the contrary, this contraction produces compression normal to the 
axis of the cylinder and therefore favourable. 

Adhesion does not only assist in resistance to shearing stresses, 
but the variations in relative volumes of the two materials in contact 
must be considered, these variations being caused by change of 
temperature or by shrinkage of the concrete. 



June 1905. FERRO-CONCRETE. 491 

Temperature has no influence, for the coefficients of expansion of 
iron and concrete are practically the same. When concrete sets in 
air it contracts, and therefore may produce considerable initial 
compression upon the embedded metal whilst being itself subject to 
an equivalent tensile stress. 

When the concrete sets under water the opposite effect is 
produced, the concrete expands and puts the metal in tension. 

It is important that the shearing stress should neither overcome 
the adhesion of the metal to the concrete nor shear the rib of the 
beam. 

To fulfil the first condition, it is necessary to form the metal 
framework of such a number of bars that their surfaces in contact 
with the concrete shall be large enough, and in consequence the 
chance of surface slips shall be reduced. To avoid shearing of the 
rib, special stirrup-shaped bars are used which join together the two 
members. It is not correct to state that these stirrups directly 
resist the shearing stress. In reality, a piece which is under 
shearing stress throughout its entire length is by that stress subject 
to bending, but the stirrups have no rigidity and are incapable of 
resisting any appreciable bending moment ; as a matter of fact, they 
fulfil the same purpose as the tension bars in the web of a lattice 
girder ; the duty of the compression bars is fulfilled by the concrete 
of the rib. It is therefore obvious that it is essential for the stirrups 
to be hooked at one end to the tension bar, and that at the other end 
they should be solidly imbedded in the concrete platform. Fig. 1 
(page 492) shows the transverse section of a beam built on the 
Hennebique system, and beside it the drawing of a single stirrup. 
This stirrup consists of a flat bent bar with the two ends bent in the 
form of claws, allowing it to hook itself solidly into the floor structure. 
The use of the flat bar in preference to the round bar facilitates the 
construction. Fig. 2 (page 492) represents a cross-section of a beam 
on the Coignet system. Here there is an upper iron framework. 
Owing to this it is possible to put the framework of a beam together 
in advance and to place it in position entire while holding it by the 
upper bar. The attachments are made of bars of round iron bent 
to U section, and the ends are twisted together so as to form an 



492 



FERRO-CONCRETE. 



June 1905. 



elongated ring, Owing to the presence of the two frameworks the 
attachment binds the two members together very effectively. The 
round section is better suited to the concrete than the rectangular 
form. 

Another means of resisting sliding consists in omitting horizontal 
bars in tension, and substituting bars fixed obliquely in the webs and 
rising to the floor structure at the ends of the beam. In this manner 
a beam of variable height is obtained approaching more or less 
closely to the parabolic form, that is to say, the ironwork will be 
subject to a constant tension upon its entire length, and the 
shearing stress will be zero, as it is neutralised by the vertical 
component of the oblique tension in the bar. In this manner, 



Fig. 1. 

Section of Beam. 
Hennehique System. 



Stirrup. 



Fig 2. 

Section of Beam. 
Coignet System. 




"^ff 



w 




U 



"^^ 



however, the difficulty is only set back a step, for the bar being in 
tension right up to the ends must lose it rapidly in a very limited 
space, whence arises a considerable tendency to slipping, which is 
met either by opening out the end into a swallow-tail, or, if this is 
insufficient, by bending the bar upon itself and placing a cross-pin 
in the bend. The simple oblique bar is not used for a beam with 
free supports. 

The author has pointed out that one of the characteristics of 
concrete is to lend itself easily to continuity of structure ; but it 
is clear that this may have the effect of displacing, towards the 
bottom, the diagram of the bending moments in such a way that 
near the supports certain reversed moments may be of higher value 



June 1906. FERRO-CONCRETE. 493 

than the moment at the centre of the span. From this cause tensile 
stresses are produced in the top member and compression stresses in 
the bottom member. In order to overcome the first set it is 
necessary to provide new reinforcement if it does not exist in the 
upper member, or raise the lower reinforcement towards the top of 
the beam, which has also the advantage of neutralising the tendency 
to sliding, as has already been pointed out. 

To overcome the second series the concrete of the lower part of 
the web is often insufficient, and then the lower reinforcement is used 
at this point in compression. Another method which is coming into 
use consists of banding the concrete of the web, which, as will be 
pointed out hereafter, considerably increases the resistance to 
crushing. 

When the reversal of the moments is due to the continuity of the 
beam and the floor platform, the use of reinforcement in the upper 
part is not always necessary ; in fact, the platform may then play 
the part of the member in tension, owing to the " distributing " bars 
which it contains, and to the resistance to tractive efforts of the 
ferro-concrete itself, which the author will discuss later. 

When the beams have no continuous platform at their upper part 
they ought to be provided with compressive reinforcement or be 
banded. 

Fig. 3 (page 494) is a longitudinal section of a portion of a beam 
on the Hennebique system. It will be observed that half the bars 
composing the lower reinforcement rise to the top at the supports, 
resisting at the same time the tendency to slide and also the moment 
of shearing at the junction. 

The other half keep their position right to the end, where they 
serve to resist the compression and to support the stirrups. 

Fig. 4 (page 494) represents a portion of abeam built upon the system 
of Perraud and Dumas. The attachments are formed of a trellis of 
flat strips bent obliquely between the lower reinforcement and a 
bar at the top, so that the body of the beam forms one mass of great 
rigidity which is of value in the erection. The lower ironwork is 
reinforced in its central part by a bar in order to obtain as nearly as 
possible a section of canal resistance. This method has, however, the 



494 



FERRO-CONCRETE. 



June 1905. 



disadvantage of giving a minimum perimeter to the section of the 
ironwork at the supports, which are precisely the points where it 
should have its greatest value. At the level of the platform there 
are bars to resist shearing at the supports and compression bars in 
the centre. 



Beams. 



Fig. 3. 

Hennehique System. 



Fig. 4. 
Perraud and Dumas System. 




Fig. 5. 

Autlior's System. 




The author of this Paper desires to advocate a simple system 
which affords a reinforcement of canal resistance and offers every 
security for resistance to sliding. Fig. 5 shows its application to a 
beam without fixed ends. For a span of moderate amount the 
reinforcement consists of bars of three different lengths. 

It will be seen that they leave the lower part of the web at 
determined distances and rise towards the floor structure, where they 
end in a hook. In these hooks, as well as in the angles formed by 
bending, there are placed transversely pins a and h. These pins are 



June 1905. FERRO-CONCRETE. 495 

of great importance, for they represent the placing between the 
oblique parts of bars which may be considered as the diagonals of 
the web of a lattice girder, and the concrete lying between a and h 
which is analogous to the compression bars. 

The reinforcement of a beam with platform is calculated 
according to the nature of its central transverse section where the 
maximum bending moment occurs. 

As the sliding efforts there are zero, the normal sections remain 
flat after deformation and merely swing upon the neutral axis. This 
condition of conservation of the plane sections, combined with the 
equations of equilibrium of movement and of moments, suffices to 
determine the forces acting upon the reinforcement. 

But for this it is necessary to know the law which connects the 
compression of the concrete to its decrease in length. This law, 
however, varies within wide limits according to the composition of 
the concrete, and especially with the manner in which it is used in 
the work, the degree of fluidity, and the ramming. Moreover, the law 
is not linear, and is greatly affected by hysteresis, so that the 
deformation is a function of the duration of the compression and of 
the former life history of the concrete so far as stresses are concerned. 

Yet when the height of the beam is great relatively to that of the 
floor platform, the variation of the law of deformation of the concrete 
does not sensibly affect the forces acting upon the reinforcement, and 
the section of the latter may be calculated with sufficient accuracy. 

It should be noted that the elements of the floor platform which 
constitute the reinforcement in compression have an efficiency which 
diminishes as their distance from the web increases and as the 
length of the beam decreases. 

The Floor Platform. — When in a beam of T section the part of 
the concrete in tension in the web has merely an insignificant 
influence and may be neglected, it is desirable to inquire if it is 
not the same with the concrete in tension in a floor platform which 
has a section larger than that of the concrete under compression. 
M. Considere has shown that ferro-concrete can extend without 
cracking ufttil the reinforcement rea.chQ8 its limit of elasticity, and 



496 FERRO-CONCRETE. June 1905. 

that tlie part of the resistance due to the concrete attains after 
the first extensions a value which remains constant until rupture 
occurs. It is evident that if this result is to occur the piece 
must be free from all cracks before the test, and to obtain this 
condition special precautions must be taken in its manufacture and 
setting. Excess of water must be avoided in gauging, ramming must 
be carefully done, and the piece must be kept moist during the first 
days of setting. 

In practice these conditions are never fully realised, and therefore 
the rapid shrinkage of the cement produces premature tensile stresses 
which cause hair cracks. But these cracks never affect any great 
depth of the floor structure, and many contractors think that one 
may therefore count upon almost the whole of the concrete in tension 
to relieve the reinforcement ; others, on the contrary, believe that it 
is always dangerous to count upon this aid. 

The tension of the reinforcement of a platform will also vary 
with the law of deformation which it may be thought desirable to 
apply to compressed concrete, a law which is all the more uncertain 
as concrete is not homogeneous ; in fact the last layer of concrete is 
not generally rammed so as to facilitate levelling. 

It is therefore obvious that the theoretical calculation of floor 
structures can only have a relative value, and that the coefficients 
which are used in these formulae ought to be obtained from practice. 

This uncertainty in calculation is besides often hidden by the 
want of precision with which the reinforcement can be fixed in the 
thickness of the platform. 

Like the beams the platforms are very often fixed or continuous, 
and this condition requires reinforcement in the upper layer, or in a 
simpler manner by the raising of the lower reinforcement near the 
supports, Fig. 2 (page 492). 

With respect to resistance to sliding, as the perimeter of the 
section of the reinforcement is almost always less than the width of 
the platform, it is useless to provide stirrups. 

Besides the principal reinforcement there are placed immediately 
above it, and in a perpendicular direction, some " distributing " bars 
which are intended to interlace with several of the principal bars, and 



Junk 1905. FFRRO-CONCRETK. 497 

offer a better resistance to a concentrated load. In order to ensure 
regularity of erection, the two systems of bars are sometimes bound 
together at some of the crossing points by means of iron wire. 

When it is a question of covering a surface approximately square 
— a case which occurs frequently in houses — there is a great 
advantage in making the two systems of bars of equal strength, so as 
to make use of all four sides of the space as supports. 

The uncertainty of the conditions of stress upon the reinforcement 
which is so great for the ordinary floor platforms is still more so for 
the square surfaces referred to above. 

The reinforcement may be fixed in lines parallel to the sides or 
parallel to the diagonals, the use of the latter system, which 
necessitates a large number of different lengths of bars, cannot be 
justified by any theoretical consideration ; a satisfactory system 
consists in placing the reinforcement parallel to the sides and 
grouping the bars closer together along the middle of the sides. 

A simple and elegant arrangement for surfaces of small dimensions 
consists in the use of expended metal as the reinforcement ; perfect 
security is then obtained for the necessary adhesion, and a poor 
concrete may be employed. This method of construction can be 
recommended for foundation blocks. 

Pieces for resisting Compression Stresses. 

Columns. — Columns are usually made of square section to 
facilitate centering ; as a rule the reinforcement consists of four 
longitudinal bars placed near the four angles, and joined at intervals 
by cross ties. The ties are intended to prevent the buckling of the 
individual bars, and also serve to retain them in position during the 
filling of the concrete. The cross ties consist of plates or round 
bars which surround the longitudinal bars. Sometimes they are 
twisted round each bar, or plates with holes may be threaded upon 
the bars. 

The calculation of the reinforcement is empirical, for the 
determination of the strain is first of all affected by the uncertainty 
regarding the law of deformation of the concrete, since it is directly 

2 N 



498 FERRO-CONCRETE. June 1905. 

proportional to the latter. Besides this the shrinkage of the cement 
communicates to the metal an initial compression which is often very 
considerable, but impossible to value beforehand. Moreover the bars 
also assist in resisting the tendency to buckling of the entire column, 
but the buckling is proportional to the coefficient of elasticity of the 
column, and this value varies with the load, and is very diflferent for 
the same load according as the preceding load has been greater or 
less than the load under consideration. 

The resistance to buckling and to crushing is much increased by 
the use of a concrete very rich in cement, gauged with only a small 
quantity of water and well rammed. 

Banded Concrete. — The methods of reinforcing concrete against 
compression have been revolutionised by the introduction of the 
banded concrete invented by M. Considere. 

The very remarkable work of M- Considere has shown that 
metal employed under the form of bands has, from the point of 
view of resistance to compression, an efficiency 2*4 times greater 
than that used in the form of longitudinal reinforcement. To be 
effective these bands ought to be circular, and should not be placed 
further apart than one-seventh of the diameter of the banded 
concrete. 

By submitting the banded pieces to an initial compression, or 
letting them harden under water, the bands become also much more 
efficient than longitudinal bars in resisting buckling. They have 
also the considerable advantage of being endowed with an absolutely 
surprising plasticity, which makes a thoroughly sound structure. 
Warning is also given of excessive compression by the shelling 
off of the concrete covering the bands long before the structure has 
reached its safe limit of compression. 

The simplest and most economical form of the bands is the 
spiral. By putting the turns of the spiral closer together it is 
possible to strengthen certain parts which are under the greatest 
stress, and give a known increase to the strength. It is also possible 
to reinforce a structure already completed by winding round it an 
iron wire which is afterwards covered with a layer of plaster. 



Junk 1905. FERRO-CONCRETE. 499 

Banded pieces prepared in the workshop with all possible care, 
and made with a very rich concrete and subjected to initial 
compression, ojffer a resistance to compression comparable with that 
of a structure of riveted steel of the same weight, and are less 
expensive. 

It is safe, therefore, to prophesy that banded concrete will be 
able in a large number of cases to replace with advantage the 
compression members made of rolled steel sections in a large metallic 
structure. The tension members would consist of bundles of round 
bars lightly banded and dipped in concrete. The connections would 
be more easily made than those of other frameworks, and could be 
reinforced by a supplementary banding. Such structures would 
have an unlimited life with practically no expense for maintenance. 
In this connection M. Considere has constructed and tested with 
complete success a bridge of 65 • 6 feet span entirely built of banded 
concrete. 

An important application of banded concrete has come into 
ordinary use for piles. Such piles are stronger than those of wood, 
and have the enormous advantage that they do not rot. 

Parts Submitted to Complex Stresses. 

The most important type of these pieces m the vault, in which 
the reinforcement is specially designed to resist bending stresses 
caused by non-symmetrical or concentrated loads, whilst the concrete 
supports in compression those that are fixed and distributed. 

The greatest uncertainty exists as to the most rational method of 
reinforcing a concrete vault ; the calculation of the reinforcement 
is only rendered possible by the use of empirical coefficients. 

For vaults of wide span one observes that different builders 
follow opposite methods ; some, like Hennebique, thinking that the 
essential character of ferro-concrete is the opportunity which it 
gives for producing monoliths and of profiting in all cases of possible 
bonding, form their vaults of a series of ribs in arcs connected at their 
upper parts by a horizontal platform ; the arc has then near the 
supports its maximum of height, and this permits a good bonding 

2 N 2 



500 FERRO-CONCRETE. June 1905. 

with the abutments. Other builders seem to be haunted with the 
fear of the effect of expansion and of shrinkage, and give to their 
vaults as much flexibility as possible by making them with a 
cylindrical lattice work. In recent years some have gone further in 
this direction and formed arches with three joints. 

As has been seen from the preceding remarks it is most 
frequently impossible to determine accurately by means of theoretical 
formulae the actual cross-sections of the reinforcement. 

The best guide is experience translated as well as possible into 
empirical rules. These rules will be all the more reliable according 
as they are based upon a larger number of structures, and from this 
point of view it is best to prefer the systems which have been the 
most employed in practice. 

On the subject of principles of calculation for structures in ferro- 
concrete, it is interesting to cite the opinion of Professor Rabut : — 
" I often hear it said that structures in ferro-concrete cannot be 
so accurately calculated as metallic structures ; in my opinion the 
contrary is true; the formulae for metallic bridges are in their 
principles just as arbitrary and just as far from the expression of 
the real strains as those for ferro-concrete ; but the latter have the 
advantage of containing twice as many constants, those of iron and 
those of concrete, which, if these constants are conveniently chosen, 
will allow of approaching the truth much more nearly." 

The author will now briefly describe some types of constructions 
in ferro-concrete which present interesting features, and will restrict 
his choice to recent works and preferably to those to be found in the 
neighbourhood of Liege. 

Dome op the Central Railway Station at Antwerp. 

This dome, Plate 18, springs from the level of the roofs of the 
mass of buildings of the station at a height of 130 feet above ground 
level and rises another 130 feet to the spire. It is entirely constructed 
of ferro-concrete by the firm of Vasanne of Brussels. The work 



June 1905. FERRO-CONCRETE. 501 

was originally intended to be built in stone, but it was discovered 
that the foundations would not carry such a weight, and therefore 
ferro-concrete was preferred, as it could be built hollow. 
However, it was necessary to follow minutely the form of the 
original design, which, rationally conceived for a massive material 
like stone, often presented serious difficulties of execution in 
ferro-concrete. 

The dome comprises four large glass lights placed upon the sides 
of a square, and upon these rests the actual dome which in its turn 
supports a campanile. Each glass light is in the form of a gallery 
with seven arcades surrounded by a demi-arch of 32*8 feet radius. 
The arches are framed by an archivault of 11-6 feet height, which 
receives at its periphery the haunches of the dome. 

The whole structure, which is 1,800 tons in weight, rests entirely 
upon the columns at the angles of the glass lights, for it was only 
at these points that a solid support could be obtained. These 
columns are Y shaped in the cross-section which has an area of 
10*7 square feet, and they are subdivided at the height of the 
centres of the arches into three beams. 

The tail of the Y is extended in the diagonal plane in the form 
of a thrust-block rising obliquely between the two shells of the 
dome. Each of the limbs of the Y forms the abutment of the beams 
in the arch, 8 * 2 feet high, situated in the archivault. In the horizontal 
plane passing through the tops of the archivaults is placed a beam 
in the form of a flat ring 4 • 92 feet wide, which is supported at eight 
equidistant points, which are the four tops of the archivaults and the 
four ends of the thrust-blocks. 

This beam serves two purposes — it balances the horizontal 
reactions due to the obliquity of the thrust-blocks, and resists the 
tensile stress created by the joists of the dome. At the top of each 
beam of the archivault are hooked two tie-rods which go down in 
the two midribs of the arches and extend into the two central 
columns of the gallery of arcades, and support in its place a horizontal 
beam hidden in the entablature of the gallery, and supporting all 
the lights ; it is obvious, therefore, that all the weight of the latter is 
supported by the columns. 



502 FERRO-CONCRETE. JuNK 1905. 

The dome consists of two superposed shells at a distance apart 
varying from 3*28 feet to 6-56 feet. The internal shell which 
forms the ceiling of the entrance hall is completely decorated with 
sunk moulded panels ; some are round and others square, and they 
diminish in depth and size progressively from the springing to the 
summit. They leave only flat bands on the inside of the shell, and 
these follow a series of meridians and parallels. Some of these flat 
bands are formed by a skeleton of joists and trimmers which are 
supported on the annular beam, and they carry the whole weight of 
the dome. This skeleton was first erected, and served to support 
the cores which formed the moulds for the panels, and then the latter 
were filled in with concrete. 

The external shell has a uniform thickness of 3 '15 inches, and 
it is relieved by six moulded ribs following meridian lines. It is 
supported upon the internal shell by small distance pieces normal to 
the two surfaces ; this method of support has been chosen to allow 
as much freedom as possible for the unequal expansion of the shell 
owing to the rays of the sun striking it obliquely. 

The most interesting feature of the construction of the dome 
of the Antwerp station is that all the mouldings and all 
the sculptures, which are so numerous and of so many different 
forms, have been executed by direct moulding, and not, as is usually 
the case, by rough applications which are trued up afterwards by 
gauge-boards. 

It would never have been possible to make in wood the numerous 
moulds which would have been needed for the latter process, 
particularly as the work had to be done upon surfaces bent often in 
two directions like the panels of the dome. 

M. Vasanne has invented a very ingenious system of moulding. 
He begins by executing in plaster the model of the sculptures 
which are to be reproduced in the concrete. He then spreads 
upon this negative mould a layer of 1*2 to 2 inches thickness 
of a paste made of sawdust and magnesium oxychloride. This 
paste hardens rapidly and gives him the desired mould, which is 
light, strong, and can be worked like wood ; the same negative 
mould can be used several times. 



Junk 1905. FERRO-CONCRETE. 503 

The manufacture of the glass lights was specially difficult, 
because of the great richness of the ornamentation and of the 
exactitude with which the moulds had to be made to obtain a 
perfect fitting of the different mouldings. The part of the glass 
lights which forms a gallery was filled with concrete at one 
operation in a mould built up in position on the site. With 
regard to the part in the form of a half rose, a complete mould 
was arranged upon a - perfectly level platform. This mould 
showed, therefore, in hollow all one face of the half rose. The 
concrete was moulded to a thickness of 2 inches, and then the 
moulded part was cut into portions specially marked. Then the 
moulding, followed by the same division into parts, was repeated to 
obtain the opposite face. In this way two corresponding portions 
placed back to back formed one hollow structure representing one 
element of the glass light. The parts were then used to build up 
the light, just as stones would have been employed, but care was 
taken to pass bars of iron into the hollow part, and then to pour in 
concrete, so as to obtain a monolithic structure which possesses great 
rigidity. 

Benommee Hall at Liege. — This hall, Plate 19, was built entirely 
of ferro-concrete by the firm Perraud and Dumas of Brussels. 

In opposition to what took place in the design for the dome at 
Antwerp, the general arrangement, the style and proportions of all 
the parts of the building were specially thought out by the 
architect, M. Jaspar, so as to be the most suitable for construction 
in ferro-concrete, and in order to use the properties of that 
material to the best possible advantage. It was desired to oppose 
the tendency to make concrete merely play the part of a servile 
imitator of stone, by the employment of a characteristic design 
which should indicate the nature of the material used. 

The principal hall is covered by three cupolas, each 55 feet 
diameter, placed at a height of about 50 feet above the level of the 
ground. Each cupola forms part of a sphere which continues in 
haunches pierced with lights and descending to the corners of the 
circumscribed square. The intersections of the spheres with the 



504 FERRO-CONCRETlE. June 1905. 

vertical planes passing through the sides of the squares are formed 
by arched beams which spring from the capitals of short cylindrical 
columns. The cupolas are 4J inches thick and are made of concrete 
composed of cement and clinker finely broken up ; they are reinforced 
by a layer of expanded metal and with a lattice work of bars. The 
centering of the first cupola was carried out upon a new design. 
In order to avoid the great expense entailed by the construction in 
wood of a spherical centering, a skeleton wgis built up of ironwork 
consisting of 16 bars, each IJ inches diameter, fixed upon the 
meridian lines like the ribs of an umbrella, and these were interlaced 
upon parallel horizontal circles by other weaker bars. The 
whole skeleton was then covered with sheets of expanded metal, 
which were designed as the first reinforcement, and afterwards the 
concrete was put on above and below so as to surround the expanded 
metal completely, which thus acted as its own centering, and it 
was merely necessary to render the surface up to the required 
thickness. 

The bars of the skeleton were then removed and used for the 
other cupolas, and they were finally intended for reinforcing the 
beams. Unfortunately this system of centering was found to be 
wanting in rigidity, and it was necessary to use the wood centering 
after all. 

The roof of the galleries and the spherical triangles between the 
cupolas form a terrace of 957 square yards area, which serves as a 
promenade. The concrete of the cupolas and the terraces is not 
rendered in cement, it is made watertight by a large, layer of 
rubberoid. 

The principal hall is lighted upon its two long sides by six 
semi-circular glass lights, each 62 • 5 feet diameter, framed by arched 
beams. The spandrels are formed by panels of ferro-concrete 
showing on the inside ornament in relief ; the moulding was done 
in the workshop, and then each panel was cut into portions that were 
erected in position. 

In spite of the complete absence of mouldings, which were left 
out to facilitate the centering, the hall is of most elegant appearance 
owing to its satisfactory design. 



Junk 1905. 



FERRO-CONCRETE. 



505 



Widening of the La Boverie Bridge at Liege. — This bridge is 
147 yards long. The width has been increased from 32-1 feet to 
46*3 feet by means of two platforms corbelled out at each side of the 
bridge. At each end these platforms are extended so as to meet the 
quays by wide quadrants. 

The corbelling, Fig. 10, is carried out by a platform 5*9 inches 
thick, supported at intervals of 6*56 feet by brackets 11-8 inches 
thick by 19*7 inches high at the back. The method of fixing 
these brackets is interesting. Cutting and pinning brackets into the 
old structure presented serious difficulties, as it would have been 



Widening of La Boverie Bridge, Liege. 
Fig. 10. — Vertical Section through the centre of a Bracket. 




^\\\\\\\^^^^^^^ 



necessary to cut deeply into the ashlar work and to destroy the 
keystones in the arches, which would have been very unwise. The 
author of the design, M. Prangey, engineer of roads and bridges, 
got over the difficulty in a very satisfactory manner by fixing the 
brackets in pairs opposite one another. The brackets were made 
independently two months before they were needed ; the three round 
bars, 1*22 inch diameter, which formed their reinforcement in 
tension, were left projecting beyond the concrete about a yard, and 
the ends were screwed. The brackets were put in position, a pair at 
a time, by means of two cranes, and as soon as their bases had been 
entered into the recesses cut in the masonry of the bridge, the ends 



506 FERRO-CONCRETE. June 1905. 

of the three pairs of bars were coupled together by means of three 
tie-bars which crossed the roadway, and these tie-rods were provided 
with long sleeves threaded to fit the ends of the screwed bars. In 
this way each pair of brackets was mutually self-sustaining, and 
the cranes could leave them fixed. It was merely necessary then to 
level them up carefully by tightening or slackening the screwed 
sleeves, and then to pin them permanently in position by running in 
cement grouting round the ends of the brackets in the masonry. 

All the tie-bars are embedded in a layer of concrete 6 • 3 inches 
thick, which extends across the roadway and serves to constitute the 
gutter to carry off the water which percolates through. 

The platform, which is supported by the brackets, was calculated 
for a distributed load of 80 lbs. per square foot, and was tested up 
to 160 lbs. per square foot without suffering any appreciable 
deflection. 

The Bridge at La Boverie Island over the Derivation at Liege. — 
This bridge, Figs. 11 and 12, Plate 20, was built upon the Hennebique 
system by the Dulac Company, under the direction of the engineer, 
M. Prax. The chief details of the work are: length between 
abutments 260 feet, comprising a central span of 180 feet and two 
side spans each 32 • 8 feet wide. The total width of the roadway is 
32 • 8 feet. The span of 180 feet is the full width of the river-bed ; 
the soffit of the arch is an arc of a circle with a rise at the crown 
of about 12 feet, or a proportion of rise to span of about 1 in 85. 

One of the special features of the bridge is that it is built upon 
foundations constructed by a comparatively new process — the method 
of mechanical compression of the soil. The piers and abutments 
rest upon a group of concrete piles driven deeply into the bed of 
gravel, which is strongly compressed by the method adopted. The 
piles are reinforced by vertical bars of steel which are continued 
into the piers and abutments, so that the whole is solidly bound 
together. The advantage of this method is to root the bridge 
solidly into the earth, so that it has a resistance amply sufficient in 
case of a floating accumulation of ice occurring which would 
temporarily transform the bridge into a dam. 



June 1905. FERRO-CONCRETE. 507 

From the point of view of its construction the bridge is 
considered as a cantilever with unequal arms, the shorter arm 
balancing the longer by the weight of the abutment. 

The roadway is a platform 7 * 9 inches thick, and forms a tension 
member by means of its reinforcement, and it is supported by the 
arches which form the struts. In the neighbourhood of the 
keystone, and for about 82 feet in length, the roadway is solid 
with the vault of the principal span. This vault is 2*46 feet 
thick at the haunches, and 1 * 97 foot thick where it dies into the 
roadway ; it diminishes further until at the crown the roadway and 
arch together are only 1 • 15 foot thick. In the parts adjoining the 
pier, where the vault is clearly separated from the roadway, the 
latter is supported by three vertical struts 7 * 85 inches thick ; one is 
the axis of the bridge, and the other two form solid spandrels. 
The width of the arch being only 18 feet, the sidewalks are carried 
upon brackets 7 • 4 feet long. The roadway is covered with asphalt, 
and the sidewalks are paved with artificial stone. 

The same firm built, also upon the Hennebique system, a bridge 
above the high road and the railway at Yal Benoit. 

This is a skew bridge at an angle of 80°, and is 39*5 feet wide. 
It has three spans, of which the two principal ones are 39*4 feet and 
59 feet in length normally to the axis. The roadway consists of a 
platform 4*7 inches thick, supported by four lines of straight 
girders connected together by transverse joints extended outwards 
by brackets, which serve to support the overhung sidewalks. 

These girders rest upon two piers, each of four columns, and also 
upon abutments at the ends. The foundations are formed by 
masonry wells carried down to a solid bearing 13*52 feet deep, 
except for the abutment adjoining the railway which is on the slope 
of a hill. At this point a solid wall work 8*2 feet thick has been 
built, anchored about 32*8 feet deep into the soil, so as to hold 
back the masses of clay which are above and prevent a landslip. 



508 FERRO-CONCRETE. JuNK 1905. 



Framework for Lead Chambers at the Chemical Works of 

THE Engis Company. 

This company, under the advice of its managing director, M. 
L. G. Fremont, who is an engineer, has adopted the system of 
reinforced concrete for the construction of various foundations and 
for the construction of tunnels, cellars, platforms, hoppers, silos, 
frameworks for lead chambers, and for various towers, such as are 
required in the Glover and Gay Lussac processes. 

Among the interesting applications should be specially 
mentioned the framework for lead chambers, Fig. 13, Place 20, 
which constitutes one of the boldest innovations, not only in the 
method of construction itself, but also in the special apparatus 
which is enclosed. 

It is known that the most modern constructions designed to 
shelter and support the lead chambers are essentially composite 
structures, generally made of brick in the lower portion and of iron 
and wood in the upper portion. 

The combination and connection of the various heterogeneous 
parts present many weak points at the lines of meeting ; joints and 
connections for all these junctions represent so many points for 
attack either by the sulphur gas, or by sulphuric acid, when an 
accident has happened. Besides this, the enormous quantity of wood 
used for this kind of structure makes them highly combustible, and 
exposes the manufacturer to serious risks. 

These disadvantages alone suffice to justify the use of ferro- 
concrete, and the structure built at Engis, from designs prepared 
by M. Faure under the guidance of Professor Henri Dechamps, 
constitutes really a monolith from the foundations to the summit 
without joints or discontinuity. 

From the point of view of resistance to gas and acid liquids the 
composition and treatment of the concrete has been the subject of 
special study, and the mixtures employed liave given perfect 
satisfaction. 



June 1905. FERRO-CONCRETE. 509 

It is impossible for the author to give either a plan or 
photographic views of the interior of these chambers, for their 
special arrangements, which are due to the investigations of 
M. Fremont, are technical secrets which cannot be divulged. 

The structure occupies in plan an area 230 feet long by 92 feet 
wide ; it has a total height of 82 feet, of which 23 feet is for the 
lower portion and 59 feet for the upper part. 

The lower portion. Fig. 13, Plate 20, is formed of piers placed 
16*4 feet apart, each consisting of four columns connected at their 
upper ends by a horizontal girder 92 feet long aided at various 
points by struts fixed to the base of the columns. The total load 
carried by the lower structure is about 3,700 tons. At certain 
points the girders are connected together by joints, and there is a 
flooring of ferro-concrete which serves as a footway all round. 

The superstructure consists of vertical' columns 59 feet high, 
which rise from the ends of each horizontal girder, and besides 
these there are also columns in the centre of the building. At the 
upper part all the columns are connected together by horizontal 
girders with a span of 44 • 5 feet, and each of these bears a distributed 
load of 45 tons. The horizontal connections of these girders form 
the roof terrace. 

The columns are solidly connected together by vertical partitions 
of ferro-concrete, which, in conjunction with various well-designed 
struts, give the desired rigidity to the entire structure. The 
building was erected in 1900, and since that date has been in 
constant use and has given no trouble whatever. 

Another marked advantage possessed by this building is that, 
owing to the careful distribution of the materials, which are 
calculated with ample factors of safety, as was proved by rigorous 
tests, it has not been more expensive to construct than the old 
unsatisfactory composite buildings already described. 

The Paper is illustrated by Plates 18 to 20 and 6 Figs, in the 
letterpress. 



510 FERRO-CONCRETE. JuNE 1905. 

Discussion. 

Professor W. E. Lilly, in opening the discussion, said lie had 
listened to the Paper with very much interest, and there were just 
one or two points on which he desired to make a few remarks. In 
the first place, in England there had been no such advance in the 
use of ferro-concrete or reinforced concrete as had taken place 
on the Continent and in America ; but, in spite of that fact, reinforced 
concrete was coming to the front, and it was therefore a matter of very 
great importance that the reinforcement should be as far as possible 
the best of its kind. In the Paper that had been brought forward 
there were one or two statements with which he did not agree, and for 
that reason he desired to make a few criticisms. On page 500 the 
author referred to the subject of the principles of calculations for 
structures in ferro-concrete, and cited the opinion of Professor Eabut, 
who said : " I often hear it said that structures in ferro-concrete 
cannot be so accurately calculated as metallic structures ; in my 
opinion, the contrary is true." From the nature of things, ferro- 
concrete did not allow of calculations being made with the same 
accuracy as could be made for metallic structures, and he thought it 
was very much an open question whether the statement as quoted 
could be said to be correct. He had had to deal with some 
concrete work, and had found the greatest difficulty in trying to 
calculate deflection of concrete beams with a fair degree of accuracy. 
He was aware of most of the work that had been done in this 
direction, but the results of the calculations did not come out at all 
comparable with the results he had obtained on experiment. From 
the nature of things, he thought it must be agreed for the present 
that the reinforcement of concrete, and particularly with regard to 
the results of deflection obtained by calculation, was more or less 
tentative. 

There was one other very important point with regard to 
reinforcing concrete, namely, whether it was advisable to use a great 
number of bars or small wires, rather than several thick bars. 
From the work he had done experimentally on the subject, he was 
inclined to favour the system which adopted the use of a large 



June 1905. FERRO-CONCRETE. 611 

number of small wires, because by the use of a large bar very 
unequal stresses were sometimes set up which were not found, or at 
least not to the same extent, when the reinforcement was distributed 
to smaller bars throughout the concrete ; and at the same time by 
the latter process far better adhesion was obtained owing to the 
increased surface of the reinforcement itself. That point had not 
been remarked upon in the Paper, and would bear some further 
discussion. 

The systems in use were also much more numerous than had 
been cited in the Paper. There were almost hundreds of systems in 
use at the present time to which no reference whatever had been 
made. There was, for instance, the Cottancin system, which merely 
used a large number of the smaller wires he had referred to, and 
that system had been used a great deal in some parts of the 
Continent. 

Mr. William H. Maw, Past-President, said there was a point 
raised by Professor Lilly on which he desired to ask a question. 
The effect of the tension bars in reinforced concrete depended largely 
on the hold which those bars obtained in the concrete, and that hold 
was obtained either by the adhesion of the bars to the concrete itself 
or by giving a mechanical form to the ends of the bars by which 
they obtained a hold. He would like to ask Piofessor Lilly whether 
he had found it at all necessary to use any special preparation of 
the bars themselves. He had been told — he had no experience of 
his own to go upon — that a much better hold was obtained if the bar 
were given a thin coating of cement by means of a cement wash 
before insertion into the concrete. He did not know whether 
Professor Lilly had made any experiments bearing on that point. 

Professor Lilly said that, so far as he had been able to determine, 
there was not very much difference whether the ordinary round bar 
were used or a bar prepared with a rusty surface. The adhesion in 
both cases was very good, and, provided the concrete had been 
rammed sufficiently well, the adhesion seemed to be as good on the 
rough bar as on the smooth bar. 



512 FERRO-CONCRETE. JuNE 1905. 

M. NoAiLLON said he was glad to learn that Professor Lilly's 
experiments had led him to prefer wires of small diameter, because 
it confirmed his own rule, which was to prevent the concrete from 
breaking away from the metal ; this was effected by distributing the 
given section of framework over a number of wires, so that their 
contact-surface with the concrete might be large enough. There 
were other systems using the same construction ; if he (the author) 
had not mentioned them it was because they were very little used in 
practice, especially in Belgium, as they involved great difficulties in 
the construction of the frame and the concrete foundation. The 
system that was theoretically perfect would employ a large number 
of bars of different lengths with the extremities reaching to the 
rough plaster. It was in order to reduce the number of bars that he 
used hooks and rivets in his system. 

With regard to Mr. Maw's question, coating the bars with a 
cement wash had been advocated to increase the adhesion. 
Theoretically it was perfect, but in practice he did not think it 
advisable, because it often happened thac the thin coating of cement 
had dried before the concrete could be put on, and the skin so 
formed, being very friable, might peel off and come undone by the 
shocks produced by the concrete foundation, and then there would be 
scarcely any adhesion. 

The President said it was his pleasure and duty to propose a 
hearty vote of thanks to M. Noaillon for his interesting Paper. 

The resolution was carried unanimously. 



Communications. 



Mr. W. Noble Twelvetrees wrote that the remarkable properties 
of reinforced concrete fully deserved investigation by members of 
the Institution. Although comparatively little known in this 



June 1905. FERRO-CONCRETE. 513 

country, ferro-concrete, or " concrete-steel " as he preferred to 
describe it, had been very extensively employed on the Continent 
and in the United States for many purposes coming strictly within 
the province of the mechanical engineer. This new material was 
extremely suitable for the construction, not only of engineering 
workshops, but also of foundations fur boilers, engines, and heavy 
machinery. Consequently it was very desirable that engineers 
should study the principles governing the design of reinforced 
concrete construction, so that they might be able to apply it without 
feeling compelled to place themselves in the hands of the various 
patentees and contracting firms who made a speciality of such work. 

The author, in the first portion of his interesting Paper, had 
remarked upon the difficulty of determining accurately by means 
of theoretical formulae the cross-section of the reinforcement, and 
he advocated the employment of empirical rules based upon 
practical experience. No doubt, as Professor Lilly said, calculations 
could not be so accurate for ferro-concrete as for purely metallic 
structures, but there did not seem to be any reason why computations 
should not be based upon well-recognised principles, so long as 
suitable values were assigned to the various factors. For instance, 
in the case of beams, it was not necessary to adopt empirical 
njethods of calculation, as a simple adaptation of ordinary equations 
would give perfectly reliable results. In adepting such formulae 
it was necessary to consider the compression and tension areas 
separately, and in using the resulting equations, matters would be 
simplified by adopting limiting unit stresses for the materials 
employed. The following unit stresses might be adopted with 
perfect safety : — mild steel in compression and tension, 12,000 lbs. 
per square inch ; 1:2:4 concrete in compression 4.00 lbs. per 
square inch, and similar concrete in tension 4.0 lbs. per square 
inch. It was usuf\l in practice, however, to allow nothing for the 
resistance of concrete to tension. Although not to be recommended 
on theoretical grounds, this omission had the effect of simplifying 
calculations and of adding something to the factor of safety. 

For the purpose of deriving simple rules from ordinary beam 
formulae, it was necessary to equate the expressions for the bending 

2 



514 FERRO-CONCRETE. June 1905. 

(Mr. VV. Noble Twelvetrees.) 

moment and the moment of resistance. Thus, for a rectangular beam 
under a uniformly distributed load, M = Wl -^ 8, and B = fhh^h. 

Whence ^- X ^, = 1 

8 fbh^h 

where W = total load, I = span, / = extreme fibre stress at ^the 
distance li from the neutral axis, h = breadth of the beam section, 
and h = distance of the extreme fibres from the neutral axis. From 
this equation the following rules were derived : — 

For extreme Jibre stress on concrete in compression 

For extreme fibre stress of concrete in tension 

•^' = 5^hl • • • • (2) 

In the case of steel used as reinforcement in the tension area, 
the total stress (F't) had to be regarded as concentrated at the axis 
of the metal, and no account was taken of the resistance offered by 
the concrete. Hence, the quantity bh had to be omitted and 
allowance made for the progressive diminution of stress from the 
extreme fibres to the neutral axis, where its value was zero. 
Further, as in practice the axis of the reinforcement was rarely 
at the distance of f ^ from the neutral axis, as implied in the 
expression for the moment of resistance, the actual distance 
between the axis of the reinforcement and the neutral axis was 
conveniently denoted by the symbol h^. With these modifications 
the following rules were derived : 

For total stress on steel in tension 

F''=i^ .... (3) 

For the width of concrete, without reinforcement, in the compression 
area 

^ = H« • • • • ^*) 



June 1905. FERRO-CONCRETE. 515 

For the height of concrete^ without reinforcement^ in the compression 
area 

''' = a-^fe-*« = V^ • (^) 

The height of the tension area for a beam of any given load 
and span depended upon the sectional area of steel (a) used as 
reinforcement, and conversely, the area of steel was governed by 
the distance of the axis of the reinforcement from the neutral axis. 

For the height of the tension area (neutral axis to axis of 
reinforcement) 

K = hj^ . . . . (6) 

For the sectional area (a) of steel used as reinforcement 
-i^. .... (7) 

Similar rules could easily be derived for dealing with 
reinforcement in the compression area, and for taking • into 
account the value of concrete in both the compression and tension 
areas of any beam. 

The accuracy of calculations made by the rules given might be 
verified by the subjoined expression for the moment of resistance, 
the first term representing the resistance of the compression area 
and the second that of the tension area. 

B = (fJ^KiK) + (fiaK) . . (8) 

where /,', = (/! 4- 0) -^- 2 = mean stress per square inch of the 
concrete in compression. 

A calculated example showed that these simple rules could be 
conveniently applied for determining the proportions of a beam to 
any given load and span. Assuming the uniformly distributed load 
on a beam to be 120,000 lbs. and the span 200 inches, and taking the 
width of the beam at 20 inches, the height of the compression area, 
without reinforcement, by rule (5) would be 

K=^ /^, = ^ /_L20,000 X 200 ^ 23-7 inches. 
V 5-3&/; V 5-3 X 20 X 400 

2 2 



516 FERRO-CONCKETE. June 1905. 

(Mr. W. Noble Twelvetrees.) 

The axis of the reinforcement in the tension area could be 
placed at, say, 20 inches from the neutral axis. Then the required 
area of steel reinforcement in the cross-section of the beam, by 
rule (7), would be 



a = 



- ^ '"^ = '!"'r"^!°" = 6-25 square inches. 



2 8/;'/i„ 8 X 12,000 X 20 

Making the width of the tension area the same as that of the 
compression area, the reinforcement could be divided into any 
convenient number of bars giving the requisite sectional area of 
metal. Adding a thickness of at least 2 inches of concrete below 
the axis of the reinforcement the final dimensions of the beam 
would become 20 inches wide by 45-7 inches deep. Of course, it 
would be easy ta reduce this depth very considerably by using 
reinforcement in the compression area and increasing the amount of 
reinforcement in the tension area. 

The opinion expressed by Professor Lilly, that a few thick bars 
were less desirable than a larger number of thin bars or rods, was 
clearly worthy of support, and in the example calculated it would 
be better to divide the reinforcement into two rows of four or five 
bars each than to employ one row of bars each having double the 
sectional area. If two rows of bars were employed it would be 
necessary to add another inch or two of concrete at the bottom of the 
beam for adequate protection of the metal. 

Provision for shear was always advisable, especially in short and 
heavily loaded beams, and in long beams with a relatively small 
area of concrete. The calculation of shearing stress presented no 
difficulty, and as the state of simple shear was one in which there 
were two principal stresses only, giving rise to stress that was wholly 
tangential on any two planes inclined at 45° to the axis of principal 
stress, it appeared very desirable that the reinforcement for 
resisting shear should be inclined at an angle of 45° to the 
horizontal. For this reason, the system of reinforcement proposed 
by the author was strictly rational. 

In connection with the calculation of columns, it might be said 
that when reinforcement was applied in the form of longitudinal 
bars with occasional horizontal ties, the strength of the member 



June 1905. FERRO-CONCRETE. 517 

merely represented that of concrete plus steel. On the other hand, 
as the author had stated, the value of reinforcement employed in 
the form of hoops or spiral winding was 2 • 4 times greater than if 
the same quantity of metal had been used in the form of longitudinal 
reinforcement. This result, first ascertained by M. Considere, had 
recently been confirmed by a series of experiments conducted by 
Professor McCaustland, of Cornell University, who found that 
columns reinforced by eight circumferential hoops developed a 
resistance of 214,600 lbs., while if the same weight of metal had 
been applied in the form of rods the resistance could not have been 
more than about 157,390 lbs. As tho resistance of a plain 
concrete column was 115,000 lbs., the differences of 99,600 lbs. and 
42,390 lbs. respectively^ were practically in the ratio stated by 
M. Considere. 

M. NoAiLLON wrote, in reply to Mr. Twelvetree's communication, 
that he thought some agreement should be come to as to what 
is "empirical formula." In setting up practical formulae, it was 
evidently necessary to approach as near as possible to theoretical 
formulae, which had been calculated for imaginary phenomena from 
observed phenomena. For the formation of an equation of a 
phenomenon, the experimental coefficients must be known. If these 
coefficients were very uncertain (as was the case with concrete), the 
formula would not give a very accurate result, and it would also be 
very complicated. In order to simplify it, care would have to be 
taken that inaccuracies did not creep in. The formula thus modified 
would be empirical, since it no longer corresponded to any observed 
law. The formula that Mr. Twelve trees gave for rectangular beams 
was certainly quite empirical, as it was supposed that the resisting 
moment of the tension stresses was equal to the resisting moment 
of the compression stresses, when, in reality, they were forces 
themselves, and it was not their moments which were equal. 
Besides, the value of ^„ was deducted from that of h^ simply by 
supposition. 



June 1905. 519 



AN INVESTIGATION TO DETERMINE THE EFFECTS OP 

STEAM-JACKETING UPON THE EFFICIENCY OF A 

HORIZONTAL COMPOUND STEAM-ENGINE. 



By Mr. A. L. MELLANBY, M.Sc, of the Municipal School of Technology, 

Manchester. 



At the present time there is much diversity of opinion as to 
whether steam -jacketing the cylinders of reciprocating steam-engines 
has any considerable influence upon their economy. Much 
information has been collected by the Committees * appointed by this 
Institution to inquire into the subject, and the results of many 
trials have been collected and tabulated by them. As the outcome of 
these and other investigations, opinion has become somewhat settled 
on the following lines : 

(1) That jackets are useful for slow revolution, but not for 
quick-revolution engines. 

(2) That jackets are useful for simple and compound engines, 
but that their efficiency is doubtful, if they are applied to triple or 
quadruple expansion engines. 

It has long been obvious that there was considerable room for 
further experimental research upon this subject. The author 

* See Steam-Jacket Research Committee's Reports : First Report, Proceedings 
1889, page 703; Second Report, 1892, page 418 ; Third Report, 1894, page 535; 
and Steam-Engine Research Committee's First Report, 1905, Part 2, page 171. 



520 STEAM- JACKETING. JuNE 1905. 

therefore determined to take advantage of the opportunity which he 
had of carrying out tests upon the experimental engine at the 
Manchester School of Technology, to see whether some definite 
knowledge upon this important point could not be obtained. 

In most of the tests that have been published, it has been the 
custom to run a trial on any available engiLC that was fitted with 
jackets, and to find out the water used per I.H.P. per hour. This 
water would include the amount that entered the cylinder, and the 
amount that was condensed in the jackets. Another trial would 
then be made with the jackets shut off, and the consumption per 
I.H.P. per hour again measured. In some cases it has happened 
that the jacket steam has had the effect of reducing the engine-feed 
or increasing the horse-power, and so improving the economy. In 
other cases it has happened that any reduced engine-feed or increased 
horse-power has been more than counterbalanced by the extra 
steam used in the jackets. In the majority of the experiments 
published, the jacketed and unjacketed trials have been run at the 
same load, and no attempt has been made to find the best conditions 
for either case. The chief reason for this has been that the engines 
experimented upon were in actual service, and their load was fixed 
by the amount of work they had to do and not with regard to their 
best performance. Comparisons from such trials are of little 
scientific value, as it might easily happen if an engine were working 
at its most economical load for the jacketed conditions, the same 
load would not be the most suitable when the jackets were off. A 
certain amount of work has been done upon the experimental engines 
in colleges and technical schools, but this work has not had much 
effect upon current practice. In most cases the engines are so small 
that manufacturers have not sufficient confidence in their results to 
apply them to the design of large commercial engines. Also few 
systematic series of trials have been published. 

After devoting considerable thought to the matter, the writer 
determined upon the following scheme of experiments. To run a 
series of trials at the same boiler pressure, revolutions and vacuum 
with different points of cut-off in the high-pressure cylinder. 

At each point of cut-off to run five trials : 



June 1905. STEAM- JACKETING. 521 

(1st) with both cylinders unjacketed. 

(2nd) with the ends of the high-pressure cylinder jacketed. 

(3rd) with the ends and barrel of the high-pressure cylinder 
jacketed. 

(4th) with the ends and barrel of the high-pressure, and the 
ends of the low-pressure cylinder jacketed. 

(5th) with the ends and barrels of both high and low-pressure 
cylinders jacketed. 

It will be seen that this gives five series of trials with various 
grades of expansion, and, if the range of expansion be sufficient, 
curves may be constructed which will allow us to determine the 
most economical number of expansions, and the consumption at that 
point for each series. From these curves the effect that jacketing 
has upon the economy can at once be determined, and the most 
efficient trials of the five series compared with one another. It 
seemed desirable that an attempt should be made not only to find 
out whether jacketing was efficient, but the reason for any efficiency 
it might have. Until the Paper on Cylinder Condensation read 
before the Institution of Civil Engineers in 1897 by Messrs. 
Callendar and Nicolson, it had been the generally accepted opinion 
that any reduced steam-consumption with steam-jacketing would be 
altogether due to reduced initial condensation. Messrs. Callendar 
and Nicolson, however, pointed out that initial condensation in 
steam-engines was much less than had hitherto been suspected, and 
that a considerable portion of the " missing quantity " or difference 
between the indicated and actual weight of the steam passing through 
the engine was due to valve leakage. So far as the author knows, 
no experimental evidence has been adduced to disprove the assertions 
of the above-named experimenters. As he had carried out some of 
the valve leakage experiments mentioned in their Paper, he 
determined to adopt their method of calculating the amount of 
condensation, and of inferring the probable valve leakage. This 
will be referred to more fully later on. 

Description of the Engines. — The engines, which were made by 
Messrs. J. Carmichael and Co., of Dundee, are of the horizontal 



522 



STEAM- JACKETING. 



June 1905. 



compound side-by-side type, with, high-pressure cylinder 11 J inches 
diameter, low-pressure cylinder 20 inches diameter, and stroke 
36 inches. The high-pressure cylinder is provided with Corliss 
valves and gear, and the low-pressure with slide-valves and Meyer 
expansion plates. The ends and sides of the cylinders are steam- 
jacketed. The jacket steam for both cylinders comes direct from 
the main steam-pipe without any intervening reducing valves, and 
there are separate supply pipes each with a shut-off valve leading 



Measuring Vessel. 



Fig. 1. 
Arrangement of Jachet Supply and Drain Pipes. 




Drains U) Measuruig Vcssds- 



to each end and each barrel jacket. Fig. 1 shows the arrangement 
of the jacket supply and drain pipes. In all the trials the steam- 
pressure in the jackets was practically equal to that at the high- 
pressure stop-valve. During the trials the condensed jacket-steam 
passed into separate calibrated vessels where it was measured and 
then discharged to waste. The condenser used in these trials was of 
ordinary surface type. The air-pump was separately driven by an 
electric motor of variable speed, and, after taking the condensed 
steam from the condenser, delivered it into calibrated cast-iron 
measuring vessels. 



June 1905. STEAM- JACKETING. 523 

The brake used was a flat iron tube passing round the fly-wbeel. 
Cold water continually circulated through the tube to keep down 
the temperature. The load consisted of dead weights applied 
through levers, but this load was directly measured by the pressure 
in a diaphragm cylinder, through which the actual pull was 
transmitted. The author, however, regrets that, when these trials 
were being carried out, the gauge attached to the diaphragm cylinder 
had not been calibrated, so that it has not been found possible to 
give the brake horse-power of the engine. As the point of cut-off 
in the high-pressure cylinder was fixed for each trial, any change of 
pressure or vacuum tended to alter the speed of the engine. An 
observer was therefore placed at the brake, whose duty it was to watch 
a tachometer and keep the speed of the engine as constant as possible 
by varying the brake load. The clearance volumes and surfaces of 
the cylinders are given in Table 1 (page 524.) 

The observations on the various trials were made by the third- 
year students at the school, under the personal supervision of the 
author. Before carrying out these trials, they had had considerable 
practice in engine testing, and all of them took the greatest care to 
get accurate readings. Observations of pressures, temperatures, and 
revolutions were taken every 5 minutes, of the air-pump discharge 
every 2 J minutes, and indicator cards were taken every 10 minutes. 
Two indicators were directly attached to each cylinder, one at each 
end, to avoid inaccuracies in the diagrams due to long connecting- 
pipes. The observations were taken simultaneously, the signal for 
making them being given by a whistle blown by one of the students. 
After the trial was completed, the records were plotted on a time-basis. 
This gives an admirable means of checking the results of the variouR 
observers, as it enables any inaccuracies to be easily detected. 

If the effect of steam-jacketing is to reduce initial cylinder- 
condensation, it would appear that the best results ought to be 
obtained when the engine is running slowly. It was therefore 
decided to run the engine at about half its normal speed — a fact 
which ought to be remembered when comparing the results here 
given with those from engines of a similar type. Had the engine 
been working at its maximum speed, the consumption in terms of the 
I.H.P. would have been much less. 



524 



STEAM-JACKETING. 



June 1905. 



TABLE 1. 

Particulars of Clearance Volumes. 





H.P. 
Cylinder. 


L.P. 

Cylinder. 


Cover 
end. 


Crank 
end. 


Cover 
end. 


Crank 
end. 


Clearance volume . . cubic feet 
Percentage of jjiston displacement 


0-203 
9-38 


0-180 
9-05 


0-488 
7-46 


0-406 
6-54 



Particulars of Clearance Surfaces, 







H.P. Cylinder. 


Cover end. 


Crank end. 




-6 1 






nd 








T3 






'a 










_2 


a> 


• 


© 


0) 








o3 


o 


o3 
O 

H 


O 

e3 


o 


o3 

o 






t-5 


P 




•-5 


P 




Clearance surface . 


. .sq. ft. 


0-88 


4-05 


4-93 


1-13 


3-80 


4-93 


(Crank on dead 


centre) 














Percentage of whole 


. . . . 


18 


82 


— 


23 


77 


— 






L.P. Cylinder. 


Cover end. 


Crank end. 




nd 






■73 








•T3 


O) 




TS 










o 


"S 


^^ 


<x> 


O 


r>^ 






4) 




oi 
o 




o 


03 

O 






o 

03 


"a 


H 


^ 




EH 






>-i 


P 


1 


•-5 


p 




Clearance surface . 


. .sq. ft. 


3-57 


8-83 


12-4 


2-07 


8-24 


10-3 


(Crank on dead 


centre) 














Percentage of wliole 


. . . . 


29 


71 


— 


20 


80 


— 



June 1905. STEAM- JACKETING. 525 

It was decided to carry out these trials with a boiler pressure of 
150 lbs. per square inch (gauge) at about 60 revolutions per minute, 
and with a back-pressure in the condenser equal to 6 inches of 
mercury. The driving gear was disconnected from the governor, 
and, by means of an arrangement which allowed the governor 
sleeve to be raised or lowered, the point of cut-off in the high- 
pressure cylinder could be fixed at any part of the stroke. 

Four sets of trials were made, each set having a different point of 
cut-off in the high-pressure cylinder as follows : — 

Ist set with 8*1 total expansions. 

2nd „ 12-3 „ 

3rd „ 18 

4tll ,, Zo ,, ,, 

In each set there were 5 trials (with the exception of the 4th, 
which contained only 3), one unjacketed and the others with 
various degrees of jacketing as explained in a previous paragraph. 
The point of cut-off in the low-pressure cylinder remained fixed for 
the whole of the trials. This was in accordance with the work done 
by Professor R. L. Weighton and the author, on the engines at the 
Durham College of Science, and the actual point of cut-off was fixed 
from the curves given in Professor Weigh ton's Paper.* 

The author would here express his regret that more trials were 
not made with a greater variety of points of cut-off. So many 
variables come into play in an engine trial, that it is practically 
impossible to run one test under exactly similar conditions to 
another. If there are a large number of experimental points, 
errors due to these differences can be generally eliminated when the 
results are reduced to diagrammatic form, and several discrepancies 
which appear in the diagrams might not have been there if more 
experiments had been made. The work involved in the preparation 
of the Paper has, however, been so enormous, since every calculation 
and indicator diagram had to be worked out by the author alone, 
that it would have been almost impossible to finish the investigation 
had more been added to it. 

* Proceedings, North-East Coast Institution of Engineers, 1899-1900, 
vol. xvi, page 63. 



526 



STEAM-JACKETING. 



June 1905. 



TABLE 2 (continued to page 529). 
General Results of Trials, 



1 
2 
3 

4 

5 

6 

7 

8 

9 
10 
11 

12 

13 

14 

15 

16 

17 

18 

19 
20 
21 
22 

23 
24 

25 

26 

27 

28 
29 
30 
31 



Number of Trial . 
Date of Trial 
Duration of Trial . 
Steam-pipe pressure lbs. abs. 
Steam-pipe temperature F.° 
H.P. Exhaust pressure lbs. abs. 
,, ,, temperature F.° 

L.P. „ pressure lbs. abs. 
temperature F.° 
inches of mercury 



Vacuum 
Barometer 
JAir-pump 
trial . 



25 Expansions. 



discharge duringl 
\ trial ." . . . .] 

Air-pump discharge per hour . 
J Steam condensed in L.P. steaml 
\ chest . . . ./ 

J Steam condensed in H.P. cover j 
\ jackets . . . ./ 
I Steam condensed in H.P. barrel) 
\ jackets . . . . j 
J Steam condensed in L.P. cover i 
\ jackets . . . . j 

I Steam condensed in L.P. barrel j 
\ jackets . . . ./ 

Total steam used during trial . 
„ „ „ per hour . 

Total revolutions during trial . 

Average revolutions per minute 
("Mean pressure in H.P. cylinder^ 
\ lbs. per sq. in.) 

J Mean pressure in L.P. cylinder 
\ lbs. per sq. in. 

|Mean pressure reduced to L.P. 
\ cylinder lbs. per sq. in. 

I.H.P 

Steam per I.H.P. per hour lbs. 

H.P. cylinder feed per revo- 
lution . . . lbs. 

L.P. cylinder feed per revo-1 
\ lution . . . lbs. ^ 

/Clearance surface temperature 
\ H.P. cover end . . F.°, 
/Clearance surface temperature^ 
It L.P. cover end . . F.°/ 



108 
21/6/04 
60 mins. 

166 

365-5 

18-2 

219 

31 

135 

25-2 

30-23 

1603 

1603 

163 

Off 

Off 

Off 

Off 

1603 
1603 
8544 
1 



59 

50 



24 

81 
19 







4 

7 


75 

452 

406 



324 

192 



109 
21/6/04 
60 mins. 

166 

365-2 

16-6 

214-4 

3-1 

135-6 

25-2 

30-23 

1503 

1503 

129 

66 

Off 

Off 

Off 

1569 
1569 
3563 



59 
50 



23 

78 
20 






421 

385 



340 
187 



105 
14/6/04 
60 mins. 

165 
366-7 

235 

141 
24-7 
29-7 

1458 

1458 

99 

46 

51 

63 

75 

1693 
1693 
3510 

58-5 

47-0 
130 

28-3 

92-0 
18-35 

0-415 
0-387 
345 
325 



June 1905. 



STEAM-JACKETING. 



527 



TABLE 2 (continued on next page). 
General Besults of Trials. 









1^ 


I Expansions. 




1 


Number of Trial . 


88 


89 


90 


93 


94 


2 


Date of Trial 


22/4/04 


26/4/C^: 


26/4/04 


3/5/04 


3/5/04 


3 


Duration of Trial . 


60 mins. 


60 mins. 


55 mins. 


60 mins. 


50 mins. 


4 


Steam-pipe pressure lbs. abs. 


161 


163 


165 


163 


163 


5 


Steam-pipe temperature F.° 


364 


366 


366-3 


365-2 


365-0 


6 


H.P. Exhaust pressure lbs. abs. 


22-6 


22-0 


21-4 


25 


26 


7 


„ „ temperature F.° 


— 


230 


228 


236 


239 


8 


L.P. ,, pressure lbs. abs. 


— 


2-75 


2-75 


2-7 


2-7 


9 


„ „ temperature F.° 


— 


133 


129 


132-5 


133-7 


10 


Vacuum inches of mercury 


24-6 


25-2 


25-2 


24-8 


24-9 


11 


Barometer „ ,, „ 


29-6 


30-2 


30-2 


29-9 


29-9 


12 


JAir-pump discharge during'^ 

\ trial . . . ./ 

Air-pump discharge per hour . 


1757 


1735 


1505 


1686 


1397 


13 


1757 


1735 


1642 


1686 


1676 


14 


("Steam condensed in L.P. steaml 
\ chest . . . ./ 
(Steam condensed in H.P. coverl 
\ jackets . . . . j 
J Steam condensed in H.P. barrel^ 
\ jackets . . . ./ 


156 


145 


110 


171 


139 


15 


Off 


59 


55 


39 


36 (hour) 


16 


Off 


Off 


54 


56 


52 (hour) 


17 


J Steam condensed in L.P. cover 1 
\ jackets . . . ./ 
jSteam condensed in L.P. barrel^ 
\ jackets . . . ./ 
Total steam used during trial . 


Off 


Off 


Off 


30 


41 (hour) 


18 


Off 


Off 


Off 


Off 


60 (hour) 


19 


1757 


1794 


1614 


1811 




20 


„ „ „ per hour 


1757 


1794 


1762 


1811 


1865 


21 


Total revolutions during trial . 


3521 


3544 


3214 


3596 


2977 


22 


Average revolutions per minute 


58-7 


59-1 


58-4 


59-9 


59-5 


23 


JMean pressure in H.P. cylinder^ 
\ lbs. per sq. in./ 


54-4 


56-5 


59-0 


55-0 


52-5 


24 


JMean pressure in L.P. cylinder! 
\ lbs. per sq. in./ 


12-0 


11-7 


11-3 


14-1 


15-5 


25 


jMean pressure reduced to L.P.i 
\ cylinder lbs. per sq. in./ 


29-7 


30-1 


30-5 


32-0 


32-6 


26 


I.H.P 


97-0 


99-0 


99-0 


107-0 


108-0 


27 


Steam per I.H.P. per hour lbs. 


18-1 


18-14 


17-78 


16-93 


17-26 


28 


/H.P. cylinder feed per revo-1 
\ lution . . . lbs./ 


0-499 


0-489 


0-468 


0-469 


0-469 


29 


/L.P. cylinder feed per revo-i 
\ lution . . . lbs./ 


0-455 


0-448 


0-434 


0-421 


0-431 


30 


/Clearance surface temperature! 
\ H.P. cover end . . F.°/ 


328 


342 


344 


336 


338 


31 


/Clearance surface temperature i 
\ L.P. cover end . . F.°/ 


205 


205 


204 


306 


318 



528 



STEAM-JACKETING. 



June 1905. 



TABLE 2 (concluded on ojpposite ;page). 
General Results of Trials. 



1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
11 

12 

13 

14 

15 

16 

17 

18 

19 
20 
21 
22 

23 
24 

25 

26 
27 

28 
29 
30 
31 



Number of Trial . 

Date of Trial 

Duration of Trial . 

Steam-pipe pressure lbs. abs. 

Steam-pipe temperature F.° 

H.P. Exhaust pressure lbs. abs. 
„ „ temperature F.° 

L.P. J, pressure lbs. abs. 
temperature F.° 
inches of mercury 



Vacuum 
Barometer 

/Air-pump 

\ trial 



discharge during! 

Air-pump discharge per hour . 
!j Steam condensed in L.P. steaml 
|\ chest . . . . . j 
u Steam condensed in H.P. cover I 
I ( jackets . . . • J 
i J Steam condensed in H.P. barrel) 
|\ jackets . . • •/ 
I r Steam condensed in L.P. cover 1 
l\ jackets . . . • / 

r Steam condensed in L.P. barrel) 
\ jackets . . . ./ 

Total steam used during trial . 
„ „ „ per hour 

Total revolutions during trial . 

Average revolutions per minute 
jMean pressure in H.P. cylinderj 
I j lbs. per sq. in./ 

IrMean pressure in L.P. cylinder^ 
i| lbs. per sq. in./ 

I J Mean pressure reduced to L.P.| 
'■ ' lbs. per sq. in./ 



12 "3 Expansions. 



evo-\ 
lbs./ 



( cylinder 

I LH.P 

! Steam per I.H.P. per hour 

/H.P. cylinder feed per 
,\ lution 

l/L.P. cylinder feed per revo-| 
J lution . . . lbs./ 

I /Clearance surface temperature 1 
l\ H.P. cover end . . F.°/ 
1/ Clearance surface temperature! 
i\ L.P. cover end . . F.°/ 



95 


96 


97 


98 


10/5/04 


10/5/04 


10/5/04 


17/5/04 


50 mins. 


50 mins. 


55 mins. 


60 mins. 


160 


160 


160 


165 


363-5 


363-8 


363-3 


365-4 


26-4 


25-6 


25-2 


29-4 


241 


239 


238 


248-3 


2-7 


2-7 


2-7 


2-9 


132 


132 


132 


135-3 


24-8 


24-8 


24-8 


24-8 


29-8 


29-8 


29-8 


29-76 


1820 


1715 


1815 


1967 


2184 


2058 


1910 


1967 


65 


93 


86 


150 


Off 


47 


42 


39 


Oif 


Off 


43 


43 


Off 


Off 


Off 


89 


Off 


Off 


Off 


Off 


1820 


1762 


1900 


2138 


2184 


2114 


2073 


2138 


3007 


3007 


3356 


3519 


60-1 


601 


61-0 


58-6 


66-0 


67-0 


67-5 


64-2 


14-3 


13-6 


13-4 


17-0 


35-8 


35-4 


35-4 


37-9 


120 


118 


120 


123 


18-2 


17-9 


17-3 


17-2 


0-605 


0-571 


0-541 


0-559 


0-584 


0-540 


0-516 


0-514 


335 


342 


342 


342 


211 


209 


211 


308 



99 
17/5/04 
60 mins. 

163 

365 

31-3 

252 

2-9 
137-5 
24-8 
29-76 

1955 

1955 

148 

38 
38 
108 

87 

2226 
2226 
3519 
58-6 

62-0 
19-3 

39-2 

128 
17-4 

0-556 
0-512 

342 

318 



Junk 1905. 



STEAM-JACKETING. 



629 



TABLE 2 {concluded from page 526). 
General Besults of Trials. 



Number of Trial 
Date of Trial 
Duration of Trial . 
Steam-pipe pressure lbs. abs. 
Steam-pipe temperature F.° 
H.P. Exhaust pressure lbs. abs. 

,, ,, temperature F.° 

L.P. ,, pressure lbs. abs. 

,, ,, temperature F.° 

inches of mercury 



discharge during 



Vacuum 

Barometer 
J Air-pump 
\ trial .... 

Air-pump discharge per hour 
I Steam condensed in L.P. steam) 
\ chest . 

(Steam condensed in H.P. cover\ 
\ jackets 

j Steam condensed in H.P. barrel ) 
\ jackets 

J Steam condensed in L.P. cover) 
\ jackets 
(Steam condensed in L.P. barrel) 
\ jackets . . . . ( 

Total Steam used during trial . 
„ „ „ per hour 

Total revolutions during trial . 

Average revolutions per minute 
(Mean pressure in H.P. cylinder\ 
\ lbs. per sq. in 

(Mean pressure in L.P. cylinder)^ 



(Mean pressure reduced to L.P.I 
\ cylinder lbs. per sq. in./ 

I. H.P. . 



lbs. per sq. in./ 
educed to 1 
lbs. per sq 



8*1 Expansions. 



Steam per I. H.P. per hour lbs. 
H.P. cylinder feed per revo- 
lution . . . lbs. 
L.P. cylinder feed per revo- 
lution . . . lbs. 
Clearance surface temperature 
H.P. cover end . . F.° 
/Clearance surface temperature 1 
\ L.P. cover end . . • F.°/ 



100 


101 


102 


103 


31/5/04 


31/5/04 


7/6/04 


7/6/04 


60 mins. 


60 mins. 


60 mins. 


60 mins. 


161 


164 


158 


161 


364-5 


365-7 


363-8 


364 


33-3 


33-3 


31-8 


350 


257 


256-5 


254 


257-5 


2-9 


2-9 


3-1 


3-1 


137 


137 


138-6 


140-0 


24-8 


24-8 


24-9 


24-9 


29-81 


29-81 


30-16 


30-16 


2757 


2672 


2557 


2580 


2757 


2672 


2557 


2580 


132 


122 


135 


145 


Off 


48 


41 


47 


Off 


Off 


49 


36 


Oif 


Off 


Off 


110 


Off 


Off 


Off 


Off 


2757 


2720 


2647 


2773 


2757 


2720 


2647 


2773 


3583 


3573 


3503 


3555 


59-7 


59-5 


58-4 


59-2 


79-0 


78-0 


77-5 


74-0 


19-6 


19-2 


18-7 


21-7 


45-3 


44-6 


43-9 


45-8 


151 


148 


143 


151 


18-25 


18-4 


18-5 


18-35 


0-769 


0-747 


0-730 


0-726 


0-732 


0-713 


0-691 


0-685 


344 


349 


344 


346 


222 


223 


221 


— 



104 

14/6/04 
55 mins. 

364-2 

268 

142-5 
24-7 

29-7 

2365 
2580 
169 

46 (hour) 

35 (hour) 

76 (hour) 

108 (hour) 

2845 
3283 
59-7 

70-0 
25-0 

47-8 

159 
17-9 

0-720 

0-669 

347 



530 



STEAM-JACKETING. 



June 1906. 



Besiiltsfrom the Trials. — Table 2 (pages 526-529) gives the chief 
readings and results from the trials, the most important of which have 
been plotted in Figs. 2, 3, 4, 5 and 6 (pages 530-534). In thesefigures it 
will be seen that the base line taken is the total number of expansions, 

stroke 



that is, it is the value of r B where r is the fraction 



Lbs. 
2,600 



2,200 



Fig. 2. — No Jackets. 



Piston travel to cut-oflf 



o 



1,800 



Lbs. 1,400 
per 
hour 
700. 



g 500 1,000 

a 

^ 300 



100 600 





20-0 P^ 
L 19-0 go 



f 18-0 



LH.P. 

155 



145 
135 

125 

(^ 

W 

115 ^ 

105 

95 

85 



12 



16 20 

Number of Expansions 



24 



in the high-pressure cylinder and B is the ratio of the volume of the 
low-pressure to the high-pressure cylinder. No correction has been 
made for clearances, as it was felt that the results would be more 
useful to engine-builders if the terms used were those generally 
employed in everyday commercial work. 

The line marked air-pump discharge in the illustrations gives 
the quantity of steam in pounds per hour that passes through the 
high-pressure cylinder. As the low-pressui'e steam-chest was 



June 1905. 



STEA1\I-JACKETING. 



531 



drained, the whole of the air-pump discharge did not enter the 
low-pressure cylinder. The steam condensed in this chest passed 
through a measuring vessel and thence to the condenser. The amount 
of this condensed steam for the different trials is given in Table 2, 
line 14. The total steam per hour line shows the air-pump 
discharge together with the steam used in the jackets. This latter 

Fig. 3. — Jackets on H. P. Cylinder Ends. 




16 20 

Number of Expausions 



24 



does not of course appear in Fig. 2 (page 530). From the curves of 
total steam per hour and of I.H.P. the curve of steam per I.H.P. per 
hour ^ has been constructed. This enables one to find the number of 
expansions at which the engine will work most economically and 
the best results in terms of the I.H.P. under the various conditions. 

From curves Figs. 2 to 6 (pages 530-534), Table 3 (page 536) 
has been constructed. 

2 p 2 



632 



STEAM- JACKETING. 



June 1905. 



It will be noticed that in each case there is a considerable range 
of expansions within which the efficiency of the engines in terms 
of the I.H.P. varies very little. The effect of the jackets was 
therefore, on the whole, to lessen somewhat the consumption of the 
engine. The best results were obtained when the whole of the 
high-pressure and the ends of the low-pressure cylinder were 

Fig. 4. — Jackets on H.P. Cylinder Ends and Barrel. 




w 

18*5 ^ & 

^ o 

17-5 I a 

+-> 



I.H.P. 
150 



140 



130 



120 



110 



100 



90 






12 



16 20 

Number of Expansions 



jacketed. The extra steam used for the barrel of the low-pressure 
cylinder (Table 3, line 6) does not seem to have been sufficiently 
compensated for by decreased cylinder-feed or increased horse-power. 
One practical result worth emphasising is that the mean effective 
pressure of a compound engine may be made much higher than it 
usually is in practice without any loss of efficiency. On this point the 
results may be compared with those published by Professor Weighton. 



June 1905. 



STEAM-JACKETING. 



533 



In that Paper* he showed that, for an entirely different type of engine 
with a much higher boiler pressure, the point of maximum efficiency 
would only be obtained with a mean pressure much higher than was 
generally adopted. Both of these sets of trials seem to show that 
engine-builders may construct smaller engines, to develop a certain 
power, than they usually do, without decreasing their economy. If 

Fig. 5. — Jacliets on H.P. Cylinder Ends and Barrel, 
and on L.P. Cylinder Ends. 



Lbs. 
2,600 



2.200 



ft 1,800 

CD 



Lbs. 1-400 
per 
bour 
700 



§ 500 1,000 

O 

bo 

fl 300 



^ 



100 




600 




16 20 

Number of Expansions 



the consumption had been measured in terms of the brake horse- 
power, then the point of maximum efficiency would be further 
moved to the left or the most efficient mean pressure would be even 
higher than that given in Table 3 (page 536). 

It is worth while to analyse the trial results, and see how the 



* Proceedings North-East Coast Institution, 1896-1897, vol. xiii, page 88. 



534 



STEAM-JACKETING. 



June 1905. 



application of the jackets does alter the efficiency. If one simply 
looks at the decrease or increase of the consumption of steam per 
I.H.P. per hour very little will be learnt. If, however, one 
observes how the indicated horse-power is increased or decreased by 
the action of the jackets, and how the steam passing through the 



Fig 6. — Jacl-efs on H.P. Cyl. Ends and Barrel, 
and on L.P. Cyl. Ends and Barrel. 



Lbs. 
2,600 



2,200 



f^l,800 



Lbs. 1.400 

per 
hour 

700 



g 500 1,000 

a 

bo 

fl 300 



100 600 





12 16 20 

Number of Expansions 



cylinders per hour is affected although the point of cut-off is fixed, 
some idea of the nature of the influence of the jackets will be 
obtained. 



Effect upon the Indicated Eorse-Power. — The application of the 
jackets to the high-pressure cylinder has the general effect of 



June 1906. 



STE AM-JACKETIN G. 



535 



decreasing the mean effective pressure of the engine, that is, of 
decreasing the I.H.P. In all cases when jackets are used in the 
high-pressure cylinder, the pressure at high-pressure release, the 
back pressure of the high-pressure and the admission pressure of the 
low-pressure cylinder are diminished. Looking at the indicator cards 
alone, it appears as if less steam were present in the engine when the 
high-pressure cylinder is jacketed than when it is unjacketed, 
Fig. 14 (page 546). 

Fig. 7. — Diagram-Factor. 























1 •! 


- 








At^^ 










o 






^ 


J^ 




Vjackd^^i^ 








r i u ^ 












a 


- 






,^ 


"^ 














bo 

•1—1 


^ 






















•0 


\ 






















Lbs. 
per 
sq.in. 
- bO 




^. 




















- 


^ 


4^/X 


^ 
















yO 
- 40 1 


- 






^ 


^ 


-^ 










- 


u 
Ph 

30 i 


- 










^'i/aJz 


^te<i. 










0) 

- 20 


- 






















10 


1 




1 


1 


1 


1 


\ 


1 


1 


1 








8 




12 


I'uinbe] 


16 
rof Ex 


pansio 


20 
ns 




M 







The jacketing of the low-pressure cylinder has a different effect. 
The release pressure in the high-pressure cylinder is further 
diminished to a slight extent, but the high-pressure back-pressuro 
and the low-pressure admission-pressure are considerably increased. 
It may therefore be said that the low-pressure jackets appear to 



536 



STEAM-JACKETING. 



June 1905. 



slightly lessen the steam in the high-pressure cylinder but to 
increase that passing through the low-pressure cylinder. 

It is interesting to notice how the jacketing affects the distribution 
of power between the two cylinders. From trials 95 to 99 inclusive 
Table 4 has been constructed. 



TABLE 3. 



Nature of Trial. 



No jackets .... 

H.P. ends jacketed 

H.P. ends and barrel jacketed. 

H.P. ends and barrel and L.P.^ 
ends jacketed . . / 

H.P. and L.P. ends and barrels! 
jacketed . . . / 



Best 

number of 

Expansions. 



Steam per 
LHP. 

per hour. 



Mean pressure 
reduced to 

L.P. cylinder 
at 60 revs. 



11 to 17 
11 „ 14 
11 „ 15 



18-1 to 18-2 
17-7 „ 17-8 
17-3 „ 17-4 



14 „ 19 I 16-95 „ 17-05 
13-4 to 19 17-25 „ 17-85 



38-2 to 30 
37-8 „ 33-5 
37-6 „ 32-8 

35-3 „ 31-1 
36-6 „ 31-4 



TABLE 4. 



No. 

of 

Trial. 


Nature of Trial. 


Mean pressure 

reduced to L.P. 

Cylinder. 


Power in 
L.P. cyl. 
Power in 
H.P. cyl. 


Diagram 
Factor. 


95 
96 

97 
98 
99 


No jackets .... 

H.P. ends jacketed . 

("H.P. ends and barrel "i 
\ jacketed ... ./ 

TH.P. ends and barrel and) 
\ L.P. ends jacketed . / 

fH.P. and L.P. ends andj 
\ barrels jacketed . . / 


PI.P. L.P. Total. 
21-5 14-3 35-8 

21-8 13-6 35-4 
22-0 13-4 35-4 

20-9 17-0 37-9 

20-2 19-0 39-2 


0-66 
0-62 

0-61 
0-81 
0-94 


0-83 
0-82 

0-82 
0-85 

0-88 



June 1905. 



STEAM-JACKETING. 



537 



The " diagram factor " is the usual relation, 



Actual mean pressure, 



P (1 + loge r . R) 



-P, 



r .R 

where P is the absolute steam-chest pressure and P^, is the 
condenser back-pressure. This constant is useful in fixing the sizes 
of engines for estimating purposes, and, as very few values have 
been published, Fig. 7 (page 535) has been drawn. This figure 
gives the theoretical and the actual mean pressures for jacketed and 
UDJacketed trials, and shows how the diagram-factor varies with the 
number of expansions. 



Effect upon the Steam Consumption. — The application of the 
jackets seems as a general rule to reduce the amount of steam 



Fig. 8. 




passing through the cylinder of the engine. It has been suggested 
in discussions on engine trials that the increase in the area of the 
low-pressure cards was due to leakage of the high-pressure jacket 
steam into the cylinders. An inspection of the results in Table 2 
will show, however, that this was not the case in these trials, as the 
air-pump discharge is generally reduced when the low-pressure 
jackets are applied. Table 2, line 13, shows these results. With 
only the high-pressure jackets on, the total steam per hour, namely 
air-pump discharge and jacket steam, is less than when the engine is 
unjacketed. When the low-pressure jackets are applied, although 
the air-pump discharge is lessened, the total steam used is increased. 
To study the action of the jackets, their effect upon the " missing 
quantity" has been observed. Fig. 8 shows how this has been 



538 



STEAM-JACKETING. 



June 1905. 



done. It represents one of the diagrams from the high-pressure 
cylinder. Assuming the length of the card A B to represent the 
volume of the cylinder, the clearance volume A C has been set off to 
the left on the same scale. At D any point after compression, the 
distance D E represents the volume of cushion steam imprisoned in the 
cylinder. Assuming this steam to be dry and saturated, its volumes at 
different pressures have been calculated, and are represented by the 
dotted line. The amount of steam passing through the cylinder per 

g^j,^j^^ /steam during trial \ ^^^ ^^^^ ^^^^^ ^^^ .^^ volumes at different 
\strokes duriDg trial/ 

pressures obtained from steam tables and set off to the same scale as the 

other volumes on the diagram. Thus at the point H near cut-off, the 

distance F G shows the volunie of the cushion steam, and G H the 

volume of steam shown by the indicator to be passing through the 

cylinder. This amount G H reduced to pounds per hour is called 

the " Indicated Weight at cut-off." G K gives the actual amount of 

steam passing through the engine as measured by the air-pump 

discharge, and the distance H K shows the difference between the 

indicated and the actual weight. This distance H K reduced to 

pounds per hour is called the " Missing quantity of cut-off." 

Similarly L M is the indicated weight at release, L N the actual 

weight at release and M N the " Missing quantity at release." The 

indicated weights at release and compression have been calculated 

for all the trials ; their values are given for the high-pressure 

cylinder in Table 5 (page 539) and for the low-pressure cylinder in 

Table 6 (page 540). These values have been plotted in Figs. 2 to 6 

(pages 530 to 534) for the high-pressure cylinder only. 

It will be at once noticed that the indicated weight at release is 

in all cases greater than that at cut-off. That is, the missing 

quantity is less or there is more steam shown to be present in the 

cylinder by the indicator at release than at cut-off. Keferring to 

Figs. 2 to 6 it will be seen that the difference between the two curves 

of indicated weight and the air-pump discharge curves gives the 

missing quantity in lbs. per hour at release or cut-off. The two 

curves showing these quantities have also been drawn. It is 

apparent that jacketing the high -pressure cylinder decreases the 



Junk 1905. 



STEAM- JACKETING. 



539 



TABLE 5. 

Particulars from High- Pressure Cards. 

















* 


Increase 


Number 
of Trial. 


Air-pump 

discharge 

per hour. 

(Lbs.) 


Indicatec 
(Lbs. p 

a 

Cut-off. 


I Weight 
3r hour) 

t 

Release. 


Missing 
(Lbs. p( 

a 
Cut-off. 


Quantity. 
)r hour) 

t 
Release. 


Indicated Weight 


of 1 
Indicated 
Weight 
between 
Cut-off 
and 
Release. 


Actual 
a 
Cut-off. 


Weight 
t 
Release. 


108 


1603 


792 


1140 


811 


463 


0-53 


0-73 


348 


109 


1503 


780 


1107 


723 


396 


0-55 


0-755 


327 


105 


1458 


768 


1128 


690 


330 


0-572 


0-795 


360 


88 


1757 


1017 


1403 


740 


354 


0-61 


0-81 


386 


89 


1735 


1035 


1358 


700 


377 


0-63 


0-80 


323 


90 


1642 


1061 


1342 


581 


300 


0-67 


0-83 


281 


93 


1686 


1069 


1329 


617 


357 


0-67 


0-81 


260 


94 


1676 


1044 


1349 


632 


327 


0-66 


0-81 


305 


95 


2184 


1431 


1781 


753 


403 


0-68 


0-83 


350 


96 


2058 


1412 


1762 


646 


296 


0-71 


0-87 


350 


97 


1980 


1470 


1760 


510 


220 


0-76 


0-90 


290 


98 


1967 


1415 


1713 


552 


254 


0-74 


0-88 


298 


99 


1955 


1395 


1634 


560 


321 


0-74 


0-85 


239 


100 


2757 


1950 


2184 


807 


573 


0-73 


0-81 


234 


101 


2672 


1957 


2160 


715 


512 


0-75 


0-82 


203 


102 


2557 


1883 


2032 


674 


525 


0-76 


0-81 


149 


103 


2580 


1905 


2060 


675 


520 


0-76 


0-81 


155 


104 


2580 


1937 


2138 


643 


442 


0-77 


0-84 


201 



* Cushion sttam included. 



640 



STEAM-JACKETING. 



June 1905. 



TABLE 6. 
Particulars from Low-Pressure Cards. 



Number 
of Trial. 


Air-pump 
discharge 
minus 
L.-P. 
Chest 
drain. 
Lbs. per 


Indicated Weight j 
(Lbs. per hour) 

at 


Missing Quantity. 
(Lbs. per hour) 

at 


Indicated Weight 
Actual Weight 

at 


Increase 

of 

Indicated 

Weight 

between 

Cut-off 




Cut-off. 


Eelease. 


Cut-off. 


Release. 


Cut-off. 


Release. 


and 
Release. 




hour. 














Lbs. per 
hour. 


108 


1440 


806 


840 


631 


600 


0-58 


0-60 


34 


109 


1374 


697 


766 


677 


608 


0-53 


0-58 


69 


105 


1359 


1090 


1136 


■209 


223 


0- 


81 





84 


46 


88 


1601 


1164 


1186 


537 


515 


0- 


68 





69 


22 


89 


1590 


1060 


1063 


530 


527 


0- 


67 





67 


3 


90 


1522 


1013 


1049 


509 


473 





68 





70 


36 


93 


1515 


1223 


1225 


292 


290 





81 





83 


2 


94 


1510 


1267 


1306 


243 


204 





84 





87 


39 


95 


2106 


1204 1251 


902 


855 





58 





61 


47 


96 


1946 


1155 


1221 


791 


725 





61 





64 


66 


97 


1886 


1140 


1220 


746 


66Q 





62 





6G 


80 


98 


1817 


1385 


1396 


432 


421 





77 





78 


11 


99 


1807 


1522 


1547 


285 


260 





85 





86 


25 


100 


2625 


1575 


1595 


1050 


1030 





57 





62 


20 


101 


2550 


1505 


1560 


1045 


990 





60 





62 


55 


102 


2422 


1443 


1552 


979 


870 





61 





65 


109 


103 


2435 


1G78 


1746 


757 


689 





70 





•73 


68 


104 


2395 


2015 


2069 


380 


326 





•84 





87 


54 



* Cushion steam included. 



June 1905. 



STEAM-JACKETING. 



541 



missing quantity both at cut-off and at release, but it appears to 
make more difference in the former than in the latter case. This 
accords with the fact already noted that the pressure shown by the 
indicator at release in the high-pressure cylinder is smaller in the 



Fig. 9. 

Expanded Indicator Diagrams. 

Trial No. 95. No Jackets. 



ISO — 



L'O 



90 



60 



30- 



oxj i_a 




jacketed than in the unjacketed trials. The indicated weight of 
steam at release appears therefore to be greater in the latter than tbe 
former trials. Fig. 14 (page 546) illustrates this point. 

The effect of the jacketing upon both the indicated horse-power 
and the steam consumption can be seen in the expanded]^ cards shown 
in Figs. 9, 10, 11. Fig. 9 is from Trial 95 where no jackets were 



\ 



542 



STEAM-JACKETING. 



June 1905. 



used. The curves showing the volume of saturated steam actually 
passing through the engine at the different pressures have been 
drawn, and the distance between these curves and the indicator 
expansion curves shows the missing quantity at the various pressures. 

Fig. 10. 
Expanded Indicator Diagrams. 
Trial No. 97. H.P. Cyl. Jacketed. 



IW 



120— 



90— 



CO — 



30 — 



oiJ 




Fig. 10 is from Trial 97 when the whole of the high-pressure 
cylinder was jacketed. Fig. 11 (page 543) is from Trial 99 when all 
the jackets were on both high-pressure and low-pressure cylinders. 
It ought to be remembered I when looking at these figures that the 
points of cut-off in both high-pressure and low-pressure cylinders are 
the same in the three trials, and that the steam passing through the 
cylinders per hour is less for Trial 99 than for Trial 95. 



Ju^K 1905. 



STEAM- JACKETING. 



543 



From an inspection of the curves of consumption, it would appear 
that if one wished to improve the economy of an engine, one ought to 
aim at reducing the " missing quantity." In order to reduce this loss, 
one ought to have a clear idea as to its cause, so that it may be known 



Fig. 11. 
Expanded Indicator Diagrams. 

Trial No. i99. H.P. and L.P. Cyls. Jacketed. 



150 



120 



90 



60 



30 — 




how best to apply the remedy. The necessity for a true explanation of 
this source of loss has long been obvious, and various theories have 
been brought forward by different writers and experimenters. The 
most commonly accepted explanation is that the cylinder walls are 
cooled by the out-going exhaust steam, and tbat the incoming steam on 
meeting these cold walls is immediately condensed and fresh steam then 
passes in to supply the place of that condensed. A careful inspection 



544 STEAM- JACKETING. JuNE 1905. 

of most text-books reveals the fact, that it is the general opinion 
that the temperature of the cylinder walls follows that of the steam, 
and that at the end of release the walls are cooled down to the 
exhaust temperature. However, Professor Ootterill has pointed out 
that, if this were true, the missing quantity would be much greater 
than it usually is. He therefore inferred that the temperature range 
of the metal was less than that of the steam, and suggested that 
although the maximum temperature of the metal and steam would 
be probably the same, yet the minimum temperature of the metal was 
higher than that of the steam. 

The calculation of the amount of heat flowing into a metal plate 
whose surface undergoes a periodic temperature change is given in 
the Appendix (page 555). It will be there seen that if T^ is half 
the temperature range in degrees Fahr. and N the number of 
periodic changes per minute, the heat flowing into the metal per 

square foot per cycle = -r^ British Thermal Units. 

In applying this formula to ascertain the condensation taking 

place per revolution in a steam-engine cylinder, it is generally 

assumed that the surface of the metal goes through the same 

temperature cycle as the steam. It is also usual to assume that the 

temperature changes follow the simple harmonic law, and that the 

clearance surfaces may be treated as if they were portions of an 

infinite plate. Consider Trial No. 95. In this case the temperature 

range of the steam is from 357° F. to 247° F. = 110° F. or T^ = 55. 

At 60 revolutions per minute the B. Th. U.'s absorbed per square foot 

, . 4 Ti 4 X 55 

perrevolution = ^^=-^^. 

Therefore the B.Th.U.'s absorbed per square foot per minute 
= 220X60 ^ 220 X 7-73 = 1700 

II one imagines that all the condensation has taken place at 
cut-off, which is not quite the case, then one may take the amount of 
surface upon which condensation takes place as being equal to the 
clearance surfaces together with half of the barrel surface exposed up 
to cut-off ; this amounts to 6 square feet at one end. 



June 1905. STEAM-JACKETING. 545 

Therefore the heat absorbed per minute = 6 X 1,700 = 10,200 
B.Th.U.'s at each end. 

Taking the latent heat at 860 ; 

Pounds of steam condensed per minute = '— =11-9 at 
each end ; 

Therefore the steam condensed per hour at both ends = 11*9 x 
60 X 2 = 1,428 lbs. 

In the actual trial the missing quantity at cut-off was only 
753 lbs. per hour, so that it is evident that the cylinder walls did 
not have so great a range of temperature as the steam. 

To give an idea of the inaccuracy involved in the supposition 
that the temperature of the steam follows a simple harmonic law, the 
indicator diagram from the high-pressure cylinder in Trial 95 has 
been re-drawn so as to show the varying temperatures of the steam 
on a base of crank-angles. The temperatures have been obtained 
from steam tables on the assumption that the temperature at any part 
of the stroke is the temperature of saturation. This cycle has been 
analysed by the Fourier method, and its equation has been found to 
tobeT- 283-5 + 52-3sin(^ + 18°-35') + 22-3 sin (2 6* - 8°-47') 
4-16-5 sin (3 ^ - 24°-43') + 2-6 sin (4 ^ - 70°-17') - 2-2 sin 
(5 (9 4-71° -34'). 

The first 3 sine-curves have been drawn in Fig. 12 (page 546). 
The dotted black curve shows the sum of these three curves and the 
amount it differs from the actual temperature curve drawn in a full 
black line will be seen. The rate at which heat would be absorbed 
and rejected has been calculated from the first three terms of 
the equation, and has been plotted in Fig. 13 where the steam 
temperature is again shown. The quantity of heat that this 
represents to be flowing into the metal up to cut-off has been 
calculated to be 30-5 thermal units per square foot per revolution. 
It is proportional to the area shown by the shaded lines. From this 
it is estimated that the heat absorbed by the metal at both ends 
is equal to 1,320,000 thermal units per hour, which represents a 
condensation of 1,530 lbs. of steam per hour. This is equivalent to 
saying that, if the metal went through the same temperature-cycle 
as the steam, the missing quantity at cut-off ought to be 1,530 lbs. 

2 Q 



546 



STEAM-JACKETING. 



June 1905. 



O 

Ps 

2 I 






O 



i^ 



I 




Pressure. 













June 1905. STEAM- JACKETING. 547 

per hour instead of the actual amount 753 lbs. It appears then that 
if the whole of the missing quantity is put down to condensation, the 
opinion of Professor Cotterill must be accepted that the temperature- 
range of the metal is less than that of the steam. 

The researches of Messrs. Callendar and Nicolson throw a 
considerable amount of light upon this question, and bring forward a 
source of loss that has hitherto been in most cases unconsidered. It 
has already been shown that if the temperature-range of the surface 
is known, the range of any depth within the metal can be found. 
Conversely, if the range at any depth of the metal can be measured, 
then, assuming a simple harmonic temperature-range at the surface, 
the surface range and also the heat absorbed per cycle can be easily 
calculated. Messrs. Callendar and Nicolson set themselves to 
measure the range of temperature at various depths within the metal 
of a steam-engine, and found that the surface temperature instead of 
following that of the steam went only through a very small range. 
Thus, at a speed of 70 revolutions per minute, whilst the steam 
temperature within the cylinder varied for 335° F. to 212° F., a 
range of 123° F., the temperature of the inside surface of the metal, 
only went through 7° F. From this it was easy to show that only 
a small proportion of the missing quantity was due to condensation, 
and they suggested that the greater part of this missing steam was 
due to direct leakage past the slide-valve into the exhaust-pipe. It 
followed also from their experiments that the rate of condensation 
of steam, instead of being practically infinite as had been generally 
supposed, must be limited. From these and other experiments,* it 
was suggested as a provisional law, that the rate at which steam 
condenses on a metal surface is proportional to the difiference of 
temperature between the steam and the metal, and may be given by 
the following equation : 

B.Th.U. given out per second per square foot = • 74 (T — B) 
where T = temperature of the steam, 6 = temperature of the metal 
in degrees Fahr. 



* Proceedings, British ABSOciation 1897, page 418. 

2 Q 2 



548 STEAM-JACKETING. JuNE 1905. 

If this law is taken to be true, then it is easy to make an 
estimation of the amount of condensation per revolution. Figs. 15 
and 16 (page 649) are the indicator cards from the high-pressure 
cylinders in Trials 95 and 97, drawn so as to show the variation of the 
temperature of the steam with the angles traced by the crank. One may 
therefore say that the base line represents time. Thermometers in the 
metal of the covers give the average temperature of the clearance 
surface. The holes for these thermometers were drilled to within a 
distance of about |-inch from the inside surface, leaving sufficient metal 
to avoid any possible cyclical change of temperature. These average 
temperatures are shown on the diagrams. Assuming then that Messrs. 
Callendar and Nicolson's experiments were correct, the fluctuation of 
temperature of the surface of the metal is so small as to be practically 
negligible. If also their law of coDdensation is taken for granted, then 
at any instant condensation is taking place on the clearance surfaces 
at a rate proportional to the diflfeience of temperature between the 
metal and the steam. Thus the difference between the horizontal 
line and the indicator line measures the rate of condensation at any 
time. Kemembering that the base line represents time, it can be 
seen that the amount of condensation, or the amount of heat 
passing into the metal up to cut-off, is proportional to the shaded 
area. If therefore this area is measured by means of a planimeter, 
the amount of condensation can be calculated from the equation : — 
Condensation per square foot per second = * 74 (^T — 0), because 
the area represents %{T — 6)dt to a scale that can be found. It 
therefore follows that this area, measured in degree-seconds, 
multiplied by • 74 times the area of the clearance surface gives the 
condensation in thermal units taking place on one end of the 
cylinder per minute. 

If the distribution of temperature along the barrel is known, the 
condensation taking place on the barrel surface up to cut-off can also 
be found. The method adopted is somewhat similar to the one 
immediately described for the clearance surfaces, but is a much more 
tedious process. The barrel surface must be divided into strips, and 
the condensation area for each strip found between different crank- 
angles. The amount of condensation taking place on each strip per 



June 1905. 



STEAM-JACKETING. 



519 



Figs. 15, IG and 17. — Temperature of Steam on base of Cran1s\. ingles. 




270 300 330 360 30 60 90 120 150 180 ZIO 240 Z70 

Crank Angles 




Tliermonuler tube 



FiG.ns. 

Section 

through 

Loiv-pressure 

'Cylinder. 



550 STEAM- JACKETING. June 1905. 

minute can thus be measured, and these amounts added together will 
give the barrel surface condensation. 

The relation between the temperature of the steam and the 
temperature of the metal, both in the clearance space and along the 
cylinder-walls, can be seen in Figs. 9, 10, 11 (pages 541-543). 
The temperatures of the steam in relation to the stroke have been 
obtained from the indicator diagrams and plotted to the same scale as 
the observed temperature of the cylinder walls. The manner in which 
the thermometers are arranged to measure the wall temperatures is 
shown in Fig. 18 (page 549). The condensation areasfor the unjacketed 
Trial 95 and the jacketed trial have been measured for the high- 
pressure cylinders. As might be expected, the mean temperature of 
the metal for Trial 97 is higher than for 95, but the difference is not 
very marked. The results obtained are as follows : — 

Trial 95.— 
Condensation on clearance surface per minute = 665 B.Th.U. 

„ „ barrel „ „ „ = 159 „ 

Total condensation up to cut-off per minute = 824 „ 

Therefore steam condensed per hour at average temperature of 
355^ F. at each end = 57 lbs. 

Therefore total condensation per hour =114 lbs. 

Trial 97.— 
Condensation on clearance surfaces per minute = 530 B.Th.U. 

„ „ barrel „ „ „ == _43 „ 

Total condensation up to cut-off per minute = 573 „ 

Therefore steam condensed per hour at average temperature of 
355° F. at each end = 40 lbs. 

Therefore total condensation per hour = 80 lbs. 

So that the effect of jacketing the high-pressure cylinder is to 
reduce the initial condensation from 114 lbs. to 80 lbs. per hour. If 
it be assumed that the law of re-evaporation is similar to that of 
condensation, at the end of expansion the steam in the cylinder will be 
dry, so that any missing quantity at release must be due to leakage. 

If one tries to picture how this leakage takes place, it will be 
seen that the action must be very complicated. Let any leakage 
there may be past the piston be neglected altogether. During 



June 1905. STEAM-JACKETING. 551 

admission to the high-pressure cylinder there is the probability of a 
direct leakage past the exhaust-valve from the high-pressure steam 
in the cylinder. If, as Messrs. Callendar and Nicolson say, the 
leakage is proportional to the difference of pressure on the two sides 
of the valve, the leakage during this period may be very great. 
After cut-off the pressure in the cylinder diminishes, so that the rate 
of leak from the cylinder to the exhaust will diminish, but at the 
same time steam will now leak past the admission-valve into the 
cylinder at a rate increasing as the pressure in the cylinder gets 
lower. It would appear then that during expansion more steam 
would leak into the cylinder than out of it, and one would expect to 
have a greater indicated weight at release than at cut-off. After 
release, steam would continue to leak past the admission-valve into 
the cylinder, and be carried out with the exhaust steam. 

Fig. 17 (page 549) shows the low-pressure card for Trial 97 set out 
to a base of crank angles and temperatures. The condensation area has 
been calculated as in the previous cases with the following results : — 

Condensation on clearance surfaces per minute = 3,010 B.Th.U. 
„ „ barrel „ „ „ = 780 „ 

Total coridensation up to cut-off „ „ - 3,790 „ 

Therefore steam condensed per hour at average temperature of 
235° F. at each end = 239 lbs. 

Therefore total condensation per hour = 478 lbs. 

In Trial 99, where the low-pressure cylinder was jacketed with 
high-pressure steam, the mean clearance surface temperature is far 
above the saturation temperature of the steam, and consequently 
there could be no initial condensation. Fig. 11 (page 543) shows the 
relation between the wall temperatures and the steam temperatures on 
the assumption that the latter was saturated. This was probably not 
the case, as it is possible that the steam in the cylinder was 
superheated at some portions of the stroke. 

In Trial 97, the missing quantity of cut-off =746 lbs. Probably in 
this case the greater part of the leakage is directly past the valve 
into the exhaust. If this be assumed then one has of the missing 
quantity at cut-off, 478 lbs. condensation per hour, which gives 
746 - 478 = 268 lbs. per hour leakage. 



552 STEAM-JACKETING. June 1905. 

In Trial 99 the missing quantity at cut-off is 285 lbs., all of which 
may be put down to leakage. 

From these examples it is apparent that the effect of the jacket 
is to reduce the leakage as well as the condensation, especially in 
the high-pressure cylinder. In connection with the low-pressure 
cylinder a complication is introduced by the fact that the steam-chest 
pressure is increased when the jackets are used. This agrees with 
the results of Messrs. Callendar and Nicolson, who found by warming 
the valve-face, a considerable reduction in the leakage past a 
slide-valve was effected, and shows that one valuable feature of 
the jackets is that they warm the valve-faces, and so prevent the 
steam leaking past them in the form of water. More direct evidence 
is, however, given by these trials that the temperature-cycle of the 
metal must differ from that of the steam. If both had the same 
ranges, then they must also have had the same mean temperature, but 
in all cases the mean temperature of the metal is considerably higher 
than that of the steam. Thus in Fig. 15 (page 549) the mean 
temperature of the metal is 335° F. and the mean of the steam is 
284° F. Since the highest temperature of the steam is 356-5, the 
maximum temperature-range of the metal could not be more than 
2 (356 • 5 — 335) = 43° F. unless the temperature-cycle of the metal 
were of an extraordinary character. Taking this temperature-range 
and assuming it to be of a simple harmonic nature, the formula : 

4 7" 

Heat absorbed per revolution = _i, may be applied to find 
the condensation. 

On working this out for the conditions of the trial, it appears 
that such a temperature-range would give a total condensation, or a 
missing quantity near cut-off of 555 lbs. of steam per hour. Now, as 
the actual missing quantity is 753 lbs., and as it seems impossible that 
the temperature-range could be more than 43° F., it appears that even 
without the evidence of Messrs. Callendar and Nicolson's experiments 
a considerable quantity of the missing quantity must be due to 
leakage. An inspection of the indicator cards shows what again seems 
to the author to be a strong proof that condensation accounts for only a 
small portion of the missing quantity. Figs. 9, 10, and 11 (pages 541, 
642, and 643) show that, both for the jacketed and unjacketed 



June 1905. STEAM-JACKETING. 553 

trials, the indicated weight between cut-oflf and release is always less 
than the actual weight shown by the air-pump. It would thus be 
generally said that the steam was wet throughout expansion in all 
cases. There is no doubt then, that, if water were present in the 
cylinder, the re-evaporation between cut-off and release would be 
greater from the hot jacketed walls than from the unjacketed ones. 
Both the indicator cards and Table 5 (page 539) show that this is not 
so. Indeed it appears that in almost every case there is less apparent 
re-evaporation during expansion with the high -pressure cylinder 
jacketed than there is when it is unjacketed. It is difficult to see 
how this can be so, unless the view be accepted that the missing 
quantity is largely due to leakage, and that the steam in the 
cylinder is dry before release takes place. 

Although it is the almost universal opinion that initial 
condensation on the cylinder walls is the prime cause of inefficiency 
in steam-engines, the author feels that the work of Messrs. Callendar 
and Nicolson and the experiments that have been described in this 
Paper afford considerable evidence that this is not the case. The 
subject, however, is one of so much difficulty and one upon which so 
varied opinions are held, that it is only by considerable experimental 
research that the actual truth can be reached. Seeing that the 
elimination of the missing quantity would so greatly reduce the 
running cost of all steam-engines, it is not staging too much to say 
that the subject is worthy of the attention of all experimenters upon 
the steam-engine. 

The author has now commenced a series of experiments to 
determine the effect of superheated steam, and he hopes soon to have 
some more evidence of what is the actual cause of the steam losses. 
If it can be proved that valve leakage is the principal cause of loss, 
then engine builders will have their attention directed to the 
improvement of the valves now used, and the costly and complicated 
arrangements that are often made to reduce initial condensation will 
be done away with. 

Summary of Eesults. 

The following is a summary of what appear to be the most 
important results obtained from this series of trials. 



554 STEAM-JACKETING. JuNE 1905. 

(1) Best mean pressure. — That compound condensing engines with 
a boiler pressure of 150 lbs. may be worked with a mean pressure 
referred to the low-pressure cylinder of about 40 lbs. per square 
inch without any loss of efficiency in terms of the brake horse-power. 

(2) Maximum efficiency. — That the jackets have their maximum 
efficiency when the whole of the high-pressure and the ends of the 
low-pressure cylinders are jacketed with high-pressure steam. 
(See Figs. 2-6, pages 530-534.) 

(3) Variation of indicated Jiorse-jpovcer. — That when the jackets 
are applied to the high-pressure cylinder the total indicated horse- 
power is slightly reduced, but when applied to the low-pressure 
cylinder the total indicated horse-power is considerably increased. 
{See Table 4, page 536, and Fig. 7, page 535.) 

(4) Initial condensation. — That the jackets have little effect in 
the high-pressure but have a considerable effect in the low-pressure 
cylinder upon initial condensation. (See example worked out from 
trials Nos. 95, 97, 99.) 

(5) Temperature-cycle of metal. — That the temperature-cycle of 
the cylinder walls next to the steam must be considerably less than 
that of the steam. 

(a) Because the actual missing quantity is much less than 
it would have been, had the steam and metal gone through 
the same temperature changes. 

(6) Because the mean temperature of the metal is higher 
than that of the steam. 

(6) Missing quantity. — That the greater part of the "missing 
quantity " must be due to leakage and not to initial condensation. 

(a) Because the application of the methods suggested by 
Messrs. Caller dar and Nicolson shows the condensation 
to be but a small fraction of the total missing quantity. 

(6) Because the apparent re-evaporation during expansion is 
less for jacketed than for unjacketed engines. 

The Paper is illustrated by 18 Figs, in the letterpress, and is 
accompanied by an Appendix. 



I 



June 1905. STEAM- JACKETING. 555 



APPENDIX. 

If heat is being steadily transmitted through a metal plate 
thickness x whose temperature on one side is T^ and jon the other 
is T2, then, if Q be the quantity of heat transmitted in unit time 

T — T 
across unit area, it is known that Q = K , — ^— — ^, where K is 9, 

constant called the Conductivity. If the flow is not steady, one can 
find the heat transmitted across any section at any time if one knows 

d T d T 

the thermal gradient -r— at that section, for Q = — K . ^^ 

Consider now the flow of heat through a long bar of section A, 

Let T be the temperature at a distance x from one end above the 

surrounding medium, and let the mean value of the thermal gradient 

d T 
over cross-section A at this point be ,— 

d T 

The flow of heat per unit time = — K A ^ ^ 

d T 
The temperature of an adjacent point x-{-Sx=T-\-^\^'^x 

Therefore flow of heat across a section at the point x -]- S x = 

If the flow of heathas attained a steady condition, the difference 

between the heat flowing into any layer thickness 8 x and that 

flowing out must be lost in radiation. 

d^ T 
This difference = K A , o . 8 x 

d x^ 

Let the perimeter of the bar = p 

Then surface of the layer = p . S x 

If ^ = the surface emissivity of the bar. 

Heat radiated = E . T . p . S x 

Therefore K . A , ^, S x = E . T . p . S x 
Therefore K . A . '^, = E . p , T . (1) 



656 STEAM-JACKETING. June 1905. 

Next consider the flow of heat before a steady state is a;Cquired. 
During this stage the difference between the heat flowing into and 
that flowing out of an elementary layer thickness 8 a; is not all 
radiated. Some of it is spent in raising the temperature of the 
layer. If in time 8 t the mean temperature of the elementary layer 
increases 8 T and if C = thermal capacity per unit volume. 

Heat spent per unit time in raising the temperature of the 

d T 
layer = C . A , h x . ^-p 

Equation (1) becomes A . K , ^. = E . jp . T + A , C , ^ (2) 

If the bar be covered with a non-conducting substance, or if 
instead the heat be considered to be flowing through a plate of 

infinite area, then E = 0, and equation (2) becomes ^ . ^-^2 = ^ 

d- T d T K 

or Tc . ^T— 2 = j7 where A: = ^, and may be defined as the 

diffusivity of the material. 

This equation gives the connection between the temperature 
at different points in the thickness of the metal and the rate of 
change of the temperature. If now we consider one face of the 
plate to be periodically heated and cooled, a thermometer placed 
within the metal would have corresponding changes of temperature, 
alternately rising and falling. 

When the periodic heating and cooling has been maintained for 
some time, the temperature-changes at a point along the bar will 
attain a fixed character, and the mean temperature will remain 
steady. 

In considering these changes of temperature it will be convenient 
to measure time by the angle turned through by a uniformly rotating 
crank. 

Let the crank make N revs, per minute. 

Then 2 tt iV" radians are traced per minute. 

Therefore the time occupied in tracing any angle 6 = t = qI^^ mins. 

Assume the surface of the plate to go through a periodic change 
of temperature so that any time T - T^ cos 0, where T is the surface 
temperature and T^ the range of temperature. Both T and T^ are 
measured from the mean temperature, that is, the mean temperature 



June 1905. STEAM- JACKETING. 657 

will be called 0, and tlie maximum temp, will be T^ and the 
minimum temp. — T^ 

The temperature changes at the surface are represented by 
T = T^ cos = T^ cos 2 TT N t, and the temperature changes within 

the metal by equation h v-^^ = jy (3) 

<>^' ^' • d^ = ^ • 2 ^ ^ Therefore ^-^, - ^- . ^^ = 

Let ^ A- = M. Then ?-^ - 2 u^ ^ J = 0. 



^- = /.. ±nen ^|-^, - ^ /x- ^^ 

The solution of this equation that satisfies the condition of the 

problem is T = T^ e~^^ cos (^ — /x, a;), which gives the temperature 

at any point within the metal at any time. 

If this be compared with the surface equation T = T^ cos Oj 

it is seen that the temperature-range is reduced from T^ to T^ e~'"^, 

and that the temperature lags behind by an amount which is 
measured by the angle fx x. 

Now cos 6 has the same value as cos {6 — 2 tt). 

Therefore when (xx = 27r the temperature waves lag behind a 
complete period, that is, they agree in phase with those at the surface. 

To find the depth at which this takes place 

ijL X = 2 77 Therefore x = — 
^ /^ 

Therefore x = ^^J^ = 

aJNtt 

From the researches of Messrs. Callendar and Nicolson it appears 
that for cast-iron the value of ^ = 1-20 [If = 5*4 0-4:*5]. 




Therefore . = ^iiU^l^Jl-L^ = ^ 



15vl 

N 



the range of temperature at this point T = T^e ^^ cos $, 

— U.X — 27r T 

SO that the maximum value of T = T^e ' = T^e = T^i. 

Thus, at the depth where the temperature-changes are in step with 
those at the surface, the range is less than 3^^^ of the surface 
range. 



558 STEAM-JACKETING. JuxE 1905. 

Since this distance has been shown to bo /vZ-W-j.^^ ^^^ 
assumes the changes to take place at the rate of 60 per minute, a; 
would equal . / = 0*5 inch, from which it can be seen that a 

thermometer placed in the metal at this depth would [give 
practically a stationary reading. 

d T 
Rate of Fluiv of Heat. Q — — K . ^-^ is the fundamental 

equation connecting flow of heat with temperature gradient. In 

d T 
order therefore to find Q at any time, one must obtain^ ^^ — at the 

surface. To do this, take the equation T = T^e~^^ cos [0 — fx x), 
differentiate with regard to x, and then put x = 
,— = — /x T^e^ cos {0 — fji x) -\- IX T^e ^ sin (0 — fi x). 
Putting x = 0, 

if (x = o) = /^^i (-^°^ ^+^^^ ^)- Therefore Q = Efi T^ (cos6-sm0). 

We must next find for what period of time or between which 
values of 6 heat is flowing into the metal. 

Q becomes zero when cos — sin 6 = 0, or when cos = sin $. 

This is the case when ^ = j or 



4 V.X ^ 

Again Q has its maximum positive value when cos — sin ^ is a 
maximum. To find the value of when this takes place, differentiate 
and equate to zero. 

^Q (cos ^ — sin e) = 
— sin ^ — cos $ = 
cos = — sin ^ 

This happens when 0= — -, and since Q is zero when 
= A and — ~T-, it can be seen that heat must be flowing into the 

metal whilst changes from — ~^ to ^ 

From this one can obtain the amount of heat received (or rejected) 
per revolution of the crank. 



June 1905. STEAM- JACKETING. 569 

Since heat flow per unit time = Q = K fx T^ (cos 6 — sin 6) 

_Q K^ 

2ir N ~ 27riV' 



Therefore the heat flow per unit angle = o"^ = o—iijf^ ^i 



(cos — sin 0) 

Therefore the amount flowing in per revolution 



2Vivr ^.^ ^1 (eos ^ - sin ^) r? ^ = 2^Z /. Ti 
Stt 



4 

(cos ^ -- sin ^) cZ ^ 



= ~K^.TA[,in0Y + [cose] 4 

I - 3 TT -Sir 

4 r 



V2/ 



— ^ XT ., T o /n _ V 2 JC /i T j 



But/i = A./^ 



X / 2 
Therefore heat flowing in per revolution = ~tl\/ ~j^ • ^i 

Substituting the values of E^= 5-4 and ^ = 1-2 

Heat flowing per revolution = -y^=; . / ? m 

° ^ Vl-2 V 3-14 N ^ 

_ ^ ^1 

Or the heat absorbed in British Thermal Units per square foot of the 

4 T 
metal plate, per revolution of the crank, is equal to y=p, where T, 

is half the actual temperature-range of the metal and N the 
revolutions made by the crank per minute. 



560 



STEAM-JACKETING. 



June 1905. 



Discussion. 

Mr. Mellanbt added that since the Paper was written he had 
carried out another series of trials, in which the low-pressure 
cylinder alone was jacketed. The main results were given in the list 
below, which ought to be compared with Table 3 (page 536) : — 



Nature 
of Trial. 


Best Number 

of 
Expansions. 


Steam per 
I.H.P.-hour. 


Mean Pressure reduced ' 

to L.P. Cylinder 

at 60 revs. 


L.P. ends only j 
jacketed . ) 


9-5 to 15 


17-4 to 17-5 


44 to 36 



Mr. Yaughan Pendred thought he had said almost everything 
he had to say on the subject in the recent discussion in London,* 
but would like to take the opportunity of bringing forward a matter 
which he thought deserved further investigation than it had yet 
received. Certain results had been obtained in America and in 
Britain in which the effects of compression on the utility of the 
jacket appeared to be so remarkable that it was worth while 
making further enquiry into the subject. He wished to guard 
against the possible misconception that he was making any assertion 
on the subject ; what he had to say was entirely by way of a 
question as to whether the matter had been worked out by others 
and made the subject of investigation. 

Taking a diagram with a very square compression corner, it had 
been found in the cases he had mentioned that the utility of the 
jacket was exceedingly small, but as soon as there was compression the 
utility of the jacket immediately increased. He did not understand 
why that should be so. He knew that, with some American 
pumping-engines of very high efficiency indeed, it was found by 
experiments carefully worked out over considerable periods that the 
jackets were of no possible utility and were productive of no good 
result ; that was to say, condensation took place in the jackets, but 
the condensation in the jackets was exactly made up by what was 

* Proceedings 1905, Part 2, page 280. 



June 1905. STEAM-JACKETING. 561 

saved in the cylinder. There were various other cases which might 
be mentioned in which that had occurred, and he wished to know 
whether any engineer had worked out the subject and considered 
what was the general effect of compression on high efficiency or low 
efficiency of the jacket. 

There was one other question he would like to bring forward, 
namely, leakage. This question appeared to go a long way towards 
upsetting various conclusions that had been mathematically 
established with much accuracy as to the quantity of leakage that 
must take place in a cylinder. Either the leakage must be wrong 
or the deductions wrong, for he could not at all reconcile the two 
conditions. Unfortunately Captain Sankey was not present, but 
both Mr. Mark Robinson and Mr. Davey were in the room and 
might be able to say how it had come to pass that all the interesting 
deductions as to cylinder condensation had been nullified — that it 
was valve leakage and not cylinder condensation that was taking 
place. 

Mr. Alfred Saxon congratulated the author on his Paper ; and 
said that on the general question of jacketing and superheating he 
was fairly in agreement with him. He took exception, however, to a 
statement made in the last paragraph of page 532 where it was said, 
" One practical result worth emphasising is that the mean effective 
pressure of a compound engine may be made much higher than it 
usually is in practice without any loss of efficiency." The accuracy 
of this statement depended, in his opinion, on the ratio of the cylinders. 
If the assertion of the author without qualification were true, then 
engine builders would be grateful for the information, because they 
could sell smaller engines for the same indicated or nominal power, 
but engine builders knew the extent of the limitations which 
undoubtedly existed. Under the summary of results (page 554), 
practically the same statement was made, and there again he joined 
issue with the author, who spoke of about 40 lbs. as the best mean 
pressure. This general conclusion had been arrived at from one 
set of experiments with an engine of a particular ratio of cylinders, 
namely, 3 to 1 (a ratio which is not usual in large engine practice) 

2 R 



562 STEAM-JACKETING. June 1905. 

(Mr. Alfred Saxon.) 

and it should not be adopted without being first carefully examined. 
Although this was no doubt the best mean pressure for this 
special engine and particular ratio of cylinders, it would not 
necessarily be so in other cases. On that point he thought it 
would be well to bring forward some data with regard to compound 
condensing engines, because the author was dealing with an engine 
of this class. All engines were not condensing engines that were 
compound, and he thought the author must have intended in 
that statement to use the words " that compound condensing 
engines with a boiler pressure of 150 lbs.," and so on. He then 
submitted a series of theoretical diagrams giving the measure of the 
volume of steam at the point of the cut-off in the high-pressure 
cylinder with various ratios of cylinders. 

In explanation of the diagrams. Fig. 19 (page 563), it would 
be observed that whilst the total I.H.P. and the average pressures 
referred to the low-pressure cylinder were the same in each case, and 
also the point of cut-off in the low-pressure cylinder, there was 
considerable variation in the length of the steam line on the high- 
pressure diagrams and also in the steam consumption, even when the 
smaller areas of cylinders were taken into account. Working out the 
steam consumption with the various points of cut-off he found that, 
taking the 3 to 1 ratio of cylinders as a basis, cylinders of the 
ratio of 4 to 1 would use 2 * 2 per cent, more steam, which increased 
to 19*9 per cent, with a ratio of 5 to 1, and 47*8 per cent, with a 
6 to 1 ratio. It was possible that the above percentages might be 
modified under actual working conditions by altering the point of 
cut-off in the low-pressure cylinder. From these diagrams and 
calculations it would appear that, with the low ratios of cylinders 
experimented with by the author, a high mean-pressure might be 
economically employed, but that, with the higher ratios of cylinders 
a lower mean pressure should be used for economical working. 
His point was that, while the author was perfectly right with regard 
to the best results for the ratios of cylinders he had been 
experimenting with, yet his advice was unsound with regard to 
compound engines generally, unless the ratio of cylinders was taken 
into account and the mean pressure varied to suit. The author 



Junk 1906. 



STEAM-JACKETING. 



563 



Fig. 19. — Theoretical Illustrative Diagrams. 

The Total I.H.P. in each of the 4 Cases = 137. 

Average Pressure reduced to the L.P. Cyl. in each case = 40 lbs. per sq. in. 



Lbs. 
ISO 



r>^ ^ '^ 

>S -2 ^ 

««i ^ ^"^ 



•^ 
^ 



1 

looJ 



^ Boiler Pressure' 



50- 



1 \ I 






V, 



'"1^. 1^- 



I "" + 






1 

0^ 



^ ATM-U^NE 

14-7 

'^ PURE VACUUM LINE 



.2 (3 to 1= 63-0 Average 72 I.H.P 



"§ )4tol= 80-9 

rt )5tol= 9S-6 

r;^ l6 to 1=115-4 

O 



69-3 
67-6 
65-9 



\Lbs 






.-"-^ 



.^«£.^ 



1 



^-t 



H - -+ -AXI^-lVLne 



Il.p 



ko 



— A-0 



_l I I 

PU 



. _) L _ -1 -_ ^ 

RE VACUUM LINE 



i^ ko 
TI 1 _ J 



Q 



[6 to 1=20-7 Average 71-1 I.H.P. 



15 to 1=20-2 
14 to 1=19-7 
1 3 to 1=19-0 



69-4 
67-7 
65-0 



Cylinder Ratios. 


3tol 


4tol 


5tol 


6 to 1 


Calculated Steam Consumption in] 
lbs. per I.H.P. per hour including 1 
waste in Clearances and 10°M 
Diagram Factor. j 


Lbs. 
10-96 


Lbs. 
11-21 


Lbs. 
13-15 


Lbs. 
16-2 



2 R 2 



564 STEAM- JACKETING. June 1905. 

(Mr. Alfred Saxon.) 

should have limited his statement to the result of his particular 
experiments and not laid down something as a general recommendation 
which might be misleading to those who adopted it. The speaker's 
firm had been supplying in ordinary practice compound engines 
with cylinder ratios varying from 4 to 1 up to 6 to 1, and it was 
generally considered in Lancashire practice at the present time that 
a 4 to 1 ratio of cylinders for compound engines of 150 lbs. gave 
the best results, with a mean pressure of about 35 lbs. referred to 
the low-pressure cylinder. The author had referred to the fact 
that the results he had offered to the meeting in the experiments 
were not very good. It was a point which might be considered in 
connection with this question of ratios and best mean-pressures, that 
whereas engineers had often to guarantee results of steam consumption 
not exceeding 12 lbs. of steam per I.H.P., the lowest result the author 
gave was 16*93 lbs., which no doubt was fairly good for a small 
experimental engine. 

Another point which had been raised was with regard to the 
missing quantity. He was of opinion that the missing quantity was 
proportionally greater in small engines than it was in large engines. 
The steam consumption in large engines could be more accurately 
accounted for, the difference between the calculated steam consumption 
from the indicator cards and the water measured in passing from the 
tanks into the boilers being approximately 10 per cent. He thought 
this was a reasonable explanation of the missing quantity which 
Dr. Nicolson and his colleagues found to be a very important feature 
in small engine experiments. 

Mr. Henry Davey, Member of Council, considered the Paper 
most interesting because it contained so many anomalies. When 
one came to discuss steam-engines, their economies, and the economy 
to be derived by various appliances connected with the engine, one 
was continually brought suddenly face to face with anomalies. 
The author started with a statement that it was advisable first to 
ascertain what was the most economical mean-pressure to be used. 
He should have pursued that investigation a little further before 
dealing with the application of the jacket. The author's 



June 1905. STEAM-JACKETING. 565 

experiments had led to economical results not more advantageous 
than 17 lbs. of steam per I.H.P. per hour. It was well known that 
there were a great many engines in use with the steam-pressure the 
author had been experimenting with, or an even lower steam-pressure, 
that had consumptions as low as 11 lbs. per I.H.P. per hour. It 
would have been much better if the author had brought his engine 
into that economic condition of producing one I.H.P. for about 11 
or 1 2 lbs., and then had tried the engme both with and without the 
jacket. He had been experimenting with an engine that was not in 
itself economical, and the results obtained from such an engine were 
misleading if applied in general practice, as had been pointed out 
very forcibly by Mr. Saxon. If the Paper were not very minutely 
studied, and the general results given in the Table were looked at 
alone, it might be concluded that eight expansions was the most 
economical condition for steam of 175 lbs. absolute pressure. Steam 
expanded eight times from 175 lbs. absolute pressure would give a 
terminal pressure of about 22 lbs., and it could not be imagined that 
any engine having such a terminal pressure could give economical 
results. The best results had been obtained with terminal pressures 
below 6 lbs. absolute. He thought there must be something 
very wrong with the engine experimented upon, if it had given 
its best result with eight expansions with such a pressure of 
steam. 

With regard to the general question of steam-jacketing, it had 
been found in practice that the more economical the engine was as an 
engine, apart from the jacketing, the less advantage was gained by 
the application of a jacket. By designing the engine properly, it 
might be got to that state of perfection that it gave an economy 
without the jacket so high that the advantage of a jacket, however 
perfectly applied, would only give a few per cent, economy, if it gave 
any economy at all. He had studied the question of steam-jacketing 
for a great many years, and had seen the results of experiments that 
had been made by various experimentalists, and that was the general 
conclusion he had como to, and he did not think it was very far 
wrong. 



566 STEAM-JACKETING. June 1905. 

Mr. Mark Robinson, Member of Council, said that Mr. Davey 
had already dealt with several points to which he had himself 
desired to call attention, but, with reference to the recommendation 
of a mean pressure of 40 lbs. (page 554), it might be of interest to 
the members to know that one firm, which had for many years 
prided itself on the high economy of its steam-engines. Lad 
practically adopted 40 lbs. mean pressure for all reasonably high 
boiler pressures, such as eleven atmospheres and upwards. 

Mr. Alfred Saxon asked with what ratio of cylinders. 

Mr. Mark Robinson said that must depend upon the steam 
pressure, and then upon the vacuum obtainable; the ratio of the 
cylinders ought to be such as to fit those conditions. The best 
mean-pressure to work at depended on the total range of pressure 
available, and if that was on modern lines the ratio of cylinders and 
points of cut-ofi' which gave 40 lbs. mean pressure would give on the 
whole the best results. No doubt 40 lbs. was rather on the high 
side, but, taking friction into account, there was an almost equally 
good result, per brake horse-power of course, when the mean pressure 
was anything from 33 lbs. to 40 lbs. ; the curve of consumption was 
nearly flat in that part. Reference had been made, as though the fact 
were remarkable, to the high-pressure card being slightly reduced 
in area when the high-pressure cylinder was jacketed. Surely that 
was to be expected, and meant merely that, owing to the jacketing, 
less initial condensation took place ; therefore a smaller quantity of 
steam was taken into the cylinder for any given cut-off, and a less 
quantity of high-temperature water remained to be re-evaporated in 
the later portion of the stroke. Hence the well-known effect of 
" flattening the toe " of the diagram was absent. The diagram was 
smaller, but the consumption of steam in the cylinder was smaller in 
a still greater degree. What the consumption was in the jacket was 
another matter. 

Mr. Pendred had raised an interesting point about compression, 
and had referred to valve-leakage. He was afraid Mr. Pendred 
thought rather lightly of the famous experiments of Mr. Willans, 



June 1905. STEAM- JACKETING. 567 

which had, as the speaker believed, completely established the 
connection between initial condensation and the " missing quantity." 
Mr. Pendred thought that valve-leakage might account for it. The 
speaker did not believe in that leakage at all, if only a fairly 
good engine were used. Professors used experimental engines that 
were nearly always fitted with slide-valves, and anything might 
happen with a slide-valve, especially in a small engine. When 
designing the Willans engine years ago, most exhaustive experiments 
were made on the leakage of piston-valves. An engine was run by 
an electric motor at different speeds, with steam admitted and cut-off 
in proper sequence, parts of the engine being disconnected in such a 
way that the leakage at successive stages could be separately caught, 
condensed, and measured. Under every condition that could be 
thought of such tests were made, which resulted in the conviction that 
it was easy to have piston-valves which were almost perfectly free 
from leakage, and which could be trusted to remain so. It was a pity 
that experimental engines were not fitted with such valves, instead 
of with the utterly unreliable slide-valve. At any rate, leakage did 
not at all affect the results of the experiments identified with Mr. 
Willans' name, nor the results of similar experiments which have 
since been frequently made. 

Professor H. Hubert said he did not intend to discuss in detail 
the author's important contribution to the study of the effect of the 
steam-jacket. He wished only to add a few observations on the 
efficiency of the jacket applied to steam-engines. He believed that, 
until recently, the trials had only been made with jackets applied to 
the external parts of the cylinder. It would, however, seem necessary, 
in order to obtain the full benefit from the jackets (an arrangement 
which dated back to the time of Watt), to apply them also to the 
piston and its rod. In fact the piston constituted an important part 
of the metallic surface on which was produced the initial condensation, 
one of the principal causes of loss in the steam-engine. Moreover, 
when the piston-rod came out of the cylinder-gland, it was cooled in 
the atmosphere, so that it was a cold body, and at the same time was 
a good conductor, which entered with each stroke into the live steam 



568 STEAM-JAOKETING. June 1905. 

(Professor H. Hubert.) 

in tlie cylinder and contributed to its condensation. The Committee 

which was formed to indicate to the manufacturers who took part in 

the Brussels Exhibition of 1897 the desiderata in respect of the 

steam-engine had referred to the heating of the piston and the 

piston-rod. A Belgian, M. Ch. Beer, overcame the difficulty by 

fitting the piston with a hollow counter-rod into which two tubes 

entered through a stuf&ng-box ; one of these tubes brought the steam 

into the piston and on to the piston-rod, and the second acted as 

a drain for the condensed steam. A steam-engine fitted with this 

and other improvements obtained the highest award. This was 

probably the first serious attempt which had been made in the 

direction indicated. Unfortunately, no comparative trials were made 

with this engine, so that it was not possible to give reliable figures 

as to the efficiency of the arrangement. 

Since then M. Georges Duchesne had applied the principle of 
heating the piston by simplifying M. Beer's arrangement and by 
using, both in the piston and in the exterior steam-jacket, saturated 
steam of a considerably higher pressure than that used in the 
cylinder. M. Duchesne first fitted the piston with a hollow rod 
placed at its lowest point and sliding in an external sleeve (through 
a stuffing-box) connected with the steam-jacket. The cross-section 
of this hollow rod was sufficient for the steam and the resulting 
condensed water to circulate without interfering with each other. 
Recently M. Duchesne had shifted the hollow rod to the centre of 
the piston, but had closed the lower part of the cavity in the piston, 
in order that water might not accumulate there. According to 
M. Duchesne, the heat conductivity of the metal was sufficient to 
maintain the piston at a high temperature throughout its mass. 
This arrangement, which had been applied to a steam-engine of the 
Bonjour type, was working in the Machinery Hall at the Liege 
Exhibition. 

Two years ago a third arrangement was practically applied by 
M. Nicolas Frangois, Jun., an engineer belonging to the Societe 
Cockerill. The device of M. Frangois consisted in the application 
of the old idea, which had not been practically utilized, of 
replacing the single piston of an ordinary engine by two 



June 1905. 



STEAM- JACKETING. 



569 



pistons placed on the same rod and kept at a distance from each 
other slightly greater than their stroke. The space between 
these two pistons was kept constantly in communication with the 
steam-jacket of the cylinder, and was thus always filled with 
steam at boiler pressure. By this arrangement, not only was the 
piston kept at a high temperature, but the cylinder surfaces were 
being constantly heated ; the condensed water was swept away by the 
movement of the piston. Fig. 20 gave a section of the cylinder 
of the Francois engine. 

M. Frangois' device was in the first instance applied to a 
Corliss engine working in one of the shops of Messrs. Cockerill's 
Works, and was tested by the late M. Vingotte, a Director 



Fig. 20. — Cylinder of the Frangois Steam-Engine. 
Vertical Section. 




of the Belgian Association for the insurance of steam plant. 
M. Vingotte found that this single-cylinder engine developed 
111-22 horse-power with a pressure of 11-03 kg. per cm.^ (157 lbs. 
per square inch) superheated 17-26° C. (63° F.), the steam 
consumption being 7-965 kg. (17-5 lbs.) per horse-power-hour. 
When the pressure was 11-14 kg. per cm.^ (158-5 lbs. per square 
inch), and 31 • 95° C. (89 - 5° F.) superheat, the power developed was 
108*42 horse-power and the steam consumption was 7-642 kg. 
(16*8 lbs.), a consumption which M. Vingotte stated in his report 
to be the lowest that he had ever obtained with a non-condensing 
steam-engine. 



570 STEAM-JACKETING. June 1905. 

(Professor H. Hubert.) 

He (Professor Hubert) continued these trials with the same engine 
after it had been supplied with a condenser. The results were very 
satisfactory, notwithstanding the fact that the condenser worked 
imperfectly owing to the air-pump, which was old, not giving a good 
vacuum. With a steam pressure of 8*934 kg. per cm.^ (127 lbs. per 
square inch) without superheat, the vacuum being only 578 mm. 
(22*8 inches) of mercury, the steam consumption at 93*48 HP. was 
6*796 kg. (15 lbs.) per horse-power-hour (4,483 calories or 17,752 
B.Th.U.). When the pressure was 10*180 kg. per cm.^ (144*7 lbs. 
per square inch), the superheat 42*45° C. (108° F.), and the vacuum 
in the condenser 665 mm. (26 inches), the steam consumption was 
reduced to 5*696 kg. (12 '5 lbs.) per horse-power-hour (3,887 calories 
or 15,392 B.Th.U.) and the horse-power developed was 90 * 12. These 
brilliant results induced the Cockerill Society to apply this system 
to a compound engine ordered by the Vieille Montagne Co., Plate 21. 
It was to be observed that, owing to the lengthening of the cylinders, 
the engine had an unusual appearance, which however did not offend 
the aesthetic sense. This engine was subjected to searching tests, 
which would be published in the " Eevue Universelle des Mines," but 
the following was the best result obtained. When the pressure was 
8*820 atmospheres (129*6 lbs. per square inch), and the condenser 
vacuum 708 mm. (28 inches), the engine developed 188*7 horse- 
power at a speed of 122 revolutions per minute, and the consumption 
was only 5*092 kg. (11 lbs.) per horse-power-hour of saturated 
steam or 3,367 calories (13,333 B.Th.U.) per horse-power-hour. He 
desired to state that this remarkable result, comparable with those 
obtained with superheated steam, was due in some measure to the 
care taken in the construction of the engine to obviate all losses of 
heat, in particular by the careful lagging of the cylinders and even 
of the valve casings, and by the great reduction of the clearance 
volumes and more particularly of the clearance surfaces. He was of 
opinion that the injurious effect of the surface of the clearance space — 
a surface on which the initial condensation was deposited — was greater 
than that due to its volume. The magnitude of this surface could be 
appreciated by comparing it with that of the piston. In the high- 
pressure cylinder of the Fran9oi8 engine the ratio between the two 



I 



June 1905. STEAM-JACKETING. 571 

surfaces was 3*78, and it was only 2-95 for the low-pressure 
cylinder. These ratios were remarkably small, whereas the clearance 
volumes were respectively 2 • 88 per cent,, and 2 • 6 per cent, of the 
volume swept by the two pistons. He added that these advantages, 
as well as the possibility of the engine running at a comparatively 
high speed, were due to an ingenious steam-distributing contrivance 
with double piston-valves ; he would not, however, describe the 
arrangement, as it had no direct reference to the subject under 
discussion, and it would be fully explained in the "Revue 
Universelle des Mines." 

The results which the author had given in his Paper were of 
such a nature as to throw a strong light on the temperature 
variations in a steam-cylinder. Mr. Mellanby had recalled the 
experiments by means of which Messrs. Callendar and Nicolson 
had endeavoured to measure the temperature of the cylinder walls, 
and their hypothesis as to valve leakage in the form of steam first 
condensed and then re-evaporated. M. Armand Duchesne (Professor 
Hubert's assistant in the Laboratory of Applied Mechanics at Liege) 
had devised a new method of measuring the temperature of the 
steam and of the cylinder walls. He had been able, by means of a 
delicate and ingenious apparatus, a description of which would 
be found in the "Revue Universelle des Mines" of July 1905, 
to measure the mean temperature of the steam and of the metal 
during each ^Qth. of the forward and backward stroke of the piston, 
and had plotted these temperatures on a diagram. The accuracy of 
the measurements was verified by the fact that the temperature 
corresponding to the pressure of saturated steam, as shown by the 
indicator, was, during the expansion, exactly the same as that 
obtained by these direct measurements. This was shown on three 
diagrams, Figs. 22 to 24 (page 572), the calculated temperatures being 
indicated by black circles. The first of these diagrams. Fig. 22, had 
been taken from a steam-engine working without a jacket, the second 
from an engine with an ordinary steam-jacket, and the third from an 
engine fitted with a steam-jacket supplied with high-pressure steam. 

These diagrams were reproduced from those prepared by M. 
A. Duchesne. They exhibited the remarkable fact that even without 



572 

(Professor H. Hubert.) 



u o 



'« 




•Si 




p 


o 


^^ 




vS Q) 


>. 


^ V. 


OS 


C s 


k-H 


S S 


F^ 


i. 'w 


(Y) 


s ^ 


1— ( 


? ^ 




;ii-^ 


OS 


^.s* 


u, 


Itq 


H 


tH 




(M 




dJ 




M 




N 





r!S ^2 O 



STEAM-JACKETING. 



TcrrvpercUtire Centigrade 



June 1905. 








'^ 




1 




/'" 


'-•-1 


r ] 










--- 


^ 










t 


















"^ «, . 


''a 








\ 




(l^ 


^ 


























\ 








■\, 
























\ 










\, 
























S. 








\ 


V- 












. ^. 








P3 






^ 


^ 






^ 






























\ 




> 


\ 




























\ 






\ \ 






























\ 








































\ 








0| 
























\ 










- " 
























\ 


\ 








A 









- 


-. 




















— X 


% 


---. 


■^■^ 


-j"^-^ 


f 
• 






^ 



Temperature Centigrade' 



ri 




S 




g 


■^ 


h 


o 


<» 


Ci 


^ 




?»» 


<A 


e 


% 


•§ 


r- 


'W 


1—1 


^ 


1—1 

«3 








M 


CO 


H 


(>< 




X*-, 








\M 





^ 


>^ 


vtl 




---- 




--- 


■--- 


-.-: 


-^ 


" «». 










^^ 


^, 
















--- 


- - 


---1 


.. .^ 


\ 


N 


































•s 


V 


N; 


V- - 














CD 1 


r 




\ 




\ 


v< 
























\, 
























N 


























\ 


o] 








•' 








> 






s 


\ — i 
























: " 







, _ - 


... \ 


_. ,. — , 











^ -\ 


•^-^ 


..__, 


iJ. 





0) TO ID 



<iMO 



Terrvperature Centigrade 



rii 



^ 


-H 


S 


s 


O 


"S 


« 


o 




<J1 


1—1 


n 


■♦J 




0) 


■^J 


>» 


o 




OS 


d 




1^ 


<a 


o 


<i 


a 


r« 


CO 


' . a (i5 


g 


'^ 




1 


M 


c* d-^3 


dA 


H 




Oi 




cS 02 f-< 
=5 O ;i 






Cut-oi 

Comp 

re 




June 1905. STEAM-JACKETING. 573 

a steam-jacket, Fig. 22, the steam was superheated in the cylinder 
towards the end of the exhaust by the action of the heat in the 
metal, the temperature of which was sensibly higher than that of 
the steam. Hence the hypothesis, so much discussed at one time, 
that water would remain in the cylinder at the end of the exhaust 
was definitely negatived. 

Diagram, Fig. 23, relating to the case of a steam-engine with an 
ordinary jacket, showed that superheating began at the beginning of 
the exhaust and raised the temperature of the steam beyond that of 
the cylinder walls ; this latter temperature varied much less than 
in the preceding case, and above all remained notably higher than 
that of the steam during expansion. 

Lastly, diagram, Fig. 24, relating to the case of a steam-engine 
with a jacket supplied with high-pressure steam, showed that the 
temperature of the metal was still higher and almost constant ; the 
steam was superheated nearly the whole time. Moreover, the indicator 
diagram made by M. G. Duchesne with this class of steam-jacket 
exhibited expansion lines which followed mathematically the 
adiabatic law for the expansion of the steam present in the cylinder 
at the point of cut-oJBf. The weight of steam calculated from the 
volume and pressure shown by the indicator, on the supposition 
that the steam was saturated and dry, corresponded exactly with the 
steam consumption experimentally measured by means of a surface 
condenser and a weighing machine (" Eevae Universelle des Mines," 
7 July 1904). 

The conclusions to be drawn from these facts were : (1) That it 
was impossible any longer to suppose that water remained in the 
cylinder at the end of the exhaust. (2) That Messrs. Callendar 
and Nicolson's leakage hypothesis must be rejected, because this 
hypothesis was in opposition to the perfectly adiabatic expansion 
observed in the experiments made in the laboratory at Liege with 
jackets filled with high-pressure steam. 

In conclusion, he asked the members of the Institution to 
examine, at the Li6ge Exhibition, the Frangois steam-engine and 
the Bonjour engine fitted with jackets with high-pressure steam 
supplied from a special boiler; this last exhibit was all the more 
interesting as tests could be made at the Exhibition. 



^74 STEAM-JACKETING. June 1905. 

M. Georges Duchesne said that he would confine himself to 
showing three indicator diagrams taken from the steam-engine in 
the Mechanical Laboratory of the University of Liege, and would 
explain in a few words what he understood to be the function of 
the steam-jacket. The first of these diagrams, Fig. 25 (page 675), 
was taken when the engine was working without jackets ; the second, 
Fig. 26, when working with ordinary jackets, and the third. Fig. 27, 
when working with hotter steam-jackets. The point of cut-off being 
at one-tenth of the stroke and the speed being low, well-defined 
phenomena were produced. The missing quantity was 43 per cent, 
for the first diagram, 25 per cent, for the second, and nil for the 
third — that was to say, that in the last case the weight of steam 
admitted per stroke was exactly the weight of steam shown by the 
indicator, which proved that there was no leakage. Moreover, 
the engine was tried at intervals of three days, and, as there 
was no leakage on the third day, it was difficult to imagine 
that there could have been any leakage on the first day. Thus 
the large missing quantity for the test without steam-jacket was 
due to initial condensation, and could not be attributed to 
leakage. Mr. Mellanby, in order to be able to ascribe this 
large missing quantity to valve leakage, in conformity with the 
researches of Messrs. Callendar and Mcolson, thought that the 
small difference of temperature between the steam and that of 
the cylinder walls during the admission period was insufficient to 
account for it. This argument omitted to take into account the fact 
that in a steam-engine the absorption of heat was greater than the 
emission. If two blocks of cast-iron were introduced into an 
atmosphere of steam vapour, one being hotter and the other colder 
than the steam, the colder block, other things being equal, would 
absorb three hundred times more heat than the hotter block would 
emit. l^See also ^age 588.] 

M. EoDOLPHE E. Mathot said that, with regard to the 
consumption of superheated steam-engines, it would be of interest 
to give a summary of results he had obtained from testing a small 
superheated steam-engine built at Magdebourg in Germany by 
Herr K. Wolf, who made a speciality of such engines. The first test, 



June 1905. 



STEAM-JACKETING. 



575 



Comparative Trials made at the 
University of Li^ge. 




Fig. 25. 
Without Jackets. 




Fig. 26. 
With Ordinary Jackets. 




Fig. 27. 
With Hotter Steam- Jackets. 



576 STEAM-JACKETING. June 1905. 

(M. Rodolphe E. Mathot.) 

Table 7 (page 577), related to a semi-portable, one-cylinder, 
superheated steam, condensing engine, which was tested at the works. 
The heating surface was 128 feet, the ratio between the grate surface 
and the heating surface being 1 to 33. The heat value of the coal was 
14,827 British thermal units, and the proportion of clinkers and ashes 
collected 6 per cent. The temperature of the superheated steam was 
about 590° F. The temperature of the gas in the funnel was only 
430° F., and the vacuum of gas in the same was | inch. The 
dimensions of the engine were : diameter of piston 5|^ inches, stroke 
11 j inches, and average number of revolutions 200. The power 
guaranteed was 25 H.P. With regard to the efficiencies, the average 
B.H.P. during the nine hours' test was 24*93; the mechanical 
efficiency was found to be 90 per cent. ; and the gross coal 
consumption was 2*37 lbs. per B.H.P. per hour. Those figures 
related to a consumption per I.H.P. of 2*16 lbs. The average mean 
pressure on the piston was 85*3 lbs. per square inch, the degree of 
admission being 31 per cent. 

A better result was shown, Table 8 (page 578), by taking a 
compound 'engine of about the same size, but working with 
condensation. The heating surface of the boiler was 103 '8 square 
feet ; the ratio between the grate and heating surface 1 to 35 ; and 
the ratio between the heating and superheating surface 0*96 to 1. 
The proportion of ashes and clinkers to the coal used was 7*68 lbs. 
The temperature of the feed-water was 97*5° F., and the average 
pressure of steam was 172 • 1 lbs. per square inch. The temperature of 
the superheated steam at the high-pressure cylinder was 657° F. and 
the temperature of the superheated steam at the low-pressure cylinder 
361° F. Under running conditions of the boiler the weight of water 
evaporated per hour per square foot of heating surface was 4*21 lbs. 
and the temperature of gas between the two superheaters was 
589° F. With regard to the efficiency of the boiler, the weight of 
water supposed to be taken at zero and evaporated at 1,142 lbs. was 
8*56 lbs. The average pressure on the piston of the high-pressure 
cylinder was 78 lbs. and on the low-pressure cylinder 27 lbs., and 
the total power in the two cylinders was 38*2 H.P. The gross 
consumption of coal per B.H.P. per hour was 1 * 33 lbs., and the steam 



June 1905. STEAM-JACKETING. 577 

consumption per hour per B.H.P. was 10*8 lbs. The degree of 
admission of steam in the high-pressure cylinders was 30 to 35 per 
cent. He thought those figures showed a very good economy, taking 
the fact of the very small dimensions of the engine into account, and 
they might be of some interest when compared with those that had 
been put forward during the discussion. 

TABLE 7. 

Test of a semi-portable Single-Cylinder Superheated Steam 

Condensing Engine, in the Works of Mr. E. It. Wolf, at 

Magdebourg-BucTcau, Germany, hy B. E. Mathot. 

Duration of Test : — 7 Jiours. 

1. Heating surface ........ sq. feet 128 

2. Grate surface . . . . . . . •mi? 3*9 

3. Superheater surface . . . . . . . „ „ 119 

4. Ratio between grate and heating surface . . . „ „ ^^^ 

Fuel. 

5. Nature : — Euhr steam coal (Germany). 

6. Chemical composition: — Carbon 81 '65; Hydrogen 4*55; 

Nitrogen 1 ' 08 ; Oxygen 7 • 52 ; Ashes 5 • 20. 

7. Heat value. ..... 

8. Gross weight of coal consumed 

9. Weight of ashes and clinkers collected 

10. Proportion of „ „ „ 

11. Net weight of fuel consumed 



Water and Steam. 

e heating 



12. Average temperature of feed water 

13. „ „ „ befor 

14. Average pressure of steam . 

15. „ temperature of superheated steam 



Running Conditions op Boiler. 

16. Number of loads of coal during test .... 
17.. Average weight of each load ..... 

18. Interval between „ „ ..... 

19. Weight of coal consumed per hour and per square foot 

of heating surface ....... 



B.Th.U. 


14,827 


lbs. 


386 


)> 


23 


per cent. 


6 


lbs. 


363 


F. 


150° 


F. 


52° 


lbs. 


172 


F. 


590° 




85 


lbs. 


5-5 


mins. 


5 


lb. 


42 


2 s 





578 STEAM- JACKETING. 

(M. Rodolphe E. Mathot.) 

TABLE 7 (^continued). 

20. Weight of coal consumed per hour and per square foot of 

grate surface . 

21. Temperature of gas at funnel 

22. Vacuum of gas „ „ 

23. „ „ in the furnace 

24. „ „ under the srrate 



June 1905. 



lbs. 


10*25 


F. 


430° 


inch 


i 


• !» 


4 


J» 


#2 



Engine. 

2o. Diameter of piston : — D : 5| inches ; Diameter of rod 
1^ inch ; Stroke llf inches. 

26. Average number of revolutions 200 ; Power guaranteed 

25 H.P. 

27. Brake lever 1 ft. 9| inches ; Weight suspended at brake 

362 lbs. 

ErrrciENCiES. 

28. Average brake horse-power .... 

29. ,, indicated horse-power 

30. Mechanical efficiency ..... 

31. Gross coal consumption per hour per B.H.P. . 

32. „ „ ., „ I.H.P. 

33. Net „ „ „ B.H.P. . 

34. „ „ „ „ I.H.P. . 

35. Price of the B.H.P. hour in coal , 

36. Average mean pressure on the piston per square inch 

37. Degree of admission ...... 



B.H.P. 


24-93 


I.H.P. 


27-54 


per cent. 


90 


lbs. 


2-37 


5» 


2-16 


?J 


2-24 


>» 


2-02 


d. 


0-141 


lbs. 


85-3 



per cent. 



31 



TABLE 8. 

Test of a semi-^ortahle Compound tandem Superheated Steam-Engine 

with condensation, made at the worTcs of Mr. E. H. Wolf^ 

at Magdehourg-BucJcau, Germany, 

by B. E, Mathot. 
Duration of Test : — 7 Jiours 5 minutes. 

BOILEE. 

1. Heating surface ...... . sq. feet. 

2. Grate surface ........„„ 

3. Surface of superheaters: H.P. 78-2; L.P. 29-1 . „ „ 

4. Ratio between grate and heating surface 

5. „ „ heating and superheating surface . . • 96 



103-8 

3 

107-3 



bV 



Junk 1905. 



STEAM-JACKETING. 



579 



TABLE 8 (continued). 
Fuel. 

C). Nature of coal : Ruhr coal (Westphalia, Germany). 

7. Chemical composition: Carbon 81*65; Hydrogen 0"G5; 

Nitrogen I'OS; Oxygen 7-57 ; Ashes 6-20. 

8. Heat value .... 

9. Gross weight of coal consumed 

10. Weight of ashes and clinkers collected 

11. Proportion of ashes and clinkers . 

12. Net weight of coal consumed 



B.Th.U. 


14,827 


lbs. 


330 


5> 


25-5 


per cent. 


7-G8 


lbs. 


303 



Water and Steam. 

13. Temperature of feed water .... 

14. „ „ „ before heating 

15. Weight of water vaporised at feeding temperature 

16. Average pressure of steam per square inch 

17. Temperature of superheated steam H.P. cylinder 



18. 
19. 



L.P. 



of testing shop 



F. 


97-5° 


F. 


45° 


lbs. 


2,658 


lbs. 


172 1 


F. 


657° 


F. 


361° 


F. 


71° 



Running Conditions of Boiler. 

20. Number of loads of coal during test .... 

21. Average weight of load ...... 

22. ,, interval between each load .... 

23. Weight of coal consumed per hour and per square foot of 

heating surface ....... 

24. Weight of coal per square foot of grate surface 

25. Weight of water vaporised per hour and per square foot 

of heating surface ....... 

26. Temperature of gas between the two superheaters . 





88 


lbs. 


3-73 


mins. 


4-8 


lb. 


0-49 


lbs. 


18-00 


>j 


4-21 


F. 


589° 



Efficiencies. 

27. Weight of water at feeding temperature vaporised per lb. 

of gross coal ....... 

28. Weight of water at feeding temperature vaporised per lbs 

net coal ........ 

29. Weight of water supposed at 0° and vaporised at 1,142 lbs 

30. Vacuum of gas at the funnel .... 

31. ,, ,, in the furnace .... 

32. ,, ,, under the grate .... 



lbs. 



inch 



2 s 2 



806 

8-74 
8-56 



^ 



_3 

3 2 



580 STEAM-JACKETING. June 1905. 

(M. Rodolphe E. Mathot.) 

TABLE 8 (continued). 
Engine. 

33. Diameter of H.P. piston 5J inches; diameter of rod and 

tail-rod 1| inches; diameter of L.P. piston 9^q inches; 
stroke 11 inches. 

34. Average number of revolutions per minute 232*7; power 

guaranteed 34 B.H.P. 

35. Average number of revolutions of brake pulley 

36. Brake lever 2 11 inches weight suspended at 

37. Corresponding brake horse-power. 5 per cent, belt 

38. Average mean pressure on piston. H.P. cylinder . 
Oi*. ff >) )j 5> >j L/.Jr. ,, 

40. Indicated horse-power. H.P. cylinder . 

41. tf jj »> i-i.tr. ,, . . . 

42. „ „ „ .... Total 

Efficiencies. 

43. Mechanical efl&ciency ..... 

44. Gross consumption of coal per hour per B.H.P. 

45. ,, „ J} >j I.H.P. 

46. Price of brake horse-power hour in coal 

47. Steam consumption per hour per B.H.P. 

48. ,, „ ,, ,, I.H.P. 

49. Degree of admission of steam 

Mr. Mellanby thanked the members of the Institation for the 
interest they had shown in the Paper, especially the Continental 
members who had taken so much trouble to go into the question of 
the heat exchanges and the temperature ranges of the steam and 
cylinder walls. He was convinced that a knowledge of the latter 
was essential before anything definite could be known of the reasons 
why losses took place in steam-engine cylinders, and he was glad 
to learn that the work of Messrs. Callendar and Nicolson was being 
followed up in the renowned School of Engineering at Liege. 

In reply to Mr. V. Pendred (page 660) he would state that he 
himself had not made any experiments upon the effect of compression, 
and was unable to explain the results cited by that speaker. He 
thought, however, that if those responsible for the trials had observed 
the mean temperature of the clearance surfaces, in the manner 
described in the Paper, more light would have been thrown upon the 





255-8 


lbs. 


230 


B.H.P. 


34-69 


lbs. 


78 


>> 


27 


I.H.P. 


20-27 


• j> 


17-93 


il ,, 


38-20 


per cent. 


90 


lbs. 


1-33 


• • )> 


1-21 


d. 


0-031 


lbs. 


10-8 


» 


10 


per cent. 


32-35 



June 1905. STEAM-JACKETING. 581 

subject. He did not agree with Mr. Pendred that the mathematical 
deductions with regard to cylinder condensation were at all nullified 
by his experiments. In fact he thought that if more attention had 
been paid to the mathematical side of the question, so many of the 
wrong ideas that were now prevalent about cylinder condensation 
would never have been formed. Mathematicians had shown that, if 
the temperature of the cylinder walls followed that of the steam, the 
initial condensation, even in very uneconomical engines, must be 
generally somewhat of the order of twice the missing quantity shown 
by trial. If they considered the extraordinary efficiency of some 
engines, the trials of which had been lately published in the 
engineering press, they would see that the cylinder-wall temperature 
ranges in these cases must have been very small in order to account 
for the minute difference between the " indicated weight " and the 
'' actual weight." 

He had to thank Mr. A. Saxon (page 561) for pointing out that 
in the summary of results (page 554) under the heading, Best Mean 
Pressure, he ought to have used the words : " That compound 
condensing engines," &c.* He agreed with Mr. Saxon that it was 
dangerous to make a general statement from one particular set of 
experiments. The results however of his own trials on many 
different types of engines, and the few series of progressive trials 
published by other writers, had convinced him that most engine- 
builders made their engines far too large for their work, if they 
wanted the best efficiency in terms of the brake horse-power. He 
was supported in his statement by Mr. Mark Robinson, whose firm 
were noted for their enterprise in carrying out engine trials, and 
whose engines were well known for their economical running. He 
was ucable to understand from Mr. Saxon's figures why the cylinder 
ratios should affect the steam consumption when the mean pressure 
was constant. From a purely theoretical point of view the horse- 
power and the consumption depended only upon the total number of 
expansions made by the steam, and not at all upon the cylinder ratios. 
Differences would occur in practice due to free drop between the 

* This addition has now been made in the Paper. 



582 STEAM- JACKETING. June 1905. 

(Mr. Mellanby.) 

high-pressure and low-pressure cylinders and to clearance losses. 
These losses could only be found by actual experiment and, unless 
one had experimental results to go upon, calculations as to the 
probable efficiencies were of very little value. Mr. Saxon himself 
evidently did not place much reliance upon his calculations, as he 
showed that the consumption was less with a cylinder ratio of 3 to 1 
than with the larger ratios, whilst at the same time he stated that 
the ratio 3 to 1 was not adopted in Lancashire engine practice. 
The question of cylinder ratios, although of the utmost importance, 
was one upon which engine-builders had considerable differences of 
opinion. It appeared that in most cases the cylinder ratios and the 
point of cut-off in the low-pressure cylinder were so arranged that 
the horse-power developed in each cylinder ought to be approximately 
equal. He had made numerous experiments upon this point and 
had found that for compound engines maximum economy was not 
associated with equal work in the cylinders. Both questions however 
of mean pressure and cylinder ratios for maximum efficiency could 
easily be settled if some firm building large slow-revolution engines 
would follow the example of Messrs. Willans and Robinson and 
carry out a series of trials on the lines of those published by the late 
Mr. Willans. 

The performance of the engine from the point of view of its 
large consumption had been criticised by Mr. Henry Davey 
(page 564). It was however scarcely fair to make comparisons 
between that engine and engines in practice, as for the sake of 
the experiments it was running at about half its normal speed. 
Since the missing quantity in lbs. per hour for any particular 
point of cut-off was practically the same for 110 revolutions as 
for 60, it would be seen that when run at its normal speed 
its performance would compare favourably with that of the 
ordinary large compound engine. Mr. Davey having objected to the 
mean pressure advocated by the author as the most economical, 
he would refer him to the answer given to Mr. Saxon. He agreed 
with Mr. Davey that the more economical an engine was the less 
advantage there was to be gained by jacketing. This, he thought, 
was another confirmation of the correctness of the valve-leakage 



June 1905. STEAM- JACKETING. 583 

theory. Engines of similar power, as built to-day by different 
manufacturers, would have little difference in the amount of their 
clearance surfaces on which the greater part of the initial 
condensation took place. It seemed absurd to imagine that the law 
of condensation of steam would vary for each manufacturer, yet it 
was known that there were considerable differences in the consumption 
per horse-power of engines turned out from different workshops. 
Surely it was easier to believe that the different " missing quantities " 
in the various cases were due more to the type and arrangement of 
valves used (factors known to have an enormous influence upon the 
quantity of steam that could leak past them) than to imagine that 
there was a special law of nature for economical engine-builders. 

He had already referred to Mr. Mark Eobinson's opinion that 
40 lbs. was a suitable mean pressure (page 566), but he would again 
draw the attention of the members to Mr. Robinson's further 
statement, which was supported by the curves in the Paper, that 
there was very little variation in the economy for a considerable 
range of mean pressure. It was obviously the more economical plan 
to build the engine for the highest rather than for the lowest mean 
pressure, and thus obtain the same result for a much smaller and 
cheaper engine. He thought that Mr. Robinson had scarcely 
understood his point with regard to the pressure at release in the 
high-pressure cylinder being reduced when the jackets were applied. 
If they referred to Trials 95 and 97 in Table 5 (page 539), they 
would see that in these two cases the indicated weight at cut-off was 
practically the same, and in both cases there was a considerable 
missing quantity at both cut-off and release. If this missing 
quantity were all due to initial condensation, it was to be expected 
that the re -evaporation from the hot jacketed walls would be greater 
than that from the unjacketed ones. It would be seen however that 
in Trial 95, with the unjacketed cylinder, the apparent re-evaporation 
was 350 lbs., and in Trial 97, with the jacketed cylinder, only 290 lbs. 
per. hour. As in Trial 97 there was a missing quantity at release of 
220 lbs. per hour, it was difficult to understand why more of this 
was not evaporated if it were present in the cylinder as water. 
None of the arguments brought forward by the different speakers had 



584 STEAM- JACKETING. June 1905. 

(Mr. Mellanby.) 

convinced him that this was the case, and he was still of opinion that 
the whole of this 220 lbs. had passed through as valve leakage. 
Mr. Robinson stated that he did not believe in the leakage theory, 
but the experiments of Captain Sankey had shown that the ordinary 
plug piston-valve as fitted by many makers might have a very large 
leakage. He would refer him to Messrs. Callendar and Nicolson's 
reply to the discussion on their Paper,* where they had calculated 
that in one of the trials of a Willans engine, out of a total missing 
quantity of 90 lbs., the leakage might be between 20 and 30 lbs., 
a quantity that could hardly be called excessive. It had been one 
of the chief objects of the Paper to show that the differences in 
economy of various types of engines were generally due to the type 
and arrangement of valves used, and not to mysterious variations of 
natural laws. 

The account given by Professor Hubert (page 567) of the 
engines working with jacketed pistons using steam at pressures 
equal to and greater than the admission pressure of the working 
steam was most interesting. He himself was sure that the members 
would be grateful for having thus had brought before them methods 
which allowed a single-cylinder engine without superheat to have 
so low a consumption as 15 lbs., and a small compound engine a 
consumption of 11 lbs. of water per h( rse-power per hour. He 
agreed with Professor Hubert that within reasonable limits the 
volume of the clearance surfaces had not much effect upon the 
consumption, and thought that if this point were more appreciated 
consulting engineers would not so often put manufacturers to great 
expense by asking for absurdly small clearance volumes. He was 
interested in the diagrams presented by Professor Hubert, especially 
as they confirmed the observations of Messrs. Callendar and Nicolson, 
that in an engine without jackets the steam in the cylinder was 
superheated during compression. These experimenters had 
observed the steam temperature-cycle with two thermometers, one 
situated in a pocket in the cylinder cover and the other attached 
to the piston. The first thermometer showed that the small 

* Proceedings, TIjg Inst, of Civil EDgineers, 1897-98, vol. cxxxi, page 147- 



June 1905. STKAM-JACKETING. 585 

quantity of steam in the pocket and close to the walls was 
superheated during practically the whole stroke. The second 
thermometer, which gave the temperature of the main body of the 
steam, showed a slight superheating during compression. He was 
under the impression that M. A. Duchesne's thermometer was 
situated in the clearance space, and would therefore give indications 
between those of Messrs. Callendar and Nicolson's two thermometers. 
These two sets of experiments effectually disposed of the theory that 
a film of water remained on the cylinder walls from stroke to stroke, 
and probably every one would agree with Professor Hubert's first 
conclusion. 

Professor Hubert's second conclusion was similar to that advanced 
by M. Georges Duchesne (page 574), namely, that the leakage 
theory must be abandoned. Their grounds for this were that 
in the Liege experimental engine there was a large missing 
quantity when the engine was working without jackets and no 
missing quantity when high-pressure steam was in the jackets. 
M. Duchesne stated that, since there was no leakage in the latter, 
there could have been none in the former case. This however 
entirely ignored the contention of Messrs. Callendar and Nicolson, 
that the leakage depended greatly upon the temperature of the 
valve-face, and that by slightly heating the valve-face it had 
been possible to reduce the leakage considerably. As had been 
mentioned in the Paper (page 552), it appeared that one of 
the valuable features of the jacket was that it increased the 
temperature of the valve-faces and thus reduced the leakage. He 
thought that this went far to explain the high economy of the 
Frangois engine. A cross-section of the cylinder showed that the 
valves were entirely surrounded by the jacket steam, and they ought 
therefore to have little or no leakage. He further believed that 
Professor Hubert's diagrams showing the temperature cycles of the 
walls were a further proof that leakage did take place in the trials 
recorded in Fig. 22 (page 572). In this example the temperature 
range of the walls seemed to be about 55° C. (131° F.). In Fig. 23, 
where ordinary jackets were used, the temperature range was less than 
5° C. It was shown in the Appendix (page 555) that if the range 



586 STEAM-JACKETING. June 1905. 

(Mr. Mellanby.) 

of the wall temperature were known, the amount of heat flowing into 
it could easily be calculated. It would be seen therefore that in 
Fig. 23 the heat flowing into the walls from the steam, which of course 
measured the initial condensation, must be only a small fraction of 
what it was in Fig. 22. According to M. G. Duchesne's figures the 
missing quantity in Fig. 23 was about half of what it was in Fig. 22, 
which did not at all agree with his contention that it must all be 
due to initial condensation. It seemed evident therefore that a 
large proportion must be due to leakage. It would be interesting 
to calculate how much condensation was due to the given temperature- 
ranges, and to compare the calculated amounts with the measured 
missing quantities. 

The author's reasons for believing that much of the missing 
quantity was due to leakage and net to condensation were not, as 
M. G. Duchesne stated, based upon the belief that the heat absorption 
was equal to the heat emission, but were founded on his opinion 
that the temperature range of the walls was only very small. His 
idea was that the steam condensed on the walls at a certain rate 
which depended upon the temperature difference between the steam 
and the walls, and that this condensed steam was probably re- 
evaporated at the same rate. When, however, the walls were dry, 
the heat flowing from the hot walls to the steam would be very 
small. This was in no way contrary to the theory of Messrs. 
Callendar and Nicolson ; indeed it was the only way to explain 
why the average temperature of the metal was higher than that 
of the steam. 

The experiments quoted by M. Mathot (page 574) were of 
great interest and would make a valuable addition to the Paper. 
Beyond, however, referring to the high economy exhibited by these 
small engines, he did not feel there was anything further for him to 
say upon them. 

In conclusion he hoped it would be generally recognised that 
there was yet much to be learned about the behaviour of steam in 
an engine cylinder. Many of the points could only be cleared 
up in a well-equipped laboratory, and he would suggest that the 
Technical Colleges in England might follow the examples of those 



June 1905. STEAM- JACKETING. 687 

in Montreal and Liege and devote some attention to the prosecution 
of scientific steam-engine research work. At present experimental 
engines were chiefly devoted to teaching students the methods of 
carrying out ordinary trials, and little research was done to explain 
the heat exchanges which took place in an engine. Although the 
ordinary routine teaching was looked upon by many as the most 
important work of a college, he thought that the colleges would 
have to show themselves capable of solving problems in their 
laboratories which could not be attacked in workshops, before 
British manufacturers would give to technical education the support 
which was being continually demanded from them. He would also 
suggest that in every engine trial thermometers should be inserted 
into holes drilled in the cylinder walls. The experimenters would 
thus be enabled to see how the temperatures of the clearance surfaces 
varied as different changes were made in the steam distribution, and 
in many cases the reason for increased or decreased efficiency would 
at once be made clear. 

On the motion of the Peesident a hearty vote of thanks was 
accorded to the author for his Paper. 



Communications. 

M. Geoeges Duchesne wrote that, although he had an opportunity 
of discussing the Paper at Liege (page 574), he would like to add the 
following observations : — 

. Those engineers who first carried out methodical steam-engine 
trials soon found out the enormous difference there was between the 
true steam-consumption of the engine, and that computed on the 
supposition that at the end of admission the cylinder contained dry 



588 STEAM-JACKETING. June 1905. 

(M. Georges Duchesne.) 

saturated steam. The reason of this considerable difference was now 
well known, and it was attributed with certainty to the conductibility 
of the metallic walls which formed on all sides the boundary of the 
volume of the steam. It was known that the surfaces of the walls 
and head of the cylinder, as well as the surface of the piston, were 
exposed during the exhaust to the atmosphere or to the condenser, 
and were therefore cooled down to a temperature lower than that of 
the steam which was introduced during admission; these surfaces 
therefore absorbed heat when their temperature was again raised to 
that of the steam, and this heat was abstracted from the steam during 
compression and admission. 

The consequence was the deposition of small drops of water on 
the metallic surfaces, as had been demonstrated experimentally in 
the most unmistakable manner by the well-known engineer, the late 
Mr. Bryan Donkin, by means of his revealer.* This condensation, 
which would be divided into pre-initial condensation (produced 
during the compression) and initial condensation (produced 
during admission), was one of the most important causes of the 
diminution of the thermal efficiency of steam-engines. The whole of 
the heat stored in the form of water distributed over the cylinder 
walls, etc., was not lost however, because a portion of this water was 
vaporised during expansion, and contributed to the production of 
mechanical work ; but it would be shown later that the thermal 
efficiency of this heat was much less than that of the heat contained 
in the steam itself. 

Inventors had striven to find means of reducing this condensation, 
and numerous solutions had been proposed. All suggestions 
involving the use of bad conductors for the walls of the cylinders or 
in which the internal surfaces were covered with insulating material, 
had failed owing to the imperative requirements of construction 
which necessitated the use of metals that were conductors of heat. 
The most effective of the proposed methods had been and was still 
the jacket, which, as was well known, heated the exterior of 
the walls either by means of hot gases or by steam. It would seem 

* Proceedings 1900, Part 4, page 509. 



June 1905. STEAM-JACKETING. 589 

that originally the jacket had been specially intended to increase the 
temperature of the walls, and ensure thereby the complete and 
rapid re-evaporation of the condensed steam during admission, and 
eventually during compression. A searching examination of the 
conditions under which the jacket worked had proved that its 
essential functions were not those which its promoters assigned 
to it. 

In order to arrive at a more intimate knowledge of the 
method of action of the jacket, the writer had studied a group of 
six trials made with a steam-engine belonging to the Mechanical 
Laboratory of the University of Liege ; these trials were made by 
Professor Dwelshauvers-Dery, whose name was well known to all 
those who were occupied with the study of the steam-engine. The 
six trials under review had been classed in groups of two trials 
each (with and without jacket), and the three groups differed in 
that for the groups A and B the point of cut-off had been varied 
without altering the other conditions of trial, and for the groups 
A and C, A consisted of two condensing trials and C of two non- 
condensing trials. The detailed study of the six trials appeared in 
the " Eevue Universelle des Mines," * and did not need to be referred 
to further here ; it would be sufficient to indicate the method of analysis 
and the conclusions derived therefrom. Let an example be taken, 
namely. Trial Al, which was a condensing trial without jacket, and in 
Fig. 28 (page 590) was reproduced the mean indicator diagram. 
The design of the engine req[uired a compression period, and as 
the point at which this phase began was well known, the weight 
of steam then present in the cylinder could be calculated. This 
calculation was made on the hypothesis that the steam was then 
saturated and dry, so that a knowledge of the pressure and volume 
sufficed to determine that the weight 

M, = 0-000981 kg. (0-02 2 lb.) 
was then present in the cylinder. From a steam table the internal 

* " Function of tlie Jacket in a Single-cylinder Steam-Engine," by 
G. Duchesne. " Kevue Universelle des Mines," 3rd series, 1901, vol. Iv, 
page 212. 



590 STEAM- JACKETING. JuNE 1905. 

(M. Georges Duchesne.) 

heat of these M^ kilos of steam was found to be 

U2 = 0"5761 calorie (2-287 B.Th.U.)* 
The indicator diagram gave the condition of this M^ weight of 



Trials made at the University of Liege. (Table 9, page 597.) 

Fig. 28. 

Trial Al, Without Jackets. 14 Dec. 1893. 

(Same diagram as Fig, 25 page 575.) 




Fig. 29. 

Trial A2, With Jackets. 22 Dec. 1893. 

(Same diagram as Fig. 26, page 575.) 




steam at the end of compression, and it was found that the steam 
was superheated, and that its internal heat, calculated by Zeuner's 
formula, had become 

ETg = 0-6629 calorie (2-63 B.Th.U.). 



* The calorie has been taken as equal to 3*97 B.Th.U. 



June 1905. STEAM-JACKETING. 591 

Since the internal heat had only been increased by the heat 
equivalent of the work of compression, namely, 

*AT= 0-0806 calorie (0-32 B.Th.U.), 
it followed therefore that the cylinder walls had supplied the steam 
with • 6629 - (0 • 5761 + • 0806) = • 0062 calorie (o • 0246 B.Th.U.). 

During the admission period M^, or 0*034072 kg. (0*0750 lb.) of 
steam entered the cylinder at a pressure P4 or 55,928 kg. per m.^ 
(79*6 lbs. per square inch), and at a temperature of 197*518° C. 
(387*5° F.), measured directly by a thermometer; this weight being 
obtained by dividing the steam-consumption during a given time by 
the corresponding number of strokes of the piston. A comparison of 
these figures showed that the steam was superheated 42*992° C. 
(109*4° -^•)' *^^ *^^* *^^ *^**-^ ^®^* must be calculated by means of 
Eegnault's formula X! = A + * 485 (f — t), where X is the total heat 
of saturated steam at the given pressure, and t' — t is the amount of 
superheat. On calculation, it was found that the M^ weight of 
steam had supplied Q = 22*9767 calories (91*2 B.Th.U.). These 
calories were subject to various phenomena : — For instance, the 
pressure increased from P3 to P4, but this compression did not appear 
to have any influence on the thermal state of the resulting fluid 
Mc + M^. It was in fact produced instantaneously and without 
doing external work. A first portion of Q supplied the work 
done during admission, of which the calorific equivalent was 
AT^ = 0*5190 calorie (2-06 B.Th.U.). 

A second portion appeared in the heat of the steam which filled 
the cylinder at the point of cut-off. The volume and the pressure 
being known, this quantity of heat could be calculated, and it was 
found that it was : 

q, = 12*2529 calories (48 * 6 B.Th.U.). 
At the beginning of admission there was 0*6625 calorie (2*63 
B.Th.U.) in the cylinder, then 22*9767 calories (91*2 B.Th.U.) 
were introduced, and at the end of admission 12*2529 calories 



A= ' 



Joule's equivalent. 



592 STEAM- JACKETING. June 1905. 

(M. Georges Duchesne.) 

(48-6 B.Th.U.) were still found in the steam after doing work 
equivalent to 0-519 calorie. The difference, 

22-9767 + 0-6629 - (12-2529 + 0-5190) = 10-8677 calories 

(43-1 B.Th.U.), 
therefore represents the quantity of heat absorbed by the walls. 

The M^i 4- ^c weight of steam was at that moment present in 
the cylinder in the form of dry saturated steam, filling the volume at 
cut-off, together with a dew covering the surfaces. The above 
calculation proved that more than 47 per cent, of the heat supplied 
by the steam admitted per stroke was absorbed by the walls. By 
*' heat absorbed by the walls " was meant the latent heat given up 
to the walls increased by the internal heat of the water which 
covered them. 

Now, with regard to the expansion period, on Fig. 28 (page 590) 
was drawn the adiabatic expansion-line corresponding to the weight 
of steam filling the cylinder at the point of cut-off. This curve was 

traced by points, and its equation is </> -|- ^ = constant, where 

(f) is the entropy of the kilogram of water ; 

r is the total latent heat of the saturated kilogram of steam ; 

T is the absolute temperature ; 

X is the dryness-fraction of the mixture. 
It would be seen that the actual expansion line was above the 
adiabatic, and it was known that this difference between the two 
curves was due to the re-evaporation of the water deposited on the 
walls. It was in fact evident that during expansion the pressure 
and the temperature of saturation were constantly diminishing, and 
therefore continually greater quantities of the water were deposited 
on the walls, which later were re-evaporated owing to the temperature 
of the water and to the heat given back by the walls. The areas 
AaBh and ABC have been measured ; they corresponded respectively 
to the work done by the steam which filled the cylinder at the 
beginning of expansion, and by the water which had been vaporised 
during expansion ; the calorific equivalents of the work were 1-3319 
and 0-1639 calorie (5-3 and 0-65 B.Th.U.). 



Junk 1905. STEAM -JACKETING. 593 

The results obtained from the point of view of the production of 
mechanical energy were as follows : of the 22*9767 calories 
(91*2 B.Th.U.) introduced into the cylinder by the ilf„ kilos of 
steam, 10*8677 calories (43*1 B.Th.U.) have been absorbed 
by the walls during admission, and of these 0*1639, that is 

10^8877 — 1 ' ^^ P®^ cent., have been converted into work. 

The remainder, namely 22-9767 - 10*8677 = 12*1090 calories 
(48*1 B.Th.U.) augmented by CTg, namely 12*1090 + 0*6629 = 
12*7719 calories (50*7 B.Th.U.), has produced, during the two 
working periods of admission and expansion, an amount of work 
equivalent to 
^r, + ^T^ = 0*5190 + 1*3319 = 1-8509 calories (7-3 B.Th.U.). 

Thus lo^^yVq = 14-49 per cent, of this heat had been converted 

into work. Therefore the utilization of the heat contained in the 

14*49 
steam was -rr^ =9*6 times greater than that returned from the 

walls during the expansion. 

This phenomenon was easily understood by means of a comparison 
due to Professor Dwelshauvers-Dery and published in his " Calometric 
Study of the Steam-Engine."* The heat absorbed by the walls was 
compared to a leak of water from the wall of a side water-wheel, 
entering the wheel at a height h smaller than the total height of fall 
H, and therefore doing less work than if it had fallen directly on to 
the wheel. Calculations showed that at the end of expansion there 
was Ifi = * 02829 kg. (0*062 lb.) of steam in the cylinder (wet 
steam because expansion reduced the dryness-fraction). Hence 
M^ + M,-\-M^ = 0*006763 kg. (o* 149 lb.) remained on the walls in 
the state of water, and they contained 0*006763 x 100-109 = 
0*6772 calorie (2-7 B.Th.U.), for at the pressure of 10,204 kgs. 
per m.^ (14*5 Ihs. per sq. inch), the heat of one kilogram of water 
was 100*109 calories (397-4 B.Th.U.). The total heat of the 
fluid was 16*5518 calories (65*7 B,Th.U.), and therefore that of 
the gaseous fluid (wet steam) was 

* " Calometric Study of the Steam-Engine," by Prof. Dwelshauvers-Dery 
(Encyclopedie des aide-memoire de M. Leaute), 2nd edition, page 70. 

2 T 



594 STEAM- JACKETING. June 1905. 

(M. Georges Duchesne.) 

16-5518 - 0-6772 = 15-8746 calories (63-0 B.Th.U.). 

At the beginning of the expansion the total heat was only 
12-2529 calories (48-6 B.Th.U.), but the work of expansion, 
equivalent to 1-4958 calories (5-94 B.Th.U.), had been done; 
therefore the heat returned into fluid was 15*8746 + 1*4958 — 
12-2529 = 5-1175 calories (20-3 B.Th.U.). This quantity of 
heat was greater than that returned by the walls, but the difference 
was obviously due to the water-heat of the kilogram of water having 
been diminished by the reduction of pressure, so that the drops of 
dew had found sufficient heat in themselves for partial evaporation. 

It had been seen that 10-8677 calories (43*1 B.Th.U.) were 
absorbed by the walls during admission, and that 5 • 1175 calories (20* 3 
B.Th.U.) were returned to the steam ; it was also known that the 
external radiation amounted to 0-7647 calorie (3*04 B.Th.U.) 
per revolution; therefore 10-8677 - (5-1175 -f- 0-7647) = 4-9855 
calories (19-8 B.Th.U.) were returned to the steam during 
exhaust. Since, at a pressure of 10.204 kgs. per m.^ (14*5 lbs. per 
square inch), the total latent heat of vaporization was 519*806 
calories per kilogram (93 5* 6 B.Th.U. per lb.), the heat required to 
vaporise the 0-006765 kg. (0*0149 lb.) of water remaining on the 
wall was: 0-006765 x 519-806 = 3-5164 calories ( 1 3 • 96 B.Th.U.), 
a figure which was considerably less than the 4-9995 calories 
(19-8 B.Th.U.) which were at disposal for this re-evaporation. 
This result clearly showed that one was authorised in assuming, 
in accordance with Hirn's hypothesis, that the walls were quite dry 
at the end of exhaust. Analysis had also been made in the same 
way of Trial A2, in which the jacket was in action, and of which 
Fig. 29 (page 590) was a reproduction of the mean indicator diagram. 
It was thus found that the heat-supply during admission was 17-6678 
calories (70-1 B.Th.U.) and that 13-1750 calories (52 -3 B.Th.U.) 
were in the steam at the beginning of expansion. Thus only 
17-6678 - 13-1750 = 4-4848 calories (17*8 B.Th.U.), that was to 
say, 25 per cent, of the heat supply was absorbed by the walls. 

The immediate effect of the jacket was therefore to diminish the 
initial condensation considerably. A comparison of the actual 
expansion line and the adiabatic showed that the difference was less 



Junk 1905. STEAM -JACKETING. 596 

than in the trial without jacket. The calculation of the heat 
exchanges showed that only 3*2445 cals. (12*9 B.Th.U.) had 
passed from the walls to the gaseous mass, instead of 5*0587 
(20-08 B.Th.U.) as in the case of Trial Al. This result led to the 
following apparently paradoxical conclusion : The jacket diminished 
the quantity of heat returned from the walls to the steam during the 
expansion period. It was however undesirable that the jacket should 
supply heat to the steam during expansion, for the thermal efficiency 
of such heat would be small compared with that of the heat absorbed 
by the walls during admission, and which was returned to the 
gaseous fluid during expansion. The study of this trial had enabled 
him to make inferences respecting the rate of heat exchanges between 
the walls and the steam. The external radiation per stroke was 
1*1316 cal. (4*5 B.Th.U.), and this had been entirely provided 
by the 1*8441 cal. (7*3 B.Th.U.) supplied by the jacket; the 
difference, namely 1*8441 - 1*1316 = 0*7125 cal. (2*83 B.Th.U.), 
had even passed from the jacket to the cylinder during exhaust, and 
therefore without utilization. The quantity of heat given up by the 
walls during exhaust was therefore 1*2403 + 0*7125 = 1*9528 cal. 
(7*75 B.Th.U.). The figure 1*2403 cal. represented the difference 
between the heat absorbed by the walls during admission and that 
returned to the steam during expansion. 

If the figure of 1*9528 cal. (7*75 B.Th.U.) were compared with 
that of 4*4848 cal. (17*8 B.Th.U.), which represented the heat 
exchange during admission, and if it were noted that in this second 
case the duration of the exchange was less, the acting surfaces 
smaller and the temperature differences less than during exhaust, one 
was led to conclude that the activity of the heat exchanges was 
considerably greater during admission than during exhaust. This 
difference was entirely due to the fact that the coefficient of heat 
transmission from a saturated vapour to a colder metal (which 
consequently was covered with dew) was much greater than the 
same coefficient for the exchange between a hotter metal and the dry 
vapour in contact with it. M. J. Nadal had calculated this coefficient 
directly by means of some experiments carried out with the help 
of the new revealer of Mr. Donkin (" Eevue de Mecanique," 

2 T 2 



696 

(M. Georges Duchesne.) 



STBAM-JAOKBTrNTG. 



June 1905. 



TABLE 9 {continued on opposite page). 

3 Jacketed and 3 Unjacbeted Steam-Engine Trials made at 



Group 
Number 



Arrangement 

of the 

Engine. 



Pre-admission 
Cut-off . 

Pre-exhaust 



Left 



Right 

Amount of compression . 

^ Condensation 

Jackets ...... 

Weight of Steam present at the end of exhaust 

Weight of Steam admitted per stroke 

Heat supplied by this Steam . 

Heat supplied by the Jacket . . 

Total heat supplied per stroke 

Weight of Steam at the end of expansion . 

Heat absorbed by the Walls during admission 

Heat absorbed by the Walls during admission and from the 
beginning of compression 

Dryness-fraction at the end of expansion 

Weight of Steam present at this moment . 

Weight of Water evaporated during expansion 

Total weight of the working fluid . 

Weight of Water on the Walls at the beginning of expansion 

Initial condensation of the weight of Steam admitted 

Weight of Water on the Walls at the end of expansion 

Consumption of Steam per horse-power per hour 

Utilization as work of the Heat absorbed by the Walls during 
admission ....... 

Utilization as work of the Heat contained in the gaseous working 
fluid . . . . 

Utilization as work of the heat absorbed by the walls during 
expansion ...... 

Value of the Jacket relative to the absolute work 

Value of the Jacket relative to the effective work 



lb. 

lb. 
Th.U. 
Th.U. 
Th.U. 

lb. 
Th.U. 

Th.U. 



lb. 
• lb. 
lb. 
lb. 
per cent, 
lb. 
lbs. 

per cent. 
per cent. 

per cent. 

per cent, 
per cent. 



Junk 1905. 



■TEAM- JACKETING . 



697 



(concluded from opposite page) TABLE 9. 
the University of Liege, by Professor Dwelshauvers-Dery. 



0-0 
0-1 

0-054 
0-070 
0-193 
with 



without 
0-0022 
0-0750 
91-21 



0-0444 
43-18 

43-12 

0-7477 
0-0577 
0-0133 
0-0771 
0-0327 

43-8 
0-0149 

24-551 

1-508 



with 
0-0022 
0-0577 

70-14 
7-32 

77-47 
0-0477 

17-81 

18-22 

0-9023 
0-0540 
0-0064 
0-0599 
0-0122 

21-2 
0-0006 

21-421 

2-15 



without 
0-0026 
0-1179 
144-29 



0-0859 
47-17 

47-20 

0-8287 
0-0996 
0-0139 
0-1204 
0-0345 

29-3 
0-0141 

25-524 

1-13 



B 

2 
0-0 
0-3 
0-054 
0-070 
0-193 
with 

with 
0-0024 
0-0983 
120-05 
: 6-03 

! 126-08 

i 

0-0889 
18-93 

18-73 

' 0-9317 
' 0-0938 
[ 0-0049 
0-1007 
0-0118 
12-0 
0-0008 
21-388 

1-77 





14-49 


14-48 


12-617 


12-568 




3-20 


3-04 


2-43 


2-74 




— 


21-83 


— 


18-3 


• 


~ 


20-5 


— 


13-3 





1 I 

0-0 
0-1 
0-000 
0-000 
0-000 
without 

without ! 

0-0044 



0-0588 
71-66 



0-0432 
26-98 

26-98 

0-9444 
0-0597 
0-0165 
0-0632 
0-0201 

34-1 
0-000 

36-016 

3-32 

14-84 

3-86 



with 

0-0044 

0-0452 

55-05 
6-67 

61-38 
0-0450 
8-36 

8-36 

superheated 
0-0496 
0-0046 
0-0496 
0-0046 

10*2 
0-000 

32-875 

5-66 
15-02 

5-66 

15-7 
14-1 



598 STEAM-JACKETING. June 1905. 

(M. Georges Duchesne.) 

December 1900), and had found respectively 36 and 0-1 cal. per 
square metre, second, and degree Cent. 

The quantity of heat absorbed by the walls, the thermal efficiency 
of which was small, amounted to 10*8615 cals. (43*1 B.Th.U.) in 
Trial Al and was reduced to 4-5901 cals. (18 -2 B.TLU.) in 
Trial A2. A comparison of these two figures clearly showed that 
the object of the jacket was to maintain the temperature of the 
cylinder at a higher point. Two antagonistic effects were the result : 
a diminution of the initial condensation, and an increase of the 
quantity of heat uselessly supplied during exhaust. These two 
phenomena were fortunately governed by different laws, and the ratio 
of their intensity was of the order 360 to 1 ; the prejudicial effect 
therefore disappeared in comparison with the useful effect, and one 
might conclude: The principal duty of the Jacket consisted in the 
diminution of initial condensation resulting from the increase in the 
temperature of the cylinder walls. 

In order to exhibit the advantages gained by the use of the 
jacket under the various conditions of its use, he had collected 
together the three groups of trials, and gave on Table 9 (pages 596- 
597) the particulars and the results. 

In Figs. 30, 31, 32, and 33 (page 599) were reproduced the 
mean indicator diagrams of Trials Bl, B2, CI, and C2. It would be 
seen that, as in the case of Trials Al and A2, the jacket had the 
effect of bringing the actual expansion line closer to the adiabatic. 
A comparison of the results of Table 9 showed that a late cut-off, 
as well as working non-condensing, reduced the gain due to the 
jacket. The reason was that both these conditions of working were 
accompanied by an increase of temperature of the walls and by a 
diminution of initial condensation. It would in fact be observed 
that in the Trials of group B, with 0*3 cut-off, the initial 
condensation was sensibly the same as for group A, with • 1 cut-off, 
notwithstanding that the time-interval of the heat exchanges was 
greater; and the initial condensation was also the same in the 
non-condensing trials with the same cut-off, namely, 0*1. 

The following principle was deduced : — 



June 1905. 



STEAM- JACKETING. 



599 



Trials made at the University of Liege. (Table 9.) 









Fig. 30. 








TW«; Rr T?/-.V7.^,./ T^^l.^*. 






\ 6 Jan, 1894. 




I 




V 


K 



t 


^^ 


Xv^^^ 


f^ 

^ 


—- 



Fig. 31. 
Trial B% With JacUts. 
12 Jan. 1894. 




Fig. 32. 

Trial CI, Without Jackets. 

30 April 1897. 




Fig. 33. 

Trial C2, With Jackets. 

1 May 1897. 




600 STEAM-JACKETING. JuNE 1905. 

(M. Georges Duchesne.) 

All causes (^superheat, late cut-off, working non-condensing, high 
speed) which tended to raise the temperature of the walls, diminished the 
useful effect of the jaclcet. 

With regard to the efficiency of jackets, as depending on the point 
of cut-off and on the number of revolutions of the engine, the writer 
would recall the results of trials made by Herr Schroter in his 
mechanical laborEitory in Munich, with a Siilzer engine having 
a cylinder of 0-280 m. (11-024 ins.) diameter and 0*650 m. 
(25 '59 ins.) stroke. 



TABLE 


10. 












1 
Speed : 53 revolutions per minute. 


Cut-off 


0-1 


0-2 


0-3 


0-4 


Improvement due to Jacket . 


per cent. 
15-7 


per cent. 
12-25 


per cent. 
8-96 


per cent. 
4-57 


Speed : .39 revolutions per minute. 


Cut-off .... 


0-1 


0-2 


0-3 


0-4 


0-5 


1 
1 

Improvement due to Jacket . 


per cent. 
18-85 


per cent. 
16-80 


per cent, 
14-00 


per cent. 

8-75 


per cent. 
6-05 



Bemarhs. — (1) The efficiency of the jacket, when the engine was 
working under like conditions, diminished with a later cut-off, and 
with full admission its action would be reduced to supplying the 
external radiation. In this extreme case the jacket would even be 
harmful, because by increasing the external surface of the cylinder 
it assisted radiation. There was a case however which had not 
sufficiently engaged the attention of steam-engine builders, a case in 
which jackets were generally necessary ; namely that of engines 
intermittently loaded, such as reversing engines for rolling mills, 
winding engines,* and blowing engines for steel works. In these 

* "Theoretical and Experimental Study of the Winding Engine," by 
R. A. Henry, lately engineer of the " Corps des Mines," Chief Engineer at the 
Hasard Colliery. " Revue Universelle des Mines," 4th series, 1903, vol. 2. 



Junk 1905. STEAM-JACKETING. 601 

engines the temperature of the cylinder walls was considerably less 
than in the case of engines working under constant load, and initial 
condensation was present in enormous proportions. Even with full 
admission the water present in the cylinder was sometimes ten, or 
even more, times the weight of steam. In such engines the jacket, 
although theoretically useless during the periods of rest, would have 
a mean efficiency depending upon the proportion of the periods 
of rest and work. The statement " theoretically useless " had been 
made because, after a stop, the Jacket prevented the initial 
condensation being sufficient in amount to produce water-hammer. 

(2) With respect to the expansion curve, it was to be remarked 
that at first it agreed fairly well with the adiabatic corresponding to 
the steam present in the cylinder at the beginning of expansion ; 
then it rose more and more above this line. Initial condensation was 
the only cause of this rise, and, other things being equal, the work 
done during expansion was greater the more considerable was the 
condensation, but, as had been shown, this increase of power was 
obtained by an enormous consumption of heat. For the same point 
of cut-off and for the same thermal value of the working fluid (same 
pressure and same temperature of the steam and vacuum) the more 
would the expansion line rise above the adiabatic and the smaller 
would be the thermal efficiency. 

(3) Engine-builders had used the jacket in different ways. 
Steam was supplied either by a special connection taken straight 
from the boiler or by admitting the steam to the jacket before 
it entered the cylinder. Of these two methods the first seemed 
theoretically the better, because the second had the disadvantage 
of introducing into the cylinder steam which was slightly wet 
and therefore liable to induce the action of the walls. Many 
builders preferred the class of jacket in which the steam circulated 
to that in which there was no circulation (^enveloppe stagnante), 
and attributed a greater value to it. The difference, if real, 
might be due to the large quantity of air that could be retained 
by the non-circulating jacket which hindered the heat exchanges ; 
but if special precautions were taken to get rid of this air, the 
non-circulating jacket became equal, if not superior, to the circulating 
jacket. 



602 ' STEAM-JACKETING. JuNE 1905. 

(M. Georges Duchesne.) 

Superheating Steam-jacket. — The principal deduction to be made 
from the preceding examination of the conditions of working of the 
steam-jacket was that the rate of transmission of the heat from 
saturated steam to a colder metal, through the intermediary of a 
thin film of water, was about 360 times greater than the rate of 
transmission of heat from a dry metal surface which was hotter than 
the steam ; and from this the conclusion would be drawn that the 
heat exchanges would become a minimum in an engine in which the 
temperature of the cylinder walls was at each instant superior to the 
saturation temperature of the steam. 

* " He had thus been led to the conception of a single-cylinder 
"steam-engine in which the jackets applied to the covers, to the 
" cylindrical portions, as far as possible to the valves and to the 
" piston, f were heated by saturated steam of a higher temperature J 
" than that of the steam entering the cylinder. Such a jacket 
" might be called a ' superheating ' jacket, because it was capable of 
" superheating the working steam ; this name was all the more exact 
" since, not only could the superheating be produced, but it actually 
" occurred shortly after the steam had begun to expand in the 
" cylinder. The walls of the cylinder, owing to the intermediary of 
" the thin film of water, absorbed very easily the heat of the hotter 
" steam of the jackets, and would be sensibly maintained at the 
" temperature of this steam, which would at all times be higher than 
" that of the cylinder steam, to which, moreover, it would only yield 



* The portions in inverted commas are contained in " Revue Universelle des 
Mines," 4th series, 1904, vol. vii, page 221. 

t See Research Committee on Value of the Steam-Jacket, Second Report, 
Proceedings 1892, page 495. [Sec, I. Mech. E.] 

X Records of previous experiments, with jackets supplied with steam of a 
higher temperature than that of the boiler, will be found in the Second Report 
of the Research Committee on the Value of the Steam-Jacket, Proceedings 
1892, page 421 (Experiments in Milan, 1886, showing an advantage of 16*8 
per cent, by use of steam at 187 lbs. in the jacket instead of 48 lbs. which was 
the boiler pressure), and page 497, where an attempt by Mr. David Thomson, 
in 1860 or 1861, to use a higher temperature in the jacket than in the cylinder 
is mentioned. [Sec, I. Mech. E.] 



June 1905. STEAM- JACKETING. 603 

"but a small amount of heat. This engine would not exhibit the 
" phenomenon of initial condensation ; the action of the walls would 
'•be reduced to a minimum, and the steam-consumption would 
" approximate to the theoretical consumption. 

" It was impossible, for the following reasons, to use any other 
" fluid than saturated steam for the heating of the jackets : When a 
" saturated vapour gave up heat its temperature did not diminish, but 
"a precipitation of dew was produced on the cooling surfaces, the 
"presence of which increased the absorbing power. A saturated 
" vapour, which maintained a rigorously constant temperature, was 
" therefore the only fluid capable of heating a volume consisting of 
" spaces so irregular and inaccessible as those which constituted a 
"judiciously studied jacket of a steam-engine. In addition to the 
"fact that no other fluid had the above property, saturated steam 
" possessed the great advantage of transmitting its heat easily by the 
" intermediary of the thin liquid film, so that a difference of a few 
" degrees was sufficient to maintain the metal at the temperature of 
" the hotter steam which bathed one of its surfaces, even when the 
" colder steam in contact with its other surface was at a temperature 
" as much as 100° C. lower. 

"The writer had had the honour of submitting these ideas to 
" Professor Dwelshauvers-Dery, who decided to make a trial of this 
" system. Professor Dwelshauvers-Dery obtained, therefore, for his 
"laboratory a small boiler of 2*50 m.^ (26*9 sq. ft.) of heating 
" surface, and able to supply steam to the jackets at a pressure of 
" 294 lbs. per square inch. At the same time important modifications 
" were made to the engine of the mechanical laboratory, and specially 
" the piston was provided with a heating arrangement. On 9th June 
" 1903 a trial was made under these conditions. The particulars 
" of this trial are as follows :— 





Pre-admission 


. 0-0 


Arrangement 


Cut-off .... 


. 0-1 


■ of the , 


Point of release (advance) 


. 0-05 


Engine. 


Compression . 
Working condensing 


. 0-0 


Volume swept by the piston 


. 0-04304 m.» (1-52 cub. ft.). 


Clearance voIud 


ae 


. 0-00180 m.'(o'o636 cub. ft.). 



STEAM-JACKETING. 



604 

(M. Georges Duchesne.) 

Trials made at the University of Liege, 9 June 1903. 

Fig. 34. 
With Jackets and Piston heated. 

(Same diagram as Fig. 27, page 575.) 



Junk 1905. 





Fig. 35. 

Without Jacket*. 

Other conditions identical with Fig. 84. 



n 



o 



The jackets and the piston were heated by steam at a pressure of 
9 kgs. per cm.^ (128 lbs. per sq. in.). 

" On Fig. 34 was reproduced a mean indicator diagram ; OK 
"represented 10,000 kgs. per m.^ (14* 23 lbs. per sq. inch). The 
"weight of ^steam • present ' in the cylinder at the end of exhaust 
"would be neglected. During admission M„' = 0*01585 kg. 
"(0*0349 lb.) of steam (computed from the observed steam- 
" consumption) entered from the steam-chest into the cylinder at a 
"pressure of 63,583 kgs. per m.^ (76*2 lbs. per sq. inch). The 



Junk 1906. STEAM- JACKETING. 605 

" saturation temperature of steam corresponding to this pressure was 
"153-591° C. (308-5° F.) and the total beat of 1 kg. (2-2 lb.) of 
" saturated steam at the same pressure was 653 - 345 calories 

"(2 594B.Th.U.). 

" Since the steam was not superheated, the heat-supply of the 
"M„ weight of steam is 0-01585 x 658-345 = 10-37 calories 
"(41-17 B.Th.U.). 

" The pressure P„, measured on the indicator diagram at the 
" point of cut-off, was 48,408 kgs. per m.^ (68 - 9 lbs. per sq. inch), 
" and the volume occupied at that instant was equal to the clearance 
" volume plus ^^ of the volume swept by the piston during the whole 
'* stroke — altogether - 006144 m.^ (0-217 cub. ft.). If the weight of 
" dry steam which occupied the volume 0-006144 m.^ at a pressure of 
" 48,408 kgs. per m.^ was computed by means of Eegnault's tables it 
" would be found to be 0*01587 kg. (0-035 lb.), which was identical 
" with the weight of steam already found for M^,. 

" From this the following conclusion was derived : the heating of 
" the jacket by means of saturated steam of a higher temperature 
" than the working steam in the cylinder suppressed the harmful 
" phenomenon of initial condensation, and at the end of admission the 
" steam present in the cylinder was identical with that in the steam- 
'• chest of the engine. 

"Let the adiabatic corresponding to this weight of steam be 
" traced. One knew how to do this by computing the position of 
" a number of points, as shown in Fig. 34 by means of small circles. 

" It was seen that the actual expansion line was identical with the 
" adiabatic ; this confirmed the statement that the transmission of 
" heat from the walls to the steam was very small, since the indicator 
" was unable to detect it. The writer had some original indicator 
" diagrams, which were at the disposal of those who desired to verify 
" this point further. With the given clearance volume and particulars 
" of the distribution, the only loss of work was represented by the 
'* area hatched horizontally ; this loss was insignificant, and it 
" could be concluded that the superheating steam-jacket permitted of 
" a practically perfect utilization of the cylinder steam." 



606 STEAM-JACKETING. JuNE 1905. 

(M. Georges Duchesne.) 

The writer said " practically perfect " because it was evident that 
various other losses, which influenced the efficiency, were either 
irreducible or difficult to diminish. 

As a verification another trial was made under identical 
conditions, but without a jacket. The admission pressure was 
slightly more, namely 60,564 kgs. per m.^ (71*9 lbs. per sq. inch), 
the pressure in the surface condenser was 1,402 kgs. per m.^ 
(2*0 lbs. per sq. inch), and the temperature of the circulating 
water was 18 • 2° C. (64-8° F.). In Fig. 35 (page 604) was reproduced 
the mean indicator diagram of this trial, which gave a consumption 
of 6,650 calories (26,400-5 B.Th.U.) per I.H.P. per hour, or 
10* 15 kgs. (22-33 lbs.) of steam, taking as unit a kilogram of steam 
at a pressure of 88*2 lbs. per sq. inch whose total heat is 655-602 
calories (2,603 B.Th.U.). The maximum thermal efficiency when 
working between the available temperatures of 151*405° and 
18-2° C. was given by the Carnot cycle, and was equal to 

424-405-291-2 

424.405 = 31 -05 per cent. 

Such a cycle would give 1 indicated horse-power hour for a 
consumption of 2,000 calories (7,940 B.Th.U.) equivalent to 3*06 kgs. 
(6*73 lbs.) of steam. 

" What was the reason of this difierence between the theoretical 
" and the actual consumption ? A first cause of loss was the back 
"pressure. It was not possible to cool the steam down to the 
" temperature of the circulating water. In the above trial owing 
" to : (1) the imperfection of the air-pump ; (2) the difference in 
" temperature between the steam and the water ; (3) the pressure 
" necessary to eject the steam from the cylinder ; the exhaust pressure, 
"which ought to be 211*49 kgs. per m.^ (0*3 lb. per sq. inch) was 
" in reality 1,402 kgs. per m.^ (2 -o lbs. per sq. inch), a pressure which 
"corresponded to the temperature of 52-291° C. (126* 12° F.), or an 
" absolute temperature of 325-291° C. (617 - 5° F.), which, substituted 
"for 291-2 in the expression for the thermal efficiency of the Carnot 
" cycle, gave as the maximum accessible thermal efficiency 

424-405-325-291 ^„ ^ 

424.405 = 23-3 per cent., 



,)UNE 1905. STEAM-JACKETING. 607 

"equivalent to a consumption of 2,720 calories (10,798 B.Th.U.) per 
" I.H.P. hour, or of 4 • 16 kgs. (9*15 lbs.) of steam. 

"A second cause of loss is due to the nature of the cycle 
"performed by the steam. The maximum economy that could be 
" expected from the process of the action of the steam in a cylinder 
" was furnished by a special cycle, of which the diagram had been 
" traced in Fig. 35 (page 604) by superposing it on to the indicator 
" diagram." This cycle, which was supposed to be effected by 
1 kilogram of water initially at the temperature corresponding to 
the real back-pressure, namely, 1,402 kgs. per m.^ (2-0 lb. per sq. 
inch) consisted of ; — 

(1) Supply of a sufficient quantity of heat to the kilogram ot 
water to raise its temperature from A^ to A^. 

(2) Vaporization of the water at the temperature and at the 
pressure of admission. 

(3) Adiabatic expansion down to the condenser pressure. 

(4) Condensation of the expanded steam at constant temperature 
and pressure until all converted into water at a temperature A^. 

This cycle was represented by ABCD, Fig. 35 (page 604), and its 
thermal efficiency was computed by means of easily established 
formulae,* and was equal to 19*8 percent., that is 8*5 per cent, less 
than that of the Carnot cycle. This diminution was due to the absence 
of the adiabatic compression of the Carnot cycle. " But there were 
" two reasons which prevented the steam-engine from realising even 
" this cycle. (1) The impossibility of continuing the expansion down 
" to the back-pressure, at any rate in a condensing engine, which was 
" the most interesting. Complete expansion was that which would 
" reduce the pressure to the exhaust pressure, and resulted in a very 
" small mean pressure, necessitating the use of enormous cylinders, 
" out of all proportion to the power of the engine. Further, it would 
" involve a considerable reduction of the mechanical efficiency. By 
" referring to Fig. 85, it would be seen that this steam-engine, supposed 
" perfect from all other points of view, only allowed expansion to the 



* Steam Tables of M. Fe Deruyst. Diagram of M. Bateau. " Annalea des 
Mines de France," Feb. 1897. 



608 STEAM- JACKETING. JuNB 1905. 

(M. Georges Duchesne.) 

"point Ej and the pressure Ee was 4,100 kgs. per m.^ (5-8 lbs. per sq. 
" inch). The diagram ABGEF thus represented the ideal cycle defined 
" by Professor Dwelshauvers-Dery. In the case under consideration 
" this cycle required 3,740 calories (14,848 B.Th.U.) per I.H.P.-hour, 
" or 5 • 7 kgs. (12-54 lbs.) of steam. The work represented by the area 
" EFD was a loss generally called the triangular loss. The clearance 
" volume also diminished the work obtainable from the steam. The 
" work ABCEF W3is reduced by the clearance volume to GHCEF, and 
" this was the maximum that could be obtained from one kilogram of 
" steam when the action of the walls was suppressed. The accuracy 
" of this statement was not complete, for it would require absolute 
" vacuum in the clearance volume at the moment admission commenced. 
" M. Armand Duchesne,* Demonstrator at the University of Liege, 
" had measured directly the considerable degree of superheat, and 
" had calculated most precisely the weight of steam present at this 
" instant ; but this weight was so insignificant as to be negligible. 

" The diagram GHGEF corresponded to a consumption of 4,260 
"calories (16,912 B.Th.U.), or of 6*52 kgs. (14*34 lbs.) per J.H.P.- 
" hour. This was the minimum figure that could be expected ; and 
" there was no valid reason why this figure could not be obtained, if, 
" by some means or another, the harmful effect of the cylinder walls 
" could be suppressed. Now in the trial with a superheating jacket 
"the consumption was 4,050 calories (16,078 B.Th.U.) per I.H.P.- 
-hour, or 6*18 kgs. (13*6 lbs.) of saturated steam of a pressure of 
"88-2 lbs. per sq. inch. This figure was somewhat less than the figure 
"of 6-52 kgs. (14*34 lbs.) which was stated as being the minimum, 
" because the pressure of the steam was slightly higher, 53,583 kgs. 
" per m.2 (7^*2 lbs. per sq. inch) instead of 50,564 (71*9 lbs. per sq. 
" inch), and the pressure in the condenser was slightly lower, namely, 
" 1,200 kgs. per m.^ (1-7 lbs. per sq. inch) instead of 1,402 (2*0 lbs. 
" per sq. inch). It followed therefore that complete jackets, supplied 
" with saturated steam of a higher pressure than the working steam in 

* Theory of the S team-Engine, and the direct and instantaneous measurement 
of the temperature of the Steam and of the Walls," by Armand Duchesne. 
" Revue Universelle des Mines." 4th series, 1904, vol. vii, page 66. 



June 1905. STEAM-JACKETING. * 609 

"the cylinder, assured an actual steam-consumption equal to the 
" minimum which was theoretically possible, taking account of the 
" imperfections of construction and of the cycle followed by the steam. 
" An addition must evidently be made to this consumption, namely, 
" the steam used in the jackets ; for although the indicator had not 
" shown it, a certain quantity of heat had passed from the jacket to the 
" interior of the cylinder ; only a small amount during expansion but 
"more during exhaust. On the other hand, however, the jacket had 
" provided the heat required for external radiation, but the excellent 
*' lagging materials which were generally adopted for good steam- 
" engines enabled the loss due to radiation to be minimised. It was 
" otherwise with the internal radiation, and it was necessary, in order 
" to obtain maximum economy, to reduce as far as possible the metallic 
" surfaces which were in contact with the cylinder steam. This loss 
" by internal and external radiation, when reckoned on the I.H.P. per 
"hour, amounted to 0*610 kg. (1*34 lbs.) in the trial with the 
" superheating jacket. This figure was very high on account of the 
" low speed of the engine (30 revs, per minute), and more particularly 
" on account of the enormous dimensions of the engine in comparison 
" with the power developed." 

Professor V. Dwelshauvers-Dery wrote that the Institution of 
Mechanical Engineers could not be too heartily congratulated on 
having appointed a Research Committee to Investigate experimentally 
economical questions relating to the steam-engine. It would be 
exceedingly difficult to obtain a steam-engine doing actual practical 
work which would lend itself to all the requirements of trials 
devised to ascertain practically useful data, and which at the same 
time would assist in the progress of science, and of " the academic 
side" of the teaching of Engineering, as it was termed by Sir 
Alexander Kennedy.* For this reason Mr. Mellanby was well 
advised in profiting by the occasion of being able to make trials with 
an exclusively experimental steam-engine in order to collect positive 
data on the subject of the usefulness of the steam-jacket. He had 



Engineering, 7th July 1905, pages 27 and 28. 

2 u 



610 STEAM -JACKETING. June 1905. 

(Prof. V. Dwelshauvers-Dery.) 

carried out these trials by an excellent method and according to a 
simple and clear classification, calculated to produce well-founded 
conclusions. All friends of science would add their congratulations 
to those which the Institution had not failed to accord to him. He 
was desirous " not only to find out whether jacketing was sufficient, 
but the reason for any efficiency it might have." 

Without doubt the first part of this programme had been carried out, 
but with regard to the second part he (Professor Dwelshauvers-Dery) 
had to make reservations. In explaining the phenomena due to the 
steam-jacket which he had experimentally verified, Mr. Mellanby was 
absolutely under the influence of the hypothesis of Messrs. Callendar 
and Nicolson, an hypothesis which he explained with unusual 
clearness and which attributed the missing quantity far more to 
direct leakage of steam than to leakage of heat due to the effect 
of the cylinder walls. Hirn's theory enabled the magnitude of the 
missing quantity to be determined as well as the corresponding 
amount of heat, without, however, providing any means of analysing 
what portion was due to the leakage of heat caused by the action of 
the cylinder walls, and what portion was due to direct leakage of 
steam. Ordinary common-sense would attribute a part to each of 
these causes, and Messrs. Callendar and Nicolson undertook a research 
to analyse the matter quantitatively. Their work relied as to one 
part on direct observations which, from his point of view, were 
unassailable, but another portion which depended on calculations and 
hypotheses seemed to him .open to doubt. The steps they followed 
were clearly shown by Mr. Mellanby in the following terms, which 
also exposed the weak part of the argument (page 547) : " It has 
already been shown that if the temperature-range of the surface is 
known, the range of any depth within the metal can be found. 
Conversely, &c. . . . following equation : — • 

B.Th.U. given out per second per square foot = • 74 (T — $).'* 
By this method, therefore, the range of temperature of the 
internal metallic surface of the cylinder was calculated by a 
roundabout method, instead of being directly observed. This was 
the first weak point. Further, the rate at which steam condensed on 
the metal surface was calculated from this uncertain data by means of 



Jl^ne 1905. STEAM-JACKETING. 611 

a formula which was of somewhat doubtful character. It therefore 
followed that the calculated result of the action of the cylinder 
walls, and consequently the amount of the leakage, could only be 
accepted under full reserve. The problem could only be solved in a 
manner free from doubt by the direct and instantaneous measurement 
of the temperature of the steam and of the cylinder ivalls at every point 
of the stroke ; this ivas the only method capable of yielding data " ne 
varietur" The words in italics were taken from a Paper* by 
M. Armand Duchesne, assistant at the University of Liege, who had 
devised a new method of measuring directly and instantaneously the 
temperature of the steam and of the metal. Although M. Duchesne 
was still far from having completed the collection of data which his 
method promised, he thought it was desirable to give the principal 
conclusion derived from the first experiments made with the steam- 
engine in the laboratory of the University of Liege, which had 
nothing special during the tests but its slow speed of 30 revolutions 
per minute. 

M. Duchesne's Paper was illustrated by diagrams, arranged 
somewhat like those taken by an indicator, for the back end of the 
cylinder, and for the whole stroke Fig. 22 (page 572) ; one of these 
diagrams (B) represented the instantaneous temperature of the wall 
as obtained by observation, the second (A) gave the instantaneous 
temperature of the steam, and the third (C), indicated only by small 
circles, gave the calculated temperature of saturated steam as 
obtained from the data given by the indicator diagram by means of 
the usual steam tables. M. Duchesne made the following remarks with 
regard to these diagrams : " We have seen that during expansion 
the two diagrams, that indicated with small circles (C) and that 
marked by continuous lines (A), coincide absolutely ; this is in 
agreement with Hirn's theory, because the initial condensation has 
covered the cylinder walls with vapour, and therefore, during the 
expansion period, a mixture of a liquid with a vapour is being dealt 
with. 



* " Revue Universelle des Mines," 4th series, 1904, vol. vii, page C6. 

2 u 2 



612 STEAM-JACKETING. JUNE 1905. 

(Prof. V. Dwelshauvers-Dery.) 

" At the beginning of the exhaust period the two curves show a 

" small temperature difference, at most 2 J'^ C, which is due to the 

" following circumstance : the ordinates representing pressure being 

" Very short during exhaust, it follows that a small error in their 

" measurement results in a considerable relative error. Moreover, a 

" steam table shows us that during the exhaust, when the pressure is 

" about 1,700 kg. per m.^ (2*4 lbs. per square inch), an increase of 

"pressure of 100 kg. per m.^ (o* 14 lb. per square inch) corresponds 

" with an increase of the saturation temperature of 1*2° C, whereas 

" during the admission an increase of 100 kg. per m.^ (0*14 lb. per 

"square inch) corresponds with 0-06° C. (0*108° F.). It also 

" sometimes happens that the indicator is working below atmospheric 

" pressure, that is to say, the action of the spring is reversed so that 

" the play of the indicator joints has to be taken into account. It can 

" thus be affirmed that during the beginning of the release the real 

" temperature of the liquid is equal to that of saturation. Continuing 

" the examination of the diagram, it is seen that /row that pointy -^QtJis 

" of the return stroke, the steam temperature rises suddenly. The steam 

" is superheated, and if the superheat be measured at the end of release, 

" it is seen that it is 45° C. Far from having water in the cylinder, 

" therefore, there is steam possessing 45° of superheat ; and this 

"phenomenon will not astonish us, because the diagram of the 

" temperatures of the cylinder walls shows us that, towards the 

" middle of the return stroke, the temperature of the metal is about 

" 50° C. higher than that of the steam. If the weight of steam which 

"remains at point No. 16 of the stroke be calculated, that is at the 

" moment at which the steam begins to be superheated, it will be found 

" that it is 0*00208 kg. (o'oo48 lb.). It is natural that 2 grammes 

" of steam, enclosed between walls which become hotter and hotter 

"and the active surfaces of which are becoming proportionately 

" greater, should bo heated by contact. Let us now consider the 

" temperature diagram of the cylinder walls. During the admission 

" the walls are colder than the steam ; it is this which produces the 

"initial condensation; during the expansion the temperature falls, 

"following a law similar to that obeyed by the temperature of 

"the steam, but less rapidly, so that there is a certain difference 



June 1905. STEAM-JACKETING. 613 

" between the two temperatures, wliich goes on increasing, reaching 
" 18° C. at the end of the expansion. 

"The Alsatian school believed that during this period the 
" temperature of the walls was equal to that of the steam, because 
" they were covered with liquid ; that is to say, we ought to have 
" observed, during expansion, a diagram which would be superimposed 
" on that found for the steam. The difference between these two 
" temperatures at first surprised us and induced us to ascertain 
" experimentally whether it were possible for the walls to have a 
" temperature somewhat higher than that of the steam. We instituted 
" the following experiments . . , ." 

In order to illustrate these conclusions he (Professor 
Dwelshauvers-Dery) referred to two of the illustrations of M. 
A. Duchesne's Paper. Fig. 22 (page 572) gave the diagrams of 
the temperatures of the walls and of the steam for a trial carried 
out under the following conditions. The admission pressure was 
6 atmospheres ; the steam was not superheated ; the engine was 
working condensing and without steam-jackets ; the cut-off was at 
■j^gth of the stroke (point No. 1) ; the release commenced at y^^ths 
of the stroke ; and the compression commenced at y^oths of 
the return stroke. It would be seen from this diagram, 
which was confirmed by a dozen other trials made under varying 
conditions, that during the expansion period the temperature of the 
steam was exactly equal to that of saturation, whilst, contrary to the 
opinion of the Alsatian school, but in conformity with that of 
Messrs. Callendar and Nicolson, the temperature of the metal was 
higher than that of the steam. 

The second illustration. Fig. 36 (page 614), established the 
comparison between the temperatures of the metal in the cases when, 
other things being equal, the engine was working without steam-jackets 
(No. 1), or with a jacket filled with steam of admission pressure 
(No. 2), or lastly, with a jacket filled with steam of high pressure 
(No. 3), M. Georges Duchesne's system. The conclusions were thus 
formulated by M. A. Duchesne : " A comparison of the three curves 
" shows clearly that the magnitude of the exchanges of heat depends 
'* on the weight of the film water covering the walls. In curve No. 1 



614 STEAM-JACKETING. June 1905. 

(Prof. V. Dwelshauvers-Dery.) 

'• of the trial without jackets a considerable amount of water remains 
" on the walls at the end of the expansion ; and, at the instant 
" communication is established with the condenser, this water produces 
" a cooling of the metal shown by the rapid fall on curve No. 1. In 
' ' the second trial there is less water, and curve No. 2 falls less rapidly 
" at the end of expansion than does curve No. 1. Lastly, in the third 
" case, the superheating of the steam shows us that there is no water 
" present at the beginning of the release, and it is clear from curve 
" No. 3 that at this moment the temperature of the walls is subject to 
" no variation." 

Fig. 36. 

^ Temperature Centlgrcule 



o 

o a 

o ce o) 

(S "i* b 

a 03 

ee Si ^ 

JZi -^ rCj 

^t: be 



■» '^ '^ '^ 

^ -^ ^ *- 






5'1 



^v- ^- ^^ 



t> 

O 

e 

SI. 



.'»> 







M. A. Duchesne came to a further conclusion which gave a high 
importance to his work, for it was the discovery of a fact hitherto 
unsuspected, as follows : '• The steam is superheated before the end 
of exhaust, even when the engine is working without a jacket, and 
the sooner the superheating commences the more pronounced is the 
action of tbe jacket." 

All the above conclusions were based entirely on experiment and 
on direct measurement. Were they in accord with the six conclusions 
with which Mr. Mellanby finished his Paper ? Mr. Mellanby's first 
four conclusions were absolutely independent and gave to the Paper 
a great practical value. The fifth, with the exception of the two 



June 1905. STEAM-JACKETING. 615 

clauses (a) and (b), was in agreement with the experiments of 
M. A. Duchesne. As regards the sixth, which was deduced by 
the method of Messrs. Callendar and Nicolson, there were strong 
reasons for believing that it was defective, because the numerical 
results given were obtained by a procedure which was not exempt 
from doubt, and because the hypothesis that the greater part 
of the missing quantity must be attributed to leakage of steam, and 
that the part due to the thermal action of the cylinder walls was 
almost negligible, and was incapable of directly accounting for the 
economy due to the steam-jacket. 

By what phenomenon was it possible for the steam-jacket to 
reduce the leakage of steam ? By what natural process could the 
jacket suppress the leakage and annul the missing quantity ? This 
was what happened when the jacket was supplied with saturated 
steam at a much higher pressure than the steam admitted to the 
cylinder — the invention of M. Georges Duchesne. This fact was 
reported by that engineer in the " Revue Universelle des Mines." * 
The weight of steam admitted per stroke experimentally determined 
was exactly equal to the weight of steam present at cut-off as shown 
by the indicator. There was therefore no missing quantity ; 
moreover, the expansion curve followed the adiabatic law, a fact which 
enabled M. Georges Duchesne to announce the following very 
important practical deduction : " The result of heating the jackets 
with saturated steam of a higher temperature than that admitted to 
the cylinder is to produce a practically perfect transformation of 
the steam in the cylinder." 

As a last question, he would ask for an explanation, by means of 
Callendar and Nicolson's hypothesis, of the following fact which 
had not yet been published, but which had been observed in the 
laboratory of Liege, and for the accuracy of which he would 
vouch. With the steam-jacket arranged in accordance with M. 
Georges Duchesne's system, and filled with high-pressure steam, 
18 per cent, reduction of the missing quantity had been obtained 
compared with the trial made with an ordinary jacket, other things 
being equal. 

* " Revue Universelle des Mines," 4th series, 1904, vol. vii, page 221. 



616 STEAM -JACKETING. JUNE 1905. 

(Prof. V. Dwelshauvers-Dery.) 

In conclusion, he thought that that part of Mr. Mellanby's 
Paper which dealt with experimental observation had a great value 
and was far-reaching ; but that that part which gave the explanation 
of the phenomena by means of the method of Messrs. Callendar and 
Nicolson was open to the same objections as the method itself. 

Mr. Mellanby wrote, in reply to Professor Dwelshauvers-Dery 
(page 609), that although at the first glance there seemed to be 
several points upon which they were not in agreement, he thought 
he would be able to show that really their differences were not so 
marked. Professor Dwelshauvers-Dery mentioned that ordinary 
common sense would attribute part of the missing quantity to 
condensation and part to leakage. Although this might seem quite 
obvious, yet it was exactly the point which most experimenters upon 
the steam-engine had overlooked. Modern text books, and even the 
accounts of engine trials published in the engineering press, contained 
the most elaborate calculations of the heat exchanges taking place in 
a steam-engine, and detailed instructions were given to show how 
these losses could be graphically represented on the temperature- 
entropy diagram. Yet the whole of this work was based upon the 
idea that there was absolutely no leakage of steam either into or 
out of the cylinder during a revolution. As the author by 
independent experiments had proved that valves into which the best 
workmanship had been put did leak considerably when running, 
he felt that these intricate calculations were, from any point of 
view, altogether useless. Professor Dwelshauvers-Dery thought 
that Messrs. Callendar and Nicolson's method of calculating the 
amount of condensation was roundabout and therefore open to 
objection. The basis of their calculations however was the direct 
temperature measurements they had made, and if these were correct 
the possible errors in their estimated amounts of condensation could 
only be very small. He would also point out that the formula 
objected to was only suggested because it appeared to give the same 
amount of condensation as was given by the calculations from the 
measured temperature ranges. It was also in agreement with the 
special experiments made to determine the rate of condensation of 



June 1905. STEAM-JACKETING. G17 

steam by an apparatus of the surface-condenser type. Considering 
the vast experience of Professor Callendar in thermo-electric work 
and his reputation as a careful experimenter, the author had accepted 
his figures and had attempted to show that the results of his own 
experiments were in general agreement with them. 

He was not thoroughly acquainted with the details of 
M. A. Duchesne's apparatus, and was therefore unable to criticise 
his measurements of the temperature-ranges of the metal. He had 
however pointed out in his reply to Professor Hubert and 
M. G. Duchesne that, if the temperature-ranges shown in Fig. 22 
(page 572) and Fig. 36 (page 614) for the unjacketed and the 
ordinary jacketed engines were correct, the whole of the missing 
quantity in the unjacketed engine could not have been due to 
condensation. It therefore followed that, to account for the 
observed differences in the two trials, there must have been 
a considerable valve-leakage in the unjacketed case. If 
M. Duchesne's curve of the metal cycle for the unjacketed engine 
were considered, it would be seen that it showed the metal to go 
through a temperature-range equal to about half that of the steam. 
It would be noticed that it was a very regular curve with no sudden 
changes. If, therefore, the mean temperature of the metal and the 
maximum temperature to which the metal could be raised were known, 
the experiments of M. Duchesne appeared to show that the range of 
temperature would not be more than twice the difference between 
these maximum and mean temperatures. It would be seen, by 
reference to the Paper (page 552), that in one example the mean 
temperature of the metal was 335^ F. The maximum temperature of 
the steam was 356*5^ F., and therefore the temperature-range of the 
metal could not have been more that 43° F. This, as shown, would 
amount to a condensation of 555 lbs. per hour. Since the total 
missing quantity for this trial was 753 lbs., it would appear that, 
even on the basis of M. Duchesne's experiment, there must have been a 
leakage of 198 lbs. per hour. In Trial 99 the low-pressure cylinder 
was jacketed with steam at a temperature of about 360° F. The 
admission temperature of the steam was only 235° F., so that this 
corresponded with what Professor Dwelshauvers-Dery called M. G. 



618 STEAM-JACKETING. June 1905. 

(Mr. Mellanby.) 

Duchesne's system. It was almost certain that in this case there was 
no initial condensation, hut there was a missing quantity of 285 lbs. 
which it appeared could not be attributed to anything but valve- 
leakage. It must be remembered that the Corliss valves in the high- 
pressure cylinder and the slide-valve in the low-pressure cylinder 
were of the highest class of workmanship and at least equal in steam- 
tightness to the valves of the ordinary engine. The author therefore 
felt that, taking Messrs. Callendar and Kicolson's, M. Duchesne's 
and his own observations together, there was the strongest proof 
that cylinder condensation was much less, and valve-leakage more, 
than was generally imagined. 

In answer to Professor Dwelshauvers-Dery's inquiry as to how 
the steam-jacketing could reduce leakage, he would repeat what he 
had before said ; that leakage had been found to be much reduced 
by warming the valve-faces, and that one effect of the jacket was to 
do this warming and consequently reduce leakage. 

From his written contribution it would appear as if M. Georges 
Duchesne were absolutely certain that the whole of the missing 
quantity was due to initial condensation. The author was however 
of the opinion that if M. Duchesne studied the work of Messrs. 
Callendar and Nicolson, the cylinder- wall temperature curves of M. A. 
Duchesne and the Paper now before him, he would at least begin to 
have some doubts as to the correctness of this opinion. He was 
unable to place any confidence in the elaborate calculations 
M. Duchesne had presented to them. These calculations were 
based upon the assumption that there was absolutely no leakage 
either into or out of the cylinder during a revolution, an assumption 
that seemed to the author to be quite untenable. 



Junk 1905. 



619 



THE GKOWTH OF LAKGE GAS-ENGINES 
ON THE CONTINENT. 



By M. RODOLPHE E. MATHOT, Blember, of Brussels. 



(^Translated from the French.') 

As large Gas-Engines have now won an important place in 
industry, it will be of interest to cast a retrospective glance at their 
evolution. So long as town gas was the fuel " par excellence " for 
industrial engines, by reason of the facility attending its use, the 
applications of the explosion-engine were limited to 50 to 75 H.P., 
beyond which the cost of working was found to be excessive. Poor 
gas produced under pressure with the old apparatus of the Dowson 
type enabled one, it is true, to venture upon 75 to 100 H.P. and even 
greater powers with more practical results ; but the complication of 
the gas-generating apparatus, the initial cost, and the space taken 
up by the producer-plant and engine rarely compared favourably 
with the steam-boiler and engine. The latter, moreover, adapts 
itself better to every class of fuel, whether gaseous, liquid or 
solid. Amongst solid fuels coal dust, peat, vegetable waste, straw 
and saw-dust constitute very advantageous fuels without requiring 
complicated and troublesome furnaces. 



6£0 LARGE GAS-ENGINES. June 1905. 

In order to compete successfully with the steam-engine, the 
explosion-motor required to he provided with cheap gas, easy of 
production by means of simple and economical apparatus. Suction 
gas-producers have decided the question for the industry in general, 
whilst the purifying and washing processes for blast-furnace gas, 
coke-oven gas, &c., have, in an unexpected manner, brought about 
the possibility of applying explosion-engines to the greatest 
motive powers required in the metallurgical industry. Whilst in 
electric-lighting stations the steam-engine, in spite of its great 
regularity in work, encountered a serious rival in the gas- 
engine, the petrol engine has decided the question of road 
locomotion. Motoring, thanks to the wonderful attributes of small 
engines, has led to the application of internal-combustion engines 
for this purpose. Submarines have already been fitted with these 
engines, and it is safe to predict that at no remote date the explosion 
engine will take its place in the mercantile marine, side by side with 
powerful steam-engines, for the propulsion of vessels. 

The development of large gas-engines can be said to date back 
no further than five to six years. Eight to ten years ago they 
were initiated simultaneously in Germany, England and Belgium, 
early attempts being made to utilise blast-furnace gas, which was 
expected to open up such a vast field for the employment of large 
engines. Although the first trials were only attempted on small 
engines, the results of the experiments soon gave encouragement 
to the efforts of the investigators. The Cockerill Co. of Belgium 
constructed a single-acting Otto-cycle engine of 200 H.P., which has 
been working regularly at their establishment for six years. This 
stage in the path of progress was strongly accentuated by the 600-H.P. 
engine on the Delamare-Deboutteville system, which the Cockerill 
Co. exhibited at the Universal Exhibition at Paris in 1900. This 
magnificent engine was single-acting, the piston having a diameter 
of 1*300 m. (4 feet 3^% inches) and a stroke of 1*400 m. 
(4 feet 7 J inches). It was designed to develop its power at 
80 revolutions per minute, which, with an initial explosive pressure 
of 310 to 325 lbs. per square inch, produced on the piston a pressure 
of 300 tons at each explosion. In a short time several famous firms 



June 1905. LARGE GAS-ENGINES. 621 

entered, in their turn, upon the construction of large engines intended 
to utilise the gases liberated by the various reactions occurring in 
the manufacture of iron, coke, etc., and the metallurgical industry 
was not long in entering upon the path of progress by replacing for 
its old boilers and engines powerful installations of explosion- 
engines. It remained, however, to conquer the vast domain of 
manufacturing industry. It is in this sphere that the struggle is 
taking place with the steam-engine, which a long career has endowed 
with improvements in methods and execution. The physical laws 
which govern the production and utilization of steam as a motive 
power have long been known, having at an early date emerged 
from the obscurity which enveloped their interpretation, and 
thermodynamic science has given them definite sanction by numerous 
investigations. _^ 

Improvements in construction advanced side by side with the 
progress of scientific theory to accomplish mechanical marvels. 
Steam is, however, a fluid much less complex in nature than 
explosive mixtures. The action of steam is governed by precise 
laws which pertain only to the sphere of physics, whilst the 
production of combustible gases and their mode of evolution under 
the form of explosive mixtures in engines, are as much within the 
domain of chemistry as of physics and mechanics. Although the 
generic theory of gas-engines has rested up to the present on a series 
of hypotheses which have not yet received experimental confirmation, 
these engines have gained ground in application to various industries 
with exceptional rapidity compared with any other kind of motive 
power. The invention of gas-producers and the improvements 
made in the last few years, and especially in those working with 
the direct suction of the engines, are manifestly most important 
factors in this success. 

Before examining the successive developments of these producers, 
the principal phases of the improvements through which the engines 
themselves have passed will be analysed. It was first of all in 
Germany and then in England that their construction underwent the 
most rapid development. Afterwards America was the country 
which produced the greatest number. It is therefore not surprising 



622 LARGE GAS-ENGINES. June 1905. 

to notice that each type has retained in its construction or design 
something which reveals its nationality. The German engine has 
always presented the appearance of a well-finished machine as 
regards constructional details — all the parts machined were usually 
polished bright, which disclosed a real anxiety on the part of the 
makers to impart a high finish to their machines. This, of course, 
affected the price, but the life of the engine was materially increased 
thereby. There are cited as examples of longevity certain " rack " 
engines of the Otto-Langen type and make, which are completing 
a career of thirty years' service. 

The English makers took up another position, namely, that of 
producing cheaply in order to produce on a large scale. Thus the 
world is indebted to them to a great extent for the propagation of 
small engines for industrial purposes. This was a fertile field for 
investigation and experiment, which continental makers, deeply 
immersed in the construction of powerful engines, havQ often failed 
sufficiently to notice. English builders of engines designed for 
the use of town gas vie with each other in ingenuity in the 
arrangement of the parts in order to attain efficient and simple 
mechanical devices. The lift of the valves and their mode of 
operation are in general obtained by cam and lever movements 
with a definite movement giving a positive action. The governor 
itself is reduced to its most simple expression, since in the " hit- 
and-miss " arrangement its action consists merely in displacing 
to a slight extent a small piece which is normally interposed 
at the point of contact between the controlling lever and the 
stem or spindle of the gas-valve, in order to open it or to leave it 
closed, and which transmits the motion by which the valve is 
opened ; when the piece is displaced the motion is no longer 
transmitted. Unfortunately, this type of governing is not compatible 
with the requirements of the working conditions of modern large 
engines. 

The English makers are familiar with the series of very special 
phenomena produced by the system of " hit-and-miss " regulation. 
In particular they are aware : — 



June 1905. LARGE GAS-ENGINES. 623 

(1.) That after a stroke with no charge the following explosion 
is more powerful in certain cases and weaker in others, according to 
the form and the arrangement of the explosion-chamber and its 
ports. 

(2.) That violent explosions are often produced, the consequences 
of which are the more injurious to the life of the working parts as 
the engines are larger and consequently oppose the inertia of heavier 
masses to movement. 

(3.) That the intermittence of the explosions produces cyclic 
variations or irregularities in the revolutions of the fly-wheel 
regarded individually, and that these irregularities are incompatible 
with the working conditions of dynamos for lighting, etc. 

(4.) That in order to overcome these difficulties it is necessary 
to use extra heavy fly-wheels, which constitute an additional load 
on the engine and cause a reduction of mechanical efficiency. 

As against these defects it will, it is true, be contended that the 
" hit-and-miss " constitutes the simplest form of governing as regards 
construction, and is the least liable to get out of order ; that it keeps 
within very small limits the variations of speed resulting from 
alterations of load ; and that it is the mode of regulation which 
secures the lowest consumptions at the different speeds, because 
with a constant charge the latter is always properly proportioned 
and is fixed once and for all. It must, however, be recognised 
that if this system satisfies the conditions required in the case 
of small engines for industrial purposes, which have for a long 
time made the reputation of the English makers, it is incompatible 
with what should be obtained from large engines of hundreds or 
thousands of horse-power, if they are to compete with the steam- 
engine as prime movers. 

The merit of having entered upon the new path which the 
construction of gas-engines has followed for five or six years 
undoubtedly belongs to the Germans. The old makers of gas-engines 
in Germany took the initiative of departing from old methods. In a 
short time their processes were themselves improved and perfected 
by the makers of steam-engines, long accustomed to circumvent or 
overcome practical difficulties in the construction of large engines. 



624 LARGE GAS-ENGINES. June 1905. 

With the exception of large gas-engines tending towards a single type, 
it may be said that they all have manifest tendencies to resemble the 
modern steam-engine from the point of view of form and valve-gear. 
Having regard to the fact that valves are the common means of 
distribution, that they are operated by a side shaft, and that large 
engines now work double-acting, it is natural and logical that 
the explosion machine should borrow from the steam-engine the 
design and methods with which it has been equipped in its long and 
victorious career, Plate 23. The introduction and growth of suction 
gas-producers and the utilization of blast-furnace gas, coke-oven 
gas, etc., which have marked the development of large gas- 
engines, have led to the creation of different designs for their 
construction. Different principles have thus been modified in their 
applications, such as the regulation, the compression, the cooling, and 
the ignition. Without dwelling upon the different stages of their 
transformation, it will be shown in what way modern methods 
differ from old methods, and the probable direction of future change, 
having regard to the knowledge and practical experience already 
acquired. 

Regulation. — For the reasons which have been enumerated, the 
*' hit-and-miss " system of regulation has been completely abandoned. 
This system, moreover, does not lend itself to working with very 
light charges, or with no charge, in the case of engines fed by suction 
gas-producers. As in these circumstances the gas-supply alternates 
with three, four or even five strokes with no charge, it happens that 
the suction which determines the supply of air to the producer 
is not sufficiently uniform, and that the fire finishes by being 
extinguished, by producing a very poor gas through the lack 
of activity in the furnace. German makers then invented the 
conical cam for the admission of the gas. Fig. 1, which, being 
displaced by the action of the governor, produced variable lift 
of the gas-valve. But this device was only a variation of the 
stepped cam or of the stepped pecker block, Fig. 2, which the English 
makers had tried in their electric types, and they soon discovered 
the uneconomical results it caused. The stepped arrangement had 



June 1905. 



LARGE GAS-ENGINES. 



625 



the advantage over the conical cam of lessening the work upon 
the governor. But, as both systems acted on the quantity of gas 
admitted, whilst the quantity of air of the mixture remained constant, 



Fig. 1. 

Conical Cam for Variable Gas 

Admission. 



Fig. 2. 

Stepped Pecker Block for Variable 
Gas Admission. 



(^=) 





Fig. 3. 
Card with Sharp Explosion owing 
to rich mixture. 




Fig. 4. 
Card with Late Firing owing 
to weak mixture. 




mixtures of variable composition — often too rich in the case of a 
full charge, Fig. 3, and always too poor with a weak charge, Fig. 4 
— were formed. In the latter cases the ignitions were tardy, the 
diagrams bad, and the efficiency less as the charge was reduced in 

2 X 



626 



LARGE GAS-ENGINES. 



June 1905. 



richness. Whilst with a good engine regulated by the " hit-and- 
miss " system the consumption at half load, which from the industrial 
point of view is the most interesting, was not more than about 
20 per cent, higher per H.P.-hour than with full load, it became 
40 to 50 per cent, higher with an engine with variable mixture. 

It is, therefore, towards the system of admission of a variable 
quantity of mixture, but of uniform composition, that makers have 

Fig. 5. 
Mixing- Valve and Throttling Butterfly- Valve. (Benz.) 




directed their attention. A few English makers, Tangye, Willans and 
Eobinson, also Westinghouse, decided the question by throttling, by 
means of a butterfly-valve or a cylindrical slide-valve, the mixture 
which is regulated in advance; this butterfly- valve or slide-valve is 
controlled by the governor and placed immediately before the valve 



June 1905. 



LARGE GAS-ENGINES. 



627 



admitting the mixture to tlie cylinder. The principle was still further 
improved by an automatic mixing-valve, preceding this regulating 
device, Fig. 5, as constructed by Benz in his four-cycle engines. 
However, the majority of the Continental makers, with the object of 
obtaining a prompter and more reliable action of the governor, have 
endeavoured to combine under one single control and in a single 
device the slide or inlet-valve and the mixing-valve, and, instead of 
throttling the charge in the passage, they have provided the mixing- 
valve with a variable stroke under the action of the governor, Fig. 6. 



Fig. 6. 
Variable Admission- Valve with Stroke controlled by the Governor. (Otto-Deutz.) 




The application of this device has not, however, been extended to 
very large engines, as it was found to be too heavy and cumbersome 
to handle and dismantle for the purposes of cleaning. Admission in 
variable quantity of a uniform mixture involves variable compression 
and the necessity for a high original compression. If for the full 
admission of the charge this compression were from 170 to 200 lbs. 
per square inch, it might fall to less than 45 to 55 lbs. per sq. inch 
for the minimum admissions, which would interfere with the prompt 
ignition of the very poor mixtures which are at the present time 
used in engines applied for industrial purposes. Further, this mode 

2x2 



628 



LARGE GAS-ENGINES. 



June 1905. 



of admission produces at the time of the suction with light charges, 
a vacuum or negative work which would become considerable in 
engines of high power. This has been avoided by combining with 
the variable admission of the constant mixture an additional 
admission of air or impoverished mixture, in order to effect at the 
same time the constant compression and minimum vacuum in the 
cylinder. The firms of Cockerill and Niirnberger Maschinenbau 
have, as the author foretold three years ago in " Power " of 
New York, already adopted mechanism for this purpose. 

Fig. 7. 
Balanced Inlet-Valve and Rolling-Path Lever. (Winterthur.) 




Control. — The lift of the valves, which is usually effected in the 
old engines by simple movements of levers at one end of which 
a cam acts upon a roller, has also undergone some alterations. 

In large engines this has been replaced by eccentrics and 
" roller-path " levers, thus imitating what is applied to steam-engines 
of the Sulzer type. Fig. 7 shows an application of this kind adopted 
by the Winterthur Co. Some special valve-gears have even been 
provided with ratchet movements with air-pistons, such as CockerilPs 
and Niirnberg's. 



June 1905. LARGE GAS-ENGINES. 629 

The main object of these improvements has been to secure a 
more gentle and silent action relative to the size and weight of the 
parts to be actuated. 

It is advisable to bear in mind with regard to English and 
American engines that the compression generally adopted for large 
gas-engines is from 170 to 200 lbs. per square inch. 

Cooling. — Cooling is one of the points which has most attracted 
the attention of makers of large gas-engines, as the effects of tension 
in the castings, due to unequal expansion, play a considerable part 
in the heavy pieces which enter into their construction. Cylinder- 
heads or explosion-chambers have been, in fact, one of the principal 
sources of difficulties and disappointment, as without any apparent 
reason they fracture at the most unexpected places. It is only 
long experience, assisted by numerous examples, which has placed 
makers on the track of the most appropriate forms for securing the 
strength of cylinder heads, rather than the selection of the materials 
to be employed in their construction. Steel itself has been tried 
without success. The principal factor for the preservation of 
cylinder heads is the manner in which the cooling water circulates 
therein, and the type which appears to be favoured at the present 
time for Otto engines is the one which places the inlet and exhaust- 
valves in the same vertical axis. Both are placed in a passage or 
ante-chamber surrounded on all sides by the cooling water. In order 
to secure for the whole of the parts an equal expansion, this chamber 
is arranged symmetrically relatively to the axis of the cylinder. The 
exhaust is allowed to discharge on the extension of the combustion- 
chamber, which is surrounded as completely as possible by water. An 
abundant circulation is thus provided around the seat and the stem of 
the exhaust-valve, which is also provided with an internal circulation 
in the case of engines of more than 60 to 100 H.P. The water which 
has entered at the bottom of the cylinder-head escapes at the top, or 
if need be in the case of engines of less than 75 H.P. completes its 
circulation in the jacket of the cylinder proper, Fig. 8 (page 630). 
It will be noticed in this example that this jacket is cast with the 



630 



LARGE GAS-ENGINES. 



June 1905. 



frame, and that the cylinder itself is independent, and free to expand, 
whilst the cylinder-head is attached by flanges and bolts. This 
arrangement is the only one which secures a large base for the 
engine, and avoids the overhanging cylinder, which is completely 
banished from modern construction. 

In order to combat the excessive temperature which attends high 
compression, certain makers, amongst whom may be mentioned the 
Niirnberger Maschinenbau Co., have even provided the piston ends of 



Fig. 8. 
60 H.P. High- Compression Gas-Engine. (Soest.) 




their single-acting engines with a water circulation. This is effected 
by a pump forming part of the engine. Fig. 9, Plate 22. This is 
evidently a complication which others have easily avoided for engines 
even cf 150 H.P., by arranging the piston so that a free access of air, 
due to the to and fro movement, alone effects the cooling. In order to 
derive every economical advantage from high compression without 
running the risk of self-ignition, the firm of Koerting has even 



June 1905. 



LARGE GAS-ENGINES. 



631 



arranged inside the explosion-chamber of their Otto-cycle engines a 
hollow casting through which there is a special water circulation, 
Fig. 10. This cooling effected in the very heart of the mixture deals 
with the excessive temperature there, and is said to have given the 
best economical results. In double-acting engines the cooling is 
still the object of great attention, and apart from the cylindrical 
jacket and cylinder ends, independent water circulation is used to 
cool the piston and piston-rods, the seats of the exhaust- valves, and 

Fig. 10. 
Internal Water-Cooling Arrangement. (Koerting.) 




the stuffing-boxes, Fig. 11, Plate 23. The general temperature of the 
surrounding parts is kept cooler than in English engines, and for this 
purpose water is delivered at certain parts, such as the piston and 
piston-rods, at pressures from 1 to 4 atms. (15 to 60 lbs. per sq. inch) 
by means of a special pump. 

It appears from several experiments which the author has made 
on double-acting Otto-cycle engines that the quantity of circulation 
water required for the different parts is as follows : — 



632 



LARGE GAS-ENGINES. 



June 1905. 



Per B.H.P.-hour for engines of 200 to 1,000 H.P. 


Litres. 


Gallons. 


Cylinders, cylinder ends and stuffing-boxes 

Pistons, piston-rods ...... 

Valve-boxes and seats, and exhaust valves 


18 to 24 
8 to 12 
4 to 6 


4 to 5^ 

i| to 2f 

|toi| 


Or a Total of . 


30 to 42 


6f to 9| 



These figures imply water admitted on an average of 12°-15° C. 
(53-6°-59° F.) and leaving the cylinder-jackets at 26°-35° C. 
(77°-95° F.) the pistons at 35°-40° C. (95°-io4° F.) and the valve- 
seats and boxes at 45° C. (113° F.) on an average. An engine of 
1,000 H.P., of the two-cylinder double-acting type, would therefore 
require about 40 cubic metres (8,900 gallons) of cooling water per 
hour. As this is an excessive quantity which is not available at 
every works, recourse is commonly had to the use of cooling towers, 
which reduce the consumption of water to about • 5 litre (|^ gallon) 
per H.P. hour, absorbed by evaporation. This method has also the 
appreciable advantage over the ordinary water circulation of 
eliminating, owing to the continuous use of the same water, the 
deposit of calcareous incrustations. Without possessing in the case 
of gas-engines the same dangers as in steam-boilers, lime scale and 
deposits still constitute a drawback. They obstruct the pipes and 
passages, and impede the regular cooling by coating (with a 
non-conducting material) the metal at the places where a high 
temperature is most injurious. At the parts cast with a double 
jacket, and which cannot be dismantled, it is necessary to arrange 
large openings covered by bolted lids in order to enable free access 
to the inside to remove these deposits. 



Ignition. — This question has for some years been solved in a 
satisfactory manner by the use of magnetos producing a spark on the 
break of the circuit, which in a short time has ousted all the old 
method of ignition, such as the incandescent tube, the spark produced 
by batteries and accumulators, by dynamos, etc. This modern magneto 



Junk 1905, LARGE GAS-ENGINES. 633 

ignition has for three years been provided with a timing gear, by 
means of which the ignition can be advanced or retarded 
experimentally during work. Still, as all working parts subject to 
the frequent and abrupt movements of the current breaker of the 
magneto are liable to get out of order and to wear rapidly, it is 
necessary to meet these objections by the adoption of very light 
parts with but little inertia. They must be easy of access and to 
handle for upkeep and inspection. With the increase in dimensions 
of engines it has been necessary to deal with greater volumes of 
explosive mixtures but of poorer composition. This has wisely led 
many makers to provide their double-acting engines with two distinct 
ignitions for each piston face, the one placed near the inlet-valve at 
the top and the other at the bottom near the exhaust-valves. 

Starting. — A few words remain to be said on the starting of 
large engines, to which the explosion " self-starter " has not 
been applicable. The apparatus, which causes behind the piston 
the explosion of a mixture previously introduced into the 
cylinder when the engine is stopped, develops a too sudden and 
violent strain, owing to the resistance of the inertia of the parts 
at rest. Moreover the action of this apparatus is too uncertain. 
Use is therefore made of air compressed to 150 to 225 lbs. 
by a compressor and stored in a reservoir. When an engine has 
to be started, the crank is placed in the starting position 
corresponding to the explosion stroke by turning the fly-wheel by 
means of a lever or a system of gearing. In the case of the single- 
acting engines a charge of compressed air is introduced by operating 
by hand the valve shown at the end of the cylinder-head of the 
engine of Fig. 8 (page 680). This operation is repeated several 
times in succession until the engine, sufficiently started, draws in the 
explosive charge and drives itself. In the best double-acting 
engines the starting-valve working by compressed air is generally 
operated mechanically by the half-speed shaft. 

Lubrication. — Lubrication has also of late formed the subject of 
important improvements. In engines of medium power, that is, up 



634 LARGE GAS-ENGINES. June 1905. 

to 150 to 200 H.P., the main bearings of the crank-shaft are 
usually lubricated by means of a revolving ring plunging in an 
oil bath. For larger engines bearings with brasses consisting of 
several parts, to take up the wear and the working stresses, 
are used. As this system renders it impossible to apply the 
lubricating ring which gives such good results in dynamos, 
recourse has been had to continuous oil-feed under pressure. This 
pressure also secures a more reliable lubrication of large surfaces 
supporting great loads, as is the case with the crank-shafts of engines 
of 1,500 to 3,000 H.P. which exceed 500 mm. (i foot 7^^ inches) 
diameter. For lubricating the pistons and the stuffing-boxes of the 
piston-rods, oil-feed under pressure is a necessity, as the oil is more 
reliably conveyed to the rings, the tightness of which depends to a 
great extent on the free play secured to them by proper lubrication. 
Excess of oil in the cylinders, which, by rendering them dirty, is 
the principal cause of " back firing," has also been greatly reduced 
by the use of a draining device to the cylinder. Fig. 8 (page 630) 
represents a valve which is fitted for this purpose at the back 
lower part of the cylinder. It is operated by hand, and enables 
surplus oil or foreign matters to be expelled whilst the engine is 
working. 

Fly-wheels. — The mode of regulation by admission on each cycle 
enables fly-wheels relatively less heavy to be used. The same may 
be said of the application of double-acting, and of multiple 
cylinders. The requirements, however, which are now expected 
from engines in large electric-light stations have given rise to 
the use of special fly-wheels, with the object of reducing as much 
as possible the degree of cyclic irregularity. For ordinary 
industrial purposes a regularity of 1/25 to 1/30 of the speed in 
one and the same cycle can readily be attained. For electric 
lighting by continuous-current dynamos it is necessary, with 
the view of obtaining a steady light, for the degree of irregularity to 
be less than 1/50 or 1/60, whilst for the working of alternating- 
current generators in parallel it should practically be about 1/150. 

The following formula enables one to determine the dimensions to 



June 1905. 



LARGE GAS-ENGINES. 



635 



be given to the fly-wheels of diiferent types of engines, having regard 
to the purposes for which they are intended : — 



P D^^ = K 



N 



N 



D'^a n^ 



hence P = K 

in which 

P = the weight of the rim (without arms or boss) in tons, 
D = diameter of the centre of gravity of the rim in metres, 
a = the degree of irregularity, 
n = the number of revolutions per minute, 
N — the number of brake H.P., 
K = coefficient varying with the type of engine, 
K = 44,000 for Otto-cycle engines, single cylinder, single-acting. 
K = 2S,000 for Otto-cycle engines, two opposite cylinders, 
single-acting, or one cylinder double-acting. 
Ill K - 25,000 for two cylinders single-acting, with cranks set at 90°, 
K = 21,000 for two twin cylinders, single-acting. 
K = 7,000 for four twin opposite cylinders, or for two tandem 
cylinders, double-acting. 
The total weight of the fly-wheel is P x 1,4 
Diagram showing the arrangement of cylinders. Fig. 12. 



I 
II 



IV 
V 



Fig. 12. 
Diagrams of Cylinder Positions relating to calculation of Fly- Wheels. 



" { 



Tm 



III 



r\ 






mm u 




636 LARGE GAS-ENGINES. June 1905. 

Consumption and Efficiency. — Modern large engines have attained 
high organic efficiency owing to the proportional reduction of their 
weight and the finish of their construction. Double-acting engines 
are usually made with a weight of at least 100 kgs. per H.P. 
(220J lbs.). It is admitted that Otto-cycle double-acting engines 
attain 90 to 92 per cent, mechanical efficiency, whereas an output of 
only 75 to 80 per cent, was attained by 2-cycle engines. This 
waste, being due to the work absorbed by the air-pump and by the 
gas-pump, cannot, however, deteriorate the value of the magnificent 
engines, of which the Oechelhauser and the Koerting are classical 
types possessing their own advantages. Double-acting Otto-cycle 
engines attain a thermal efficiency of 28 to 30 per cent, relatively to 
the effective work, that is, the H.P.-hour is attained with about 2,200 
calories (8,734 B.Th.U.). This consumption converted into the 
volume of the different gases used industrially would be as 
follows : — 

Coke-oven gas 585 litres (20*7 cubic feet). 

Mond producer-gas 1,760 litres (62 '2 cubic feet). 

Anthracite producer-gas 1,850 litres (65 -4 cubic feet). 

Blast-furnace gas 2,500 litres (88*3 cubic feet). 
This implies the mean chemical compositions and the average 
calorific values indicated on the synoptic drawings of the table of 
Fig. 13 (page 637), which shows the normal content of each gas in 
hydrocarbons (GJij,), marsh gas (CH4), carbon monoxide (CO), 
hydrogen (H), carbonic acid (CO2) and nitrogen (N). 

The author will now examine in their main features the 
improvements in the construction of large engines during the 
last three or four years. In order to proceed chronologically, 
he will analyse simultaneously the present engines of the 
Oechelhauser and Koerting 2-cycle type, and the double-acting 
Otto-cycle engines of the Deutz, Cockerill, Niirnberg, types. 
Amongst the numerous systems of gas-engines designed for the 
utilization of the waste gases of blast-furnaces and producer- gas may 
be mentioned the remarkable 2-cycle engines designed by Herr von 
Oechelhauser. 



June 1905. 



LARGE GAS-ENGINES. 



637 



Fig. 13. 
Synoptic Table of Compositions and Calorific Values of Gas, 



1-5^3 25-35 



50-55 




Coke-oven Gas. 
1. 3,700 Cal. per . ms = 414 
:•• ■■ -"^ B.Th.U. per cub. ft. 



c H 



1-3-5 3-16 25-30 



CO -N 



% 




CH CO H 

1-2-5 16-28 10-20 






CO *N 



Mond Producer-Gas. 
,250 Cal. per ms = 140 
B.Th.U. per cub. ft. 



% 







AnthraciteProducer-Gas. 
1,200 Cal. per m3 = 134 



CH CO 



CO'*N 



•••• r'--" -r'- ':\ B.Th.U. per cub. ft 



H 


20-28 


3 






% 


m 


p 


V »* V 






n 




1 


1 





Blast-furnace Gas. 
4. 900 Cal. per ms = 101 
B.Th.U. per cub. ft. 



CH CO 



CO 'N 



OecJielhduser Engine, — This system was one of the first to be 
applied to high-power engines. It was put into practice in the early 
part of 1898 in the shape of a 600-H.P. engine, and on that 
occasion disclosed its excellent qualities which have continued 
throughout a period of seven years' work. The main feature of the 
system consists in the employment of two trunk pistons working in a 
single cylinder, which is somewhat similar to the device already 
applied for several years by Robson and made by Scott Brothers of 
Halifax (England). This engine was of the Otto type with 
distribution by valves, whilst the von Oechelhauser engine is of the 
2-cycle type, and the distribution is effected without the intervention 
of valves applied to the cylinder. This feature is the more 
interesting since, in large engines especially, the valves constitute 
delicate organs of difficult upkeep, having regard to the high 
temperatures to which they are subject in the explosion chambers. 



638 



LARGE GAS-ENGINES. 



June 1905. 



The diagram Fig. 14 shows the mode of working the two single- 
acting pistons. The front one is attached by a connecting-rod to 
the centre crank of a triple crank-shaft and works by thrust, at the 
time of the working explosion, whilst the back piston is attached by 
a system of swing-bar and counter-rods to the side cranks and works 
by pulling at the time of the explosion. The first advantage of this 
arrangement is the attainment of perfect balancing of the working 
parts, and particularly of the crank-shaft, with regard to which the 
driving efforts neutralise the reactions at the bearings. Again, the 
explosion-chamber is formed by the cylinder itself and the ends of 
the pistons when the latter approach each other at the time of 

Fig. 14. 
Diagram of the Engine. 




the ignition of the gaseous mixture. This explosion-chamber thus 
forms a chamber presenting the minimum of surface cooled by water 
circulation, and free from passages and recesses which on the ignition 
of the explosive mixture impede the movement of the pistons. These 
devices should have a manifest influence on the thermal output of the 
Oechelhauser engine. The distribution is secured by the pistons 
themselves forming slide-valves which uncover, and in their movement 
again cover the ports provided in the cylinder for the admission of 
air and gas and the discharge of the products of combustion. 

Fig. 14 shows a diagrammatic section of the engine through the 
cylinder and the feed-pump. The openings in the sides of the 
cylinder serve for e the exhaust, a the admission of air, g 



June 1905. 



LARGE GAS-ENGINES. 



639 



the admission of gas. Before the pistons attain their dead centre or 
outer end of stroke, the front piston uncovers the exhaust port e, 
then the back piston uncovers the air-inlet ports a, and the air, 
thus entering under pressure, sweeps the cylinder before it has 
uncovered the ports g admitting the gas, which then mixes with the 
air to form the mixture. The pistons then return inwards, and, as soon 
as they have again covered the openings which have been described, 
compress the charge in order to commence the operation anew. This 
therefore comprises two principal strokes — compression and expansion. 
The other operations, which consist of admission and exhaust, and 
correspond to one revolution in the Otto-cycle engines, only 
occur in this case during a fraction of the cycle, owing to the fact 
that the gas and the air are introduced under pressure. 





Lbs. 


Fig. 15. 


per □' 




- 355 


Av^. Tress. - ^1 Ihs.feru" / 




121 lUvs. / 




B.H.P. 605 / 





-170 



ATM. 



The double-acting pump p, whose piston is fitted on the 
extension of the back piston of the engine, supplies it with air 
and gas respectively through the passages a^ and ^^, terminating at 
the front face and at the back face of the pump piston c. The 
distribution is regulated in such a way that the passages d^ and g^ 
always contain air and gas under a pressure of 5 to 6 lbs. for filling 
the working volume of the cylinder, which corresponds to about 
• 70 of the total volume of the pump. The indicator diagram thus 
takes the form of Fig. 15, showing the exhaust, the air-scavenging 
and the gas-inlet from a to h. The regulation is efiected by 
variation of the quantity of mixture. The compression is variable, 
but in order not to produce a vacuum the entire charge is introduced 



640 



LARGE GAS-ENGINES. 



June 1905. 



in the cylinder. The governor then controls the opening of a return 
valve which enables a part of the mixture admitted to be returned 
to and stored in the piping. 

Fig. 16 shows a section across the distribution ; Fig. 17, Plate 24, 
a longitudinal section, and Fig. 18, Plate 25, a plan. The two latter 
represent an engine with blowing apparatus (on the right) for blast- 



FiG. 16. 
Section across the Distribution. 




furnace. The gas- and air-pump is shown in a dotted line under 
the surface, on the horizontal view. The Oechelhauser engines 
have been made since 1899 by the Deutsche Kraftgas Gesellschaft. 
Several important firms are making them under licence or for their 
own account after having acquired the patents. Thus some sixty of 
these engines, representing a total power of 50,300 H.P., have found 
their way into difi'erent countries and particularly for use in 



June 1905. LARGE GAS-ENGINES. 641 

metallurgical establisbmonts. Fig. 19, Plate 26, shows an installation 
of an Oechelhauser twin-cylinder engine combined with a fly-wheel 
dynamo. Fig. 20 shows an engine driving direct a blast-furnace 
blowing cylinder. 

Koerting Engines. — Messrs. Koerting Brothers, of Hanover, have 
made gas-engines since 1881 and producers since 1889. Up to 1896 
this firm had turned out about 3,500 engines, representing 15,000 
H.P. Since that time the Koerting Works, now carried on under 
the name of Koerting Brothers, Limited, has produced 7,200 new 
engines. In recent years in particular 50,000 H.P. in two-cycle 
engines of their special type have been supplied. Side by side with 
the construction of the two-cycle engines the construction of the four- 
cycle type has developed, and the sale has attained 100,000 H.P., 
whilst at the present time engines are under construction representing 
several thousands of horse power. The Otto-cycle engines are all 
provided with the Koerting valve-gear previously described with the 
special mixing-valve regulating the simultaneous passage of air and 
gas, and contributing to their homogeneous mixture before admission 
into the cylinder. This valve is automatic and is lifted by the 
suction action of the piston only, falling back by its own weight as 
soon as the suction ceases, that is, as soon as the inlet-valve closes 
through the compression. 

The Koerting Co. makes Otto-cycle engines of the single-acting 
type in about 15 sizes, from 2 to 175 H.P., the latter type having 
a diameter of 650 mm. (25*5 inches) with a piston-stroke of 
955 mm. (37*6 inches) and performing 135 revolutions. The 
60 to 175 H.P. engines are provided with the cooling device 
shown in Fig. 10 (page 631). The external arrangement is seen in 
Fig. 21, Plate 27, which illustrates the 125-H.P. Otto-cycle engine. 
The Koerting Company makes on the same system, but with two 
twin cylinders, a 350-H.P. engine, being 650 mm. (25*6 inches) 
in diameter, by 955 mm. (37*6 inches) stroke, running at 
160 revolutions per minute. It will be seen that these dimensions 
are amply large, inasmuch as they effect the work specified 
with a mean pressure of 4: * 6 to 4*7 kg. per cm.^ (65 • 4 to 

2 T 



642 



LARGE GAS-ENGINES. 



June 1905. 



66 '8 lbs. per square inch) on the piston, even if a mechanical 
efficiency of 80 per cent, only is considered. As a matter of fact 
single-acting Koerting engines usually exceed that output. Above 
250 H.P. the Koerting Co. frequently applies its double-acting 2-cycle 
type, Fig. 22, which is used both with producer-gas, with blast- 
furnace and with coke-oven gas. 



Fig. 22. 
Double-acting 2-cycle Engine. 




.JM* 



These types being doubled enable twice the power to be obtained, 
and if necessary an engine of 3,000 H.P. can be made with two cylinders 
each of 1,500 H.P., with a common shaft and one fly-wheel only. In 
this engine the work is ejfifected in the same manner on the two sides 
of the piston, which is supported by rods working in stuffing-boxes. 
The double-acting 2-cycle engine is represented in elevation and in 
horizontal and vertical sections in Figs. 23 and 24 (pages 643 and 
644). its action is as follows : — 

(1.) The ignition of the mixture and the development of the motive 
pressure take place after the introduction of the charge and its 
compression quite close to the back dead-centre of the piston. 



Junk lOOf) 



LARGE GAS-ENGINES. 



643 



(2.) The expansion of the ignited mixture and the transmission 
of the power to the crank-shaft take place during the forward motion 
of the piston. 

(3.) When the piston has reached its front dead-centre, the 
expulsion of the products of combustion and the admission of the 
new mixture take place. 

Fig. 23. 
Double-acting 2-eycle Engine. 




A. Cylinder. 

B. Working Piston. 

C. Air Pump. 

D. Ga,s Pump. 

E. Cam Shaft. 

H. Pumps Connecting-rod. 



a a'. Air Passages, 
c c'. Gas Passages. 

e. Exhaust. 

i. Exhaust Porta. 
m. Explosion Chambers. 

o. Water Jacket. 



(4.) During the backward stroke of the piston the latter 
produces the compression of the explosive mixture. 

This work, which is accomplished in the Otto-cycle single-acting 
engines during four strokes of the piston, is effected in the Koerting 
engines in two strokes, the expulsion of the residuum of the 
combustion, and the intake of the new charge being effected during 

2 Y 2 



644 



LARGE GAS-ENGINES. 



June 1905. 



the very short time in which the piston is in the vicinity of the front 
dead-centre. The working piston B does not effect the suction of 
the mixture ; the latter is delivered to the cylinder A by pumps D 
and'C, and the new charge itself expels the products of combustion 
which it afterwards replaces. 

The introduction of the new mixture therefore takes place in the 
following way : — 

(1.) It distributes itself at once over ^the whole section of the 
cylinder, and thus drives out uniformly the old mixture. 

(2.) A stratum of neutral gas interposes itself between the new 
mixture and the hot residuum of the preceding explosion, preventing 



Fig. 24. 
Double-acting 2-cycle Engine. 





the contact of the two mixtures and premature ignition of the new 
charge. 

The exhaust ports i for the products of combustion are placed 
in the centre of the cylinder and form annular orifices ; the piston, 
of a length equal to its stroke, less the width of the exhaust 
openings (-xV^^ ^^ *^® stroke), itself closes these openings alternately 
to the back and front ends of the cylinder. The exhaust of the 
products of combustion and the introduction of the new charge 
expelling the last traces of gas of the previous explosion, take place 
in a time equal to that of the opening and the closing of the ports 
by the piston. The inlet- valves, which are of the flat type, closing by 
a spring, are operated by cams and are situated one at each end of 



June 1905. 



LARGE GAS-ENGINES. 



645 



the cylinder. Gas and air are supplied by separate pumps D and C ; 
the mixture takes place at the entrance to the cylinder only. These 
pumps in the first instance deliver pure air to the cylinder and then 
a mixture of uniform composition, the quantity of which alone 
varies to regulate the power of the engine. The pumps introduce 
the mixture at a pressure of about 5 lbs. Their pistons are 
therefore at the end of the stroke as soon as the engine-piston has 
again covered the exhaust openings, and the compression of the 
mixture commences. The pistons of the two pumps are on the same 
rod, they have the same movement, and the composition of the 
mixture depends on the relation between the two pistons only. 

Fig. 25. 
Admission- Valve admitting Air before Mixture., 2,-cycle Engine. 






»a.5..-' 




Air 



m 



The air passages aa' and gas passages cc' terminate at the inlet- 
valve. It follows that, on the opening of this valve, the fluid which 
is nearest it will enter first into the engine-cylinder. If then, by a 
device of some kind, air can be passed into the gas passage driving 
out the gas, air will then be on both sides of the valve, as shown in 
Fig. 25 ; it is certain that on the opening of the valve it will first 
cause pure air to pass into the cjlinder until gas has reached the 
valve. Taking into account this material point, it is easily understood 
how the pumps act in order to introduce the charge and regulate it, 
Fig. 26 (page 646). As has already been said they draw and deliver 
at the same time. What has to be done is first to introduce pure air 
and then an intimate mixture always of equal composition ; the gas 
pump should work without delivering gas for a certain time during 



646 



LARGE GAS-ENGINES. 



June 1905. 



which the air-pump is at work. The gas-pump afterwards supplies the 
gas, so that a mixture of the desired composition reaches the engine 
in the place of pure air. The air-pump acts as a steam-engine with 
full admission. In the gas-pump the closing of the suction takes place 
after the piston has performed 40 to 50 per cent, of its stroke only. 
During this period the delivery remains closed, and the gas, drawn 
in during the preceding stroke, re-enters the suction pipe. 

Fig. 26. 
Pumps regulating Inlet-Valve. 




During the second period of the piston stroke, the pump delivers 
the gas, and from that time the two pumps deliver together gas and 
air with the same speed per unit of time, so that the composition of 
the mixture remains the same. The inlet-valve of the engine is not 
open at the commencement of the delivery stroke of the piston of the 
air-pump ; it opens only after the pump-piston has performed about 
half its stroke. During this time air accumulates in the passage and, 
in consequence of its pressure, passes into the gas passage forcing 
back the gas, as the gas-pump is not at this time delivering. Pure 
air only therefore enters into the cylinder ou the opening of the 



Junk 1905. LARGE GAS-ENGINES. 647 

inlet-valve until the gas pump begins to deliver. In this way the 
cylinder is scavenged, and an insulating stratum formed before 
introducing the explosive mixture, which is always of a uniform 
composition. 

The regulation of the engine is obtained by two means : — 

(1) By retarding the delivery of the gas-pump. This retardation 
is obtained by operating the slide-valve of the pump by a link 
operated by two different eccentrics. The link is shifted by the 
action of the governor which raises and lowers it. This mode of 
regulation requires the use of a very heavy governor. 

(2) By establishing a communication between the delivery 
and suction of the gas-pump. This passage is provided with 
a throttle-valve (butterfly) actuated by the governor. When this 
valve is partly open, a varying quantity of gas repasses, 
during the suction period and during the first period of the 
return of the piston, from the delivery passage to the suction 
chamber. The gas column recedes into the passage near the inlet- 
valve, and is replaced by air. On opening the inlet-valve a 
quantity so much richer in air and weaker in gas enters the 
cylinder, according to the opening of the valve. The regulation is 
effected, as in the first case, by means of a link. The ignition of the 
mixture is effected electrically by means of two magnetos. For 
starting the engine, compressed air is used. 

It is the Koerting Co. which has so far fixed the most 
important installations, which necessarily consist of double-acting 
engines. Mention may be made of the power-station of 5,500 H.P. 
of the Gutehoffnungshiitte establishments at Oberhausen (Kheinland) 
equipped with seven engines, four of which are 1,000 H.P. and 
three of 500 H.P. Fig. 27, Plate 27, shows two of these 500-H.P. 
engines. The Koerting engines were the first explosion-engines 
used in America for the utilization of blast-furnace gas, which 
innovation was the more remarkable as it involved, at the outset, 
a total of 42,000 H.P. It was carried out by the De la Vergne 
Refrigerating Machine Company of New York for the Lackawanna 
Iron and Steel Company of Buffalo (N.Y.), and consists of ten 
coupled dynamos and two-cylinder engines of 1,000 H.P. each, for 



648 LARGE GAS-ENGINES. June 1905. 

electrical service, Fig. 28 ; and sixteen engines of tlie same power 
which drive from their crank-shaft blowing engines for the blast- 
furnaces, Fig. 29, Plate 28. This installation constitutes up to the 
present the most important power-house in the world. 

Several important installations have also been made for the 
utilization of lignite, which is very abundant in many parts of 
Europe and America. This lignite is treated by special producers 
supplying gas under pressure. Fig. 30 is a view of two 2-cycle double- 
acting engines of 300 H.P. fed with gas produced from lignite. 
Fig. 31 represents an electric station installed at Julienhutte in 
Germany, and driven by coke-oven gas. It consists of three 
Otto-cycle engines, single acting, of which three of 300 H.P. are 
of the twin type with fly-wheel dynamo, and one is a single 
cylinder of 125 H.P. with dynamo on the crank-shaft. The latest 
important installations of Koerting's 2-cycle engines have been 
undertaken by the parent company of Hanover, and consist of 104 
engines developing in all 89,125 H.P. 

Cocherill Engines. — Mention has been made of the successful 
trials of the Cockerill Co. which, as far back as 1894, made its first 
experiments on a small engine driven by gas from its blast-furnaces. 
Four years later this firm constructed from the details furnished by 
the late Professor Delamare-Deboutteville the 200 H.P. engine which 
at the present time is still driving an alternating-current dynamo 
for the production of motive power. Then came the famous engine 
of 600 H.P. single acting, which drives a blowing apparatus 
for the blast-furnaces delivering 500 m.^ (17,600 cubic ft.) of air at a 
pressure of 40 cm. mercury (15I ins.). This engine, started in 1900, 
formed the subject of memorable experiments undertaken under the 
direction of Prof. Hubert, with the collaboration of some experts, such 
as Professor Aime Witz and the late Mr. Bryan Donkin. In order to 
increase the power of the engines without adding to the dimensions 
already involved by a piston of 1*300 m. (4 ft. 3y\ ins.) diameter x 
1*400 m. (4 ft. 7 J ins.) stroke, the technical department of the 
Cockerill Co. then designed an engine with tandem cylinders which 
enabled the power to be doubled whilst increasing the regularity. 



JuxK 1905. LARGE GAS-ENGINES. 649 

The Cockerill Co. was one of the first to take up the construction 
of double-acting engines, and at the end of 1901 built its first single- 
cylinder Otto-cycle engine of 1,200 H.P. for the blast-furnaces of their 
own works. The construction of double-acting engines has become 
common, and they have replaced the single-acting engines, which 
are now only constructed in exceptional cases. Before arriving at 
their present type the Cockerill Co. created quite a series of engines, 
all of which are remarkable for a manifest tendency to depart from 
common principles in order to attain original forms and devices of 
real interest. As makers of machinery for the iron industry the 
Cockerill Company is, in fact, in a better position than most people 
to appreciate the many requirements to be satisfied by engines 
suitable for this industry. 

Fig. 32, Plate 29, represents one of their latest types of single- 
cylinder double-acting engines driving a " Southwark " blowing 
engine. The principal dimensions are : — 

Diameter of the engine piston •. . . 1*300 m. (4 ft. 3^^^ ins.) 
Diameter of the blower piston . . . 2*250 m. (7 ft. 4 J ins.) 
Stroke of the pistons l'400m. (4 ft. 7^ ins.) 

This blowing engine has to furnish air at a pressure of 60 cm. 
(23*6 ins.) mercury, and develop 1,200 H.P. at 80 revolutions per 
minute. Fig. 33 is a photograph of the double-acting tandem 
engine of 500 H.P., the particulars of which are: diameter of 
piston D = 0-600m. (23-6 ins..) X stroke C = 0*800 m. (31*5 ins.) 
at 130 revolutions. The engines present a characteristic aspect, 
and careful investigation of the tensile strains and other stresses 
to which the different parts are subject have given rise to 
the creation of special forms peculiar to the Cockerill engine. 
The strength and simplicity of the engine have thereby been 
improved, but the neat appearance to which one is accustomed has 
to some extent been lost. Still, commercial considerations are 
the most important. The frame consists of two strong cast-iron 
girders which extend the whole length of the engine, between 
which the cylinders are placed, and which are terminated by 
the cross-head guides and the bearings of the crank- shaft, 



650 LARGE GAS-ENGINES. June 1906. 

Plate 29. The cylinders being fixed bj bolts and keys, are 
independent of the frame, and are cast with their jacket and 
connections for the inlet- valves at the top and exhaust-valves at the 
bottom. 

The piston consists of two parts bolted and secured to the rod by 
collars, thus dispensing with screwing, which is unsuitable for 
withstanding great strains. As in the case of all double-acting 
engines, the piston is cooled by a circulation of water under pressure 
passing through the rods. Apart from the cylinder-jacket, the ends 
are provided with a water circulation, as also the valve-chests, 
the brasses of the crank-shaft, and the crosshead guides. The 
regulation is effected in two ways, according to the type and the 
application of the engine. For high-speed engines, where the masses in 
movement are considerable and for which great regularity is required, 
a high compression is necessary to assist in giving uniformity. In 
this case the regulation is effected by varying the composition of the 
mixture and by keeping the air-supply constant. The total volume 
being invariable the compression likewise remains constant. The 
valve-gear is shown on Fig. 34, Plate 29, in a cross-section of the 
cylinder. 

In the case of engines working with less speed and whose 
variable revolutions, as in the case of those driving blowing 
engines, contribute to the balancing of the working parts of the 
engine, regulation by variable admission of variable quantity 
of charges of uniform composition is used. As a consequence 
the compression varies. In both cases the valves are operated by 
cams and springs. The inlet-valves are returned to their seats by a 
ratchet movement with air-pump. Fig 34. The types constructed 
by the Cockerill Co. are divided into — 

(1) Single-cylinder engines. 

(2) Engines with two twin-cylinders. 

(3) Engines with two cylinders tandem. 

(4) Engines with four cylinders twin-tandem. 

These types can be seen from Fig. 12 (page 635) which shows the 
diagram of the arrangement. The two kinds of regulation are 
applicable, according to each case, to the different types enumerated. 



June 1905. LARGE GAS-ENGINES. 651 

The double-acting single-cylinder engines are made up to 1,500 H.P., 
but their coefficient of cyclic irregularity with standard fly-wheels 
is fairly high, so that this type of engine is rarely used except for 
driving blowing apparatus or other work not requiring absolute 
regularity. With their engine of the 4-cylinder twin-tandem type, 
double acting, the Cockerill Co. can obtain a prime mover of 5,000 
to 6,000 H.P. 

The total number of engines constructed up to the present by the 
Cockerill Co. and their concessionaires has reached 148, representing 
in all 102,925 H.P., or an average power per engine of about 
695 H.P. The purposes for which they are employed are 
approximately the following : — 

Electrical ...... 45 per cent. 

Blast-furnace blowers . . . 52 ,, ,, 

Boiling mill . . . . . 2 ,, ,, 

Various . . . . , . 1 ,, ,, 

Otto-Deutz Engines. — The brilliant record of the Deutz Gas-Engine 
Works is well known. Continuing the traditions of its illustrious 
founders, Messrs. Otto and Langen, the Otto-Deutz. Co. has for 
40 years constructed the four-cycle engines, bearing the name of 
the inventor. Since its formation this Company and its licensees 
have built 75,000 engines. It necessarily commenced with the small 
single-cylinder horizontal and vertical engines, but since 1895 it 
has been engaged in the construction of high power blast-furnace 
single-acting engines attaining 125 H.P., 500 H.P., 1,000 H.P. and 
upwards, with one, two and four cylinders, twin or double-twin, 
Plate 30. A large number of these engines were in a short time 
working in industrial establishments with generator gas, blast- 
furnace gas or coke-oven gas, whilst others took the places of 
steam-engines in water-works, electric stations, and so on. 

Fig. 53, Plate 11, Proceedings 1901, illustrates one of the 1,000 
to 1,200 H.P. engines with four cylinders twin and opposite, installed 
at the Horder Bergwerk Works at Horde, Germany. This engine is 
combined with a fly-wheel dynamo, and works with blast-furnace 
gas. 



652 LARGE GAS-ENGINES. June 1905. 

The 1,200 H.P. 4-cylinder engine, exhibited at Diisseldorf in 
1902, won the admiration of experts as being a type of the most 
powerful engine of the time in Germany. It was constructed for the 
" Gutehoffnungshiitte " for driving a blast-furnace blower to furnish 
1,000 m.^ (35,316*6 cubic feet) of air per hour at a pressure of 
0'5 atm. (7*4 lbs.). The total weight of this engine was 219 tons, 
the fly-wheel accounting for 19 tons. This weight is manifestly 
stupendous compared with the weight of steam-engines with which 
poor-gas engines have to compete. Those large engines, constructed 
according to the single-acting type, thus attained the excessive 
weight of 180 kgs. (396*8 lbs.) per horse-power. Their price was 
necessarily high and the space occupied excessive. It was therefore 
expedient to find a solution of this difficulty, in order not to impede 
the development of large engines. Such a solution is found in the 
double-acting engine which is now the current type of the Deutz 
Works. Reference will be made to this engine later, and in the 
meantime the author will examine the characteristics of the small 
engines now made by this firm. 

The single-acting type has been retained, and for about two years 
the makers have been working for the unification of the systems of 
valve-gear. The system shown on Fig. 6 (page 627) is now applied 
to all sizes, except for engines of 120 to 150 H.P. per cylinder, in 
which the mixing-valve and the gas-valve are fitted separately, with 
the object of facilitating access to the latter for the purpose of 
cleaning. This device is again seen in the large double-acting 
engines described later. In order to appreciate the simplicity of the 
parts and arrangement of the new valve-gear for admission in variable 
quantity of a charge of uniform composition, compared with the 
old system of variable mixture by conical cam, it is sufficient to 
refer to the new Otto type made not only in small sizes but up to 
250 H.P. For 200 to 300 H.P. the old type of engine was constructed 
with twin cylinders, that is, with two cylinders side by side or 
opposite, that is, with two cylinders one facing the other. The latter 
arrangement has, however, been abandoned by the principal European 
firms, experience having shown that it is subject to considerable 
wear and tear. In fact, one of the pistons, in consequence 



June 1905. LARGE GAS-ENGINES. 653 

of the reverse working, is subject to a reaction from the bottom to the 
top at the time of the explosion, which tends to lift the piston and 
produce shocks, if there be the least play due to wear and tear. For 
engines of 400 to 1,000 H.P., the arrangement of double-twin 
opposite cylinders was adopted. 

With regard to the form of the parts exposed to heat and the 
proper utilization of the explosive mixtures, the Otto-Deutz Co. has 
benefited from the long practice acquired in the construction of the 
different forms of cylinder-heads for its single-acting engines, in 
designing the double-acting type, for which they have adopted the 
general arrangement and external form preferred for good modern 
steam-engines. Since 1901, the Otto-Deutz Co. has engaged in the 
making of these double-acting engines and maintained its preference 
for the 4-cycle type, which is, moreover, the type most generally 
adopted nowadays by its competitors. 

In 1902 this Company started its first double-acting engine of 
200 H.P. at its own electric station, where it first did duty as an 
experimental engine and afterwards in the works. In the meantime 
46 engines on this system, representing in all 31,500 H.P., had been 
constructed by the Otto-Deutz Co. at its Deutz Works and its 
concessionaires. Plate 23 represents a 200-H.P. engine and Fig. 38, 
Plate 31, a 300-H.P. engine of the latest double-acting type, of which 
the principal dimensions are : diameter of piston 620 mm. (24 '4 ins.) ; 
stroke of piston 780 mm. (30*7 ins.), at 150 revolutions per minute. 
The cylinder is constructed so as to allow for the independent 
expansion of the jacket and the liner. The latter is cast in one 
piece with the ends of the jacket, whilst the annular middle portion is 
connected and made tight by a flexible joint. The ease with which 
this part can be dismantled facilitates the thorough cleaning of the 
jacket. In the case of engines of the largest sizes the cylinder 
is even made in three parts, the two ends of which carry the valve- 
boxes. The latter are in each case fitted in the same vertical axis, 
the inlet-valve at the top and the exhaust-valve at the bottom, 
and in such a way that the surplus oil or moisture is driven 
out. The lower part is provided with water circulation, and 
fixed by bolts, with springs which enable them to move slightly 



654 



LARGE GAS-KNGINES. 



June 1905. 






o 



05 hn 
CO "^ 



. 


o 


o 


<-> 






^ 






H 




o 




rO 



CJ5 

2 




June 1905. 



LARGE GAS-ENGINES. 



655 



» 



on expansion. The piston is supported by the piston-rods, without 
its weight resting on the bottom of the cylinder. The cooling-water 
inlet and outlet is effected through the rods being connected to a 
pump by articulated pipes. 

The channel made in the piston-rods contains, throughout its 
length, a tube having an inside partition opposite the piston-head. 
The water is supplied to the piston along the passage concentric 
to the tube and leaves by the tube itself. The stuffing-box, being one of 
the delicate parts of the double-acting engine, has been the subject 
of special care. It will be seen from Fig. 39 (page 654) that it consists 
Fig. 40. Fig. 41. 

Cylinder and Valve-Gear. Exhaust-Valve with Water Circulation. 

CyUnder 





of a movable casing fixed to the cylinder by an external flange. This 
casing is provided with a longitudinal oil- way. At its extremity, near 
the cylinder, is a chamber containing a series of collars carefully 
adjusted, in the recesses of which the rings, of a special metal, 
which bear on the rod are placed. In a second chamber there ie 
a series of babbit metal collars of triangular section, forming a 
wedge when tightened up, an operation which is effected by means of 
the outer flanged sleeve forming a gland. 

Fig. 40 is a section across the cylinder showing the valve-gear. The 
regulation is effected by acting on the quantity of charge admitted to 



656 LARGE GAS-ENGINES. June 1905. 

the cylinder, as in the case of the single-acting engines, but with the 
difference that the charge is regulated by an auxiliary double-seat 
valve 8 whose lift is determined by the governor. This valve receives 
its movement from the lever operating the main inlet-valve p which 
has a fixed lift. This lever has, as a fulcrum, a rolling path ; it is 
jointed on the one hand with the stem of the inlet-valve and on the 
other hand to the oblique connecting-rod h, receiving its movement 
from a cam keyed on the side-shaft I of the engine. The valve is 
depressed, that is, opened by the operation which has just been 
described, and returned to its seat by the spring attached to its stem. 
With regard to the regulating valve, the levers actuating 
it have a variable fulcrum, formed by roller a which is displaced 
to vary the lift of the valve. This roller, which is movable, is 
fitted at the end of a rod attached to the governor. In order to 
regulate in advance the composition of the mixture, the quantity of 
which admitted to the cylinder is determined by the governor, a cock 
has been fitted on the gas inlet-pipe and a throttle on the air-pipe. 
Both these are operated by hand. The exhaust-valve is also operated 
by a connecting-rod worked off the side shaft by a cam and a lever 
with rolling path. It is provided with a water circulation which is 
operated through its stem as in the case of the piston-rod. A water 
circulation is also set up in its box, which carries the seat of the 
exhaust- valve. In the largest engines, such as the 2,000 H.P., Fig. 42, 
Plate 31, with two tandem cylinders, with a diameter of piston of 
1,100 mm. (43*2 inches) and 1,300 mm. (51*1 inches) stroke, the 
exhaust- valve has a seat of D = 380 mm. (14*9 inches) which, on 
the exhaust, would have to overcome a pressure of about 2,000 kgs. 
(4,409 lbs.) when lifted. It has been constructed with two seats, 
Fig. 41 (page 655), with a communicating passage which establishes 
the same pressure underneath as above the disc of the valve. The 
disc is therefore balanced and only offers to the lift the resistance 
of the reaction springs. 

The ignition of the charge is effected by magneto-electric 
apparatus, the starting is by means of compressed air and the 
lubrication is under pressure, as is also the water circulation for the 
piston, the stuffing boxes, &c. 



June 1905. LARGE GAS-ENGINES. 657 

The number of double-acting engines which the Deutz Works 
has supplied or has at present in course of execution amounts to 
sixty-four, ranging in size from 200 to 2,000 H.P. with an aggregate 
of 34,660 H.P. 

Niirnherg Engines, — The " Vereinigte Maschinenfabrik Augsburg 
and Maschinenbaugesellschaft Niirnberg A.G. " has always devoted 
itself to the construction of high-power engines and won a well- 
earned reputation. The experience and practice acquired in this 
construction, as also the use of powerful and improved tools, 
placed this company in the best position for constructing large gas- 
engines. This company quickly discovered that the use of the single- 
acting engine, the dimensions of which were being increased and the 
cylinders multiplied in order to obtain a sufficiently powerful engine 
for modern requirements, was only transitive, the space occupied, the 
enormous weight and low efficiency of this type of engines being 
defects which caused it to be abandoned in favour of the 
double-acting type. Still, for small powers, up to 160 H.P., the 
Niirnberg Company makes the single-acting engine and even of 
double this power, if necessary, by using twin-cylinders side by 
side. This firm, moreover, commenced to construct these in 
1889. The photograph. Fig. 43, Plato 31, refers to a single-acting 
engine of 70 H.P. From the very simple general appearance and 
solidity of the whole, the features of the German construction will be 
recognised. 

Plate 22 shows the way in which the piston and the exhaust-valve 
are cooled in large single-acting engines of the Niirnberg type ; 
its present construction differs slightly, as regards the valve- 
gear, from what is shown in this section. The circular slide-valve 
which regulates the air and gas-inlet is, in the new valve-gear, 
fixed to the stem of the inlet-valve, as in Fig. 44 (page 658), which is a 
diagrammatic representation of this arrangement. In Plate 22 it is 
seen that this slide-valve is independent. Its motion is regulated 
by the governor, whilst in the present arrangement, the governor acts 
on the stroke of the inlet- valve, increasing or reducing it as shown in 
Fig. 45. The variation of the travel of the inlet^valve is effected 

2 z 



658 



LARGE GAS-ENGINES. 



June 1905. 



Fig. 44. 
Circular Slide-Valve fixed to Stem 
of Inlet- Valve. 



Fig. 45. 
Governor acting on Stroke of Inlet-Valve. 




Fig. 46. 
Piston. 




T^TTTPr. 



722^ — ^^ 



Ezzzzzzzzzzzza 



Jrus. IZ 

Lull I III 1 1.1.1., 



4- Feet 
J 



June 1905. LARGE GAS-ENGINES. 659 

by the displacement of the fulcrum of the operating lever L. 
This fulcrum is formed by a roller r, attached to the governor by- 
means of a rod T. In order to prevent the accumulation of surplus 
oil and dirt, the seat of the exhaust-valve has been slightly depressed, 
under the bottom level of the cylinder. An original device consists 
also in the mode of attaching the connecting-rod to the piston. It 
is, of course, well known that the piston pin is placed at the two 
ends in bosses cast with the piston sides, where they are then kept 
by screws ; the bearing at the end of the connecting-rod is fitted 
between these bosses. The Niirnberg construction is different. It 
consists, according to Fig. 46, in holding the two ends of the pin in 
a fork at the end of the connecting-rod bottom end, and in causing 
it to turn, at the centre, in a bearing carried by a support cast in 
the vertical axis of the piston. In short, although the single-acting 
engine of the Niirnberg Company has formed the object of several 
substantial improvements, it is in the construction of double-acting 
engines that this company especially excels. 

Fig. 47 (page 660) shows this mechanism for the admission of the 
double-acting engines of the Niirnberger Maschinenbau. T is the 
eccentric rod receiving its movement from the side shaft, the end 
of which is maintained by the connecting-rod B ; at the same time it 
is connected to the ratchet-piece D. The latter engages the end of 
the lever L, which is connected also to the valve stem. The lever 
L rests on the lever Z as a point forming a rolling path, and it is 
also jointed at the valve stem, whilst its free end can move downward 
owing to the diplacement of the pivot p controlled by the governor 
by means of the crank lever Z. A is an air piston which deadens 
the fall of the valve on its seat when the levers D and L disengage. 
It will be seen that, according to the position which the governor 
causes the pivot j? to assume, the fulcrum advances or recedes, and 
the valve opens more or less late relatively to the stroke of the 
piston. The opening of the air inlet-valve being constant and that 
of the gas being variable from the point of view of the time when it 
takes place, air only is first admitted which follows the piston, 
and more or less gas is afterwards admitted in proportion to the 
work to be developed. The device therefore effects the admission of 

2 z 2 



660 



LARGE GAS-ENGINES. 



June 1905. 



a mixture of variable composition but with constant compression. 
These engines are made, according to the power required, with 
one, two or four cylinders. The single-cylinder engine is made 
up to 1,500 H.P., the two-cylinder up to 2,800 H.P. and the 
four-cylinder up to 6,900 H.P., forming in the last case two 
tandem engines arranged side by side and driving the same crank- 
shaft. 

Fig. 47. 
Automatic Cut-off Valve-Gear. 




Fig. 48 shows the succession of the strokes of the twin-tandem 
arrangement in two revolutions of the fly-wheel. 

Guided by experience, the Niirnberg Company advises the valves 
and cylinders to be dismantled and examined after some months, 
because, despite the perfection which the processes of washing 
and purifying producer or blast-furnace gas have attained, 
accumulations of impurities are still to be feared. This company 
has therefore endeavoured to arrange the parts, so that they lend 
themselves to the easy dismantling of the heaviest pieces and, from 



June 1905. 



LARGE GAS-ENGINES. 



661 



this point of view, the construction is a remarkable one. Fig. 49 
(page 662) shows, in a tandem engine, how the back covers are removed 
to afford access to the corresponding valves. The same operation is 
effected with the front covers of the cylinders, but by placing the 
crank at the end of the forward stroke. Fig. 60 shows the way in 
which the pistons are removed, that is, by disconnecting the 
connecting-rod from the piston-rod and taking out the latter with 
the front cover, whilst the rods themselves are disconnected at the 
centre in order to liberate the back piston. 



Fig. 48. 
Diagram shmcing succession of Strokes. 




POWER Comp. 
POWER Comp. 

Exl-u POWER 
Exh. POWER 

Suet. Exh. 
Suet. Exh. 

Ck)mp. Suct>. 
Slid. 



Slid/. Exh. 
Suet. Exh. 

Comp. Siut. 
Comp. Suet. 

POWER Comp. 
POWER Comp. 

Exh. POWER 
Exh POWER 



» 3^^ 

— *■ V 



vystrou 



ni 



8 POWER STROKES /N 2 REVOLUTIONS 



The longitudinal section of the single-cylinder engine is given 
in Fig. 61, Plate 32, which shows the details of the inlet- valve and 
the gas-regulating valve, balanced with double seat, also exhaust 
valve boxes formed by a vast chamber with water circulation, whilst 
the cross-section. Fig. 62 (page 663), through the gas and air-inlet 
piping shows the mode of attachment of the cylinder to the frame, 
the operation of the valves by eccentrics and rolling paths, also the 
water circulation round the cylinder which is cast in one with its 
jacket. The cooling of the piston is effected by a water circulation 
introduced through the rod by the back crosshead, by means of 
articulated piping which follows its to-and-fro motion. Fig. 53 



662 



LARGE GAS-ENGINES. 



June 1905. 



(page 664). This circulation, after having traversed the piston, 
Fig. 54 (page 664), passes through the rod and the front crosshead. 
The stuffing-box of the piston-rods is shown in a longitudinal 
section in Fig. 55. 

In order to give an idea of the huge dimensions attained by the 
large engines made by the Niirnberg Company, a few of the separate 
parts are reproduced, photographed at the Works : — Fig. 56, Plate 33 : 

Fig. 49. 
Bach Covers removed to afford access to Valves. 




Fig. 50. 
Pistons removed. 




Crank-shaft, diameter 600 mm. x 9-570 m. (23*6 ins. X 31 feet 
4| ins.) ; length, weighing 20,000 kgs. (19-68 tons), for 3,000 H.P. 
twin-tandem engine, diameter of pistons D = 950 mm. (37*4 ins.) 
and stroke 1*200 m. (47 '2 ins.). Fig. 57 : Frame, one solid casting, 
weighing 27,000 kgs. (2 6 J tons). Fig. 58 : One of the cylinders of 
the same engine. These parts are intended for the Eombach Steel 
Works, Lorraine, where they are to complete an installation of 
12,500 H.P. for the electric service and blowers, for which an engine 
of 1,900 H.P. tandem is shown in Fig. 59. 



.Tune 1905. 



LARGE GAS-ENGINES. 



663 



Fig. 60 refers to two engines with twin cylinders of 1,100 H.P. 
each, working at 100 revolutions and driven by blast-furnace gas ; 
they are used for generating continuous current for the electric 
service of the Steel Works at Micheville (France). 

Fig. 52. 
Single-Cylinder Engine. Cross-Section through Gas and Air-Inlet Piping. 

A* 




Figs. 61 and 62, Plate 34, show the front of a frame and the 
crank-shaft of a 1,200 H.P. engine, of 870 mm. diameter of piston X 
1,100 mm. stroke, 107 revolutions. The diameter of the shaft at 



664: 



LARGE GAS-ENGINES. 



June 1905. 



the fly-wheel is 670 mm. (22-45 i^^s.), and its weight 15,000 kgs. 
(14I tons). These parts are intended for engines to complete an 
installation of 3,700 H.P. for the Burbacher Hiitte (Alsace). Fig. 63, 
Plate 34, shows the exhaust valves, with their seats, belonging to an 

Fig. 53. Fkj. 54. 

Piston and BodjOooling Arrangement. 





Fig. 55. 
Piston-rod Stuffing-Box. 




engine of 2,000 H.P., diameter of piston 1,030 mm. (40*5 ins.) 
X stroke 1,300 mm. (51-1 ins.), 90 revolutions. 

Amongst the largest engines may be mentioned those of 
3,600 H.P. twin-tandem, forming part of a group of 9,100 H.P. of 



Junk 1905. LARGE GAS-ENGINES. 665 

the Schalker-Gruben and Huttenverein of Gelsenkirchen, a coupled 
dynamo and engine of wliich is shown in Fig. 64. Mention 
may also be made of the installation of 12,000 H.P. in six units 
executed for the electric station of the Sociedad de Gasificacion 
Industrial of Madrid, which is driven, as also a 350-H.P. engine, by 
fuel gas generated by Duflf producers. From 1903 up to the end of 
August 1905, the Niirnberg Company had supplied and had in 
hand 137 engines of its double-acting type, representing a total of 
157,000 H.P. These engines are driven by blast-furnace gas, coke- 
oven gas and producer-gas and are employed as follows : — For the 
electrical service, continuous or alternating current, 99 engines, 
equivalent to 108,000 H.P. For blowing apparatus, for blast- 
furnaces, steel works, rolling mills and various applications; 
88 engines representing 4=9,000 H.P. 

Ehrhardt and Sehmer Engine. — The firm of Ehrhardt and 
Sehmer of Saarbrucken in Alsace has also won distinction for the 
construction of large engines, and has built, in the short space of 
three years, engines representing a total of 25,000 H.P., on their 
double-acting system, in which they also arrange the cylinders, as the 
case may be, either twin or tandem. Fig. 65, Plate 35, shows in a 
longitudinal and cross-section through the inlet-valve, a cylinder of 
the 700 H.P. tandem engine. The valve-gear is arranged in order 
to obtain a mixture of constant composition, admitted in a variable 
quantity under the action of the governor. For this purpose there 
is one mixture inlet- valve of variable lift and a lantern-shaped 
valve, the variable stroke of which determines the quantity of 
mixture admitted. The cylinder is cast in one piece with its jacket, 
but the latter, owing to the large amount of space, can expand 
freely, having regard to the elasticity of the casting at the places 
where these two parts join. The makers have also endeavoured to 
obtain symmetrical forms, in order to secure equality of tension in all 
the parts. It will be noticed that large man-holes are provided 
everywhere, to give easy access to the jacket. At each end of the 
cylinder there are two independent contact-breakers for the electro- 
magnetic ignition. There will also be found in the Ehrhardt and 



666 LARGE GAS-ENGINES. June 1906. 

Sehmer engine arrangements for the cooling, controlling, etc., wliich 
are common to good types of gas-engines. 

Dingier Engine. — The Dingier engine, Figs. 66 and 67, and 
Plate 36, made by the Dingier Engineering Works of Zweibriicken 
(Palatinate), differs materially, from the point of view of the system, 
from the devices employed by other makers of large engines. 
Instead of obtaining the double action in a closed cylinder by 
exploding the mixture alternately on each face of the piston, in the 
Dingier engine two cylinders, open at one end, are united at 
their explosion chambers. As is shown in Fig. 68, Plate 36, 
each of these cylinders contains a piston, the two pistons being 
connected together by an internal rod. The explosion is therefore 
produced alternately on each inside face of these pistons. The rod 
common to the pistons is provided with rings, and works in a casing 
passing through the division common to the two parts of the cylinder. 
This division, as also the jacket of the double cylinder, is provided 
with a water circulation. The valves are arranged in the division 
piece, as in the single-acting engines. The system avoids the use of 
piston-rod stuffing-boxes. It lends itself to the free expansion of the 
jacket and inside cylinders, which are only solid at one end. 
As the cylinders are open their supervision and upkeep are as easy 
as in the case of an ordinary single-acting engine. The governor 
acts on the admission. Fig. 70, Plate 37. The valve, with a constant 
stroke, always opens at the same place, but closes later or sooner so as 
to effect the admission of variable charges with constant composition of 
mixture. The device. Fig. 71, which effects this movement is somewhat 
complicated. Besides the ordinary side-shaft, which carries the cams 
h, there is a secondary shaft a receiving its movement from the first 
shaft but turning twice as quickly, that is, at the speed of the crank- 
shaft. The shaft governor is fitted on this secondary shaft, also a 
tube to which the governor imparts a slight rotary movement. 
This tube carries a cam d, which is consequently movable. Another 
cam k is fixed on the ordinary shaft h. The simultaneous contact 
of the two cams, through the medium of a set of cranked levers 
controls the opening of the valve, and the latter closes earlier 



H 



June 1905. 



LARGE GAS-ENGINES. 



667 



Fig. 66. 

Elevation. 

(For Longitudinal and Cross-Sections, see Plate 36.) 

(Governor, Plate 37.) 




Fig. 67. 
Plan. 




5 

lu-LjJ 1 \ 1 I \ I L. 



loEeet 






668 



LARGE GAS-ENGINES. 



June 1905. 



or later according to the position of the movable cam, under the 
action of the governor. 

The principal makes of large gas-engines constructed on the 
Continent have now been described, and reference has been made to 
the methods and principles which guide the makers. It would, 

Fig. 72. 
Diagram from 1,000 H.P. Engine. (Oechelhauser.) 

34-8 

Mhs. 

per a" 

Avp. Press. 4-6 -8 lbs. per a" 
12 2 Re/vs. 
Load. 612 HP 




ATM. LINE 



Fig. 73. 
200-B'.P. Double-acting Engine. 180 Bevs. (Otto-Deutz.) 

Diagrams from to Full Load. Tachometer Diagram. 



355 

Lbs. 
per 




therefore, be of interest to analyse concisely a few of the details and 
the indicator diagrams taken. Of the two-cycle engines the diagram. 
Fig. 72, taken on an Oechelhauser engine of 1,000 H.P., is 
reproduced. The engine was working with blast-furnace gas. The 
mean pressure corresponding to the work of 612 H.P. shown by the 
diagram is only 3*3 kg. (47 lbs.). It will be seen that the exhaust, 



June 1905. 



LARGE GAS-ENGINES. 



669 



the air scavenging, and tlie admission of the charge are effected 
in about one-eighth of the total stroke of the two pistons. 

The superposed diagrams and the tachogram, Fig. 73, are taken 
from an Otto-cycle double-acting eiagine of Otto-Deutz make, 
driven by a suction gas-producer. The indicator diagram shows 
the variations of the compression and corresponding explosive 
pressures. The tachometer diagram shows that the sudden alterations 
of load between working light and the maximum work produced, 
over or under, cause variations in speed of 3J per cent. 



Fig. 74. 



Mathot Explosion Bscordfrom a 200-R.P. Double-acting Engine. (Otto-Deutz.) 




No 
vUh ^plosimiErpiosionsadms" 



Period of OsdUaMon 
of the Oove-rrwr 



Normal Rumiuig ronrlrtions 

wiUivut Load 

after 9D seconds 



Fig. 75. 
Diagram from 1,000-H.P. Gas-Engine. 100 Eevs. (Niirnberg.) 



Avg. Press =865 lis per o" Avg. Press. .92-5 U>s.per a" 





Table 1 (pages 672 and 673) gives the interesting details of a test 
made with this engine. It is remarkable for the low consumption of 
coal of 0-704 lb. per B.H.P., especially when it is considered 
that the test took place under regular working conditions, that is, 
taking into account the fuel consumed at night for keeping the 
producer hot during the stoppage of the engine. Fig. 74 is the 
reproduction of explosion record taken at the time of starting; 
it shows the phases of this operation. It will be seen from the 
diagrams, Fig. 75, taken at the front and at the back of one of the 



670 LARGE GAS-ENGINES. June 1905. 

two cylinders of a Niirnberg engine of 1,000 H.P., that mean 
pressures of 6 to 6^ kg. (85 to 93 lbs.) are obtained, even with 
blast-furnace gas, the mean heat value of which is only about 
110 B.Th.U. Such mean pressures are not however usual, and 
it is customary amongst the large firms on the Continent to 
calculate the working dimensions of engines (diameter, stroke of 
piston, and number of revolutions) for mean pressures not greater 
than 70 lbs. per square inch. A wide margin is thus left to 
cover fluctuations in the quality of the gas or temporary defects of 
regulation caused by lack of care or negligence on the part of the 
attendant. 

The mean piston speed has also been increased, and has now 
reached without difficulty 800 to 850 feet per minute, whilst it was 
a few years ago 650 to 700 feet. 

It is advisable, before closing this Paper on the new motive power 
which utilises every kind of fuel gas produced commercially, to 
point out the principal difficulties which remain to be overcome in 
order to enable large engines working with poor gas in general to 
work with safety and facility of control and upkeep ; such attributes 
appear to have remained the special feature of the steam-engine. 
Economically the gas-engine is evidently superior, but low 
consumption is not the only quality which users require from a 
motive power. It is necessary above all that it should be free from 
the risk of sudden stoppages which lead to great expense, as they 
hamper production and may be at times fatal in certain industries 
where breakdowns must be specially avoided ; such is the case with 
electric stations, pumping, hoisting and ventilating work in mines, 
etc. The principal cause of breakdowns in engines fed by gas, 
other than town gas which undergoes a complete purification, 
is the fouling of parts, such as the pipes, the valves and the 
cylinder, by impurities in the gas. 

In large installations where it is extremely important to avoid 
breakdowns, recourse is had to special means of washing and 
purifying, by the use of centrifugals and other apparatus. But, 
although these succeed in reducing the amount of dust accompanying 
the gas to 4 gramme per cubic metre, they still allow the 



June 1905. LARGE GAS-ENGINES. 671 

passage of tar, vvhicli is the principal element destructive to the 
engine. This tar adheres to the sides of the passages, and causes the 
valves and piston-rings to stick, thus preventing their proper action. 
It also deposits itself in the cylinder where it finally gives rise to 
premature ignitions. Certain coke-oven gases contain as much as 
1 gramme of tar per cubic metre. It is this very question of tar 
which makes it necessary, in order to avoid great complications, to 
employ exclusively non-caking anthracite coal for suction gas- 
producers which are now so largely in use owing to their simplicity 
and economical efficiency. As tar is a product of distillation of the 
volatile carbons (carbides) which escape on combustion and on the 
conversion into fuel gas, it is expedient to reduce them in the producer 
itself. Different systems have been proposed, amongst which is the 
system of inverted combustion, which is said to have the result of 
burning the volatile matters as they are given off and the distillation 
of the hydro-carbons in the independent reducers. It must be 
admitted that so far the efforts of the inventors have not met with 
the success expected. The caking of the fuel and the formation 
of cavities injurious to the regularity of combustion constitute the 
principal difficulty to combat. The use of bituminous coal, especially 
in suction gas-producers, is not therefore a question which has been 
industrially accomplished. 

Whether it be producer-gas, blast-furnace gas, or coke-oven gas 
which is used, purity is an important factor in the proper working 
of the engines. The organic defects of large engines do not appear 
to play in this respect a preponderating part. They should, however, 
be perfected as regards facility of access and upkeep of the parts, 
such as valves and piston. The oil consumption should be low, and 
they should be made less susceptible to variations in the quality of 
the gas. The skill of experts and makers will, doubtless, soon 
overcome these difficulties. Large gas-engines will then constitute 
one of the most remarkable industrial victories of the present age. 

. The Paper is illustrated by Plates 22 to 37 and 39 Figs, in the 
letterpress. 



672 



LARGE GAS-ENGINES. 



June 1905. 



TABLE 1. 

Test made on a Gas Plant of a 4- Cycle Double- Acting Engine of 
200 R.P. and a Suction Producer in the Works of the Gasmotoren 
FahriJc, Deutz- Cologne, the 14:th and 15 th of March 1904, ly 
Messrs, A. Witz, B. Mathot, and de Herhais de Thun. 



TABLE AND DATA OF THE TESTS AND FIGURES. 



Piston Diameter : 21J inches x Stroke 27^^ inches. 
Diameter of Piston-Rods : Front, 4| inches ; Rear, 4^^^ inches. 



Full Load Tests. 



Engine. 

14 March. 15 March. 

1 Average number of revolutions per minute . . 151 "29 150*20 

2 Corresponding effective load ; B.H.P. . . . 214-22 222-83 

3 Duration of tests ; hours ...... 3 10 

4 Average temperature of water after cooling the piston 117' 5° F. 

135° F. 



t, J Average temperature of water after cooling the cylinder! 
\ and valve-seats . . . . . . . / 

^ r Water consumption for cooling the piston per hour; 
\ gallons ........ 

6a Water consumption for cylinder and valves ; gallons . 



386 
1003 



Producer. 

r. ("Nature and Origin of Fuel : Anthracite Coal, " Bonne 
' \ Esperance et Batterie," Herstal, Belgium. 

8 Heating value of fuel ; B.Th.U 

I Consumption of fuel per hour (plus 35 lbs. during the, 
night of the 14th inst. for keeping the generator lired 
during 14 hours, the engine being stopped) ; lbs. . ) 

10 Water consumption per hour in the vaporiser ; gallons 

11 Water consumption per hour in the scrubbers; gallons 



14,650 



199 



160 

14 
315 



June 1905. LARGE GAS-ENGINES. 673 

14 March. 15 March. 

,Q (Average temperature of ^as at the outlet of the\ kkoo ■c 

^- t generator / 558 F. 

,o r Average temperature of gas at the outlet of thel A9-f^op' 

\ scrubbers . . . . . . . . / bZ o J? . 

Efficiencies. 

J. ("Gross consumption of coal per B.H.P. and per hour , 'I a. 097 n*79n 
\ lbs. ... ...../ 

.J. JConsumption of coal per B.H.P. after deducting the^l 0'Q07 0-'7n4 

\ moisture . . . . . . . ./ 

,^ JTherrnal efficiency relating to the effective H.P. and"^ -.q 04-4 
\ to the dry coal consumed in the boiler; per cent. . ) 

Water consumption per brake horse-power hour : — 

For the cylinder, stuflBng-boxes, valve-seats, and jackets ;\ . ^. 

gallons ■ . .f ^'^^ 

^ For the piston and piston-rods ; gallons ... 1-75 

For the vaporiser ; gallons ..... 0*0625 

For washing the gas in the scrubbers ; gallons . . 1 • 42 

ic /Water converted into steam per lb. of fuel consumed in\ ^ no^n 

^^ \ the generator ; gallons j u u»/o 



Discussion. 



Mr. W. J. Crossley, in opening the discussion, thought he would 
be expressing the feeling of the meeting if his first words were 
devoted to thanking the author for the very interesting record he had 
given of the present position of the gas-engine industry on the 
Continent. It was a most interesting Paper, and very well illustrated 
— a Paper which, as it necessitated frequent references to the diagrams 
and a thorough study, might perhaps be read with more profit than 
it could be listened to. That the development of large gas-engines 
on the Continent had been truly marvellous nobody could doubt for 
a moment, but that development appeared to be almost entirely 
in connection with blast-furnace gas. Very few producers were 
mentioned in the Paper. He thought the author rather gave the case 
away in connection with bituminous-gas producers, and did not say 

3 A 



674 LARGE GAS-ENGINES. June 1905. 

(Mr. W. J. Crossley.) 

very much for tliem. England had done a great deal with reference 
to bituminous-gas producers, but as far as blast-furnace gas-engines 
were concerned Germany certainly was far and away ahead of 
England. It was not for him to explain why that was so, but he 
was bound to acknowledge the wonderful pluck and enterprise with 
which Continental blast-furnace owners had gone in for that new 
form of power. Possibly English manufacturers were waiting for 
others to make the experiments. Looking at the case from a 
British engineer's point of view, he thought everyone would agree 
with him that the number of orders executed and in hand on the 
Continent was truly tantalising. 

He did not exactly agree with all that had been said in the Paper, 
which, taken in conjunction with a visit to the Liege Exhibition, 
struck rather a note of sadness in the mind of the English 
engineer ; he could not help hearing to some extent the death-knell 
of many of the principles that were very dear to him. With 
regard to the " hit-and-miss " governor, for instance, he could not say 
more than that the author had ordered its funeral, but his own view 
was that the body was not yet quite ready ; it was still alive, and 
the " hit-and-miss " governor would live for some time, although 
not in connection with large gas-engines. The fact was that for 
large gas-engines it was out of place, but he ventured to submit to 
all those who knew anything about the question that the combination 
of the " hit-and-miss " with the various forms of cut-off mentioned 
in the Paper was a good thing. He could give as an illustration 
several instances where his own company were working the " hit- 
and-miss " in connection with automatic cut-off. Their principle 
was to use the automatic cut-off down to a little below half power, 
and then, if it were required to run the engine below half power, to 
use the " hit-and-miss," because in that way far greater economy 
was obtained in the working. It was also necessary to consider 
the various troubles which arose, and to adopt one of two principles : 
either to limit the amount of combustible and leave the space to be 
made up with air — in which case the combustion was made somewhat 
difficult and it was necessary to rely entirely on a high compression 
in order to get combustion ; or to limit the mixture, in which 



June 1905. LARGE GAS-ENGINES. 675 

case low compression was used. The maximum compression was 
fixed so high that a large margin was available, making it possible 
to fire a small quantity of mixture. Neither of those systems was 
very economical, but the " hit-and-miss " would be more economical 
if it were used below a certain point. The fact was that neither of 
the principles now adopted by German engineers would have been 
of any use had not high compression been brought in instead of the 
old-fashioned low compression. 

Another interesting point was whether stratification existed or 
not. This was especially interesting to those who had been in 
the trade for a long time, and were familiar with the ancient 
actions fought by Otto against several people, such as Steele and 
others. It took sixteen days in Court, with the present Lord Chief 
Justice as the advocate for Messrs. Crossley's firm, to prove that it 
did exist ; it was now universally admitted that it existed, and that 
engines worked on stratification. 

Another thing mentioned by the author as having disappeared 
was tube ignition. He ventured to say, however, that tube ignition 
had not yet quite gone, and that it was still the most convenient 
form of ignition for gas-engines where town gas was used and the 
engines were a fairly small size. He thought, however, that tube 
ignition had served its generation. He was glad to say that it was 
introduced by his own firm, the first patent having been taken out 
by his colleague, Mr. James Atkinson, before he joined the firm. 
It was tube ignition which made higher compression in gas-engines 
possible. On the old-fashioned slide it was impossible to get more 
than 45 lbs. of compression in the cylinder, but with tube ignition 
it was possible to go to 60 or 70 lbs., and thus a great step forward 
was made. He thought that tube ignition would disappear now, but 
it had done its duty. 

Another change which he foresaw, involving the abandonment 
of the old principle which, in his opinion, was not in 
the direction of improvement, was the giving up of liners in 
cylinders, which were always put in by English engineers. Of 
course he knew that to cast the liner in gave a simple form of 
cylinder. He ventured to think that the abandonment would be a 

3 A 2 



676 LARGE GAS-ENGINES. June 1905. 

(Mr. W, J. Crossley.) 

cause of regret by-and-bye, as dnst would come in witb the gas and 
air, and thai dust at higb temperature would wear away the cylinders. 
With a liner it was only a question of a few hours to pull the old 
imperfect liner out and put in a new perfect liner. But to take 
out a big cylinder and repair it, or make a new cylinder, or new 
pistons to suit, would be a costly piece of work, which, however, 
under the circumstances would have to be done. 

He noticed in Fig. 8 (page 630) an arrangement which certainly 
had never found much acceptance in England, namely, the arrangement 
of the valves. He thought to have two valves one on the top of 
the other, as shown, would necessitate giving up English notions with 
regard to valve-settings. In England, the custom was first to slightly 
open the air-valve, say 20°, before coming to the end of the stroke, 
and then to open the gas-valve 10° later while the exhaust-valve was 
still open, and finally to close the exhaust, say, 20° late, relying on 
the momentum in the column of the exhaust to keep the exhaust 
moving out while the charge was beginning to come in, and in that 
way getting a full charge. If the arrangement shown were adopted, 
it seemed to him it would be necessary to close the exhaust- valves 
before opening the admission-valve, otherwise there would be firing 
in the exhaust-pipe, and mixture would not be obtained. If it were 
necessary to close the one before opening the other, time would be 
lost and the cylinder not properly filled. That was the idea of 
Koerting, because in the next illustration. Fig. 10 (page 631), 
it could be seen that, while adopting the same arrangement with 
regard to valves, a water-piece was introduced between the two 
valves to prevent the incoming gas passing out of the exhaust port. 
Therefore that arrangement was not quite as good, perhaps, as it 
should be, and might have to be changed. 

He felt rather surprised at the diversity of the positions of 
the valves in the engine as shown in the Paper. No standard 
appeared to have yet been reached by any German makers. If the 
author could give some definite answer as to the best positions for 
valve-openings and valve-closings for engines, he would be conferring 
a great service upon gas-engine builders. Doubtless, with his great 
knowledge and access to German works, he would be able to say 
something on that point which might be taken as a standard. 



Junk 1905. LAKGE GAS-ENGINES. 677 

It seemed, according to the author, that scavenging, of which no 
instance was found in the German engine at all, had been dropped. 
When compression was at 50 lbs., scavenging was useful, but 
now compression had gone up to 200 lbs. the small residuals left 
were hardly worth considering. It surprised him, however, that, 
considering the enormous loads on exhaust-valves, no attempt seemed 
to have been made in Germany towards the production of equilibrium 
exhaust-valves. His own company had done a lot in that direction, 
and with very considerable success, but the only attempt mentioned 
in the Paper was that of the Otto-Deutz firm, Fig. 41 (page 655), 
who introduced a sort of double-seated valve, which he was sure 
would never answer its purpose. It was not possible to keep two 
faces tight ; and the two faces shown were subject to pressure not 
only during an exhaust, but during the whole of the compression 
and ignition strokes, and therefore the waste through such a valve 
would be very considerable indeed. 

He noticed also in Fig. 44 (page 658) an arrangement which he 
thought would lead to breakdowns. If there were any tar at all in 
the gas, it would fix itself on the sliding block in the particular 
position shown, and would be sure to dirty it and cause sticking, 
which must break something. He had had a good deal of experience 
of that kind, and had been obliged to give up that particular form 
of valve long ago. It was rather interesting to notice the precautions 
which the Niirnberg people took. It was said (page 660) that, 
guided by experience, the Niirnberg Company advised the valves 
and cylinders to be dismantled and examined after some months. If 
that were necessary with that particular form of valve, he contended 
that his fears with regard to the sticking of the valve were to 
some extent justified. 

M. Mathot said he could give some examples of Koerting 
engines that had been running in Westphalia for seven months 
without stopping In another place he had seen an engine working 
for two and a half months, and at another place close to Hanover 
another engine was working in the same way. 



678 LARGE GAS-ENGINES. JuNB 1905. 

Mr. Crosslet noticed that Continental engineers followed 
steam-engine practice very closely, and in that way he thought 
their engines became too complicated and too expensive. With 
regard to gas-engines, he did not think that it was quite necessary 
to follow steam-engine design so closely. He thought they made 
a great mistake in putting too much work into their engines. 
He had heard of trouble often occurring from broken cylinder- 
heads, as they were called, and he thought the author would do 
good service if he stated whether the breaking of those heads 
was a thing of ancient history — whether it did really occur — 
and whether the heads ought to be made of steel or of cast- 
iron. By a device he himself had invented the trouble of breaking 
of heads had been overcome, although of course his firm had not 
made engines of quite the same size as those made in Germany. 
Still it was a satisfaction to think that in England that difficulty 
had been met. In these days, when England needed all the reputation 
it could get, and when it was being so far outstripped by Germany, 
it was necessary to preserve carefully such reputation as it had, and 
in that connection he had just a few remarks to make. 

He noticed (page 651) it was said that the Otto-Deutz firm had 
made 75,000 engines. He ventured to dispute that figure, which 
he thought must be a combination of the makes of some of their 
concessionaires and must include a large number of engines made 
by Crossley Brothers whilst they were themselves concessionaires. 
He remembered that the last time he discussed the matter with 
some of his friends of the Otto-Deutz Co., he found that 
Messrs. Crossley Brothers had made far more engines than they 
had. Messrs. Crossley Brothers had made 50,000. 

There was no question at all that the advent of gas-power as a 
great power had been enormously assisted by the advent of the 
producer plant, and that that convenient form of modern plant had 
done much to produce its popularity. The economy of that plant, it 
seemed to him, chiefly arose from regeneration. The fact was that 
the sensible heat passing from well-made gas was just about sufficient 
to raise the necessary steam for proper working, and he found that 



JVMS 190i. LARGE aAS-£NGIN£S. 67^ 

in his own producers the gas passed away almost cool. In ancient 
forms, of course, the heat was wasted, and a separate boiler was 
needed in order to blow in steam, and in that way economy was 
lost. 

The author gave a most interesting case (page 672) of a plant he 
had just recently tested, and the thermal efficiency was given as 
on dry coal. He desired to ask him what dry coal was ? The 
calorific value given in the test of the coal was evidently of the coal 
as bought. English practice was to give the efficiency of the coal as 
bought, and if that were done in the present case the efficiency would 
be only 23 • 85 instead of 24 * 4 per cent. He thought it would be more 
convenient if the moisture were not deducted, because the buyer of 
the coal had to pay for the moisture and required the efficiency of 
the stuff he bought without the moisture being taken out. Such 
results as were given in the Paper were not uncommon, and it was 
only fair that it should be known what was being done in England. 
He did not like to mention his own experiments, but unfortunately 
he was compelled to do so as he had not the experiments of other 
people. He would therefore give one or two English tests to 
compare with the German tests. Recently they had tested an engine 
giving only 16 B.H.P. which showed an efficiency of 25 per cent, of 
the whole of the heat. That was not taking the coal as dry, but 
simply as it was received. His firm would expect with an engine 
as big as the author's an efficiency of about 25*5 per cent. 
He thought with regard to the producers it might be fairly claimed 
that England was quite as far advanced as Germany, and certainly 
had not been passed. There were many makers of really good 
producers in England now. 

With regard to bituminous coal, he thought England had made 
as many producers as any other part of the world, and they were 
working extremely well. The author stated that the suction 
principle was inadmissible in connection with bituminous coal, and 
he fully agreed with him ; but it was quite easy to have a fan, and, 
by using a fan and a pressure plant instead of a suction plant, a 
greater amount of power could be obtained out of the engine, because 



680 LARGE GAS-ENGINES. June 1905. 

(Mr. W. J. Cross] ey.) 

a certain amount of vacuum was saved which was necessary in the 
suction plant. His firm had no difficulty in getting rid of tar in the 
producers, and he did not know that there need be any difficulty in 
well-made plant. It might be interesting to mention an English 
test of gas-coke. The gas-coke was very much saturated owing to 
the common practice at gas-works of throwing buckets of water over 
it. It was always bought well saturated, but, even so saturated, 
an efficiency had been obtained of 26*54 per cent., that is, energy in 
engine divided by heat units in fuel, and a consumption of 0*91 lb. 
per B.H.P. He thought that was very interesting. The economy of 
gas-plants was now truly astonishing considering their small size, 
and he ventured to conclude by saying that the gas companies 
would have to look to themselves and act very carefully indeed, if 
they were to withstand the competition of the electric light and that 
of gas-plants. The gas industry was assailed in both directions. 
He sincerely hoped the gas companies would take the sensible course 
of endeavouring to reduce their price to meet the market, instead of 
trying to bully the makers of producer plants and engines by telling 
them that they would not push their engines if they ventured to put 
up producer plant in their towns. That was not the way to deal with 
the matter. They should act in a proper, fair and businesslike 
way, by bringing the cost of coal-gas down to meet the engine. He 
ventured to congratulate the author upon his Paper and assured him 
that it had given him very great pleasure to speak upon it. 

Mr. John Fielding said the English position had been very well 
stated by Mr. Crossley, and he had nothing to add to it. With regard 
to English practice as compared with German practice, referring to 
the system of governing, the author would probably remember an 
engine that was running at the Brussels Exhibition some nine or 
ten years ago, in which, the system of cut-out governing was entirely 
abandoned and the mixtures of gas and air were independently 
regulated by the governor, so that there was no cut-out at no load 
to full load. The English engine, he thought he was right in 
saying, was one of two engines out of eight running in the Brussels 
Exhibition working the electric light, and the only gas-engine that 



Junk 1906. LARGE GAS-ENGINES. 681 

ran the whole time without a hitch. That was one little testimonial 
to the English make of gas-engine. Whilst no doubt the Paper 
would be welcomed by all interested in the development of the 
internal-combustion engine as a very complete collection of 
descriptions which had already been more or less completely 
published in the engineering journals, it was perhaps to be 
regretted that an author with such opportunities as those enjoyed 
by M. Mathot had not been able to present to the Institution 
more information with regard to the working of the Oechelhauser 
and Koerting engines, which were without doubt the most interesting 
types dealt with in the Paper. 

The author very properly made special reference to the questions 
of up-keep and freedom from stoppage, and as a valveless engine 
seemed to possess great advantages in these respects, it would be 
interesting to hear whether, in actual working over several years, the 
Oechelhauser engine had enjoyed the freedom from stoppages and 
wear and tear which might be expected. It would be most interesting 
to know whether the working of large pistons over inlet and exhaust 
ports had tended to produce any greater wear either in the pistons 
or cylinders. Obviously, all other things being equal, an engine of 
the Oechelhauser type must possess enormous advantages over all 
others by the complete absence of valves subject to heat or pressure, 
leaving only such valves as work under favourable conditions, and 
therefore presenting no difficulty in regard to up-keep. 

Then again the question of the use of piston-rods and stuffing- 
boxes might perhaps be considered of the utmost importance. The 
British gas-engine builder as a rule had not yielded to the temptation 
to adopt a type of engine involving their use. On the other hand, 
their Continental friends seemed to have no qualms about the matter, 
and it would almost appear that they introduced them in some cases 
where they were not absolutely necessary, as for instance, in Fig. 38, 
Plate 31, where a tail-rod was shown. This no doubt served a double 
purpose in helping to support the weight of the piston and acting as 
a water channel, but to the British mind the stuffing-box was a serious 
objection to set against these points. It would be instructive to 



682 LARGE GAS-ENGINES. Junk 190&. 

(Mr. John Fielding.) 

learn from M. Mathot whether, in his opinion, stuffing-box packings 
and piston-rods had given any serious trouble, as one heard 
occasionally, as might be expected, that they were a source of extreme 
anxiety to the men in charge. 

M. Fred. Kraft thought that the subject had been so 
exhaustively dealt with that it was difficult to say anything new. 
The Paper was a very able one, and clearly showed the present 
position of the large gas-engines on the Continent. Probably the 
reason the author did not refer to tube ignition was because he 
was dealing only with large gas-engines, in which it was not used. 
The " hit-and-miss " system of governing had the great advantage 
of simplicity, but it had been abandoned altogether for large gas- 
engines. The governing of the latter engines had to be solved by 
variable admission, because the stresses and the work to be stored 
in the fly-wheel were enormous in large engines, and necessitated 
very heavy and large fly-wheels. Electricians were the best 
customers for gas-engines, and they required very high regularity. 
For large engines which were not run at such high speeds, and 
could not run at such high speeds, as small engines, it was a 
question of fly-wheel and governing which had brought every maker, 
he thought, to adopt variable admission. On the advent of variable 
admission it seemed to be complicated, and was indeed so in 
comparison with the " hit-and-miss " system, but now a study of 
the various devices invented by makers of large gas-engines would 
show that it was not so complicated as it had appeared to be. 

With regard to governing generally, the " hit-and-miss " system 
had the disadvantage of low regularity, and that was the reason why 
it had been abandoned. Then came the question of the cut-off, for 
which there were different systems already devised. The first idea 
was, as had been done in steam-engines, simply to let the best mixture 
of gas and air enter the cylinder, and limit the quantity of mixture 
admitted to the power required. This had the disadvantage of at the 
same time varying the compression of the charge. He called it a 
disadvantage from the mechanical point of view — the disadvantage 
to the running of the engine. It had another another drawback, that 



Juwx 1906. LAEGE GAS-ENGINES. 682 

of high consumption at low loads. The disadvantage of the variable 
compression from the point of view of the running of the engine 
was that it was impossible at low loads to balance properly the 
inertia of the reciprocal parts of the engine, especially as engines 
were always run as fast as was compatible with good running. 
Those facts brought some makers — and his firm was amongst them — 
to look for a method of getting compression as constant as possible 
for all loads. The idea was to take in a charge always the same 
quantity of air, or almost always, and to vary the quantity of gas 
only. That meant, generally speaking, a very weak charge at low 
loads. His own view was — and he thought there was reason for it — 
that this might lead to miss-fires at light loads. To overcome this 
difficulty, some makers took the air into the charge first and then at 
a given point of the stroke, depending on the governor, opened the 
gas-valve so that from that point to the end of the stroke there was 
a mixture. With that method there was almost constant compression, 
because the air-valve was open all the time to the free air and there 
was practically, though not absolutely, the same pressure in the 
cylinder — something a little below atmospheric pressure at the end 
of the suction stroke. Another advantage of the method was that, 
although the theory of layers of gas was not perhaps absolutely true, 
there was certainly truth in it to some extent, and it was found 
advantageous to take the charge at the end of the suction stroke, 
in order to keep on the igniter a rich, good, easily inflammable 
mixture which insured ignition at all loads. That was a thing his 
firm managed to get, so that even when running light, miss-fire was 
not a thing to be feared. Having almost constant compression, 
economy of the engine in the consumption of gas was also secure. 
Unfortunately he had no data of any serious trials with different 
loads of an engine governed in that way, but that was the method of 
governing his firm generally adopted for all their engines, and it 
had been referred to by the author in the Paper. He knew, however, 
from some trials made privately that at about three-quarter load the 
consumption in calories per I.H.P. must be something very nearly 
2,000. He considered that to be certainly a satisfactory figure for a 
load under the normal. 



684 LARGE GAS-ENGINES. JuNB 1906. 

(M. Fred. Kraft.) 

The author said that for high-speed engines, where the weight of 
the reciprocating parts was considerable, and where regularity was 
required, a high compression was necessary to assist in giving 
" uniformity." He supposed the author meant " smooth running," 
because he felt himself inclined to think just the contrary. It 
assisted in dealing with the inertia of the reciprocating parts, thus 
getting over the dead centres without any shocks. That was one of 
the reasons why, especially for electrical engines running at a 
constant speed, the constant compression valve-gear was used, in 
order to be sure of having a smooth running on light load, and for very 
large engines smooth running meant sometimes not being able to run at 
all, because the shocks would be so terrific that the engines would not 
be accepted by any customer. For variable-speed engines, running 
slowly, where there was a specification to run at different speeds and 
as low as half speed, the advantage of the constant compression 
(which was certainly more economical) had been discarded, because 
the high compression, if it were constant, would interfere with the 
regularity in turning and sometimes create difficulty in running 
over the dead centres. As the blowing engines were always 
designed for blast-furnace plants, the consumption of gas played 
no important part, and it did not matter if they were a little less 
economical. As far as the customers would accept it, his firm 
thought it would not be necessary to take low consumption too 
much into consideration, but it would be better to make engines 
still more simple than those which had been designed for constant 
compression engines. 

The President very much regretted the time was getting too 
short to continue the discussion and asked those who wished to add 
anything to the discussion to send their views in writing to the 
Secretary. Before calling upon the author to reply he asked the 
meeting to accord him a hearty vote of thanks for his most 
interesting Paper. 

The resolution was carried. 

M. Mathot expressed gratification at the discussion which had 
been elicited by the Paper. With regard to the " hit-and-miss " 



June 1906. LARGE GAS-ENGINES. 685 

system of governing, there was a national tendency in England to 
keep to it, due to the fact that it was born there and that it was 
almost always applied only to small engines. When he had only to 
deal with the matter as a consulting gas-engineer for small engines 
his experience was limited, but when he came to deal with larger 
gas-engines he was such an amateur that in dealing with them at first 
he would have made their governing on the " hit-and-miss " system, 
because, being a very simple arrangement, it would have saved much 
trouble. He had pointed this out to the firm of Messrs. Bollinckx of 
Brussels, but after seeing the working of the very large engines in 
Germany and on the Continent generally, he agreed it was necessary 
to give up that kind of regulation in spite of all its recommendations. 
At any rate, as the growth of large gas-engines was especially due 
to the utilization of blast-furnace gases for suction plants, he was 
bound to say that the system of " hit-and-miss " regulation was 
not of much use. When the engine was running at light load, 
the intermittance between the different explosions and the misses 
allowed the generator to become cool and generate bad gas. In 
fact, the hotter the generator was running, the better was the gas 
produced, and the more economical it was, because it dissociated 
more water. When the engine was running at full load, the 
generator was kept very hot, but when running at half load it became 
cool and made worse gas. For that reason it was necessary to try 
and make the gas suction as regular as possible, and diminish the 
number of misses. When running at a very light load, one explosion 
might be followed by five or six misses without any kind of suction, 
which would give a very bad result in running the generator itself. 

With regard to ignition, he thought it was of no use to enlarge 
on what had already been said in the Paper, which had been 
specially written with the view of giving a description of large gas- 
engines rather than discussing technical points ; and the Paper he 
had written for the Mining Congress to be held the following week 
contained more technical matter and a full discussion of the valve- 
setting, in which Mr. Crossley was much interested. 

Mr. Crossley seemed to think the scavenging valve-setting had 
been entirely neglected in the new method of commanding the inlet 



686 LARGE GAS-ENGINES. Junb 1905. 

(M. Mathot.) 

or exhaust valves, when they were fitted on the same vertical axis. 
That would be true, if both the inlet and exhaust valves were driven 
by the same cam, which, however, was not the case. It was easy to 
make an arrangement such as that of Messrs. Koerting, where one 
cam was specially devoted to driving the inlet valve and the other 
cam devoted to driving the exhaust valve. The Niirnberg firm tried 
to drive the two valves by the same cam ; but he was not very much 
in favour of that method. With regard to the valve setting, it was 
not neglected in Germany, and the first experiments carried out by 
the Crossley Company were still much in favour in Germany. 

With reference to scavenging, which consisted in opening the 
air-valve about 10 or 15 degrees beforehand whilst the exhaust 
valve was not closed, it was still in favour and was used in almost 
all the gas-engines. 

Another thing which was of more interest was the question of 
the stratification of gases to which Mr. Crossley and M. Kraft had 
referred. His own opinion about it was that much must be taken 
and much must be left. It was what was called in France " des 
choses qui souvent ne font pas de hien, mais ne font fas de mal non 
plus" which meant that if things did not go well they need not 
necessarily be very bad, and therefore one had better try and get the 
best result out of them. He had made a great many tests ; in fact, 
in fifteen years he had made about 400 tests of different gas-engines, 
and all of them had been carefully put on record. A full discussion 
of them would be found in the Paper he had written for the Mining 
Congress. He ventured to state that in only a very few cases were 
the effects of scavenging found to have taken place, although the 
engines were regulated with the view of performing that function. 

With regard to Mr. Crossley's remarks (page 678) concerning 
the production of the Otto-Deutz firm, he agreed that the number 
given included those constructed by Messrs. Crossley Brothers, 
while they were concessionaires, and other licensees. The Otto- 
Deutz Co. stated that the 75,000 engines represented 535,000 H.P. 

There might be something wrong in connection with the test 
on the 200-H.P. engine, because the figures had been translated very 
hurriedly from French measurements into English figures. In the 



Junk 1905. LARGE GAS-ENGINES. 687 

hurry and difficulty of the work a mistake was often overlooked, and 
Continental engineers were not accustomed to the very difficult 
English system of figures. He hoped that this would be a hint 
to English engineers to improve their system of measurements as 
well as their engines. With regard to dry coal, it meant coal as it 
came from the pit, but very often it happened that before being 
stored it received a lot of water while in the car, and when that 
happened the coal was dried before an analysis was taken to 
discover its calorific value. Dry coal did not mean coal dried for 
the purpose of analysis, but coal reduced to the natural conditions in 
which it came out of the pit. 

Eespecting bituminous coal producers, without mentioning any 
name, he might state that he had been called three or four times during 
the last year to assist different inventors in France. He was quite 
sure that bituminous coal producers could work successfully on 
the suction principle. 



Communications. 



M. Adolphe Greiner wrote that, with reference to doubts 
which had been raised as to the avoidance of pre-ignition with gas 
of a certain calorific value and large compression — when in fact 
there were comparatively high temperatures to deal with — and with 
reference to the importance which many engineers attached to 
scavenging as a means of overcoming these assumed difficulties, 
it might be of interest to bring forward the following facts : — 

The Cockerill engines were of a 4-cycle non -scavenging type. 
With coke-oven gas, in an engine of 24 inches diameter with a 
32-inch stroke, there was no pre-ignition with compression as high 
as 9 kgs. (128 lbs.) above atmosphere. In another engine fed with 
producer gas (145 B.Th.U. and 20 per cent, hydrogen), the 
compression was brought to 12 kgs. (170 lbs.) without any difficulty. 
The cylinder of this engine was 32 inches diameter with a 40-inch 
stroke. Pre-ignition was not feared from high compression alone, 
provided the gas were free from dust and the combustion chamber 



688 LA.RGE GAS-ENGINES. June 1905. 

(M. Adolphe Greiner.) 

were well designed — that was to say, simple in shape and without 
any recesses to keep the hot burnt gases of the preceding explosion 
in the combustion chamber. This confidence was due to the fact 
that compression up to 15 kgs. (213 lbs.) would not bring the 
temperature of the charge higher than 880° F. (the initial 
temperature being taken at 212° F.), whilst the most inflammable 
gas in the mixture, namely hydrogen, ignited only at 1,022° F. 

Mr. Mark Eobinson, Member of Council, wrote that the fulness 
of information in M. Mathot's Paper made it of great value to all 
students of large gas-engines, but he (Mr. Eobinson) ventured to 
hope that the author would express an opinion upon one subject not 
dealt with in the Paper. In taking up the manufacture of large gas- 
engines, his firm (Messrs. Willans and Kobinson) had been influenced 
by the belief that, with gas of higher calorific value than blast- 
furnace gas, very large engines required more efficient cooling than 
was possible under the ordinary Otto cycle, even when both pistons 
and valves were efficiently water-cooled. In England there had 
hitherto been considerable unwillingness to construct large engines 
to work with producer-gas on what might be called the plain Otto 
cycle. It was recognised that such engines were largely employed 
on the Continent, but nearly always, so far as his information 
extended, in conjunction with the use of blast-furnace gas. English 
makers had either avoided making large engines, or had designed 
them with some form of air-scavenge, which, between working 
strokes, should both sweep out the remaining products of combustion 
and serve more or less to cool the cylinder as well. It was true 
that engines working on the plain Otto cycle were stated to be used 
on the Continent with producer-gas of good calorific value, but it 
was also said that the number of these cases was small, and no 
certain information had been published as to whether such engines 
were working continuously under full load. 

Omitting reference to engines of the two-stroke cycle, the tendency 
of recent designs appeared to be towards higher compressions and 
higher mean pressures — in other words, towards the development 
of a greater number of heat units in a cylinder of unit size in unit 



June 1905. LARGE GAS-ENGINES. 689 

time. It would seem to follow that to maintain efficient working 
the means of cooling must be correspondingly increased, and that 
beyond a certain point the ordinary methods of cooling by water 
might become insufficient to prevent pre-ignitions. The necessity 
under modern conditions for governing by throttling rather than by 
" hit-and-miss " added to the difficulty, because an engine governed 
under the *' hit-and-miss " system, and not having an explosion at 
every working stroke, had the benefit of a thorough scavenging 
charge whenever an explosion was missed. When an engine working 
below full load was governed by throttling, the number of heat units 
evolved in the cylinder in unit time was of course reduced, but 
there was no opportunity for such an occasional complete scavenging 
by cool air as took place when the governing was by " hit-and- 
miss." For all these reasons the opinion had widely prevailed in 
England that if large Otto-cycle engines were to be worked with gas 
of high calorific value, they should be given means of cooling beyond 
those which water circulation affi)rded — in other words, they should 
have an air-scaveuging arrangement. 

Unfortunately a scavenging arrangement was a somewhat costly 
addition to a large engine. If the scavenging charge was to be 
effective, it must be large in volume, so that the pump, or that part 
of the already existing mechanism which was utilized to serve as a 
pump, must be of large size. As the air pressure should be small, 
to avoid loss of power, the valves and passages through which the 
air was supplied needed to be very large, and such additions added 
something to the complication, and very much to the cost, of the 
engine, so that the question forced itself upon the engine builder, 
whether his engine, however satisfactory and efficient it might be, 
could practically be made at a cost low enough for competition with 
steam-engines. Throughout the Paper the author seemed to treat 
all types of engines — the plain Otto without scavenge, and the 
Koerting and the Oechelhauser with scavenge — as equally suitable 
for practical work, and did not treat one type as limited, any more 
than another type, by the size of the engine or by the quality of the 
gas to be used. It might be very useful to many students of this 
subject if those who had experience of large gas-engine practice 

3 B 



690 LARGE GAS-ENGINES. JuNE 1906. 

(Mr. Mark Robinson.) 

would say if they considered the plain Otto cycle, supplemented 
only by effective water cooling of the valves and pistons, suitable 
for engines of really large size (and if so, up to what horse-power 
per cylinder) when using producer gas or coke-oven gas of high 
calorific value, and when working with high compression and with a 
high mean-pressure — say in the region of seven atmospheres mean- 
pressure. In asking this question he had assumed that everything 
turned upon the nature of the gas, and that conclusions drawn from 
the behaviour of large engines using blast-furnace gas might be 
inapplicable as a guide to the design of engines using gas of a hotter 
nature ; it was further assumed that blast-furnace gas was more 
difficult to ignite, and therefore to pre-ignite, than good producer-gas. 
On these assumptions, absence of pre-ignitions in engines using 
blast-furnace gas, or even in engines using producer-gas but working 
at comparatively low compression and mean pressure, would seem 
to give no assurance that this disastrous form of trouble would 
not occur in similar engines when using strong producer-gas in 
combination with high compression and high mean-pressure. 

Mr. Hal Williams wrote that the Paper dealt mostly with the 
question of design, and Mr. Crossley had also confined his remarks 
largely to that question. The writer desired, however, to make 
some remarks regarding the commercial aspect of the gas-engine. 
He had had opportunities of observing the behaviour of a considerable 
number of plants working on bituminous producer-gas, on anthracite 
producer or Dowson gas, and on suction gas, and he felt that the 
author had hardly done the producer-plant manufacturers justice, 
as in England at any rate they had been able to give some very 
satisfactory results. One plant which he was acquainted with had 
been working for over two years, supplying gas to a number of 
small gas-engines scattered over the works, the load on which was 
constantly varying, and it worked with perfect satisfaction. The 
annual coal consumption, including all stand-by losses, was less 
than 2 lbs. of bituminous slack per B.H.P.-hour. 

In his opening remarks, the author had referred to the 
complication of the gas-generating apparatus, its initial cost and the 



June 1906. LARGE GAS-ENGINES. 691 

space occupied being much against it as compared with steam- 
engines and boilers. He did not at all agree with this. In the first 
place, a bituminous producer plant, where ammonia recovery was not 
aimed at, was a very simple apparatus. For moderate power, it was 
cheaper in first cost than boilers of the same power. A 500-H.P. 
plant cost, erected complete and at work, not more than £1,500, 
whereas the cost of two boilers of equal power, seatings, chimney 
shaft and the necessary accessories would exceed this, the difference 
rapidly growing greater as the power increased. The up-keep of a 
gas plant was much less than that of a boiler. The amount of water 
required was very small, and there was no need for boiler-cleaning, 
boiler fluids, etc. Another advantage was that whereas boilers 
had to be placed as near their engines as possible, gas plant could 
be put down in any out-of-the-way corner and the gas could be 
transmitted without loss to the engines anywhere in the works. 
There were some points, which had to be carefully watched, but apart 
from this, a gas plant and gas-engine installation presented no 
dijQSculties and offered wonderful economies. 

He believed that he was one of the first consulting engineers to 
introduce the suction plant into this country, and in early days had 
travelled on the Continent specially to investigate it. He now had 
suction plants and engines working under his supervision which 
were consuming less than a pound of anthracite coal per B.H.P.- 
hour, and which, even on five-eighths load, were delivering their power 
on to the second motion shaft at a cost, including permanent charges, 
oil, etc., of 0*2916?. per B.H.P.-hour, while the cost at full load sank 
to 0*2086?. per B.H.P.-hour, this cost being equivalent to electrical 
energy at 0*436cZ. and 0'Sl2d. per B.Th.U. respectively. He had 
recently reported to a large manufacturing concern in the Midlands 
on their existing power, and had found that after charging £70 a 
year dead interest on their existing engines and boilers, the first cost 
of which had not been quite written off, and after allowing 5 per cent, 
interest on the cost of the new plant, 5 per cent, for depreciation, and 
1 J per cent, for repairs, on a capital expenditure of some £7,000, he 
was able, by utilising gas-engines working on a bituminous gas 
plant driving generators and transmitting power to the mills 

3 B 2 



692 LARGE GAS-ENGINES. JuNE 1905. 

(Mr. Hal Williams.) 

electrically, to show a saving of over £500 a year. The chief 
drawback to the development of gas power to his mind was the 
excessive cost of the gas-engines, and the comparatively large space 
they occupied. He hoped that, when the manufacturers had 
obtained greater experience and had recouped themselves for their 
experimental losses, they would be able to reduce the price 
considerably. 

On the question of governing, he felt that a "hit-and-miss" 
system, though to his mind quite satisfactory for powers up to 
say 100 H.P., was not satisfactory beyond this. The practical 
gain to a gas-engine by missing a stroke every now and then, and 
so scouring itself, was very great, and he felt that, from a practical, 
as well as an economical point of view, the best results would be 
obtained from an engine which governed by variable admission from 
say full to half load and " hit-and-miss " below this. 

He did not quite agree with Mr. Crossley's remarks about the 
position of the inlet and exhaust valves, as he had carefully observed 
the behaviour of engines running with valves in this position and 
had not found any difficulties arising, while the advantages from a 
practical point of view of being able to get at the exhaust-valve 
through the air- valve opening were, to his mind, very great. 

M. Mathot wrote, in further reply to the discussion, that the 
Paper dealt with Large Gas-Engines, and therefore he had only 
slightly referred to gas-producers. 

With regard to cylinder liners. Continental makers had not 
abandoned their use, as Mr. Crossley seemed to think (page 675). 
German trunk-piston engines were all provided with independent 
liners, but the cylinder or water-jacket itself was cast in one 
piece with the frame, whilst the English method consisted in 
bolting the overhanging jacket to the frame. The former 
arrangement, he thought, was more advantageous because it 
rendered unnecessary the overhanging of a bulky piece of casting, 
consisting of the cylinder, support of side-shaft, etc. 

With regard to Mr. Crossley's enquiry as to a general valve- 
setting which could be regarded as a standard for gas-engine 



June 1905. LARGE GAS-ENGINES. 693 

makers, the author saw no possibility of such an arrangement being 
effected, because the moment and duration of the opening and 
closing of both the inlet and exhaust valves depended very closely 
upon their method of operation, their relative position and 
dimensions, the shape of the explosion chamber, the piston speed, etc. 
Mr. Crossley thought that the double-seated exhaust- valve of the Otto- 
Deutz engines shown in Fig. 41 (page 655) would not remain tight 
for any length of time. It had however passed the experimental 
stage, and was now at work on several large engines. As steam- 
engines of the Sulzer type were provided with double-seated valves, 
and worked satisfactorily even with superheated steam, there could 
be no reason why the Otto-Deutz valve should not remain tight, 
especially as it was kept at a low temperature by the water 
circulation. Only one objection could be made, namely, the effect 
of dust or tar sticking on the valve, but it must be borne in mind 
that this particular valve was used with large gas-engines dealing 
with coke-oven or blast-furnace gas specially cleaned and washed. 
Concerning the possibility of the sticking of the sliding block at 
the inlet arrangement. Fig. 44 (page 658), the author quite agreed 
with Mr. Crossley. The illustration was diagrammatic, and some 
of the details for withdrawing the tar were not given. Accidents 
to cylinder heads rarely occurred, unless there were some defect 
or weakness in the casting. European makers had adopted 
the arrangement described under the heading of " Cooling " 
(page 629). 

With respect to Mr. Fielding's remark (page 681) about the 
lack of information in the Paper concerning the working of the 
Koerting and Oechelhauser gas-engines, it must be borne in mind 
that the Paper was written from a descriptive standpoint, therefore 
personal opinions about the one or the other make had not 
been stated. Many engineers considered the two-cycle engines 
too complicated, their mechanical efficiency too low, and that 
they were liable, on account of the absence of exhaust-valves 
and to the use of an air-pump, to a certain waste of gas 
immediately after the scavenging had taken place. A considerable 
amount of trouble arose some years ago with regard to piston-rods 



694 LARGE GAS-ENGINES. Junk 1905. 

(M. Mathot.) 

and stuffing boxes, owing to the makers not having had personal 
experience with them, but, since the new features, described in the 
Paper, had been applied, no particular objections had arisen when 
kept in good order. 

In answer to Mr. Mark Robinson's remarks on the question of 
the efficient cooling of the jackets and cylinder heads (page 688), the 
author wished to state that the pistons, piston-rods, valves and 
stuffing-boxes were cooled by water supplied by special pumps under 
a pressure of 15 to 30 lbs. Large gas-engines were not affected by 
the absence of air-scavenge, which was obtained by pumps or by the 
simple effect of the cut-out in a " hit-and-miss " governed engine. 

In regard to the calorific value of the gas used, it did not matter 
whether it was low or high, as it did not interfere with the heat 
evolved. In fact, although the mechanical efficiency of good double- 
acting Otto-cycle engines attained from 28 to 30 per cent., the 
consumption in heat-units per brake horse-power remained nearly the 
same whatever the nature of the gas, as stated in the Paper (page 636). 
The quantity of water used for cooling was therefore not in 
proportion to the calorific value of the gas fed to the engine, but only 
to the heat-units contained in the explosive mixture. For the same 
reason, the possibility of pre-ignition did not increase in proportion to 
the richness of the gas, but rather to the hydrogen or hydrocarbons 
contained in the gas, and the presence of these considerably 
increased the velocity of propagation of flame in the mixture. 

In replying to another question of Mr. Mark Robinson, the author 
considered that the two-cycle engine met the special requirements 
of a good motive power quite as well as the four-cycle engine, but 
it must be taken into consideration that the low mechanical 
efficiency to a certain extent detracted from the two-cycle engines. 
Notwithstanding this, both systems had been successfully worked 
in important central electric stations where the engines were 
supplied with producer gas. 

The following list gave some examples of large engines worked 
with different kinds of producer-gas : — 

Two single-cylinder Oechelhauser engines of 250 and 400 B.H.P. 
fed with Fichet-Heurtay producers, at the Fabrique Nationale d' Armes 
de Guerre, Herstal, Belgium. 



June 1905. LARGE GAS-ENGINES. 695 

Three twin-cylinder Oeclielliauser engines of 2,000 H.P. each, 
fed with Duff producers, at the Municipality of Johannesburg. 

Six tandem cylinders Niirnberg engines of 350 B.H.P. each, fed 
with Pintsch suction producers, at the Tramways Central Electric 
Station, Scheveningen, Holland. 

Six twin tandem 4-cylinder Niirnberg engines of 2,000 B.H.P. 
each, fed with Mond producers, at the Sociedad de Gasification Indus- 
trial, Madrid. 

Six Otto-Deutz engines : one 2-cylinder double-acting of 
600 H.P. and four single-acting twin cylinders of 250 H.P., fed with 
Deutz suction-producers and Dowson blown producers, for the 
Municipality of Munster, Germany. 

Two double-acting single cylinder Otto-Deutz engines of 
250 H.P. each, fed with Deutz suction-producers, for the 
Municipality of Kazan, Russia. 

One double-acting 2-cycle Koerting engine uf 750 H.P., fed witn 
Koerting producers, in the cornmill of Eolandsmiihle at Bremen, 
Germany. 

Two single-cylinder 2-cycle Koerting engines, fed with coke- 
oven gas, at Julienhiitte, Germany. 

It might also be added that engines of the Otto type worked 
satisfactorily either with fuel gas or coke-oven gas, and that they 
attained 500 B.H.P. per cylinder and upwards with no more 
difficulty than with blast-furnace gas. 

Information was given in the Paper (pages 669 and 670) with 
regard to the mean pressure on the piston adopted for working these 
large engines. Mr. Robinson, however, mentioned a mean pressure 
of seven atmospheres, which the author ventured to say was never 
practically maintained or endured in working conditions, even with 
town gas. With regard to the risk of pre-ignitions in large engines 
using other than blast-furnace gases, it must be remembered that 
the compression should be lowered when gases containing a large 
proportion of hydrogen are utilized, for the reason just stated 
regarding the velocity of the propagation of flame. 



JvNB 1906. 697 



THE STRENGTH OF COLUMNS. 



By Professor W. E. LILLY, Member, op Trinity College, Dublin. 



The design of a column of a given length to carry a given load 
involves the determination of the area and radius of gyration of its 
cross-section, and thus indirectly of its thickness. The formulae in 
general use do not take into consideration the ratio to be adopted 
between these quantities, it being left to the designer to assume 
empirically some values which seem most suitable under the 
circumstances. Considered from a theoretical point of view, there 
is for every column of given length and load a definite area and 
radius of gyration for the most economicul cross-section, and any 
departure from these involves waste of material. For instance, take 
the case of a hollow mild-steel column of circular cross-section, if 
the diameter is large and the thickness small, the column fails by 
wrinkling of the sides of the column or by secondary flexure ; if the 
diameter is small and the thickness great, it fails by primary flexure 
or bending ; hence there is some diameter and thickness which will 
give the most economical result. The experiments contained in 
this Paper were carried out with a view of determining experimentally 
the conditions under which failure takes place either by primary or 
secondary flexure and to obtain indirectly some definite information 
as to the values of the areas and radii of gyration of the economic 
cross-sections. 



698 



STRENGTH OF COLUMNS. 



June 1906. 



For this purpose it was decided to test a number of mild-steel 
tubes similar to those now so generally used in cycle construction, 
and of as nearly as possible a uniform quality throughout. A 
consignment of these was obtained, of which the following Table 1 
gives particulars of the diameters and gauges of the tubes that were 
tested : — 

TABLE 1. 



Exteroal 
Diameter. 


Thickness, 


Standard Wire Gauge. 


inch. 






i 


10 17 


18 19 20 22 


i\ 


16 17 


18 19 21 22 


i 


17 18 


19 20 


1 


18 19 


20 


f 


18 19 




i 


18 19 




1 


18 19 





The experiments were carried out in the Engineering Laboratory 
of Trinity College, Dublin, on a 10-ton Wicksteed testing machine 
of the vertical type. Unfortunately, owing to the small size of the 
machine, it was impossible to make the experiments as complete as 
the author could have wished ; it was decided for this reason to 
limit the investigation for the present to round-ended columns with 
a view of in the near future continuing the experiments on columns 
with fixed ends, when a machine of larger size will be available. 

The tubes were prepared for testing in the following manner. 
The dimensions of the tube having been ascertained, measured 
lengths were cut off; these were then trued up in the lathe to ensure 
straightness, the ends at the same time being faced off square. Hard 
steel round-ended pins with square shoulders were fitted to the ends 
of the tubes as shown on Fig. 1 ; they were then ready for testing. 



Junk 1905. 



STRENGTH OF COLUMNS. 



699 



The mode of testing was to place the tube vertically in the 
testing machine, the round ends bearing on steel discs as shown on 
Fig. 1, the surfaces being very slightly concave to enable the 
specimen to be easily adjusted. The test was then carried out in 
the usual manner and the load producing failure noted. 

The number of experiments carried out on various lengths of 
tubing exceeded eight hundred ; of these only a selection is given 
on Figs. 2-8 (pages 700-701) which bear more directly on the 
investigation. 

Fig. 1. — TvJbe ready for Testing, fitted with hard steel pin and hearing on steel disc. 

SectioTixtV 
ElevatioTv . 



Pl^ft 





The author considers that no useful purpose would be served by 
giying more of the results of these experiments in detail. As was 
to be expected, owing to the variations that occurred in the quality 
of the material, some difference was shown in a few cases from the 
results as given, more especially on the short lengths tested. The 



700 



STRENGTH OF COLUMNS. 



June 1905. 



Fig. 2.— Tube. 
Diameter, ci = | in. Standard Wire Gauge, g = 16. 
Area of Cross-Section, A = 0-063 sq. in. Kadius of gyration, p 
Thickness of Tube, t=z0'17 d. (p/t) ^ - 3. 

Lbs. 

per d" 



1 



0-11. 



SQ 





^^ 


%• 






















4aooo 






^^ 


\ 


























^ 


^ 














n 














"~^ 




■ =. 


— — 


;- 


— 1 — 



1-0 



80 120 160 

Length tiinded by radius ofgira/wn 



200 



240 



Fig. 3.— Tw6e. 
d='^m. g = \Q. ^ = 0-076 sq. in. p = 0-13. ^ = 0-14^. (p/ty = ^'S. 



Lbs._ 
perQ 
80.000 



'J'OlDOO 






40 



80 120 160 

Length divided by radius of ^yralion 



200 



240 



•Fig. \.—Tvhe. 
rf = ^ in. g = 17. A - 0-077 sq. in. p = 0-16. t= O'll d. (p/t);^ 



8 -.3. 



Ibs.^ 
per a 
8Q000 



4aooo 






40 80 120 

Len^lh divided by radius of ^ration 



160 



June 1905. 



STRENGTH OF COLUMNS. 



701 



Fig. 5.— Tube. 

rf = I in. ^ = 18. A = 0-086 sq. in. 

p = 0-21. = 0-077 d. 

Us. 

Pfro' 

SOWO 



^03300 



Fig. 6.—Tiibe. 

d = ^in. (7 = 18. ^ = 0*107 sq. in, 

p=0-25. t = 0-065 d. 

Cp/tr = 27. 



\ 
















■\. 














N^^ 












^ 


^ 










1 






~"^ 


^-^ 


















4 










"< 














**-j 



4-0 80 120 

Length divided hy mdius of gyration 



40 



80 



120 



d^l 



Fig. 7.— Tube. 
in. g — l9. J. = 0*11 sq. in. 
p^O-31, t = 0'0i6d. 

(p/ty = 60. 

Us. 

pern" 

80.000 



■o«4-0.000 

1 



40 80 

Length divided by radius of gyration 



Fig. 8.— Tube. 

d = lin. g=l9. ^ = 0-12sq. in. 

p = 0-34. t = O'Oid, 

(p/tr = 72. 






ir- 










\. . 










%\^ , 


S?-' 






i~~^ 






I 













80 



results of these were averaged, giving a close approximation to the 
curves of the tubes shown on Figs. 2-10 (pages 700-701, and 
702-703). 

The data given on the diagrams. Figs. 2-8, refer to the individual 
tubes tested. On the base line values of Z, the length, divided by p the 
radius of gyration are plotted, the breaking load for these values 
being plotted vertically. The numerical values of the diameter = d, 
standard wire gauge = g, area of cross-section = A, radius of 
gyration = p, and the thickness — t oi the tube, are given on each 
diagram respectively. The curves on Fig. 9 (page 702) show the 



702 



STRENGTH OF COLUMNS. 



June 1905. 




■c^ oo 



Breaking Load 



Junk 1905. 



STRENGTH OF COLUMNS. 



70J 



I 



8 



C3» 









Cj> 



8 






rO 


^■j 


» 


►« 


F^ 






■^ 


<4> 




<U 


« 


cc 




'« 


!5 



« s 






fel 



F«i 



TS 



S 

ns 










^ "^ 



Breaking Load 



704 STRENGTH OF COLUMNS. JuNK 1905. 

influence of the thickness in determining the load which produces 
failure ; they have been deduced from the curves shown on Figs. 2-8, 
the diameter and thickness of the tube being varied and the area of 
the cross-section kept constant, the uppermost curve of the series 
being for the limiting case when the tube becomes a solid bar. If now 
the curves on Fig. 9 be plotted with the base line representing length 
and the breaking load for these values vertically, a series of curves 
is obtained as shown on Fig. 10 (page 703) ; from the inspection of 
these the influence of the diameter and thickness upon the length is 
very clearly shown, and from which it will be seen that the economic 
ratio of the diameter and thickness for a given cross-section depends 
upon the length of the column. 

The breaking strength to tension of several specimens cut from 
the tubes under test was determined, giving results which varied 
from 62,000 to 85,000 lbs. per square inch ; the average of these 
tests was nearly 72,000 lbs. per square inch. 

The determination of the strength to compression of the mild- 
steel tubing, that is, the strength of very short specimens, was a matter 
of some difficulty, owing to the bulging and upsetting of the 
specimens when thick, and to the wrinkling or doubling up of the 
specimens when thin ; the average of the . breaking strengths to 
compression was 80,000 lbs. per square inch. This value has been 
adopted throughout in plotting the curves on Figs. 9 and 10. 

Experiments were also made for the determination of Young's 
modulus of elasticity, giving an average value of 30,000,000 lbs. per 
square inch ; this value has been adopted throughout in plotting the 
curves on Figs. 9 and 10. 

The following phenomena were observed during the carrying out 
of the tests : 

That for all columns there is a certain value of Ijp beyond which 
the failure is sensibly due to elastic bending, the load producing 
failure varying inversely as the square of the length. 

That, as the value of pjt increased, the value of Ijp increased for 
the load producing failure by elastic bending ; the curves on Fig. 9 
show this. 

When the failure of the column takes place by secondary flexure 
for a large range of the length, there is little variation in the 



June 1905. STRENGTH OP COLUMNS. , 705 

breaking load. Hodgkinson in his experiments on wrouglit-iron 
columns has previously remarked on this. 

As the load approached the breaking load, there was on the short 
lengths a decided yielding of the tube before failure. 

Before proceeding to the discussion of the results of the 
experiments as given on the diagrams, some of the formulae in 
general use for the design of columns will be examined ; of these 
Euler's and the Kankine-Gordon formulae are the most important. 

tt'^ E I 
Euler's formula * P = — j^- 

where P = the load on the column. 

E = Young's modulus of elasticity. 

I = moment of inertia of the cross-section. 

I = length of the column. 
This formula, derived by Euler from the theory of elasticity, is 
only applicable to columns which fail by elastic bending, that is, to 
long columns. For practical columns with the usual ratio of Ijp it 
fails to give the breaking load. As will be seen from the tests of 
the tubes on Figs. 2-8, it applies with a fair degree of accuracy to 
long columns, the influence of the ratio pft causing but little variation 
in the strength. The experiments of Christie f on long columns lead 
to a similar conclusion. 



Eankine-Gordon formula J P = 



1 + e 



© 



where / = about two-thirds the compressive strength of the material, 
for wrought-iron = 36,000 lbs. 
A = area of the cross-section. 
c = -^Q^ for columns with round ends. 
= ^ei-Qo" ^^^ columns with fixed ends. 
I = length of column. 
p = radius of gyration of the cross-section. 



* A proof of this formula is given in Eankine's Applied Mechanics. 

t Transactions, American Society of Civil Engineers, 1884, vol. xiii, page 85. 

X A proof of this formula is given in Kankine's Civil Engineering. 

3 



706 STRENGTH OF COLUMNS. JuNE 1905. 

This formula gives approximately the breaking load for both 
short and long columns, the value of / and c being determined from 
experiments. For short columns the strength depends mostly on/, 
the strength of the material, the value of c (J/pY being small, and for 
long columns the strength depends on elastic bending, that is, the 
term c {l/py involving the modulus of elasticity. 

The formula takes no account of the ratio p/t, nor does it in any 
way point out the most economical values of l/p to be used, the 
ratios adopted in practice being the result of experience. 

On the diagram. Fig. 9 (page 702) the breaking load for the tubes 
calculated from Euler's formula is shown, for values of l/p > 120 for 
the solid bar to values of l/p > 240 for the thinnest tubing the 
failure is sensibly elastic, and that the experiments are practically in 
accordance with the values as calculated. In all cases of long 
columns some elastic deflection was apparent before failure ; the 
columns were then in a state of equilibrium, and it was not till the 
deflections exceeded a certain amount that collapse took place. 

It has already been remarked that the curves shown on Fig. 9 
were for purposes of comparison drawn so as to refer to tubes of 
constant cross-section. The manner of deducing these curves from 
the curves on Figs. 2-8 was as follows : — The Eankine- Gordon 
formula was used in the form 

-I- f 

by giving suitable values to / curves similar to those shown on 
Fig. 9 were obtained. From a critical examination of these curves, 
it was evident that the value of/ depended on the failure of the tube 
by secondary flexure. Now the experiments show that this secondary 
flexure does not depend on the length of the column, but only on the 
thickness and the radius of gyration. 

To determine by analysis from the mathematical theory of 
elasticity the variation of the strength in terms of these quantities is 
a complex problem, which has not hitherto been solved. It was 
therefore attacked by an induction process from the experiments ; 
from the inspection of the values obtained, it was evident that the 



June 1905. STRENGTH OF COLUMNS. 707 

load producing failure was less when the diameter was varied and 
the thickness constant, and greater when the thickness was varied 
and the mean diameter constant, showing that the strength depended 
in some way on the ratio pjt ; by plotting values of pjt and the 
breaking load the following formula was obtained : — 

F 



f = 



1 + h 



(?) 



where F = the strength of the material to compression. 

h = B. constant for mild steel = J^. 
When pjt approaches the limiting value of 0*5 for the solid bar 
the value of / is sensibly equal to F. This formula closely gives 
the required values and appears to be of the correct form, for it 
shows that the probable value of/, if deduced from the mathematical 
theory of elasticity, would be of the form 



f-^& 



B being some constant depending on Young's modulus of elasticity. 

Substituting for / in the Eankine-Gordon formula, it now takes 
the form 

putting h = q\ and c = g qL-q for columns with round ends 

This formula was used to determine the curves shown on Fig. 9, 
the value of F used being 72,000 lbs. per square inch, which may be 
considered as being sensibly the strength to compression of the 
material. 

These curves having been determined, the remaining curves, 
showing the actual failure of the columns, were plotted from the 
curves on Figs. 2-8. They relate to tubes of the following sizes 
and of constant cross-section of 0*12 square inch. 



3 2 



708 



STRENGTH OF COLUMNS. 



June 1905. 



TABLE 2. 
Particulars of tubes shown on Figs. 9 and 10 (pages 702-703). 



External 
Diameter. 


Internal 
Diameter. 


t 


P 


Reference 

to 

Curves. 


inch. 


inch. 








1 


0-92 


0-04 


0-34 


P 


0-875 


0-78 


0-048 


0-3 


E 


0-75 


0-636 


0-057 


0-25 


D 


0-625 


0-481 


0-072 


0-2 


C 


0-5 


0-3 


0-1 


0-15 


B 


0-4 





0-2 


0-1 


A 



The curves shown on Fig. 10 were derived from the curves on 
Fig. 9, as already described. The diagram, Fig. 9, thus obtained 
may be used for mild-steel columns with round ends having the same 
ratio of l/p and p/t. 

The formula takes into consideration both the primary and 
secondary flexure of the column. Since the column can fail in either 
of these two ways, it will be of maximum strength when the strength 
to primary flexure is equal to that for secondary flexure. Now the 
diagrams show the variation of p/t with the length for a constant 
cross-section, and the question at once arises if the ratio between 
these quantities can be put in a simple form so as to be of 
practical use. 

On Fig. 9 the economic values of l/p and p/t were plotted from 
the curves on Fig. 10 giving the dotted curve as shown ; this curve 
may be considered approximately to be a parabola, having the 
equation 

(i)'= 1000 (^-0-5) 

for the usual values of l/p in practice the term 0-5 can be omitted, 
giving the simplified form 

P' = tdVtt I'l- (1) 



June 1905. STRENGTH OF COLUMNS. 709 

This equation applies to the column of hollow circular cross-section ; 
for columns of other cross-sections the constant xt^Vxt requires to be 
slightly modified. 

Let A be the area of the cross-section, d the mean diameter of the 

d A 

column ; then tt dt^A approximately. Also p = -^-- ; hence t = g.^ 

approximately. Substituting in the previous equation — 

p' = "^1 
^ 94 

for hollow circular columns and for other cross-sections 

P' = ^-P (2) 

li being a constant depending on the value of pjt of the cross-section 
required. 

This equation shows that p varies as V ^ i^ ^^ ^^^^ of the cross- 
section is constant. Now A represents the area of the cross-section, and 
is supposed to be known ; to determine it, values oi p and IJp for the 
economic section were plotted, giving the curve as shown on Fig. 9, 
the equation of which is approximately — 

P F 

^ = ^ = MrwK-^ (3) 

By substitution for A from equation (2) a cubic equation is 
obtained, giving the value of p in terms of P and I and the 
constants ; this equation does not admit of easy solution, and it is 
better to proceed by a tentative method. Assume a probable value of 
Ijp for the column, usually about sixty to ninety for the ordinary run 
of columns, and solving equation (3) the value obtained is 

^ = 2 to 3 C 
F 

where P is the load on the column and F the breaking strength to 
compression of the material ; this must be divided by the factor of 
safety for the practical column, giving 

^ = 2 to 3 (^) 
where S = the factor of safety. 



710 STRENGTH OF COLUMNS. JUNE 1905. 

Substituting for A in equation 2 the economic values of p can be 
immediately determined. 

The dimensions of the column having been thus ascertained, the 
value of the safe loads P can be checked by the general formula, 
owing to the curves of the formula not coinciding with the curves of 
the actual failure of the columns ; the value obtained will, in general, 
be less than P, an error on the side of safety. 

The experimental work on columns hitherto carried out has 
been incomplete, owing to the limited range of the experiments ; this 
has led to considerable diversity of opinion, more especially by 
American observers, as to the value of the constants that should be 
used in the Rankine- Gordon formula. This is to be expected in view 
of the results of this investigation, many of the experiments being 
made on columns of different proportions, the results varied giving 
different values for the constants in each case. 

The general formula applies to columns of mild steel and 
wrought-iron ; the constants for wrought-iron given by Rankine are 
/ = 36,000 lbs. and c = goVu ^^^ columns with round ends. 
Assuming the strength of the wrought-iron to compression to be 
24 to 25 tons per square inch, this formula gives a value of djt of 
about 16 and for p/t of 5*5 to 6. Now these constants were derived 
from Hodgkinson's experiments on columns, of which the average 
ratio of d/t was approximately 16 ; hence the formula may be 
looked upon as being confirmed by these experiments. For the 
ordinary ratios of p/t in practice the Rankine-Gordon formula gives 
fair results ; it does not however apply with the same accuracy to 
solid bars or columns in which the ratio of p/t is large ; this will be 
evident from the inspection of the curves on Fig. 9, one of these for 
which / = 42,000 lbs. gives good average values, the remaining 
curves giving values which are either too high or too low respectively 
for values of Z/p, at which failure takes place by elastic bending. 
This discrepancy arises from the fact that in the formula p varies 
directly with /, that is, the load which a column will carry is 
supposed to be proportional to the strength of the material ; this is 
not true for long columns, it is the modulus of elasticity which 
governs its strength, and it is the constant c wLich involves this 



June 1905. STRENGTH OF COLUMNS. 711 

quantity. Now there is not much difference between the modulus of 
elasticity of wrought-iron and mild steel, hence the load which 
produces elastic failure will not differ much for either material. The 
experiments of Christie and others confirm this. It has already 
been remarked that Euler's formula is sensibly true for long struts ; 
granting this, it would be more logical to assume that 

i + Hp) 

for some particular value of //p, in which the failure takes place by 
elastic bending ; this gives 

where m is some particular value of Ijp. If this is done, fairly 
accurate values for the curves shown on Fig. 9 can be determined 
from the equation 

where m and a are constants and the value of 

F 



f = 



1 + 



^(^) 



For the present the author is of the opinion that it is better to 
use the more well-known form of the Eankine-Gordon formula 
already given until more experimental data has been obtained for the 
determination of the constants in the above formulas. 

In the literature of the subject referred to in the Appendix (pages 
712-713), various formulae are given for determining the strength of 
columns, none of which take into consideration the economic values 
of the radius of gj ration and thickness for a given length ; tested by 
the direct appeal to experiment this investigation shows that the 
values obtained from their use can only apply to comparatively few 
cases. 

The Paper is illustrated by 10 Figs, in the letterpress, and is 
accompanied by an Appendix. 



712 



STRENGTH OF COLUMNS. 



June 1905. 



APPENDIX. 



Bibliography. 
(arranged chronologically.) 

Transactions, Royal Society, 1840, page 227. 
On tlie Strength of Iron and Steel. Proceedings. 

The Institution of Civil Engineers, 1869-70, 

vol. XXX, page 215. 
Iron Bridges of very large span for Railway 

Traffic. Proceedings, The Institution of Civil 

Engineers, 1877-78, vol. liv, page 179. 
Investigation on the Strength of Columns. 

Civil Ingenieur, 1878, vol. xxiv, page 17. 
On the Constants in Gordon's formula. Journal, 

Franklin Institute, 1882, vol. cxiii, page 58. 
Experiments on the Strength of Wrought-Iron 

in Struts. Transactions, American Society 

of Civil Engineers, 1884, vol, xiii, page 85. 
On Practical Strength of Columns. Proceedings, 

The Institution of Civil Engineers, 1885-86, 

vol. Ixxxvi, page 261. 
A new Formula for Compression Members. 

Transactions, American Society of Civil 

Engineers, 1886, vol. xv, page 537. 
Struts. " The Engineer," 10 and 24 December 

1886, pages 464 and 513. 
Struts. « The Engineer," 14 and 28 October 

1887, pages 303 and 345. 

Wheatley and Wood Stiffness of Struts. " The Engineer," 6 January 

1888, page 1. 
On the Strength of Struts. Annales des Ponts 

et Chaussees, April 1894, page 498. 
New Formula for the Strength of Columns. 
Journal of the Association of Engineering 
Societies, 1898, vol. 20, page 239. 



E. Hodgkinson 
G. Berkley . 



T. C. Clarke 



E. Winkler 



M. Mbrriman 



J. Christie 



T. C. FiDLER 



R. Krohn 



Ayrton and Perry 



Robert H. Smith . 



De Pr^audeau 



C. G. Barth 



June 1905. 



STRENGTH OF COLUMNS. 



713 



J. M. MONORIEFF 



F. WiTTENBAUER 



J. M. MONCRIEFF 



F. WiTTENBAUER 



E. A. Neville 



APPENDIX— cowimiiec?. 

The Practical Column under Central and 
Excentric Loads. Transactions, American 
Society of Civil Engineers, 1900, vol. xlv, 

* page 334, 

Strength of Struts. Zeitschrift des Vereines 
deutscher Ingenieure, 1902, vol. 46, page 501. 

The Practical Strength of Columns or Struts. 
« Engineering," 6 June 1902, page 731. 

Strength of Struts. Zeitschrift des Vereines 
deutscher Ingenieure, 1903, vol. 47, page 245. 

Note on Euler's Formula. Technical Paper 
No. 129. Government of India. 



Discussion. 

Mr. J. Hartley Wioksteed, Past-President, as there was no 
time for the discussion of the Paper, moved a vote of thankd to his 
friend Professor Lilly, who should feel very gratified that such a 
purely scientific Paper had been brought before the meeting. It was a 
communication that was not evanescent, and the fact that there was no 
time for its discussion was of very little moment considering the 
permanent importance the Paper possessed. The 800 experiments had 
been so thoroughly made that they would no doubt be of great value 
for a long time to come. He hoped the author before long would be 
able to repeat similar experiments with fixed ends. He had never come 
across any such complete experiments in comparisons made between 
pillars with rounded ends and pillars with fixed ends. After such 
experiments the author's ratio between the diameter and the thickness 
of the sides to prevent crumbling might have again to be changed 
and restudied for columns with fixed ends. He had seen Professor 
Lilly's method of conducting tests, and had never seen anyone make 
a greater use of the specimens going through his hands. They were 
tested in every sense : they were tested for modulus of elasticity ; 



714 STRENGTH OF COLUMNS. JUNE 1905. 

(Mr. J. Hartley Wicksteed.) 

yield point and reduction of area ; th» y were tested in tension, in 
compression, in deflection, and in torsion. Such a method not only 
made a most exhaustive investigation into the characteristics of the 
material — an investigation, he ventured to say, which would render 
new tests of impact unnecessary, because the characteristics of the 
material had been exhausted — but they formed of course an excellent 
training for the students of Trinity College, Dublin, where Professor 
Lilly had charge of the mechanical engineering. 



Communication. 



Professor Egbert H. Smith wrote that in the text of his Paper 
Professor Lilly referred to the Euler and the Eankine-Gordon 
formulae, the latter of which he used in a modified form intended to 
take account of the danger of wrinkling or buckling of over-thin 
tubes. This modification must necessarily make it more representative 
of actual results obtained by experiment and constructional practice. 
Whatever merit the Lewis-Gordon rule might have, it must be regarded 
as entirely empirical. Few formulae had been used in engineering 
whidh were equally irrational in their deduction. A minor initial 
fault was the assumption that, because in each special case of purely 
transverse, or beam, bending the deflection bore a definite proportion 

to the stress multiplied by - (following Professor Lilly's 

nomenclature), therefore this was also the case in bending by purely 
end thrust. This would be true for small deflections on the assumption 
that the shape of the curve of the bent strut — and therefore the shape 
of the bending moment diagram — remained the same for all loads, all 
lengths and all sizes of section, and therefore for all deflections ; 
but only under this assumption. Dealing with perfectly straight struts, 
this assumption apj)roximately corresponded with fact, so long as the 
end load was accurately applied in the centre of each end section ; 



June 1905. STRENGTH OF COLUMNS. 715 

but no manner of jointing the ends of struts to the parts they 
supported would secure this true centering of the end thrusts, except 
very special devices sometimes used in testing machines. In 
structures and in machines the actual eccentricity of the end thrusts 
unavoidably varied with increase of load. 

If p be the severest stress produced by bending, the total stress 

f = -\- P z= -^ }\ -\- c -A if the deflection be taken in constant 
proportion to -, because /5 = bending moment x 2j = P8 X 27 while 

i'^ Ap' This is the actual rational (or irrational) derivation of 

the Gordon formula. In pure transverse-load bending, the severest 
stress due to bending ought, of course, to be the safe stress for the 
material, and should, therefore, be constant under all loads, lengths 
and sections. But in struts it was the sum of this and the 
average -j or p that must not rise above " safe " stress, and which 

sum, therefore, should be constant. The greater the ratio of ^ the lower 

P 

must - be put, and, therefore, the higher may (3 be made. And 
since the deflection is proportional to /^ -, its proportion to - 
increases with - instead of remaining constant. 

Indeed, since (3 = PS ' —, where 8 is the deflection and I is 
the moment of inertia of the section, by inserting the above value 
of 8, the result yielded is /3 oc /3 P , which is a " high equation " 
as regards y8, and gives P oc ^^, or the same result as Euler's formula. 

Otherwise viewed, putting the deflection in constant proportion to 
- is equivalent to making /? constant, and if both /3 and / are 

constant, then must - be also constant. Thus three results, 

A 

absolutely inconsistent with each other, are obtained from the 
formula viewed in three different ways. It is honeycombed with 
illogical inconsistency. 

In Euler's formula there is an initial fault in the premises, and 
there is, at the end of the investigation, an error of interpretation 



716 STRENGTH OF COLUMNS. June 1905. 

(Professor Robert H. Smith.) 

of the mathematical result; but the mathematical reasoning and 
deduction of this result are strictly accurate. It is an illustration of 
how greatly superior is the mathematical ability devoted to such 
investigations over the physical knowledge of engineering fact 
contributed to their consideration. The fault in the premises is the 
assumption of zero eccentricity of application of the'end thrusts and 
the overlooking of the fact that only a very minute, but fiuite, 
eccentricity changes the mathematical result very largely. This 

Fig. 11. — Compression of a Column. 
Diagram showing influence of eccentricity. 




Fig. 12. — Two variations of form if h\— 0. 



can be most simply explained by the diagram, Fig. 11. Here A B 
represent the centres of the end sections, each lying S away from 
the line P P of the end thrust, and the length of the strut A B is 
marked I. The Euler formula P = -tt^ ^ is correct if there be 
inserted in its divisor, not the actual length I of the strut, but the 
length Li between the two intersections on the thrust line P P of 
the curve to which the strut bends produced beyond A and B. If 
the strut is bent excessively, as shown by the thick curve-line 
intersecting P P on a length L^, then, since its end inclinations 
are large and 8 is small, therefore L^ does not greatly exceed Z, and 
using I in the formula instead of L^ does not involve very large 
error. But if the strut be bent only very slightly, as shown by the 



June 1905. STRENGTH OF COLUMNS. 717 

fine curve giving L2 as the intersecting length, then L2 is greatly 
in excess of I even though 8 may be very minute. In this case 
putting I instead of L2 in the formula would give P, the breaking 
load, many times greater than it really is. The diagram, Fig. 11, 
does not give room to show L2 very long; but it will be easily 
recognised that in an ordinary stiff column that is not supposed to 
buckle measurably, and in which the bending curve is extremely flat, 
an eccentricity 8 of not more than say ^V*^ ^^^^ ^^7 i^aake L2 five 
to ten times as long as Z, and thus Lg^ twenty-five to one hundred 
times greater than P. 

The Euler investigation was carried out with the eccentricity or 
8 reckoned as zero ; but this is a special and ambiguous case of the 
general mathematical case which can only be correctly interpreted 
as the limit of the general result when 8 becomes smaller and 
smaller. When the] general case is worked out and the special case 
so interpreted, it is found that the case 8 = only represents the 
condition, two variations of which are shown by the sketches in 
Fig. 12 (page 716). The curve of the bent strut is a sinuous and 
entirely unstable one, cutting through the P P line at a number 
of pairs of nodal points. The slightest shock would make the strut 
spring from such a position into a single bent bow, after which it 
would certainly collapse under the load prescribed by the formula. 
It is a serious mistake to apply a mathematical result, the real 
meaning of which is a condition of fatal instaoility, as a guide to the 
safe load to place upon columns. 

These objections to Euler's and Gordon's formulfe were stated in 
a lengthy mathematical 'analysis made by the present writer in a 
Paper read before the Edinburgh and Leith Engineering Society 
in 1878. After re-examination he does not find any error in the 
analysis then given. 

In three articles in " The Engineer" of Idth and 28th October 
and 25th November of 1887, a resume of the formulae obtained by 
different investigators was given. Professor J. Perry's, Professor 
Fiddler's and the writer's methods being compared ; and on 
6th January 1888, in the same journal, some account of numerous 
experiments made at Mason College to test these theories appeared. 



718 STRENGTH OF COLUMNS. June 1905. 

(Professor Robert H. Smith.) 

Professor Perry's theory is practically identical witli that of the 
writer, except that he substitutes for the trigonometrical function an 
approximately equivalent algebraic function. This substitution, 
however, does not make the formula really easier to use for practical 
calculation in design. 

Using the same mathematics as Euler's, corrected to take account 
of eccentricity in the end thrust, for the construction of curves with 

mean stress as ordinate and - as co-ordinate, curves of the same 

P 

shape are obtained as those from the numerous tests made by many 
experimentalists in various countries. The formula for this mean 
stress is 

p / 

P = 2 = 



1+^14-0 Viz) 



To show how the thrust-eccentricity 8 modifies this from Euler's 
result, let P^ represent the load by the Euler formula ; then the 
above is 

If the actual load P were made equal to P^, then, since sec - = oo , 

/ 1 A ^ 

it would be found that /=p ( 1 -1- ^ 8 oo j or the maximum stress 

infinitely greater than the average stress, unless 8 be zero with 
mathematical exactitude, in which case the ratio is mathematically 

indeterminate. 

h A 
It may be noted that in the above -j- is, for each prescribed 

shape of section, proportional to -, so that the factor of the secant is 

proportional to the ratio -. For a given amount of want of 

homogeneity in the material, or a given degree of want of exactitude 

in the workmanship, there is a probable constancy in the ratio - 

in designs where 8 is meant to be zero. 

This formula in this shape shares with all the other formulaa 
that have been suggested the defect that, while it enables one to 



June 1905. STRENGTH OF COLUMNS. 719 

calculate the maximum stress imposed by a given load upon an 
already designed column, it fails to afford any direct means of 
calculating the required size of safe section for the load. But in 
1 887 the writer showed how it might be transformed so as to enable 
this to be done. It is most convenient to write the sectional area 
and moment of inertia as equal to the square and the fourth power 
of the outside cross-dimension multiplied by numerical factors, these 
factors depending solely on the shape and not on the size of the 
section. Using a and i for these factors, or A = ah'^ and I = i /i*, 

and writing ^ = e, the transformed formula is 

X = X sec ^ (o- (x - 1)} and ^ = X J 

where — 

P = load to be carried, 

/ = safe maximum stress on the material, 

2i 
o- = — , 

ea 

Y = ratio of/ to mean stress = <-, 
X = -/^ 

A/PJL'i 

If the shape of section be selected, and e be either known or its 
maximum probable amount estimated, then both A. and or can at once 
be directly calculated, and from them the above equation gives x by 
direct calculation. From x tlie necessary sectional area is at once 
obtained. 

Fig. 13 * (page 720) is drawn to facilitate this calculation. X is the 
horizontal and x ^^^ vertical ordinate, the curves being drawn for a 
regular series of values of cr between which it is easy to interpolate. 

Figs. 14 and 15 (page 720) set out the results upon a base 

equal to -, which, or , has been most usually adopted in diagrams 

/v P 

of results of tests. In Fig. 14 each curve is for a specific value 
of 0-, and the vertical ordinate is the ratio of mean to maximum 
stress. In Fig. 15 each curve is for a specific value of this ratio 



* Fig. 13 is reproduced from a large-scale working diagram, and many 
intermediate curves are omitted from the reproduction for the sake of clearness. 



720 STRENGTH OF COLUMNS. 

(Professor Robert H. Smith.) 



Fig. 13. — Diagram giving x **» terms of \ in the equation for Struti. 



June 1905. 




500 



Junk 1905. STRENGTH OF COLUMNS. 721 

of mean to maximum stress, while the vertical heights measure the 
corresponding values of o-. But since is not among the data of 

any practical problem in design, these curves with ^ as base cannot 
be used directly. Fig. 13 alone can be so used. 

Professor Lilly's investigation of the influence of ratio of thickness 
to diameter of round tube columns was most interesting, and seemed 
now to have been attacked by him experimentally for the first time. 
In his Fig. 9 (page 702) the cross-sectional area is stated to be the 
same for the different curves. For this shape of * varies from ^^ for 

a solid bar to |- for a tube whose thickness is very minute in 
proportion to its diameter, which makes o- vary in the proportion 
of 1 to 2. Professor Lilly's Fig. 9 might therefore be compared 
with any pair of curves in the writer's Fig. 13 (page 720) with o- in 
this ratio ; only on page 704 Professor Lilly stated that his uppermost 
curve corresponded with the case of a solid bar. This was in 
accordance with his equation on page 707, but it seemed strange that 
the solid bar, reckoned the least economical, should bear without 
breaking the greatest load per unit of sectional area. Tlie 
experiments tabled in " The Engineer " of 6th January 1888 were 
also mostly upon steel tubes with various ratios between thickness 
and diameter, and their results should furnish further data for 
Professor Lilly's investigation. The writer's curves did not give 
the most economical ratio of thickness to diameter, because they 
took no account of liability to failure by wrinkling. The most 
interesting point in Professor Lilly's Paper was that his experiments 
showed that this failure by wrinkling did not depend upon the 
length but upon the ratio of thickness to cross diameter. This was 
what one would expect, except for very short lengths. 

Professor Lilly wrote that it was gratifying to him to have had 
Professor Smith's criticism on his Paper, and to know of the valuable 
work done by Professor Perry and himself. Since the publication 
of the Paper other communications on the work done by Moncrieff 
and others had been received, particulars of which were given in the 
Appendix (pages 712-713). 

3 D 



722 STRENGTH OF COLUMNS. June 1C05, 

With regard to the remarks that had been made on the Rankine- 
Gordon formula, he was of the opinion, in spite of its defects and 
of its being more or less empirical, that it was still the best formula 
to use for the tentative design of struts or columns, and further it 
had the great advantage of being easy in its application for the 
ordinary run of columns in practice. Mathematical reasoning which 
was not borne out by experimental data required that the hypothesis 
on which the reasoning was based should be carefully scrutinised. 
The formula of Professor Smith and others did not give results which 
were consistent with his experiments. The eflfect of secondary 
flexure had not been considered, and the hypothesis on which the 
formulae were based required revision in this respect ; for that reason 
he preferred to retain for the present the Rankine-Gordon formula. 
When more complete experimental data had been determined, he 
hoped to return to the subject and would then give careful attention 
to Professor Smith's work. 



June 1905. 723 



EXCURSIONS.* 

On Tuesday Afternoon, 20tli June, after luncheon at the 
Restaurant Lisansky, in the Exhibition grounds, a visit was made 
by a large party of the Members to the Works of the Societe John 
Cockerill at Seraing, under the guidance of M. A. Greiner, General 
Manager, the other officials of the Company, and Professor H. 
Hubert. Another party visited the Works of the Societe d'Ougree- 
Marihaye at agree ; whilst a third party visited the engineering 
exhibits at the Exhibition. The following Works were also open 
for inspection : — 

Ateliers de Construction de la Mouse, Sclessin. 
Station Centrale Electrique, Sclessin. 
Ateliers Fr^de'ric Recq de Malzine, Sclessin. 
Societe Anonyme des Fonderies Ketin, Sclessin. 
Acieries d'Angleur, Tilleur. 
Maison Beer, Sclessin. 

The following Works were open during the Meeting ; — 

Ateliers de Construction Me'canique de Longdoz. 

Ecole Professionelle de Mecanique. 

Societe' Anonyme Liegeoise pour la Construction de Machines. 

In the evening the Members were entertained at a Banqnet in 
the Renommee Hall, Liege, by the Liege Association of Engineers, 
the company numbering over 350. The chair was occupied by 
Professor Alfred Habets, President of the Association, who was 
supported by the Past-Presidents, Vice-Presidents, Council, and 
Members of the Association. After the Chairman had proposed the 

* The notices here given of the various Works, &c., visited in connection 
with the Meeting, were kindly supplied for the information of the Members by 
the respective authorities or proprietors. 

3 D 2 



724 BANQUET IN LIEGE. June 1905. 

joint toast of " King Edward VII and King Leopold II," M. Louis 
Canon-Legrand, Vice-President of the Liege Association, proposed 
the toasts of the " President of the Institution of Mechanical 
Engineers " and of the " President of the South Wales Institute of 
Engineers," which were suitably acknowledged. M. Charles Thonet, 
Vice-President of the Liege Association of Engineers, then proposed 
the toast of " The Ladies." The final toast of " The Liege Association 
of Engineers " was proposed by the President of the Institution of 
Mechanical Engineers, and enthusiastically honoured. 

During the evening, the ladies accompanying the British 
engineers were entertained at Dinner at the Hotel d'Angleterre by 
the Committee of Belgian Ladies. Subsequently they proceeded to 
the Renommee Hall to listen to the speeches at the Banquet. 



On Wednesday Afternoon, 21st June, after luncheon at the 
Hotel d'Angleterre, a visit was made, under the guidance of 
Professor H. Hubert, to the Fabrique Nationale d'Armes de Guerre, 
at Herstal. Subsequently the following Works were visited : — 
Chaudronneries Piedboeuf (Boilers) ; Ateliers de St. Leonard 
(Machines) ; and Compagnie Internationale d'Electricite. 

In the evening a " Fete de Nuit " was held at the Exhibition, to 
which the Members and Ladies were invited by the President, M. 
Emile Digneffe, and Executive Committee of the Liege Exhibition. 
A Concert by the Grand Orchestra took place in the grounds, and 
was followed by a Venetian Fete on the River Mouse. 



On Thursday, 22nd June, three alternative Excursions were 
made. 

One was to Spa, whence a large party drove to the Barrage de la 
Gileppe. After luncheon at the Hotel du Lion, a visit was made to 
the Barrage (page 774), the construction of which was described by 
M. E. Detienne and M. F. Leclercq. The return journey to Spa 
was made via Goe, Verviers Fontaine, and Sarister. 



June 1905. EXCURSIONS. 725 

Another visit was by train to Ans to visit the Charbonnage de 
I'Esperance, where electric winding-machines are installed, as 
described in Professor P. Habet's Paper (page 429). Some of the 
party visited also the Cream Separator Works of M. Jules Melotte, 
at Kemicourt. 

A third party visited the Charbonnage du Hasard, at Eetinne. 

In the afternoon, Professor H. Hubert conducted some of the 
Members over the Mechanical and Electrical Laboratories of the 
University of Liege. Subsequently a visit was made to the ficole 
Professionelle de Mecanique. 

In the evening the Institution Dinner was held at the Hotel 
Britannique, Spa, and was well attended by Members and Ladies ; 
the Eeception Committee of the Liege Engineers and their Ladies 
were the Guests of the Institution on this occasion. The President 
occupied the chair; and the following Guests accepted the invitations 
sent to them, although those to whose name an asterisk (*) is 
prefixed were unavoidably prevented at the last from being present : — 

Liege Beception Committee. — President, Professor Alfred Habets ; 
Vice-Presidents, M. Louis Canon-Legrand ; M. Charles I'honet ; 
Secretary, M. Eene d'Andrimont. Professor Henri Dechamps ; 
*M. Xavier de Spirlet ; M. Adolphe Greiner ; Professor Herman 
Hubert; *M. Charles Legrand; M. Jules Magory; M. Auguste 
Eaze ; M. Constant Eenson ; M. Louis Schaeffer ; *M. Carlo 
Spruyt ; Professor Armand Stevart ; M. Gustavo Vandewyer. 

M. Louis Fraigneux, Echeviu des Travaux Publics, Liege ; 
Professor Ernest Mahaim, Special Government Commissioner for 
the Exhibition Congress ; *M. Paul Forgeur, Secretary of the 
Executive Committee of the Exhibition ; M. L. Lonneux, Technical 
Director of the Exhibition ; M. Aug. Dumoulin, Director-General 
of the Exhibition; *M. G. Simonis, Assistant Secretary of the 
Exhibition Congress ; M. Nyst, Member of the Executive Committee 
of the Exhibition. 

M. Pol Boel; M. C. Bourgy; *M. F. Claessens; *M. Evence 
Coppee ; *M. E. Detienne ; M. le Baron Edgar Forgeur ; M. 
FrankignouUe ; M. Marcel Habets; Mr. E. M. Hann, President of 



726 INSTITUTION DINNER. June 1905. 

the South Wales Institute of Engineers ; *Mr. Imre Kiralfy ; M. F. 
Kraft ; M. Felix Leclercq ; M. Lhoest ; M. Renault ; M. J. Souheur ; 
M. Tonneau ; M. G. Trasenster. 

Professor Paul Habets ; Professor W. E. Lilly ; *M. Eodol])he 
Mathot ; Mr. A. L. Mellanby ; M. Ed. Noaillon. 

The President was supported by the following Officers of the 
Institution : — Past-Presidents, Mr. William H. Maw, *Mr. E. Windsor 
Richards, and Mr. J. Hartley Wicksteed. Vice-President, Mr. 
Edward B. Ellington ; Members of Council, *Mr. John F. Robinson 
and Mr. Mark Robinson. 

After the toasts of " King Edward VII and King Leopold II " 
had been proposed from the chair, and most enthusiastically 
applauded, the Pbesident proposed that of the " City of Liege," 
which was acknowledged by M. Louis Fraigneux, Echevin des 
Travaux Publics, in the absence of Bourgmestre Kleyer. 

The toast of " The Ladies and Gentlemen of the Reception 
Committee of the Liege Association of Engineers " was proposed by 
Mr. Edward B. Ellington, Vice-President, and acknowledged by 
Professor Alfred Habets. 

Mr. William H. Maw, Past-President, proposed the toast of 
" L'Exposition Universelle et Internationale de Liege," which was 
acknowledged by M. Nyst, Member of the Executive Committee of 
the Exhibition. 

The concluding toast of " The Institution of Mechanical 
Engineers " was proposed by M. Adolphe Greiner, and acknowledged 
by the President. 



On Friday, 23rd June, the Members proceeded to Antwerp by 
special train. On arrival at the South Station, they were met by M. 
G. A. Royers, Chief Engineer of the Municipality of Antwerp, and 
M. Carlo Spruyt, under whose guidance they visited the Docks and 
Quays (page 763). The Ladies were conducted in parties to view 
the Cathedral, Museums, and other places of interest. The following 
Works were also open for the inspection of Members through the 
arrangements of M. Spruyt and M. Cruysmans. 



June 1905. VISIT TO ANTWERP. 727 

Central Railway Station and Electrical Signal Appliances. 

Cockerill Co.'s Shipyard, Hoboken. 

Vicinal Bailway Station. 

Minerva Motor Works. 

Taillerie Populaire Anversoise (Diamond Cutting Works), 



On Friday, 23rd, and Saturday, 24th June, the following Works 
were open : — 

Brussels. 

Ateliers de Construction H. Bollinckx. 

Electric Station, Rue St. Catherine. 

Societe Anonyme du Canal et des Installations Maritimes. 

Tramway Central Power Station. 

Vve. Louis de Naeyer et Cie (Paper and Boiler Works). 

Chakleroi. 

Societe Anonyme des Ateliers Germain, Monceau-sur-Sambre. 
Society Anonyme des Forges, Usines, et Fonderies de Haine St. Pierre. 
Societe Anonyme de Marcinelle et Couillet, Couillet. 
Socie'te Anonyme des Constructions Electriques. 

Ghent. 

Societe Anonyme des Anciens Ateliers de Construction Van den Kerchove, 
Society Anonyme du Phoenix. 

Malines. 
Locomotive Works of the Belgian State Railways. 

Mons. 

Ateliers de Construction et Chaudronneries B. Lebrun, Nimy. 
Charbonnages et Ateliers de Construction du Grand Hornu, Hornu. 



Junk 1905. 729 



ANGLEUR STEEL WORKS, 
TILLEUR, NEAR LIEGE. 

The TiLLEUR Factory comprises the following : — Coke 
ovens; blast-furnaces; Thomas steel works containing rolling 
mills for Vignole rails, girder rails, beams, various gauged and 
merchant steels; workshops for the manufacture of permanent 
way material, bridges and frame work, and basic slag works. 
There are also coke ovens comprising 84 regenerative ovens on 
the Evence Coppee system, having an annual production of 
120,000 tons. The blast-furnaces are four in number, of which 
three are in blast. Their annual production amounts to 150,000 
tons of pig for Thomas process steel. The steel works have three 
converters of 9 J tons with circular pits, and the annual production 
of these is about 130,000 tons. A new steel works, with square 
pit and four converters of 12 tons, is being constructed and will be 
working by the end of 1905. The estimated production is 250,000 
tons. 

The Rolling Mills contain two 30-inch diameter thr^e-high 
trains, one 22-inch three-high train, one 12-inch three-high train 
with roughing rolls, and one 10-inch three-high roughing train. 
The annual production at present exceeds 120,000 tons, but the 
estimated yield by the new steel works is expected to give 200,000 
tons. The factory for the construction of permanent- way material 
comprises machinery for producing railway and tram switches, 
bridges and constructional work. The annual production amounts 
to 6,000 tons. The basic slag works have a capacity of 70,000 tons 
annually. 

Angleur Factory comprises Bessemer acid steel works for 
special articles, and Siemens basic steel works for axles, tyres, 
springs, &c. Also the following : — Small converters for steel ; 
foundry ; forge ; rolling mill for tyres ; one 22-inch three - 
high train for sections and bars. One two-high train for 
spring-plate bars. The annual production amounts to 30,000 
tons. The number of men employed at the two Works is about 
3,000. 



730 June 1905. 



ATELIERS DE LA MEUSE, 
SCLESSIN, NEAR LIEGE. 

This Company (Societe Anonyme des Ateliers de Construction 
de la Meuse) had its origin in the old establishments of M. Charles 
Marcellis, founded at Liege in 1835. It was converted into a 
company (Societe Anonyme) in 1872. The works extend over an 
area of 13^ acres and are organised for the manufacture of general 
articles of machinery. Employment is given to 1,000 workmen 
and about fifty engineers and draughtsmen. Amongst the most 
important productions of this firm are locomotives, steam- 
engines, winding engines, air compressors, pumps for mines, 
engines for metallurgical works, for blast-furnaces, iron and steel 
works, etc. The average annual production amounts to a turnover 
of about 6 million francs (£240,000), and the greater part of this 
production is exported to other countries. The Society of the 
Meuse has obtained the highest awards at the Universal 
Exhibition of Brussels in 1897, at the Paris Exhibition in 1900, 
and at the St. Louis Exhibition in 1904 it was classed ho7's concours 
and formed part of the jury. 

This Company is exhibiting the following machines at the 
International Exhibition in Liege : — 

(1) A compound tandem-engine and generator at 110 
revolutions, with equilibrium valves and variable expansion, wit]^ 
a governor at both cylinders. This engine was built to work with 
superheated steam and actuates a generator of 1,875 amperes at 
240 volts, placed immediately upon the shaft of the engine. 

(2) A winding-engine with patent valves of the company's 
design and make for the raising of 7,714 lbs. at a velocity of 
49 feet per second from a depth of 3,936 feet. 

(3) A dry air compressor with superposed compression with 
steam compound engine of the type special to the firm, with 
distribution by valves and by variable expansion. 



June 1905. ATELIERS DE LA MEUSE, SCLESSIN. 731 

(4) A high-pressure pump for deep mines, actuated by a 
belt-driven three-phase electric motor. 

(5) A 4-cyliuder high-pressure express locomotive, with 
6-coupled wheels of 6 feet 6 inches diameter with leading bogie, 
provided with a steam superheater. 

(6) A tank locomotive with 6-coupled wheels of 3 feet 3| inches 
diameter for ordinary gauge and provided with a screw and steam 
brake. Its weight when empty is 29 tons. 

(7) A small tank locomotive for railway engine for narrow 
gauge (19f inches) with 4-coupled wheels of 19 J inches diameter. 
Its weight when empty is 5 J tons. 

(8) A portable oil-engine with electric generator for provisional 
lighting purposes. 

At the time of the visit of the Members to these works, there 
will be seen in various stages of manufacture several locomotives 
as well as blowing-engines, converters, winding-engines, etc. 



CENTHAL ELECTRIC TRAMWAY ROWER-STATION, 

LIEGE. 

This large station, situated on the left bank of the River 
Meuse, near Sclessin, is under construction, and will contain 
several steam-turbines of 2,500 H.P. each. It is hoped that the 
power-station will be ready for inspection during the visit to Liege. 



JOHN COCKERILL SOCIETY'S WORKS, 
SERAING. 

(See Plan, page 733.) 

These Works are situated on the River Meuse, about six miles 
above Liege. They were established in 1817 by John Cockerill, 
who in that year received from the King of Holland the palace of 



732 JOHN COCKERILL SOCIETY, SERAING. JuNE 1905. 

Seraing and its dependencies, with permission to start new 
workshops there for the manufacture of machinery, and of flax- 
spinning by the processes which were then being introduced into 
the country. The establishment at Seraing is the development of 
the work done by Cockerill the father, at Liege, from 1802 to 
1813, and by James and John Cockerill after that date. Between 
1818 and 1823 many engines were made for spinning mills, also 
colliery winding and pumping engines. In 1824 the first rails 
made on the Continent were rolled here, and the first locomotive 
constructed. In 1842, two years after the death of John Cockerill, 
a limited company was formed to carry on the establishments. 
From 1849 to 1851 about 4,000 men were employed, and a large 
number of steamboats, stationary and locomotive engines were 
turned out. The i^eriod 1857 to 1861 was marked with a greatly 
increased output of the foregoing products, also the manufacture 
of the boring machinery for the Mont Cenis Tunnel. In 1863 
Bessemer converters were started at Seraing, the company being 
the first to introduce this process on the Continent. 

Since that time the works have steadily increased, and at the 
present date the company employs nearly 10,000 men, runs 381 
engines developing 23,000 H.P., and has an output yearly of about 
250,000 tons of coal, 122,000 tons of coke, 295,000 tons of minerals, 
216,000 tons of pig-iron, 157,000 tons of various steels, tyres, rails, 
gun tubes, etc., 39,000 tons of joists, bars, sections, plates and shell, 
6,900 tons of iron and steel castings, 1,600 tons of forgings, 8,900 
tons of marine and stationary engines, gas-engines, locomotives, 
and guns, 8,600 tons of boiler and bridge work, etc., and 4,000 tons 
of forged iron and steel wheels, axles, castings, gun mountings, 
projectiles, boiler mountings, etc. 

Recently the Director- General of the Company, M. A. Greiner, 
on the advice of Messrs. Bailly and Kraft, two engineers of the 
Company, in collaboration with M. Delamare-Deboutteville, has 
installed gas-engines of great power, utilising directly' the gases 
from the blast-furnaces. The great value of this innovation consists 
in the fact that not only does it reduce the cost of pig manufacture 
but it obviates the use of boilers, and utilises the enormous quantity 



June 1905. 



JOHN COCKERILL SOCIETY, SEUAIXG. 



733 




1. 


Mansion. 


23. 


Blowing Engines. 


44 


2. 


Park. 


24. 


9 t >» 




3. 


Workshop No. 1. 


25. 


Large Foundry. 


45 


4. 


New Warehouse. 


26. 


Small „ 




5. 


Gun Factory. 


27. 


Copper ,, 


46 


6. 


Small Erecting Shop. 


28. 


Blast Furnaces, Old Whilwell 


47 


7, 


Loco. Factories. 




ovens. 


48 


8. 


Large Electing Shop, 


29. 


Bessemer Pits, Bknug. Eng. 


49 


9. 


Bolt Shop. 




and Rail-Mill. 


5Q 


10. 


Fire Pumps. 


30. 


Canister Store. 


51 


11. 


Electric Light Station. 


31. 


Tyres. 


52 


12. 


Offices. (^Iron Works). 


32. 


Steel Rolling Mills. 


53 


13. 


Iron Stores. 


33. 


Siemens-Martin Fhirnaces. 


54 


14. 


Repairing Shop. 


34. 


Repairing Shop. 


55 


15. 


Iron Stores. 


35. 


Steel Works Offices. 


56 


16. 


Plate Mill. 


36. 


Pharmacy. 


57 


17. 


Rod Mill. 


37. 


Laboratory. 


53. 


18. 


Puddling Furnaces. 


38. 


Coke Store. 


59 


19. 


New Boilers for Blast 


39. 


Mineral Store. 


60 




Furnaces. 


40. 


New Cupolas. 


61. 


20. 


Offices (Blast Furnaces). 


41. 


New Blast Furnaces. 


62. 


21. 


Horiz. Blowing Engine. 


42. 


Coppersmiths. 


63 


22. 


Old Boilers for Blast 
Furnaces. 


43. 


Offices. 


61 



200- fOM Fffrging Press 
Hammer (Large Forge). 

Crucible Ovens (Small 
F'orges). 

Bridge Shop. 

Coppersmiths' Hearths. 

Forwarding Waj-ehotise. 

General Stores. 

Pumps. 

Girder Yard. 

Bridge Erectivg Yard. 

Turbine Ventilat&r. 

Marie Coal Pit. 

Rail Yard. 

Mineral Elevators. 

AppuU Coke Ovens. 

Wood Stores. 

Workmen's Houses. 

Stores. 

Shooting Ground. 

Slag Heap. 

Casting Stores. 

Dolomite. 



734 JOHN COCKERILL SOCIETY, SERAING. June 1905. 

of heat produced by tlie combustion of the gases, the thermal 
efficiency of which is considerably higher than that of steam- 
engines. At the present time the boilers heated by blast-furnace 
gas supply steam for about 2,500 H.P., while it is estimated that 
the same gases, supplied to gas-engines, will supply 12,000 H.P. 

Chateau de Seraing. — The buildings of the ancient Castle of the 
Prince-Bishops of Liege have been utilised for the residence of the 
Director-in-Chief, and the various offices. Arranged round the 
Court of Honour, one wing of which has been recently rebuilt 
and enlarged, are the manager's and secretary's offices, sales and 
buying departments, the accountants' offices, the archives and 
the library. The engineers' offices occupy the first floor of a 
building forming the right wing of the old service court, which 
has been covered with a roof and transformed into a workshop. 
The ancient council-room of the States-General of the Principality 
of Liege is used for the general meetings of the shareholders. 

Workshops. — Coming out from the Court of Honour, No. 1 
Workshop is the first seen. It is established on the site of 
the ancient Service Court of the castle. This workshop, 
393 feet long by 148 feet wide, where the various parts 
are machined as they come from the foundries, is lighted from 
above. One side of it is occupied by fitters' benches, and two 
electric travelling cranes of 35J cwts. transport the material. 
The ground floor, forming the right wing of the ancient Court, is 
occupied by a workshop for cocks and valves, and a polishing shop. 
The left wing of the building, surrounding workshop No. 1, 
contains pattern-makers' and joiners' shops, a workshop for packing 
goods, and storerooms. 

Workshop No. 3, one of the oldest of the establishment, was 
completely modernised in 1900. It is used for erecting fixed 
engines, stern-wheel engines, pumping plants, travelling steam- 
cranes, etc. The principal shop, 328 feet in length by 52J feet in 
width, is provided with an electric overhead traveller of 40 tons 
and two of 5 tons. 



June 1905. JOHN COCKERILL SOCIETY, SERAING. 735 

An annex of this shop has been for several years past used for 
the manufacture of quick-j&ring guns (Nordenfelt system), which 
is now carried on in a new shop adjoining the old one. The 
new shop has an area of 3,588 square yards, and is fitted with 
powerful machinery, including lathes for finishing guns 49 feet 
long, also gun mountings and two large machines for boring the 
gun tubes, and a rifling machine for guns up to a calibre of 
9J inches. 

The locomotive shop was partially rebuilt in 1895, and is now 
provided with electric travelling cranes, capable of lifting weights 
up to 40 tons ; 100 locomotives can be turned out annually. 
The majority of the locomotives on the Belgian State Railways 
have been designed and built at Seraing, as well as locomotives 
for use in foreign countries. A second shop in which lighter 
work is carried out completes this department. 

The large erecting shop, a workshop specially reserved for the 
construction of the largest engines, consists of three bays, the one 
in the centre being 65J feet high. A 40-ton electric traveller runs 
from one end to the other of the central bay, and on the same rails 
is a 5- ton electric traveller. Among the machine tools to be seen in 
this shop are slotting and vertical planing machines, various lathes 
for turning turntables, flywheels, grooved pulleys, and pistons, etc. 
An important feature of this shop is the large accurately planed 
floor bed-plate, with an area of 1,620 square feet, on which the 
heavy castings are machined by a boring-mill and other tools. 
The boring-mill has two movable independent heads which will 
take 24^ feet between them. The other tools are moved up to the 
castings and finish accurately any part, however complicated. 

In front of the workshops is a spacious yard with a network cf 
railway lines. A 40-ton electric Goliath crane serves for loading 
and unloading railway trucks. There will also be seen in this yard 
a 500-ton hydraulic press. In this department is a factory for 
making bolts, screws, rivets, etc. For protection against fire 
a system of reservoirs has been installed on the upper floors,, 
with the necessary pipe mains and hose, which are examined 
every day, and from among the workmen a voluntary fire brigade 



736 JOHN COCKERILL SOCIETY, SERAING. June 1905. 

has been organised, which is provided with a steam-pump and 
several hand-pumps. 

Forges. — These contain a powerful hydraulic forging press, but 
the steam-hammer is still kept for forging iron. The hydraulic 
press is of a maximum power of 2,000 tons, sufficient for the 
production of the large forged pieces, such as gun tubes, or 
marine-engine shafts. In an adjoining building are large steel 
pumps working a series of three large accumulators for the forging 
press. The ingots are conveyed between the furnaces and the 
hydraulic press by a 70-ton travelling crane. 

In the forging department there are ten single-acting and 
double-acting steam-hammers, the most powerful of which, of the 
weight of 30 tons, has a stroke of 10 feet. For some years oil 
tempering has been adopted, especially for hoops and tubes of 
guns. The installation necessary for this operation includes 
special furnaces and an oil reservoir of 6,600 gallons capacity. 
Machinery for roughing and finishing locomotive crank-axles is 
about to be erected in an adjoining building. 

Boiler Worlcs. — The work done in these shops is of two kinds, 
namely, that for the manufacture of steam-boilers, tanks, 
converters, etc., and the structural work for bridge building. 
Electrically driven tools and hydraulic riveters are employed, the 
latter especially in riveting bridges, boilers, etc. Steam-boilers are 
manufactured in all types and sizes for marine, locomotive, and 
stationary engines ; and among works of another kind may be 
mentioned the gasholders for the City of Brussels and the Intze 
tanks so frequently used in railway stations. 

Foundries. — These occupy two buildings for moulding in cast- 
iron, bronze, or steel. The majority of mouldings are made in 
sand, particularly for repetition work; but for articles of large 
dimension loam is used. The four statues in bronze which stand in 
the court of the castle were copied from those of the Cockerill 
Monument in the square, and were produced in the foundry. 



June 1906. JOHN COCKERILL SOCIETY, SERAING. 737 

Blast- Furnaces. — This department has six blast-furnaces built 
at different periods. The older of them, four in number, are being 
replaced by the latest type of furnaces. The new blast-furnaces 
are 79 feet high, with an output each of 250 tons daily; a 
large part of the ore is brought by boat on the Meuse to Seraing, 
and thence transported by a funicular railway of about 1,000 yards 
in length. The blowing-engines are nine in number, of which 
six are steam-driven and three others are directly actuated by the 
blast-furnace gases. 

Steel WorJcs. — Up to the year 1863 steel was manufactured at 
Seraing by the old crucible process. Since then, great changes 
have taken place by the introduction of the Bessemer and the 
Siemens-Martin processes. Some Siemens-Martin furnaces and 
Bessemer converters are established side by side, the former using 
to a large extent the scrap of the latter. The Bessemer 
department is provided with five converters of a capacity sufficient 
for a daily production of 800 tons of steel ingots. The molten 
metal coming from the blast-furnaces is first received in a 
large reservoir or mixer with a capacity of 100 tons, for 
rendering uniform the quality of melted metal treated in the 
converters. The converters as a rule receive about ten tons at a 
time. They are blown by two vertical compound engines of 
600 H.P. each, placed in buildings which also contain the 
pressure pumps and the accumulator. The five Siemens-Martin 
furnaces can receive charges of from 15 to 20 tons each, and 
ingots can be cast of forty-five tons weight. The Siemens- 
Martin steel is exclusively used for boiler plates, tubes and 
hoops of guns, tyres, axles, etc. The process used at Seraing 
consists in treating on acid and basic hearths a mixture 
of pig and scrap-iron — mostly scrap from the manufacture 
of Bessemer steel — and it is easy to obtain from four to five 
casts per furnace in the course of twenty-four hours. Owing 
to the great increase in the use of steel, only ten of the 
thirty-six puddling furnaces that were formerly in use have been 
preserved. The puddling and reheating furnaces are of the 

3 E 



738 JOHN COCKERILL SOCIETY, SERAING. JuNE 1905. 

Bicheroux type, the waste heat of which is used for the production 
of steam by means of boilers of different types. The rolling mills 
contain one train of roughing or cogging rolls, two guide mills, 
one merchant train, two large section mills, sheet mills, and plate 
mills. The largest plate-mill, dealing with 5-ton ingots (for 
locomotive frame plates and marine boiler plates of all sizes), is 
worked by a geared, horizontal reversing-engine of 1,000 H.P. 
The other roll trains are worked by variable expansion, condensing, 
vertical engines provided with heavy fly-wheels. 

Wheel and Axle >S%ojp.— This department was started in 1885, 
and is provided with machinery for the manufacture of wheels 
and axles for locomotives, wagons, etc. Originally all types of 
wheels were made on the Arbel method, consisting of hammer- 
blowing, stamping and welding together the boss, spokes, and rim to 
form a wheel centre. Adjoining this shop, forged steel wheels were 
made from ingots prepared under the hammer and then rolled in 
a circular mill. These methods have been abandoned in favour of 
cast-steel wheels. Certain munitions of war are made here, also 
shell bodies, cartridge cases for quick-firing guns and fuses by 
stamping. 

Coal Mines. — The Cockerill concession, which extends particularly 
on the right side of the Eiver Meuse, occupies an area of 750 acres. 
Coal is found in numerous seams, and is obtained from depths 
reaching as low as 2,082 feet at the Colard pit. 

Colard Coalpit. — This is the most important of the pits, and is 
situated on the south of the concession. There are two principal 
shafts, namely the " Cecil " and the " Marie." The v/inding 
engine at the latter shaft is fitted with the Society Cockerill's 
equilibrium valves, and has a spiral drum, the rope of which 
is made with a uniform decreasing section. The correct position 
of each spiral has been so carefully calculated that an almost 
perfect equilibrium has been obtained. About 2,000 tons of coal 
are obtained per day from the two shafts. There are two large 



June 1905. JOHN COCKERILL SOCIETY, SERAING. 739 

beam pumping-engines and crank-pumps; also an -underground 
pumping-engine of 175 H.P. and a hydraulic pumping-engine 
of 450 H.P. 

The ventilation of the collieries' works is effected by a Mortier 
ventilator, capable of exhausting 1,400 cubic feet per second. An 
old Guibal fan serves as a reserve. These ventilators are driven 
by electric motors which receive the current from the central 
electricity station. In addition to the engine buildings and necessary 
workshops, bath rooms have been provided for the men working in 
the mine, and are much frequented. The other mines in the 
concession are those of the "Marie" and the "Caroline" Collieries. 

Iron Ore Mines. — The Society owns in the Grand Duchy of 
Luxemburg and at Rumelange important workings which yield 
about 120,000 tons a year of ore fit for the manufacture of refined 
iron. With a view to the large development of the basic process 
(Thomas system) an option has been taken in the German part 
of Lorraine, and a second one in a concession of the French part of 
the great oolithic deposit of Lorraine. For the manufacture of the 
Bessemer steel, which requires ore that is free from phosphorus, 
the Society has secured two-sevenths of the concessions of the 
Franco-Belgian Co. of the mines of Sommorostro, near Bilbao. 

Internal Means of Transport. — The network of railways within 
the precincts of the establishment reaches a total length of about 
47 miles of standard gauge, and 35 locomotives are employed 
thereon. The vertical arrangement of the boiler in these engines 
allows the two axles to be placed near each other. The rolling- 
stock comprises trucks of different sizes for transport of coke, ore, 
slags, etc., to the number of 529 vehicles; there are also a 
steamtug and 8 barges, two of which are provided with their own 
engines. The barges have a capacity of 300 tons, and the tug is 
provided with a 100-H.P. engine. 

Lighting and Transmission of Power. — The establishment 
at the present time disposes of an electric power of 3,300 effective 

E 2 



740 JOHN COCKERILL SOCIETY, SERAING. June 1906. 

H.P., 1,800 H.P. of which is produced by Delamare-Deboutteville 
and Cockerill motors, utilizing the gas from the blast-furnaces, and 
1,500 H.P., by steam-engines. This latter installation, which 
forms the original central-station, established in 1889, will also 
be replaced by an electric generating group, the gas-engine for 
which, of 1,500 H.P., is under construction. By utilizing the blast- 
furnace gas the cost of electric lighting is reduced to a minimum. 

Various Institutions. — Among the interesting and useful 
departments attached to these Works are the schools. The 
industrial school of Seraing owes its creation mostly to the 
initiative of the Cockerill Company, and several of its engineers 
take part in the teaching there. A naval school with preparatory 
classes has also been founded at Hoboken on the Scheldt, and a 
miners' school has been established at Seraing. The hospital was 
founded in 1849 in consequence of an epidemic of cholera, and was 
completed in the year 1866 by the addition of an orphan asylum. 
It contains 230 beds, and the sick and wounded are treated 
gratuitously. The medical service is performed by five physicians 
who call daily at the factory and in the dwelling-houses of the 
sick. There is an infirmary for the first aid to be given in case 
of accidents, and a surgery which supplies medicines gratuitously 
to the workmen and their families. For those workmen who live 
at a long distance from the Works, and who cannot get home to 
have their meals, mess rooms have been established in each 
department. 



CUT-GEAE WORKS OF F. RECQ BE MALZINE, 
SCLESSIN, NEAR LIlfiGE. 

These Works are situated at 15 Quai de I'lndustrie, Sclessin, 
and were established in 1900 for the sole manufacture of cut 
gears. The machine tools in use were supplied by the firms of 
Brown and Sharpe Manufacturing Co., Pratt and Whitney, 



June 1905. CUT-GEAR WORKS OF F. RECQ DE MALZINE. 741 

Fellows and Bilgram, Gleason, Eead, etc. Spur gears are made 
by Brown and Sharpe rotary cutters, or by the Fellows process. 
Bevel and mitre gears are planed by Gleason planers or by 
Bilgram's process. Helical gears are cut by rotary cutters or 
bobbed. Worm-wheels are bobbed on Pratt and Whitney's 
machine ; and worms are made by cutters or on a screw-cutting 
lathe with a tool ground correctly by the Gisholt grinding- 
machine. Amongst the most important gears made by this firm 
are those for hoists, machine tools, engines, turbines, motor-cars, 
&c., also raw hide pinions. Gears are also made in special steels, 
such as nickel steel, nickel-chromium steel, manganese-silicon 
steel, &c., and small tools are made of high-speed steel. The 
workmen are paid on the Halsey-Eowan premium system. 



HASARD COLLIERIES, TROOZ, 
NEAR LI:6GE. 

The Collieries of this Company (Societe Anonyme des 
Charbonnages du Hasard a Trooz) have two pits, one at Micheroux 
and the other at Fleron, and the annual yield amounts to 240,000 
tons of household coal and 100,000 tons of patent fuel. The 
plant for sorting and washing the coal is capable of dealing with 
350,000 tons annually, including that brought from other collieries. 
There are two condensing steam-engines with variable expansion- 
gear and jackets, winding from a depth of 1,968 feet; also an 
engine working electrically on the " Creplet " system, continuous 
current, at a depth of 1,027 feet; and a central condenser, for 
40,000 lbs. per hour. The central power-station comprises three 
sets of generating machinery of 300 H.P. each, three-phase current. 
The alternators are driven by horizontal steam-engines on the 
" Bonjour " system, with a special piston- valve arrangement, which 
is one of the first applications of an entirely new principle in 
steam distribution. There is also a hydraulic plant with a new 
form of regulator, and a laboratory with plant for testing the 
amount of washing required by the coal. 



742 June 1905. 



INTEENATIONAL ELECTRIC CO., 
LIEGE. 

The manufacture of electrical machinery and appliances was 
started in 1883 on a small scale in one of the fire-arms 
workshops, founded by M. Henry Pieper at Liege. This trial 
proved very satisfactory, so M. Pieper decided to establish a 
company for this new industry, and it was soon found necessary 
to build special works. In 1889 the Compagnie Internationale 
d'Electricite was established under the management of M. Henry 
Pieper. This Society succeeded rapidly, and the capital, which 
was only one million francs in the beginning, was soon increased 
to nearly seven million francs. During the early years branches 
had been opened in several countries, and some of them developed 
later and became the embryos of other important companies. 

The Works in Liege are in the form of a quadrangle, and are 
situated between Quai Coronmeuse, Eue des Bayards, and Eue St. 
Leonard, with exits in the streets named, the main entrance being 
in Quai Coronmeuse. The Works consist chiefly of the dynamo- 
building shops, the lamp and instrument shops, the hoisting 
machinery works for electrically-driven cranes, winches, pumps, 
locomotives, etc. Besides the direct-current dynamos and motors, the 
company manufactures generators from the smallest type to 1,000 
kilowatts or more, and also direct-current motors for belt driving or 
direct coupling. They also build alternating-current machinery 
on a large scale, and polyphase generators and motors of all sizes 
have been constructed for installations in Belgium and other 
countries. The shops in which the Pieper arc lamps are 
constructed have had to be repeatedly enlarged, and at the same 
time improvements have been made in the lamp, such as the 
adoption of an aluminium frame, which gives an elegant 
appearance. The construction of electrical transmission machinery 
is one of the firm's special features ; and the demand for electric 



June 1905. INTERNATIONAL ELECTRIC CO. 743 

carbons necessitated the construction of a special factory in 
1899. 

In the last few years the Company has been specially occupied 
with the construction of electrical plant for collieries and 
ironworks, and their electrically-driven capstans and cranes, 
erected for the Antwerp Docks, show some great improvements. 

With regard to electric traction, the first overhead trolley 
electric car in Belgium was built by this company at Herstal, 
near Liege. They have also erected central power-stations for 
lighting and traction in different towns in Belgium and elsewhere, 
notably the Liege station for the lighting and tramways of the 
town, built in 1893, which possesses to-day a plant of 1,400 H.P., 
besides powerful storage batteries. Most of the stations have been 
built with a view to the large extension of their respective plants. 



KETIN CO.'S WORKS, SCLESSIN, NEAR Li:&GB. 

This firm, established in 1835, was turned into a limited 
company in 1899, with a capital of £80,000. Its manufactures 
comprise rolls for rolling mills cast in air furnaces, chilled rolls, 
grain rolls, and all kinds of work up to the largest sizes ; also 
condensers, cylinders for steam engines, hydraulic presses, 
tubbing for mines, wheel gears, flywheels, and pulleys 
completely finished. A speciality is made of heavy castings 
up to 100 tons, cast in green sand or loam. The Works are 
situated in the district of Sclessin, occupying an area of about 
8^ acres and having about 5 acres covered; a railway line 
connects the works with tbe Nord-Belge Railway. 

The iron foundry contains a large casting shop, 475 feet 
long by 131 feet wide, spanned by two electric overhead travelling 
cranes of 40 tons, and four of 12^ tons capacity. In this 
department are installed five cupolas, one of the melting capacity 
of 11 tons per hour, another of 8 tons, two of 6 tons, and one 



744 KETIN CO.'S WORKS, SCLESSIN. JuNE 1905. 

of 3J tons per hour. They are fed by blowers working at a 
pressure of 24 to 32 inches of water. There are three air 
furnaces, two of 20 tons and one of 15 tons capacity, serving 
specially for the manufacture of rolls. The core ovens are 
thirteen in number. In the annexes of the buildings there are 
machines for the preparation of sand, a patternmaker's shop 
and a fettling shop 195 feet long by 39 feet wide, served by 
an electric overhead traveller of 25 tons. 

The machine shop comprises four powerful lathes for turning 
large flywheels to 37^ feet diameter, and weighing up to 80 tons ; 
a horizontal milling machine for machining pieces of largest 
dimensions; a milling machine for the trefoil necks of rolls; 
25 roll turning lathes ; an electric overhead crane of 25 tons, and 
one of 1 5 tons capacity. The motive force is furnished by an electric 
central station where there are two steam-engines, compound 
(with central condensing), each of about 180 H.P. This 
arrangement of tools and appliances enables a monthly production 
of 1,000 tons to be obtained. 



WORKS OF THE SOCI^Tlfe ANONYME Li:feGEOISE, 

LI^GE. 

The productions of these Works comprise engines and 
apparatus employed in coal mines and in metallurgy. The 
foundry, machine and erecting shops cover an area of about 
2 J acres. The shops are served by seven overhead electric travelling 
cranes, from 3 to 20 tons capacity, dealing with the lifting and 
unloading of the pieces and engines that are being constructed. 
Among the specialities of these Works are winding-engines, 
electric pumps for mines, ventilating fans of great capacity, 
rolling mills, steam hammers, steam, hydraulic, and electric cranes, 
large shears and all kinds of machines. About 250 men are 
employed, and the technical staff comprises 20 engineers and 
draughtsmen. 



Junk 1905. • 745 



MAKCINELLE AND COUILLET WORKS AND 
COAL MINES, COUILLET, NEAR CHARLEROL 

The Works of this company date from the beginning of the last 
century, when the first coalpits were started at Marcinelle. In 

1821 a forge and some puddling furnaces were established at 
Couillet for manufacturing iron on the English process ; and in 

1822 a blast-furnace was constructed at Hauchies, a branch 
establishment of the Couillet Works. It was there that the first 
trials were made in Belgium to produce iron by means of coke. 
The first blast-furnaces established at Couillet date from 1828; 
and about 1834 the rolling mills were greatly extended. In 1835 
the Societe Anonyme de Marcinelle and Couillet was founded, and 
in 1866 was considerably enlarged by the purchase of the factories 
at Chatelineau. In 1889, with a view to the construction of the 
armoured turrets destined for the forts on the Meuse, there was 
built at Couillet a steel foundry producing ingots and steel 
castings of all weights and dimensions, on the Siemens-Martin 
system. At the same time special workshops were built at 
Couillet for finishing and mounting the cupolas and armour. 
An artillery polygon was built on these premises for testing war 
material. 

In 1892 works for the production of iron on the Thomas system 
were started ; and at the same time a new blast-furnace was 
erected, and also large rolling mills for the manufacture of blooms, 
billets, bars, rails, sleepers, etc. With a view to producing iron 
for the Thomas Steel Works, two blast-furnaces were constructed 
possessing all the latest improvements. This necessitated the 
demolition of the Siemens-Martin Steel Works built ten years 
previously. Railway lines of narrow and standard gauges were 
laid, connecting all the various branches of the works. The latter 
extend over an area of 178 acres, and the coalpits over 4,940 acres. 



746 MARCINELLE AND COUILLET WORKS, COUILLET. June 1906. 

By means of a series of continued extensions the company has been 
able to consolidate the following installations : — 

At CouiLLET : — (]) Coke furnaces with machinery for recovery 
of the by-products ; (2) Blast-furnaces ; (3) Thomas Steel Works ; 
(4) Siemens-Martin Steel Works; (5) EoUing mills for the 
manufacture of blooms, billets, rails, bars, sleepers, etc. ; (6) Iron 
and brass foundries, boiler and smiths' shops, large forges and 
shops for locomotive construction, stationary engines, and war 
material for artillery and the construction of fortifications. 

At Chatelineau : — (1) Boiling mills for black iron sheets and 
steel of flitch plates ; (2) Iron foundry, coppersmith shops, 
and shops for the construction of stationary engines, locomobiles, 
bridges, constructional ironwork, etc. It is intended to transfer 
these works to Couillet. 

At Marcinelle : — The coal mines of North Marcinelle with five 
pits, one of which is situated at Couillet (Fiestaux). 

In Belgium, France and in the Grand Duchy of Luxembourg 
the company has several concessions for working iron ores. 



Couillet. 

(1) Coke furnaces : This installation contains three batteries 
of 25 furnaces with machinery for recovery of the by-products on 
the Semet-Solvay system. The dimensions of the furnaces are : 
length 31 feet, width 1 foot, height 5^ feet, duration of the 
operation 22 to 23 hours for one charge of 3^ tons of coal, 
yielding 78 per cent. 

(2) Blast-furnaces : These include a group of two blast- 
furnaces producing 100 tons of basic pig every 24 hours, and a 
second group of blast-furnaces able to produce each 175 to 180 tons 
of metal every 24 hours. 

(3) Thomas Steel Works : This installation contains four 
12-ton converters, two melting shops with cupolas for melting and 
for Spiegel, and it is constructed to yield a monthly production of 
at least 10,000 to 12,000 tons of steel. 



June 1905. MARCINELLE AND COUILLET WORKS, COUILLET. 747 

(4) A Siemens-Martin Steel Works : This contains gas-producers 
on tlie Wilson system necessary for the working of its two 15-ton 
furnaces, a casting pit for ingots, and a large shop for moulding 
with an additional finishing shop. 

(5) Kolling-mills : These mills contain the following rolls : — 

(a) two large roll trains of 29 J inches diameter driven by a single 
horizontal non-condensing engine, having three cylinders of which 
the diameter is 43 inches and length of stroke 4 feet 8f inches. 

(b) A Merchant train of 23J inches diameter driven by a vertical 
condensing engine with fly-wheel. Diameter of the cylinder 
45 inches and length of stroke 4 feet 6 J inches, (c) A 12-inch 
diameter roll train driven by a vertical condensing engine with 
fly-wheel, diameter of cylinder 31 inches, length of stroke 2 feet 
7 inches, and nominal horse-power 600. (d) A roll train of 9f inches 
diameter driven by a vertical condensing engine, with diameter 
of cylinder 25 J inches, length of stroke 2 feet 7 inches, and 
nominal horse-power 450. 

(6) Engine construction shop : These shops are arranged in 
two divisions, one for locomotives and the other for various 
stationary engines. Four different types of locomotives are made, 
namely : — narrow-gauge locomotives with weight varying from 
2^ tons to 12 tons ; metre-gauge locomotives with weight 
varying from 10 to 25 tons ; locomotives for standard gauge 
weighing from 27 to 60 tons and more, and locomotives for 
tramways. 

The workshops for stationary engines are able to turn out at the 
shortest notice engines for coal mines, winding engines, horizontal 
and vertical engines of various horse-power with one or two 
cylinders or compound with slide-valves ; fans of all types direct 
driven or belt-driven ; exhaust-engines, air-compressors (simple or 
compound) ; engines for blast-furnaces, steel works, rolling mills ; 
hydraulic presses, stationary engines with double and triple 
expansion for breweries, spinning and weaving mills, ice factories, 
etc. ; electric-light and power engines ; steam and hydraulic 
cranes, crane locomotives, steam travellers, and electrical or 
hydraulic capstans. 



748 MARCINELLE iiND COUILLET WORKS, COUILLET. June 1906. 

Bridge and boiler yard ; bridge construction, constructional 
ironwork, etc. ; boilers of all description, tanks, beating apparatus, 
cylinders for storing gas and liquids under bigb pressure. 

Mecbanical workshops, specially for stamping; shells for all 
kinds of boilers ; flanged sheets for locomotive fire-boxes of all 
dimensions ; flanged sole-bars for carriages, etc. 

War material; common and armour-piercing projectiles in 
chilled and cast-iron, hard steel and chromium-steel ; shrapnel 
and case-shot ; projectiles for quick-firing guns ; torpedoes ; 
revolving cupolas for quick-firing guns, searchlights and 
armoured observatories ; armour and gun carriages ; material 
for the engineering corps and for artillery. 

The Company has constructed dwellings for its workmen, 
which it lets to them at low rents. It also supplies those of its 
workmen who desire it with flour and bread at cost price. It has 
provided infant schools, elementary schools, schools for adults, for 
housewifery, apprenticeship, for drawing and music. It has 
established a bank, which provides medical attendance to all the 
families of the workmen, advances money in case of illness, and 
for old age pensions. There is also a hospital, specially provided 
for accidents. 



MECHANICAL ENGINEERING SCHOOL, LIlfeGE. 

This School, situated in Rue Vertbois, Liege, was founded 
in 1902, and is under the patronage of the Belgian Government, 
the Province and Town of Liege, the Chamber of Commerce, 
and other representative bodies. The idea of establishing this 
School was started in 1898 by M. Wathoul, who was requested by 
several manufacturers to draw up a scheme to be submitted to the 
various authorities, who finally gave it their support. The 
building was opened on 24th March 1902, and contained a 
millwright's shop and a foundry. It started with 80 pupils and 



Junk 1905. MECHANICAL ENGINEERING SCHOOL, LI^GE. 749 

six professors and foremen. In October in the same year tlie 
numbers had risen to 130 pupils and nine professors. The 
machine-tool shop was opened on 1st March 1903, the tools being 
driven by a 7-H.P. gas-engine, which is about to be replaced 
by a 20-H.P. gas-engine. In October 1904 there were 350 
pupils undergoing a three years* course of tuition and 50 motor- 
car drivers taking a half year's course of instruction, and the number 
of professors had risen to 22. Since the date of opening the 
School more than 2,000 candidates have presented themselves 
at the Entrance Examination, of whom about 400 have been 
admitted. The objects aimed at in establishing this School were 
to train pupils between the ages of 13 and 16 years for entrance 
into mechanical engineering works, to instruct them in the theory 
and practice of engineering, and to mould their characters at a 
critical age. 

The Mechanical School contains the following departments : — 
Two millwrights' shops, with 55 benches; cycle and automobile 
shop with 40 benches, machines, and an 8-H.P. automobile ; 
turners' shop ; machine-tool shop with a 20-H.P. motor and 
a set of 50 up-to-date machines ; a foundry with two double 
forges, six anvils, and a steam-hammer ; class-rooms and drawing 
offices; a meeting hall with accommodation for 600 persons; 
kitchens, dining-rooms, &c. The School is managed by a Council 
representative of local industries, with M. Beer as President and 
M. Wathoul as Secretary and Director. 



MELOTTE CKEAM SEPAKATOR WORKS, 
REMICOURT, NEAR LlfiOE. 

These Works, situated about twelve miles from Liege, were 
established in 1893 at Remicourt for the sole manufacture of 
cream separators. The proprietor, M. Jules Melotte, had previously 
been in the agricultural works of his father, and on the death of 



750 M^LOTTE CREAM SEPARATOR WORKS, REMICOURT. June 1905. 

the latter he assisted his mother and brother (M. Alfred Melotte) 
in carrying on the bnsiness. On the introduction of the De 
Laval separator, he determined to make a careful study of it, and 
produced in a few months' time a novel and improved machine for 
the separation of cream, which obtained the first prize at Brussels 
in 1888 after a long series of trials. Encouraged by his success, 
he, on the death of his mother, took over the Remicourt Works, 
while his brother established himself at Gembloux as a specialist 
in agricultural machinery. The village of Eemicourt is now a 
small town with well-built cottages and comfortable villas, and 
the works themselves have architectural features which are well 
in keeping with the general plan of the factory. 

The Works, only a few minutes' walk from the railway station, 
have been built with the idea of saving as much time as j^ossible. 
The machine tools are of English and American manufacture, but 
except for the turret lathes and a few other machines, they have 
all had to be modified for the special work to be undertaken. 
The capstan lathes, which are run at a very high speed with high- 
speed tool-steel, have had the headstock spur-wheel gear replaced 
by belts on account of the former breaking under the strain. The 
holders and slide-rests of the planing and gear-cutting machines 
are of special design, and every part is tested before being 
accepted. 

The devices employed for turning out the parts rapidly and 
accurately are often very ingenious. In fitting the two halves of 
the enamelled cast-iron cover for the bowl, they are first placed on 
a frame which holds them in a particular position, and two or 
three holes are drilled corresponding with studs on the table 
where the cover is to be turned or drilled. In less than five 
seconds it is placed on the studs and fixed by a wedge, and the 
tool starts boring. The edges of the half covers are then grooved 
to receive the rubber band which is to make a hermetically tight 
fit when the two parts are closed. In manufacturing the catch 
which fastens the cover, a flat bar is first wedged in the holder, 
and, after being shaped on one side, is similarly treated on the 
other, and is then cut in half to make two fastenings. Six 



Junk 1905. MELOTTK CREAM SEPARATOR WORKS, REMICOURT. 751 

operations are needed for making the bowl. A steel plate is first 
heated and placed under a press, where discs are stamped out, and 
these are pressed in the form of a half-bowl. Each half is then 
fixed in the headstock of a lathe, which turns it by a special tool 
operating simultaneously against the side and bottom. Each 
half-bowl is then threaded to allow of its being screwed on to 
the other half. Eifty-five bowls are manufactured from the raw 
material in a day, which is considered very fast, seeing that only 
one machine-tool is engaged on each operation. 

In the tinplate department each man is engaged on a special 
operation. At the end of the bay the separators are fitted and 
painted. All the parts are carefully tested to gauge after 
manufacture, there being an average of six gauges for each part, 
and their total number is 1,500. When the separators are finished, 
each one passes through a controlling department, where its 
number, date of manufacture, and destination are recorded, so that 
by quoting the number the buyer is able to receive any spare 
parts of the exact dimensions he requires. 

In order to cope with the increasing business, further 
enlargements are to be carried out, when a foundry will be added 
for the casting of frames. The latter are the only part of the 
separators which are not manufactured at Eemicourt. There are 
220 machine tools in use, and about 300 men are employed. Fifty- 
five complete cream-separators are turned out daily. 

Above the offices is the laboratory, where every new separator 
(of other make) is tested with a view to measuring the power 
required to drive it. For this purpose there is a dynamo and 
recording gauge. This testing is further continued by analysing 
the milk, and the most careful experiments are carried out to 
ascertain whether any new separator possesses an advantage over, 
that of M. Melotte. 



752 Junk 1905. 



NAJREOW-GAUGE RAILWAYS IN BELGIUM. 

The following particulars have been furnished by M. H. 
Cruysmans of Antwerp. Two laws passed in 1884 and 1885 
regulate the concession of all narrow-gauge railways in Belgium. 
These laws grant to the " Societe Nationale des Chemins de fer 
vicinaux " (National Company of District Railways) the sole 
concession of narrow-gauge lines. The capital required for the 
purpose of building the lines and getting them into working order 
is subscribed by the State, the province (county) and municipalities 
through which the lines run, and a small part only by private 
contribution. Each line has its private capital, quite distinct 
from the capital of other lines. This capital is not paid in cash 
by the subscribers (except by the private ones), but the public 
authorities are allowed to write off their subscription in 90 years 
at the rate of about 3J per cent, yearly. On the other hand, the 
" Societe Nationale " is allowed to issue bonds to bearer, at 
3 per cent, interest and guaranteed by the Belgian Government. 
By these means the " Societe Nationale " is able to build the lines, 
namely, the permanent way, station buildings and rolling stock. 
When a line is completed and ready for working, the Society 
Nationale has to find a firm (or company) who will work the 
railway. This is done by public tender. The firm or company 
which offers the best 'terms receives the concession, and 
is allowed to work the line subject to a code of general 
regulations. The Societe Nationale supervises the management 
of the lines, has control of them all, and sees that the general 
conditions are fulfilled. The table of rates is fixed by the Belgian 
Government. 

The general practice is that 65 to 60 per cent, of the gross 
receipts belongs to the concessionaire, and the balance goes to 
the Societe Nationale ; with that sum (40 to 45 per cent, of the 
receipts) the latter body must pay their general expenses and if 



JUNB 1905. NARROW-GAUGE RAILWAYS IN BELGIUM. 753 

possible give a dividend to the shareholders of the line. This 
dividend varies with the value of the line ; in some cases it 
amounts to over 5 per cent. The concessionaire, having nc 
plant of his own, needs but a small capital to work the 
railway. The Societe Nationale requires, however, a security, 
being at the rate of 2,000 francs per kilometre (£128 
per mile). The concessionaire must pay the expenses of working 
the line, and that which is left over is his profit. The result ot 
this legislation is that, as the public authorities have agreed to 
pay yearly 3 J per cent, on their subscription, if the dividend, for 
instance, amounts to 4 per cent., instead of paying, they receive 
the difference, say, J per cent, in this case. Thus, instead of making 
an expenditure, they receive, on the contrary, a part of the profit, 
and besides that get all the benefit of the small gauge railways. 

Begun in 1885, there is a length of 3,000 kilometres (1864 
miles) of narrow-gauge railway in working order or nearly ready for 
use; this is divided into 127 different lines. The gauge is generally 
1 metre, and the rails of the usual continental description weigh 
23 kilograms per metre (46 lbs. per yard). The rolling stock is 
very much alike on all the lines ; the locomotives of a similar 
pattern weigh from 12 to 30 tons apiece. Nine lines are worked 
by electricity, the others by steam-power. The total receipts in 
1904 were nearly 13,500,000 francs (£540,000), for 2,400 kilometres 
(1491 miles) in working order. That makes 5,600 francs per 
kilometre (£360 per mile) and 89 centimes per train-kilometre 
(Is. 2d. per train-mile). 



NATIONAL SMALL-AKMS FACTORY, 
HERSTAL, NEAR LIEGE. 

These Works (La Fabrique Nationale d'Armes de Guerre) were 
constructed in 1889 by a company formed in 1886 by a group 
of Liege small-arms manufacturers to enable them to be in a 
position to manufacture the largest orders rapidly. The works 

3 F 



754 NATIONAL SMALL- ARMS FACTORY, HERSTAL. June 1905. 

came into full working order in 1891, and started the manufacture 
of the Mauser repeating (magazine) rifle of 0-301 inch calibre 
type which was to be adopted by the Belgian Government for 
the infantry. The Herstal establishment also manufactures 
interchangeable parts for rifles. 

The Factory is situated about a league from Liege on land of 
22 J acres in area, of which 17 J acres is covered by buildings, &g. 
There are three distinct divisions : one for the manufacture of 
arms, another for the manufacture of cartridges, and a third for 
manufacture of bicycles and motor cycles, branches that have 
been greatly extended and which employ more than a thousand 
men. A railway line joins the main line at Herstal station, 
and runs round the establishment connecting the various 
departments. 

The production of the motive power necessary for the 
requirements of the factory has been centralized. In the 
Central Station are installed a Corliss S3'stem engine of 500 H.P., 
and two high-speed Willans simple engines, one at 400 H.P. and 
the other at 350 H.P. In the adjoining shop are four boilers 
of Piedbceufs system occupying 1,614 square feet; three boilers of 
Mathot's system of 1,938 square feet heating surface, which supply 
the above engines and furnish steam necessary for heating the 
various buildings. The feed-water is taken from the Mouse, and is 
purified by Gaillet's purifier and afterwards filtered through wood 
shavings. 

More recently a gas-producing station has been added, composed 
of two groups of producers of Fichet and Heurtey type. Each of 
|;hese two is capable of furnishing 17,650 cubic feet of gas per 
Jiour, which is used for heating the difierent furnaces ; they are also 
employed for tl^e production of motive power. The gas-engine 
of 250 H.P. is of the Oechelhauser system, furnished with his 
latest improvements. From the start, electricity has been adopted 
as being the most economical method of distribution of power, andi 
the best for putting in motion the innumerable machine-tools in the 
different shops. A new engine of the same type and of 450 H.P. 
Ijap subsequently beer^ ^44®^ t*^ t^® above-me^itioned power-station. 



Junk 1905. NATIONAL SMALL-ARMS FACTORY, HERSTAL. 755 

The portion of the works devoted to the manufacture of arms 
comprises a large number of distinct shops. The Wood-Working 
Shop includes a store capable of receiving 250,000 sets of wood 
parts for rifles, two chambers for drying wood, and a shop for 
mechanical work on wood, where those sets destined to form the 
stocks, etc., of the rifles, are squared and roughed out. They 
then pass to the shaping machines, after which they are gauged, 
and the limits allowed vary from • 004 to • 002 inch. The stocks 
are then finished, oiled and polished in the machine. 

The Machine Shop for working on the metallic parts of the 
rifle is about 2J acres in area; in it are drilling machines, 
polishing machines, lathes and other machines necessary for 
finishing the parts of the guns. These perfect machines are 
so easily handled that the greater part are attended to by 
women. The jigs and cutters and other tools are made in the 
same works. There is also a shop for polishing these parts and 
for blueing and bronzing certain metallic parts of the gun, and a 
tempering shop. 

The Smithy contains twenty-six steam-hammers of various 
sizes, and six Bradley rapid-blow hammers of 44 lbs. and 66 lbs. for 
forging gun-barrels. The bars are heated red, the piece is put 
through under the first hammer, finished on the second at the 
same heat, and the "fins" cut off whilst warm. This part of 
the works contains a firing range of 217 yards in length, a 
gun-erecting shop, with offices for viewers and inspectors; a 
shop for packing and despatching various stores, for articles of 
general merchandize, for the stock of steel for manufacturing 
purposes, for the steel destined for the tool department or for 
the works use, and for the spare parts of the guns. Near 
this building are several workshops for the manufacture of 
cartridge-cases, and of bullets and charges, and at the side the 
toolshop for making dies and punches and tools necessary for 
manufacturing rifle cartridges. Large rooms for the viewing of 
cartridge-cases, bullets and loading, a room for charging and 
another for inspection and despatching, a range for ballistic trials. 
And a water range, a ropm for nickelling, gas-furnaces, annealing 

3 F 2 



756 NATIONAL SMALL-ARMS FACTORY, HERSTAL. June 1905. 

furnaces, a shop for washing and cleaning and brightening, and 
offices for the staff and the members of the committee, complete 
the preceding installations. 

For the manufacture of bicycles and motor cycles there are 
several shops, some for the making of the framework and forks, 
others for the construction of motors, carburetters and accessories. 
At the Erecting Shop all the motors are tested before being sent 
out. The brake test is effected by a dynamo, from which the 
current generated is absorbed by liquid resistances. This large 
and complete establishment has enabled this company to take a 
prominent position in the motor cycle trade. 

For the purpose of avoiding the momentary stoppages that 
occur in the manufacture of small-arms, the management of the 
National Factory has added the manufacture of sporting guns, of 
which 150 are turned out daily. For some years it has 
undertaken the production of the Browning automatic pistol, 
which has been adopted by the Belgian Army, and already 
100,000 have been manufactured. Amongst smaller articles made 
by this company may be mentioned those of steel capsules for 
sparklets, of which over 100,000 daily are made; also bicycle 
spokes, parts of the Bowden brake, etc. 

The number of workpeople (men and women) varies from 
1,500 to 2,000. Many women are engaged in the mechanical work 
of gun parts, of cartridges and other articles, in viewing these 
pieces and in the loading of cartridges. There are many 
inspectors for the various departments, having about 50 foremen 
under them, and the staff is completed by 66 accountants, clerks 
and draughtsmen. 

The daily output of these Works reaches on an average to more 
than 500 guns and 250,000 cartridges, 150 bicycles and 50 motor 
cycles. In order to ensure the use of the right material, a 
splendidly equipped testing plant has been erected where the 
various materials are tested before being used, and the results 
obtained are followed up in the various stages of manufacture. 
All the parts which compose the different arms made in this 
establishment are examined with great care ; and to obtain 



June 1905. NATIONAL SMALL-ARMS FACTORY, HERSTAL. 767 

interchangeability of parts, the allowance in the dimensions 
varies only between 0*004 and 0-002 inch. Among the large 
orders carried out by this Factory may be mentioned that of 
150,000 Mauser rifles for the Belgian Government, also of 
50,000 Mauser rifles for the Brazilian Government, and about 
100,000 for China, Uruguay, Columbia, &c. 



OUGRl&E-MARIHAYE IRON AND STEEL WORKS 
AND COLLIERIES, NEAR LifiGE. 

The Steel Works and Blast-Furnaces of this Company are 
situated at Ougree, and the Coal Mines at Marihaye. The 
original Company was established in Brussels in 1835, and in 
1892 it amalgamated with the *' Societe Anonyme de la Fabrique 
de fer d'Ougree" ; the capital of the joint concern at the present 
time amounts to over £850,000. 

The Steel Works, extending over an area of 91 acres, 
manufacture the following : — Grooved rails for tramways, rack rails 
for mountain railways, and hollow ingots for the manufacture of 
tubes, the process of which is a speciality of the Company. This 
department employs about 2,100 workmen, and produced last year 
266,600 tons of Bessemer, Thomas and Siemens-Martin steels, 
which were converted into beams, rails, sleepers, angle bars, axles, 
plates, slabs, etc. The plant comprises : — 

(1) One 100-ton mixing machine, and three cupola melting 
furnaces for recasting, four Thomas converters of 12 tons each, 
one condensing blowing engine of 1,500 H.P., and a spare one in 
case of an accident. 

(2) Two Siemens-Martin furnaces of 15 tons each. 

(3) One large reversing rolling mill, 33 inches diameter, 



758 OUQRl^E-MARIHAYE IRON AND STEEL WORKS. June 1905. 

followed by two finishing roll trains for Vignole rails, grooved 
rails, sections and beams up to 19 J inches in depth; also a 
condensing steam-engine of 10,000 H.P. 

(4) A 25i-inch diameter rail mill for rails, sleepers, slabs, etc., 
also a condensing steam-engine 1,200 H.P. 

(5) A plate mill capable of rolling plates up to 6 J feet in width, 
worked by a 1,000 H.P. condensing engine. 

(6) Merchant mills, one driven by a condensing engine of 
800 H.P. and the other by a condensing engine of 500 H.P. 

(7) Small rolling mills, composed of roughing train and two 
finishing trains. These mills are driven by two electric motors, 
one in line with the roughing rolls, the other in line with the 
finishing rolls. The fly-wheels of these two motors are rope 
driven. 

(8) A shop for axles and tyres. 

(9) A finishing shop for rails, cross sleepers, etc. 

(10) Smithy. 

(11) A bridge erecting shop. 

(12) A shop for crushing basic slags, containing six crushing 
machines of 30 H.P. and four of 75 H.P. each, driven partly 
by electricity. 

(13) An electric power-station, subdivided into three 
sections, supplying light and power to three different parts of 
the works. The current is generated in the blast-furnace 
department. 

(14) Coal mines, of which this department owns five-eighths of 
the " Six-Bonniers " coal mines, of an area of about 434 acres. 

Blast- Furnaces. — The area covered by these furnaces, including 
the ground for the ores, is about 54 acres, excluding 81 acres 
which are either lying waste or are occupied by the dwelling 
houses of the employees and workmen, and 133 acres of woodland. 
The number of workmen employed is about 1,230, and the last 
annual production of Thomas basic pig was 136,000 tons ; 119,000 
tons of coal were used for the manufacture of coke; and 154,000 
tons went to the blast-furnaces. 



June 1905. OUGRIEE-MAIIIHAYE IRON AND STEEL WORKS. 759 

This department contains the following installations : — 

(1) Five blast-furnaces, one of which is provided with four 
large Cowper stoves, and fed by inclined top railway. The fifth 
blast-furnace of medium capacity has just been put in blast. Four 
of these blast-furnaces are situated in a straight line and are 
fed by a large electric overhead traveller. They are blown by two 
compound steam-engines of the Cockerill type and by two gas 
blowing engines. 

(2) A coal mine of about an area of 692 acres. 

(3) Coke furnaces : ten vertical furnaces (Appolt system) each 
with sixteen retorts ; three horizontal furnaces (Bernard system) 
each with forty retorts. 

(4) A coal store, joining the coke furnaces. 

(5) An iron foundry. 

(6) A central electric-station, provided with powerful engines. 
Two additional gas-engines of 1,200 H.P. are about to be 
installed. The current is continuous and at 550 volts. Three 
transformers reduce the voltage to 250 and 125 for the lighting. 
The number of workmen employed at these works is abou"*" 3,000. 
The annual production is about the following : — 700,000 tons 
of coal for kitchen use, rolling mills, puddling furnaces, zinc 
works, glass works and foundries ; 40,000 tons of washed coal for 
foundries and kitchens ; 110,000 tons of special coke for foundries 
and metallurgical works. 10,000 tons of briquets for industries 
and domestic use. 

This department comprises the following installations : — 

(1) Five collieries, covering an area of 4,025 acres and 
extending underneath Flemalle, Yvoz-Eamet, Chokier, Val 
St. Lambert and Seraing. 

(2) A group of 46 coke furnaces (Smets system), situated at 
Flemalle-Grande. 

(3) A group of 68 coke furnaces (Collin system), 36 of which 
with regenerators are situated at Seraing. 

(4) A group of 25 coke furnaces (Smets system). 

(5) A workshop for the preparation and washing of the 
coals. 



760 OUGR^E-MARIHAYE IRON AND STEEL WORKS. June 1905. 

(6) A factory for washed briquets. 

This company which gives employment to thousands of 
workmen has its offices at Ougree, near Liege. The managing- 
director is M. A. Eaze, and the general manager is M. G. 
Trasenster. 



PIEDBGEUF BOILEK WOEKS, 
NEAR LIl&GE. 

The Piedboeuf establishment was founded in 1812 at Jupille, 
near Liege. It comprises the works at Jupille, Aix-la-Chapelle, 
and Diisseldorf. The Jupille works are situated near the 
railway line from Liege to Maestricht, to which it is joined by a 
railway siding. 

The works occupy an area of 36,000 square yards. All systems 
of steam boilers are constructed there, more particularly of 
the internal-flue type and of the compound type of boiler. 
All the work is executed mechanically. The arrangement of 
the machine tools is such that handwork is reduced to a minimum. 
An electric overhead-traveller discharges the material brought in 
by the railway wagons, and conveys it to the marking-out shop 
and to the machines, which will take the largest sizes rolled. 

The boilers are made up with rings in one piece. The work 
is done on the plates by planing machines with double tool-boxes, 
automatically disengaging, planing vertically, etc., which take 
plates of 31 feet length. For small diameters the plates are bent 
hot, but cold for large diameters. A vertical bending-roU curves 
plates 6 feet wide cold. Mills having produced larger plates, 
these rolls were found insufficient, and to be able to deal with any 
future sizes an hydraulic cold-bending press was installed to 
take plates 12 feet wide. The plates, after being bent, are 
assembled by means of large overhead cranes, and then are drilled 
in position by a quadruple-drilling machine, of which all the 
movements are mechanically controlled. The boiler, having 



June 1905. PIEDBCEUF BOILER WORKS. 761 

been drilled, is then taken apart, and the drilling burrs, &c., 
cleared off. It is then re-assembled progressively to enable the 
riveting to proceed. This riveting is done by large hydraulic 
riveters, with pressures (proportioned to the thickness of the 
plates) up to 118 tons. There are five riveting machines of various 
sizes to suit the different widths of plates, up to 12 feet, one 
machine being specially used for " Adamson " rings. 

There are likewise special machines for boring the Galloway 
tubes upon the flues, for forming the flanges, &c., and for stamping 
manholes and oval pieces. A vertical lathe for oval work 
enables oval holes to be bored in the plates. Hammer work 
and chipping, as far as possible, is done away with. The 
" Adamson " rings are made by a special machine, and the 
strengthening rings for manholes, covers, and mountings are 
stamped in the hydraulic press. The large erecting-shop is 
provided with a 3 5 -ton electric overhead-traveller to move the 
boilers from place to place, and to load them up on the siding, 
leading out of the shop. Compressed air and electricity are used 
for drilling in position. 

For the workmen on the establishment the administration has 
created an assistance fund, and a pension fund, independently of 
the obligation of affiliation with the State's retirement fund. The 
assistance fund is supported by a retention made upon salaries 
and by donations. The pension fund is supported entirely by the 
company. 



ST. LEONAED CO., LIEGE. 

These Works were established in 1825 by Eegnier-Poncelet for 
the manufacture of steel files and tools. In consequence of 
expansion of business, the proprietor decided to turn the firm in 
1836 into a joint stock company. From this period the company 
specially directed its energy to mechanical construction. It 



762 ST. LlSoNARD CO., LIJ^GE. June 1905. 

undertook successively tlie manufacture of macliine tools, engines 
(principally marine), pumps for mines, and finally locomotives, 
which became a speciality, a number of types adapted for all 
kinds of service being designed. 

In the history of the company two striking dates are those 
of 1836 when the first steam-engine was delivered, and 1840 when 
the first locomotive was produced. In 1860, locomotives had 
already been constructed with articulated trucks corresponding 
to bogies. In 1877 the company was the first to make trials of 
traction by steam locomotives on urban tramways, and since then 
it has constantly improved the construction of the locomotive 
both for standard and narrow-gauge railways. For some years 
the company has undertaken a new line of business, namely the 
construction of gas-engines ; to this end, it has completed 
its equipment and made arrangements enabling it to erect and test 
in its shops engines up to 2,000 H.P. in size. Several engines 
from 25 to 600 H.P. are installed at the Liege Exhibition, where 
they are working with lighting gas or with crude gas from 
producers. 

The company possess two establishments, the one at Herstal 
where an iron foundry has been installed in the most up-to-date 
fashion ; the other is at Liege where are the machine shops, and 
including forges, steam-hammers, boiler work, machine tools fitting, 
erection of locomotives and gas-engines, etc. The motive force 
is obtained at Herstal as at Liege, by Koerting gas-engines, 
constructed in the workshops of the company. The installation at 
Liege includes a two-cycle double-acting gas-engine of 250 H.P., 
fed by a gas-producer plant of Fichet and Heurtey's system. This 
engine has worked the shops since the commencement of 1902. 

The number of men employed by the company is about 750. 



June 1905. 763 



ANTWERP DOCKS AND QUAYS. 

The following description has been prepared by M. G. A. 
Eoyers, Chief Engineer to the Municipality of Antwerp. 

The quays of Antwerp built in 1880 have a length on the River 
Scheldt of 2 J miles, with a depth of 26 feet of water at low tide and 
40 feet at high tide. The tide rises about 13 feet, and the edge of 
the quays is about 8 feet above the medium high tide. In the 
central part there exists a pontoon 328 feet long by 65 feet wide. 
Along the quay walls are sheds, and between them and the river 
are about 100 hydraulic travelling-cranes, of a capacity of about 
1 J to 2 tons. Between the sheds and the edge of the wall and 
also at the back of the sheds are two or three railway lines. 
Turning and travelling platforms worked by electric capstans put 
these tracks into communication with one another. Near the 
pontoon is an ancient strong Castle " Le Steen," which has been 
left standing and has been restored ; at the present time it is used as 
a Museum of Antiquities. On both sides of the pontoon jetties have 
been built with railings and flights of steps. Before reaching the 
old docks will be seen on the right the Customs House, and at the 
entrance of the sluice the Pilots' House where there are several 
oJBfices connected with the dock administration, as well as the 
School for Navigation. 

The docks are the property of the city ; their level is always 
maintained at about 1 foot below that of high tide. Passing over 
the sluice there will be seen the small dock built by Napoleon, 
and to which a short time ago the name of Bassin Bonaparte was 
given. It communicates with the Scheldt by means of a sluice 
head, and is separated from the Dock Guillaume by another sluice 
head, which like the former has an opening of 59 feet. This is a 
half-tide dock, and is 189 yards in length with a width of 



764 ANTWERP DOCKS AND QUAYS. June 1905. 

159 yards. About 650 yards further on is the entrance to the 
Kattendijk Dock, which was built in 1860 and lengthened in 1881. 
It has a length of 1,050 yards and a width of 153 yards. This 
dock communicates with the Dock Guillaume by an intermediate 
dock. At the other end of the dock it communicates with the 
Dock Lefebvre. 

Six dry docks adjoin the Kattendijk Dock and communicate 
with it. The first of these, dating from 1863, is now able to 
accommodate ships of 508 feet length. The other dry docks have 
a smaller capacity. On the other side of the Bassin are two 
hydraulic cranes of 40 tons and a derrick of 120 tons. The Dock 
Lefebvre is of an irregular polygonal shape, and was constructed 
in 1887. 

The north, south and east quays are provided with sheds and 
with hydraulic travelling-cranes. At the south quay there is also 
a 10-ton crane for unloading wood for cabinet makers. At the 
south quay, on its western side, is a large granary, able to hold 
120,000 quarters. The Lefebvre and America Docks were 
constructed in 1888, and have a water surface of 54 acres. Li 
consequence of the increase in the draught of ships, these docks have 
been made 6J feet lower than that of the Kattendijk. The America 
Dock was formerly used for the petroleum trade, and is now being 
transformed for general merchandise. Passing along the opposite 
side of the Kattendijk Dock, one comes to the Timber and Asia 
Docks, which are used for the trade in wood and ores. 

The Timber Dock is 570 yards long by 164 yards wide, and 
the Asia Dock is 809 yards long by 104 yards wide. The east 
side of the Timber Dock is used for the traffic in ore, and is 
provided with hydraulic cranes of the portico type. On the south 
quay there is an automatic coal-tipping machine, capable of 
lifting a railway truck with full load up to 25 tons. The docks 
on the north of the river are provided with a general system of 
hydraulic power transmission. There are four steam-engines of 
150 H.P. each and six Cornish boilers. 

Near the Guillaume Dock (formerly called the Grand Dock) is 
a large storehouse belonging to the city, and recently rebuilt in 



June 1905. ANTWERP DOCKS AND QUAYS. 765 

ferro-concrete. The building covers an area of about 7J acres, 
and comprises four blocks quite separated from each other by 
courts, for the purpose of fire prevention. Each of these blocks is 
four storeys high, and is well provided with cranes and electric lifts. 

On the south of the city are three docks given up to small 
shipping. The Central Dock communicates with the Eiver 
Scheldt by a lock 42^ feet in width. The South Dock is used for 
unloading bricks and for discharging the mud and refuse of the city. 
The North Dock is used more particularly for the coal and beer 
trade. The total area of these docks is about 15 acres, and the 
length of quays is about 3,000 yards. Close by is installed a 
hydraulic power-supply station, similar to that on the north side 
of the river. Still further to the south about 3,800 yards higher 
up the river, a line of new quays nearly 2,000 yards in length has 
recently been constructed. These quays difier from the old ones 
in the width of the open ground, which is about 270 yards. A 
completely covered shed 197 feet high extends over 1,200 yards of 
their length. Fifty hydraulic cranes have recently been erected 
there. Further up the river are the petroleum storage tanks. A 
flying bridge is used when vessels come alongside. Fi\e large 
pipes convey the oil towards the vast settling tanks situated on 
lower ground, which is divided into rectangular plots. Eeservoirs 
with all necessary appliances have been constructed by the lessees 
of the ground. 

The traffic of the Port of Antwerp has been gradually and 
constantly growing, as will be seen from the following figures : — 
In 1866 it was about one million tons (tonnage of entering 
sea-going vessels), in 1872 two million tons, in 1881 three 
million tons, in 1884 four million tons, in 1894 five million tons, 
and at the end of 1904 it had amounted to 9 J million tons. 

The fortified enclosure of Antwerp built between 1860 and 
1865 has just been opened towards the north, in order to make 
room for the development of the docks on that side. When the 
docks, the construction of which is now started, are finished the 
Port will have about If miles of additional quays, thus making 
provision for the constant growth of its trade. 



766 June 1905. 



CENTEAL RAILWAY STATION, ANTWERP. 

This magnificent railway station is quite new. The central 
dome has been constructed of ferro-concrete, and is illustrated in 
M. Noaillon's Paper, Plate 18. Some of the offices have not yet 
been completed. The electric signalling has been laid out by the 
Siemens and Halske Co., and the semaphore signals for in-coming 
and out-going trains are operated electrically from a signal box. 
Hydraulic or " glycerine " buifer-stops have been placed at the 
end of each line, within the station, to absorb the momentum of 
any in-coming train, in case the automatic brakes might fail to 
bring the train to rest at the proper place. 



JOHN COCKERILL SOCIETY'S WORKS, 
HOBOKEN, ANTWERP. 

Dockyard. — The Dockyard at Hoboken on the Scheldt is 
2 J miles above Antwerp, and occupies an area of 15 acres, having 
a river frontage of 328 yards. It has four slips suitable for ships 
of all dimensions and another dry slip 400 feet in length, closed 
towards the Scheldt by dock-gates. There is also a large 
crane capable of lifting boilers and engines into the holds of 
ships, and two coal tips on the river. The workshops are 
provided with all the machinery necessary for the preparation 
of metal plates, forged iron and steel parts used in the ship- 
building industry. The Wood-working Shop contains all 
necessary and up-to-date appliances. An electric power-station 
has recently been erected for the transmission of power to the 
various machine-tools, and the smoke is taken away from the 
forges by means of electric fans. 



Junk 1905. JOHN COCKERILL SOCIETY'S WORKS, HOBOKEN. 767 

The splendid mail steamers that run from Dover to Ostend 
were built at these Works, a description of the "Yille do 
Douvres" having already been given in the Fourth Eeport of the 
Eesearch Committee on Marine-Engine Trials.* Steamers for 
river navigation in Eussia have also been constructed here in 
large numbers, as well as stern-wheel steamers for the Congo. 
A boat of this type, which has recently been constructed, is able 
to carry 500 tons, having a maximum draught of 5 feet and a 
minimum speed of 7 knots. Owing to the great success of this 
steamer, an order for a second vessel was immediately given by 
the Congo Government. A dredger of great power has recently 
been constructed on the "Bates" system of hydraulic suction. 
This dredger is able to remove in a sandy soil 3,930 cubic yards 
per hour and ram it into floating pipes 984 feet in length. 

Fitting-out Shops. — This department deals with the maritime 
branch of the Works, and is equipped with a fleet of nine ships, of 
which two are of 2,400 tons, four of 4,000 tons and three of 300 tons 
of 17-knots speed. These steamers call regularly at different ports 
on the north and east coast of Spain, Italy, Tunis, and Algeria. 
Besides the ores which they bring for the blast-furnaces of the 
Cockerill Company, they carry the productions of the Works and 
of other firms. The three new and rapid steamers of 300 tons 
(above mentioned) have been built for a daily service between 
Ostend and London (Tilbury Docks), carrying goods of a 
perishable nature. 



DIAMOND CUTTING WOEKS, 
ANTWEEP. 

This Company (La Societe Anonyme la Taillerie Populaire 
Anverspise) was established in 1898 with a capital of £40,000. 
There are two factories, employing 750 lapidaries, outside diamond 



* Proceedings, 189g, pa^e 136, 



768 DIAMOND CUTTING WORKS, ANTWERP. June 1905. 

setters, etc., and everything connected therewith is of the latest 
design, the health and comfort of the workmen being especially 
provided for. The machines in the principal factory in the Rue 
Van Immersed are driven by a 450-H.P. Bollinckx engine. 



MINERVA MOTOR WORKS, 
BERCHEM, NEAR ANTWERP. 

This company was originally established in 1897 in the Rue 
Jacobs, Antwerp, under the name of S. de Jong and Co., for the 
purpose of manufacturing cycles. In 1898 these premises having 
become too small, it was decided to acquire a plot of land in the 
Rue de la Pepiniere at Berchem (a suburb of Antwerp) and to 
build thereon new and more commodious works. These works, 
completed at the end of 1898, consist of a two-row building ; 
the rows are 164 feet long, 33 feet wide and three storeys high 
with a large gangway between them. The works are equipped 
with a modern machine plant (chiefly American) in order to 
produce popular priced bicycles cheaply and quickly, the old Rue 
Jacobs Works still being used to help the new works in producing 
different parts of these bicycles. As during 1899 competition 
in the bicycle business became more and more keen, the Minerva 
Works started the construction of motor-bicycles and motors for 
bicycles. Some special machines were added to the existing tools 
in order to produce these motors, and gradually the Minerva 
motor found its way throughout the world. Since then nearly 
25,000 motors and motor-bicycles have been produced at these 
Works. 

Whilst producing motors for bicycles, the Minerva Works have 
been preparing during the last few years for the manufacture of 
motors for cars and motor-cars, and early in 1904 it was decided 
to start this manufacture. New ground opposite the bicycle motor 
works having been acquired, commodious and modern works were 



June 1905. MINERVA MOTOR WORKS, BERCHEM. 769 

erected. These cover an area of 3,229 square yards, and are 
composed of a building four storeys high. The different shops 
are high, well lighted and ventilated. The power is generated 
by a 200-H.P. steam-engine made b}^ BoUinckx of Brussels, on 
the main shaft of which a dynamo of 1,150 amperes is coupled, 
and the electricity so formed, after having passed through the 
distribution board, is sent through the whole works and used 
for power, and lighting the shops. The different tools are driven in 
groups by dynamos receiving their current from the central dynamo. 
Of two lifts driven by electricity, one conveys the men to their 
different shops, and the other brings the motor-cars and bulky 
goods to their respective departments in the different storeys. 

The ground floor contains on one side the general store room, 
assembling shop, repair shop and engine room, and on the other 
side the boiler room, wood store and drying room, a forge with 
pneumatic hammer, an electric testing-plant for car-motors, a 
brazing shop, and the office. The first floor contains the main 
machine toolshop, toolmakers' shop, draughtsmen's office and 
revising shop. On the second floor is found the motor and gear 
box assembling-shop, car-erecting shop and case-hardening 
furnaces. The third floor contains the wood- working department, 
machine shop, wheelmakers' shop, bodymakers' shop, painters' 
shop and upholstering-room. 

The old Eue Jacobs shop is still used for the production of 
accumulators and coils, steel rim making, and also for the 
production of hoods, water and spirit tanks, etc., for motor-bicycles 
and cars. 

There are about 900 men employed at the Minerva Works, 
which turn out about 15,000 motors and motor bicycles, 300 small 
cars, and 300 large cars per annum, and their goods are exported 
to all parts of the world. 



3 G 



770 June 1905. 



.MESSES. BOLLINCKX AND CO.'S WOEKS, 
BEUSSELS. 

These Works, situated between Chaussee de Mons and Eue 
Heyvaert, with the Eiver Senne dividing them, were established 
in 1863 for the manufacture of boilers, steam-engines, etc. 

The foundry is 328 feet long by 82 feet wide. It contains 
three 20-ton electric overhead-travellers. The whole output 
of this department is used by the firm's workshops alone. 
The erecting and finishing shop, also driven by electricity, is 
164 feet in length. The planing- shop, 105 feet long by 36 feet 
wide, has an electric overhead-traveller, slotting machines, and 
hydraulic presses for pressing in cranks, pins, piston-rods, etc. 
The heavy machine-tool shop, 210 feet long by 26 J feet wide, 
contains tools of the latest pattern driven by electricity. The 
number of men employed is 300. 

In 1903 this firm established Works at Buysinghen for the 
manufacture of gas-engines, and at the present time a new 
Foundry is being erected, which will be able to turn out four 
million castings per annum. 



ATELIEES GEEMAIN, 
MONCEAU-SUE-SAMBEE, NEAE CHAELEEOI. 

These Works, established in 1857 by the late M. Brisson, made 
until 1897 a speciality of the manufacture of railway and tramway 
rolling stock of all kinds, for ordinary and narrow gauge lines. 
Their productions are found in all parts of Europe and South 
America. Eecently a considerable extension of workshops has 
taken place, and the firm can now undertake to carry out at 
short notice the most extensive orders. In 1869 the firm changed 
its name when it passed into the hands of the late M. Germain, 



June 1905. ATELIERS GERMAIN, MONCEAU-SUR-SAMBRE. 771 

and in 1897 it was turned into a company bearing the present 
name. The manufacture of motor-cars was one of the objects for 
which this transformation was made. Towards the end of the 
year 1895, soon after the first motor-car race from Paris to 
Bordeaux and back, a small syndicate of engineers was formed to 
provide means for the construction and the trials of a motor-car. 
In order to acquire knowledge in this novel industry, the syndicate 
purchased motor-cars of various types. A car was built, tested, 
altered, and ultimately rebuilt altogether ; and it was this series of 
trials, research, and hard work which ultimately resulted in the 
formation of the Societe Anonyme des Ateliers Germain. At the 
time of its formation, the firm succeeded in acquiring the 
ownership of the Belgian patents of Emile Levassor and of 
Gottlieb Daimler, the latter joining the board of the Company as 
a Director. It is on this solid basis that the prosperity of the 
Ateliers Germain and their renown as constructors of motor-cars 
has been built up. 

The Works at Monceau-sur-Sambre are remarkable for their 
methods of organisation, their tools and machinery. The motor- 
car factory has a tall frontage with numerous windows. A large 
hall forms the central part. Galleries of three floors run all 
round, and numerous machines and tools are distributed in a 
systematic manner. The erecting shop is on the ground floor, 
where there are also special workshops for joiners and coach 
builders. A vast space underneath is used for the storing of all 
the parts manufactured as they come from the various workshops ; 
there they are carefully classified and placed in the different 
divisions provided with labels in large letters, upon which are 
given the number of the piece, the date when it was stored and 
when it left. The number of manufacture is put not only on 
each of the pieces, but also on each group and on each car as 
erected ; and to each of these numbers corresponds a sheet in 
the counting-house on which is written day by day the cost of the 
workmanship. The number of the various pieces manufactured 
and used as a rule in an establishment of this kind amounts to 
nearly 4,000. 

3 G 2 



772 ATELIERS GERMAIN, MONCEAU-SUR-SAMBRE. June 1905. 

The macliinery, which is being constantly improved, amounts 
at the present time to more than 200 machines and machine tools 
of every description. The innumerable small parts constituting 
the small tools are kept and classified in a special hall under the 
care of an employee who watches over their maintenance and 
distributes them, always in a perfect condition, among the workmen. 
The latter are not allowed to alter or repair any tool. The 
selection and the reception of the raw materials are likewise an 
important feature at this establishment. 

The new steels, of nickel, chromium and manganese, are 
now in constant use in spite of their high price and the difficult 
operations which they have to undergo in the laboratory. It is 
there that the molecular state of the same piece of steel is modified 
in different ways in its various parts, some of them acquiring 
hardness and the others malleability and flexibility according to the 
use to which they will be put. In this same factory, omnibuses are 
constructed, for instance, those supplied to the London Koad Car 
Company; also trucks, etc., and large motors for barges. 

The company has, with its reserve capital and sinking funds, 
a capital of £100,000. The machinery and tools alone cost over 
one million francs, and constant additions are being made to them. 
The number of workmen employed is about 700. The factory 
covers an area of more than 3J acres. The total horse-power of 
the motor-cars built by the company exceeded 10,000 in the year 
1904, and all these are still in good condition, none of the cars 
which have left the factory for the last eight years having yet 
been definitely declared to be out of service. 

The factory, excluding the offices, laboratories, room for the 
motor engines, wood stores, and annexes of various kinds, contains 
a workshop of 229J feet by 98J feet of four storeys specially used 
for the manufacture of motor-cars. On the lowest floor are the 
store-rooms ; on the ground floor the motor-cars are put together ; 
in the first gallery are the machine tools ; and on the second 
is the fitting shop. 

There are also the following : — A shop for the manufacture 
of railway trucks ; a forge with four steam-stamps ; three joiners' 



June 1905. ATELIERS GERMAIN, MONCEAU-SUR-SAMBRE. 773 

shops ; a wheelwrights' shop ; two workshops for painting and 
finishing railway material ; a wheel- erecting shop ; and numerous 
erecting shops. The workshops have a siding to the railway 
stations of Marchienne-au-Pont and of Monceau-Usines. 



WORKS OF VAN DEN KEECHOVE, 
GHENT. 

These Works were founded in 1825 by M. Emmanuel Van den 
Kerchove. In 1860 M. Prosper Van den Kerchove, son of the 
founder, became chairman and manager of the Company, when a 
rapid impulse was given to the business which showed itself by many 
successes at the Paris Exhibition of 1867. At that time he foresaw 
that the Corliss engine had such possibilities that he promptly 
made a contract with the inventor, by which he secured to 
himself the manufacture and sale of this engine throughout 
Europe. From that date onward he applied himself to make this 
engine perfect in the smallest details of its construction. Since 
1900 the firm has developed a new system of distribution by piston- 
valves, with which the engine exhibited at the Paris Exhibition 
of 1900 was furnished. When working normally, its power is 
1,000 H.P., with a steam-pressure of 133 lbs. per square inch and an 
expansion of 13 times the volume introduced. The diameter of the 
small cylinder is 24^ inches, that of the large cylinder 43 inches, 
and the stroke of the pistons 3 feet 11 inches, and the revolutions 
are 85 per minute, but the normal velocity is 100 revolutions. The 
cut-off valves are simple piston-valves absolutely tight and as 
durable as the piston; being in perfect equilibrium, they work 
with ease, and reduce the consumption of steam to a minimum. 
This firm also manufactures single-cylinder engines from 75 H.P. 
at 88 lbs. per square inch upwards, compound engines from 150 
H.P. at 133 lbs. per square inch upwards, and multiple-expansion 
engines. The normal speed of their engines varies according to 
the power from 140 to 90 revolutions per minute, and even less. 



774 June 1905. 



LA GILEPPE EESERVOIR DAM.* 

{See 'page lib,) 

This barrage, whicli is a triumph of modern engineering, was 
constructed in the years 1867-1878 by Messrs. Braive, Caillet and 
Co., from a plan drawn up by M. Bidaut for the purpose of forming 
a reservoir of pure soft water for the use of the town of Verviers, 
not only for drinking purposes but for the town's cloth factories. 
It consists of an immense embankment, 90 yards long and 72 yards 
thick at the base, and 256 yards long and 16 yards thick at the 
top, carried across a narrow part of the valley of the Gileppe. 
The height of the parapet is 154 feet, and the normal level of 
the reservoir is 148 feet, its area being 200 acres, containing 
2,700 million gallons. 

The water is taken from the reservoir by two subterranean 
galleries, which lie in the shape of a horse-shoe round the dam. 
At the entrance end of each is a grating for the purpose of 
filtration. Passing through this the water flows through the 
gallery till it reaches the sluices. Here the gallery is closed by a 
mass of masonry, through which are laid two cast-iron pipes, 
34 inches diameter ; each pipe is closed at the lower end by a 
self-acting valve. After passing the working chamber the water 
arrives at another masonry dam ; immediately below this is the 
safety-sluice. Before entering the aqueduct the water passes 
through a measuring apparatus. On the top of the embankment 
is a colossal lion, 43 feet in height, constructed by M. Felix Boure 
out of 187 blocks of sandstone. The total cost of these waterworks 
amounted to seven million francs (£280,000). 

* A more detailed description can be found in Proceedings 1 883, page 553. 



June 1905. 



LA GILEPPE RESERVOIR DAM. 



775 



General Plan of Gathering Ground. 
Watershed covers both Permeable and Impermeable Strata. 




776 Junk 1905. 



THE WORKS OF B. LEBRUN, 
NIMY, NEAR MONS. 

The products of this firm comprise machinery for ice-making 
and brewing, steam-engines and pumps, ventilating machines for 
mines, air compressors, and general engineering work ; but a 
speciality is made in their refrigerating machinery, which has 
been supplied to firms in all parts of the world. 



June 1905. 777 



MEMOIRS. 

William Black was born at Airdrie on 9th February 1823. He 
commenced his commercial career with the Jarrow Alkali Co., in 
conjunction with Messrs. Cookson's Chemical Works, at South 
Shields. Later on he was offered a partnership in the firm, but, 
owing to ill-health at the time, he was obliged to decline it. His 
attention was then directed to the iron trade, and he started a 
foundry at Fatfield, near Washington, Co. Durham. Eventually he 
extended his business operations by taking over the North Eastern 
Foundry at South Shields, and under his management the works 
were greatly enlarged. In conjunction with the late Mr. Hilton 
Philipson and others, he founded in 1865 the firm of Messrs. 
Black, Hawthorn and Co., of Gateshead, locomotive, marine, and 
stationary engine builders ; and in 1869 he founded the St. Bede 
Chemical Works, East Jarrow, which have since been absorbed by 
the United Alkali Co. He was also one of the promoters of the 
North Eastern Marine Engineering Works at Sunderland and 
Wallsend. His death took place at his residence in Newcastle-on- 
Tyne, on 12th July 1905, in his eighty-third year. He became a 
Member of this Institution in 1879 ; and was a Member of several 
learned and technical societies, including the North-East Coast 
Institution of Engineers and Shipbuilders, and the Institution of 
Mining Engineers. 

William Carter was born at Eotherham on 6th February 1849, 
and received his general and technical education at the Manchester 
Mechanics' Institution. He attended the science classes there with 
distinction, and ultimately became one of the directors of that 
Institution. He entered the works of Messrs. Sharp, Stewart and 
Co., locomotive builders, of Manchester, at the age of fifteen, and 



778 MEMOIRS. June 1905. 

having passed through the various shops and drawing office was, at 
the early age of twenty-four, made assistant works manager. In 
1876 he was appointed manager of the Patent Tube Works at 
Smethwick, Birmingham, and eight years later went to Ghent, in 
Belgium, where he became general manager of the Societe Anonyme 
du Phoenix, general engineers and textile machinery manufacturers. 
Returning to England in 1886 he was appointed manager and 
secretary of the Hydraulic Engineering Co. of Chester, and in 1897 
became its general manager. This position he held at the time of 
his death, which took place suddenly on the 9th August 1905, at 
his residence in Helsby, near Chester, from heart failure, in his 
fifty-seventh year. He became a Member of this Institution in 1877, 
and was also a Member of the Liverpool Engineering Society. 

David Evans was born at Aberdare on 17th October 1839. His 
engineering training was obtained at the Aberdare Iron Works, 
where he succeeded his father in 1866 as blast-furnace manager. 
In 1870 he became blast-furnace, forge, and mill manager at the 
Ehymney Iron Works, and in 1875 he accepted the post of works 
manager of the Ebbw Yale Iron and Steel Co.'s blast-furnaces, 
forges, and mills. He returned to the Ehymney Works to take up 
the general managership in 1878 ; and from 1885 to 1891 he was 
general manager of the Barrow Haematite Iron and Steel Works. 
In 1891 he became general manager of Messrs. Bolckow, Vaughan 
and Co., Middlesbrough, where, under his wise management, 
through periods of good and bad times, the firm enjoyed continued 
prosperity. That he was a successful manager of men is evidenced 
by the fact that no stoppage or strike occurred during the thirteen 
years of his management. He greatly interested himself in promoting 
the social welfare of the employees. He was a Justice of the Peace 
for Monmouthshire and for the North Riding of Yorkshire, 
Chairman of the Urban District Council of Eston, a Member of the 
Tees Port Sanitary Authority, a Director of the Cleveland Salt Co., 
and of the New Cransley Iron and Steel Co. He was a Past- 
President of the Cleveland Ironmasters' Association, a Vice-President 
of the Institution of Cleveland Engineers, a Member of Council of 



June 1905. MEMOIRS. 779 

the Iron and Steel Institute and of the British Iron Trade 
Association. He became a Member of this Institution in 1884, 
and was also a Member of the Institution of Civil Engineers, of 
the Institution of Naval Architects, the American Institute of 
Mining Engineers, and the representative of the Steelmakers of 
England and Wales on Lloyd's Committee. His death took place 
at Saltburn-by-the-Sea on 8th August 1905, in his sixty-sixth year, 
after a long illness. 

James Foster was born at Sunderland on 2nd October 1845, 
and came of an old and well-known North of England Quaker 
family. His apprenticeship was served with Messrs. Boyd, 
Thompson and Co., Newcastle-on-Tyne, and George Clark, 
Sunderland. In 1867 he went out to China, remaining there six 
years, and living at Hong Kong, Saigon, and Bangkok. In 1877 he 
proceeded to Java as the representative of Messrs. George Fletcher 
and Co., Derby and London, and remained there until 1884, when 
he came home, only to return again in the following year on his own 
account. He left Java finally in 1888, and shortly after entered the 
service of Messrs. Duncan Stewart and Co., Glasgow, in whose 
interests he visited Nicaragua, Antigua, and Bahia, Brazil. He 
joined Messrs. Fullerton, Hodgart and Barclay, of Paisley, in 1895, 
and remained with them till his death, which took place at Glasgow 
very suddenly on 10th March 1905, in his sixtieth year. He made 
a special study of sugar machinery and evaporating plant of all 
kinds, and was the inventor of numerous devices connected therewith. 
In 1890 he compiled and published a work on " Evaporation by the 
Multiple System," which has run to several editions. He became a 
Member of this Institution in 1888. 

Archibald Potter Head was born at Coatham, Redcar, on 
4th August 1866, and was educated at Clifton College. At the 
age of seventeen he entered the North Eastern Steel Works, 
Middlesbrough, for a few months until, in January 1884, he went to the 
neighbouring works of Messrs. Fox, Head and Co. There he spent 
six months in the pattern shop and three months in the smiths' shop. 



780 MEMOIRS. June 1905. 

In September 1884 lie became a pupil of Messrs. E. and W. Hawtborn, 
Leslie and Co., of St. Peter's Works, Newcastle-on-Tyne ; and in 
1886 was sent by them to St. Petersburg for six months, where he 
assisted in the completion and trials of the engines of a Eussian 
war-ship. On his return to this country in January 1887, he entered 
the drawing office of the same firm, where he remained until the 
completion of his apprenticeship in September 1888. He next 
commenced a two years' engineering course at University College, 
London, under Professor Kennedy and Professor Beare. During 
this period he was successful in obtaining the Gilchrist Scholarship 
of the value of £80, and in the next session was appointed 
demonstrator to the mechanical drawing class. In August 1890 he 
became assistant to his father, the late Mr. Jeremiah Head,* Past- 
President, in Middlesbrough, and was taken into partnership in 
1893, when they moved to London and commenced business there 
as Messrs. Jeremiah Head and Son, having also a branch at 
Middlesbrough. On the death of his father in 1899 he carried on 
the business alone until 1904, when his brother — Mr. Benjamin W. 
Head — was taken into partnership. A considerable portion of his 
time was spent abroad, chiefly in the United States, where his 
business engagements took him, for he was managing director, as 
well as consulting engineer, of the Otis Steel Co., of Cleveland, 
Ohio. It was his practice to visit these works at least once every 
year, and it was while he was returning home from the last of these 
visits that he met with his death. His firm was also sole European 
representative of the Wellman, Seaver, Morgan Co., of Cleveland, 
Ohio. In addition to reading Papers before the Institution of Civil 
Engineers, the Society of Arts, and other societies, he contributed 
one to this Institution in conjunction with Colonel Cubillo, on " The 
Manufacture of Cartridge-Cases for Quick-firing Guns," f the reading 
of which did not, however, take place until after his death. As 
mentioned above, he was returning from Cleveland by the " Twentieth 
Century " Express on the New York Central Eailway when the 



* Proceedings 1899, page 134. 

t Proceedings 1905, Part 4, 20 October. 



June 1905. MEMOIRS. 781 

engine, travelling at 75 miles an hour, left the rails. The momentum 
was such that the tender was hurled over the engine, and the 
combination coach in which he was seated was thrown on the engine. 
The injuries he received were so serious that, after lingering for 
twelve hours, his death took place in the Cleveland Hospital on 
22nd June 1905, in his thirty-ninth year. He was elected a 
Graduate of this Institution in 1885, and was transferred to full 
membership in 1892. He was also a Member of the Institution of 
Civil Engineers, of the Iron and Steel Institute, and of the Institution 
of Electrical Engineers. 

James Kirkwood was born at Southampton in July 1848. He 
served an apprenticeship with Messrs. J. and G. Thomson, 
shipbuilders, of Glasgow, and, on its termination, went to South 
Africa in 1868 as engineer of the Glasgow and Limpopo Co. to erect 
machinery in the gold-fields. In the next year he proceeded to 
Singapore as assistant engineer at the Docks and New Harbour, and 
in 1870 he became foreman and afterwards superintendent engineer 
of the Kowloon and Aberdeen Works of the Hong Kong and 
Whampoa Dock Co. In 1872 he received an appointment in the 
Chinese Imperial Maritime Customs ; and in this position he 
remained until 1880, when his services were lent to the Chinese 
Navy. He was first appointed to the 16-knot cruiser "Chao Tung" — 
the flagship of Admiral Ting — then being built at Elswick Works, 
Newcastle-on-Tyne. Soon afterwards he was promoted to the rank 
of Inspector of Machinery Afloat, and Superintendent of Works at 
Weihawei. For his services during this period he was decorated by 
the Imperial Chinese Government with the Order of the Double 
Dragon. In 1892 he returned to the Chinese Customs Service, in 
which he remained until his death, which took place at Chinkiang, 
Chinaj on the 5th February 1905, in his fifty-seventh year. He 
became a Member of this Institution in 1875. 

James Mansergh, F.E.S., was born at Lancaster on 29th April 
1834, and was educated at Harmony Hall, Hampshire. One 
of the masters was Dr. (afterwards Sir) Edward Frankland, the 



782 MEMOIRS. June 1905. 

distinguished chemist, and another was John Tyndall, to whom 
he owed his first impetus towards engineering. In 1849 he began 
an apprenticeship with Messrs. McKie and Lawson, engineers 
and surveyors, of Lancaster. At the age of twenty-one he 
went to Brazil, where he remained for four years as engineer to 
Mr. E. Price, contractor for the Dom Pedro Segundo Railway from 
Eio de Janeiro to the interior. In 1859 he returned to England 
and rejoined his old master, Mr. McKie, at Carlisle, doing general 
engineering work and laying out the first sewage farm in England. 
From 1862 to 1865 he was contractor's agent for Messrs. John 
Watson and Co., being engaged first on the Mid- Wales Eailway, and 
then on the Llandilo and Carmarthen Railway. In 1866 he 
entered into partnership with his brother-in-law, Mr. John Lawson, 
at Westminster, his first work being the laying out of a gravitation 
scheme of water supply for Carlisle. The partners, before Mr. 
Lawson's death in 1873, designed and carried out either sewerage or 
waterworks schemes or both at twenty-five towns in England. In 
1870 they were associated with Mr. (afterwards Sir) Robert 
Rawlinson in reporting upon the water supply of Birmingham. 
Mr. Mansergh recommended the scheme adopted twenty years later 
for utilising the valleys of the Elan and Claerwen. In 1878 he was 
requested to prepare a scheme for the sewerage of the Lower Thames 
Valley, and was awarded one of three premiums. The accepted 
scheme having been finally rejected, he was engaged with Mr. J. C. 
Melliss in 1883 to devise another, which was approved by the Local 
Government Board, but defeated in Parliament. He also was one 
of the chief witnesses before Lord Bramwell's Commission on 
Metropolitan Sewage Discharge, and suggested the mode of 
treatment since adopted. In 1889 he went to Australia to advise 
the Government of Victoria upon the sewerage of Melbourne and its 
environs. The complete scheme which he drew up was estimated 
to cost nearly six million pounds. Whilst in Melbourne he advised 
the Government upon the draft Metropolitan Board Bill, which 
provided for the incorporation of the districts of twenty-three local 
authorities within the City of Melbourne. The bill ultimately became 
law, and the scheme has been carried out almost as he recommended. 



June 1905. MEMOIRS. 783 

His name however will best be remembered in connection with 
the Birmingham water scheme. In 1890 he was again consulted by 
the Corporation of Birmingham, and reported once more in favour 
of the Elan and Claerwen scheme. He utilised completely a 
watershed area of 71 square miles, which suffices to provide 27 
million gallons a day as compensation to the Eiver Elan, and 75 
millions for the supply of Birmingham and the towns adjacent to the 
aqueduct. The water is conveyed by a conduit having 13 J miles 
of tunnel, 23 miles of cut and cover, and 37 miles of iron or steel 
pipes crossing valleys, with service reservoirs, filter beds, and 
accessory works. The estimate for this scheme amounted to nearly 
six million pounds. The supply was inaugurated on 21st July 1904 
by the King and Queen, who were conducted over the works in the 
Elan valley by Mr. Mansergh. 

During the last ten years of his life he was engaged in carrying 
out twelve more schemes of sewerage and sewage disposal, including 
those for Coventry, Derby, Exmouth, and Plymouth, and twenty 
works of water supply, also making ninety-two reports on sewerage 
and water. His services were much in demand, as may be shown 
by the fact that he acted for no fewer than 360 municipalities, local 
bodies, &c., and he prepared upwards of 250 reports on sewerage 
and waterworks alone. As a witness before committees of both 
Houses of Parliament he appeared upwards of six hundred times ; 
and he was appointed a member of the Koyal Commission on the 
Metropolitan Water Supply, 1892-93. He lectured from time 
to time before various bodies, generally on the subject of water 
supply, and his Presidential Address to the Institution of Civil 
Engineers dealt with the history of water engineering. In March 
1903 he was made an honorary freeman of his native town of 
Lancaster. He was a Justice of the Peace for Eadnorshire, of which 
county he was High Sheriff in 1901. He was elected a Member of 
this Institution in 1875, and served on the Council from 1902 till 
his death. He was also a Member of the Institution of Civil 
Engineers, and was elected President in 1900. In 1901 he was elected 
a Fellow of the Koyal Society. His death took place at his residence 
in Hampstead, London, on 15th June 1905, at the age of seventy-one. 



784 MEMOIRS. June 1905. 

William Sumner was born at Altrincham on 6th October 1830. 
He was educated privately, and started business with Messrs. 
Richard Eoberts and Co., of Manchester. Subsequently he was for a 
few years with the firm of Messrs. Piatt Brothers and Co., of 
Oldham. In 1855 he founded, with his brother John, the firm of 
John M. Sumner and Co., engineers and machinery exporters, of 
Manchester, which firm is still in active operation. In 1864 he 
joined the board of the Broughton Copper Co., of Manchester and 
Ditton, brass and copper manufacturers and smelters, becoming 
chairman and managing director. This company was then only a 
small concern, and he lived to see it take a high position amongst 
similar industries. He was also at difi'erent times director of the 
Ebbw Vale Steel Iron and Coal Co., the Standish Co. (now incorporated 
with the Bradford Dyers' Association), and of the Clayton Aniline 
Co. His death took place at his residence in Prestwich, Manchester, 
on 23rd August 1905, in his seventy-fifth year. He became a Member 
of this Institution in 1861, and was also a Member of the Iron and 
Steel Institute, and other kindred Societies. 



Oct. 1905. 785 



Cjje Institntian 0f P^ecljankd Engineers, 



-^ 



PROCEEDINGS. 



October 1905. 



The first Ordinary General Meeting of the Session was held 
at the Institution on Friday, 20th October 1905, at Eight o'clock p.m. ; 
John A. F. Aspinall, Esq., Vice-President, in the chair. 

The Chairman regretted the absence of the President, and 
sympathetically referred to the loss the Institution had suliered by 
the lamented deaths of three prominent Members who had passed 
away since the last Meeting, namely, Sir Edward H. Carbutt, Bart., 
Past-President, Mr. William Dean and Mr. James Mansergh, F.R.S., 
Members of Council. Sir Edward Carbutt joined the Institution in 
1860, was elected a Member of Council in 1875, and occupied the 
Presidential Chair in 1887-1888. He always evinced great interest 
in the welfare of the Institution, being a regular attendant at the 
various Meetings, and frequently taking part in the discussions. 
Mr. Dean was elected a Member in 1868, and served on the Council 
from 1892 continuously to the end, though his failing health latterly 
had prevented his usual regular attendance. Mr. Mansergh joined 
the Institution in 1875, and was appointed a Member of Council in 
1902. The Council had resolved to send a letter of condolence to 
Lady Carbutt expressing their deep regret at the loss of so valued a 
Member ; and similar letters had been forwarded to the son of 
Mr. Dean and the family of the late Mr. Mansergh. 

3 H 



786 



COUNCIL APPOINTMENTS. 



Oct. 1905. 



The Chairman also announced with great regret the resignation 
of Mr. Henry D. Marshall, of Gainsborough, who had been a Member 
of Council since 1889. 

For supplying the vacancies caused by the deaths of Mr. Dean 
and Mr. Mansergh, and the resignation of Mr. Marshall, the Council 
had appointed Mr. H. F. Donaldson, of Woolwich, Mr. J. Eossiter 
Hoyle, of Shefiield, and Mr. James Kowan, of Glasgow, as Members 
of Council. These three gentlemen would retire at the next Annual 
General Meeting, but would be eligible for re-election. 

The Minutes of the previous Meeting were read and confirmed. 

The Chairman announced that the Ballot Lists for the election 
of New Members had been opened by a committee of the Council, 
and that the following one hundred and twenty-five candidates were 
found to be duty elected : — 



MEMBERS. 

Anderson, John, . . . . 

Bridges, Walter, . . . . 

Browne, Eobbrt Jamieson, 
Brunton, James Forrest, . 
Everett, Wilfred Hermann, 
Fox, William, ..... 
Gray, James, ..... 
Henderson, James Francis, 
Hutohins, Walter James, . 
Lloyd, Frederick Lindsay, Major E.E., . 
McBride, William, .... 
McCoLL, Hugh, ..... 
Mills, Frederick, .... 
Morton, Thomas Morton Gray, . 
Nugent, Charles Hugh Hodges, 

Capt. E.E 

Parker, John Henry, 
Quince, William John, 



Christchurch, N.Z. 

London. 

Calcutta. 

Karachi. 

Howrah. 

London. 

Edinburgh. 

Chinde, Brit. C. Africa. 

Paris. 

London. 

Hartlepool. 

Newlands, Cape Colony. 

Ebbw Vale. 

Errol, Perthshire. 

London. 

West Hartlepool. 

Pietermaritzburg. 



Oct. 1905. 



ELECTION OF NEW MEMBERS. 



787 



KoBiN, Frederick Alexander Garibaldi, 
KuNDALL, Charles Frank, Capt. R.E. 
Spencer, John Aubrey Berkley, 
Stirling, David Edward, . 
Sykes, Joseph Charles, 
Walker, Robert Lea, 
WooDHOusE, Richard, 

associate members. 

Allcock, Harry, 
Barrett, Harry, 
Beard, William Keith, 
Bergersen, Sidney Hill, . 
Berry, Frederick Laurence, 
Berry, Henry Frank, 
Blaker, William Herbert, 
Brown, Walter Lindahl, . 
Bull, Edward Joseph, 
Calvard, Frank, 

Camargo, John Antonio da Rocha, 
Campbell, Andrew, . 
Chambers, Edward Haward, 
Clapham, Frederick Thomas, 
Clapham, Thomas Alborn, 
Clayton, Charles Henry James, 
Cooper, Joseph Albert William, 
Craven, John Ernest Holdswortii, 
Crosier, Edward James, . 
Cross, Edward, 

Cross, Reginald Thomas Green, 
Dalby, William Ernest, . 
Darlington, Seymour Nance, 
Davson, Simon Silver, 
Divecha, Ramchundra Nursey, . 
Engholm, Alexander Goldie, 
Farnsworth, Frank Smedley, 



London. 

Chatham. 

Newcasfcle-on-Tyne. 

Taltal, Chile. 

London. 

Wigan. 

Brighouse. 



Johannesburg. 
Durban. 
Bristol. 

Fenchuganj, Assam. 
Sheffield. 
London. 
Durban. 
Southampton. 
Chupra, Bengal. 
Newport, Mon. 
Lisbon. 
London. 
Tipton. 
Birmingham. 
Keighley. 
London. 
Karachi. 
Taunton. 

Newcastle-on-Tjne. 
Rotherham. 
Cachar. 
London. 
London. 
London. 
Ahmedabad. 
London. 
Yokohama. 
3 H 2 



788 



ELECTION OF NEW MEMBERS. 



Oct. 1905. 



Gilchrist, James Archibald Robertson 

Glen, 
Grose, Francis Howard, . 
Hall, Arthur Henry, 
Hardman, Frederic, . 
Harpur, Samuel John, 
Hartley, Richard Frederick, 
Hawkins, John Charles, . 
Hayes, Herbert Charles, . 
Hbald, Harry, 
Hearn, Charles Hanbury. 
Hill, Walter Hereward, . 
Hinchliff, Henry Walton, 
HiNDLE, Robert Sidney, 
HoPKiNSON, Austin, . 
Kerr, Peter Wyllie, 
KiRBY, Edward James, 
KiRLEw, Richard Leopold, Jun. 
Laird, Stanley Morrison, . 
Laurenson, George Henry, 
Lawson, John Charles Staveley, 
LiMPUs, Arthur Edward Jewers, 
Marshall, William Johnstone, 
McDonald, William, 
Merrett, John Alfred, 
MiLLETT, Charles Walter, 
Newman, William Boughton, 
Nicholson, Edward, . 
Polden, Francis Charles, . 
Potter, Henry Samuel, 
PuLSFORD, Frederick Charles, 
Quick, Albert Hedley, 
Ritchie, Marriott Claude, 
RoBSON, John Henderson, . 
Schiller, Frederick William, 
Shaw, Thomas Raynor, 



Tientsin. 

Chatham. 

Woolwich. 

Manchester. 

London. 

Woolwich. 

Uitenhage. 

London. 

Chorley. 

Durban. 

Bangkok. 

Bradford. 

London. 

Manchester. 

London. 

Derby. 

London. 

London. 

Liverpool. 

Reading. 

Calcutta. 

Reading. 

Singapore. 

Wellington, N.Z. 

London. 

Preston. 

Manchester. 

Bulawayo. 

London. 

Leicester. 

London. 

London. 

London. 

London. 

Manchester. 



Oct. 1905. 



ELECTION OF NEW MEMBERS. 



789 



Slater, Ernest, 
Smith, Arthur Joel, 
Stmons, Diogo Andrew, 
Taylor, John Thomson, 
ToDER, John, . 

TUNNIOLIFFE, WaLTER GbORGE, 

Wans, Oswald, 

Warbrook, Frederick George, 

Wilson, John, . 

Wright, Thomas George, . 



graduates. 
Abbott, Alfred, 
Algar, Stanley Curtis, 
Beatson, Archibald Meade, 
Chenhall, Stanley Sanders, 
Craig, George, 

Cranswick, Charles Tennyson, . 
Cronk, Norman Tylee, 
Crouch, Henry, 
Cruse, George Henry Irons, 
Dallow, Walker Horton, . 
Edmonds, Harold Montagu, 
Felton, George John, . 
Fitzherbert, George French Windbatt, 
Florey, James William, 
Hanford, Rupert, 
Hebden, George Brentnall, 
IvATT, Henry George, 
Lawrence, John James Butter, . 
Mason, Herbert Richard, . 
Palmer, Edward William, 
PiRRiE, Robert Hamilton George, 
Priest, Clarence Sidney, . 
Putnam, Percy, 
Roberts, Oswald Willan Burra, 



London. 

Birmingham. 

London. 

Leith. 

Sydney. 

London. 

Howrah. 

Sheffield. 

Northampton. 

BristoL 



Rangoon. 

London. 

Birmingham. 

Dulverton. 

Perak. 

Cachar, India. 

Crewe. 

London. 

Basingstoke. 

London. 

London. 

London. 

London. 

London. 

Wolverhampton. 

London. 

Crewe. 

London. 

London. 

London. 

Lincoln. 

Birmingham. 

London. 

London. 



790 



ELECTION OF NEW MEMBERS. 



Oct. 1905. 



EoNALD, Lionel Eodgeb, 
Shearman, John, Jun., 
Thomas, Frank George, 
Thorpe, Harold Thomas Emmerson, 
Young, Geoffrey, 



Liverpool. 

Crewe. 

London. 

Johannesburg. 

London. 



The Chairman announced that the following thirteen Transferences 

had been made by the Council since the last Meeting : — 

Associate Members to Memhers. 

Carnegie, Francis, ...... Woolwich. 

Fletcher, Harold Clarkson, .... Bulawayo. 

Fletcher, William Charles, .... Cairo. 

Garratt, James Herbert, ..... Colombo. 

Ham, Frederic George Sison, .... London. 

Longbottom, John Gordon, .... Glasgow. 

Middleton, Harry Herbert, .... Durban. 

Morris, Egbert Edmund, ..... London. 

Peters, Lindsley Byron, ..... London. 

Williams, Hal, . . . . . . London. 

WooDEsoN, William Armstrong, .... Gateshead. 

Graduates to Associate Memhers. 

Shepherd, James Horace, ..... Swindon. 

Waring, Henry, . . . . . . Dublin. 



The following Paper was read and discussed : — 
The Manufacture of Cartridge-Cases for Quick-Firing Guns " ; by 
Colonel Leandro Cubillo, of Trubia, Spain, and the late 
Mr. Archibald P. Head, Member, of London. 



The Meeting terminated at Ten o'clock. 
155 Members and 62 Visitors. 



The attendance was 



Oct. 1905. 791 



THE MANUFACTURE OF CARTEIDGE-CASES 
FOR QUICK-FIRING GUNS. 



By Colonel LEANDRO CUBILLO, of Trubia, Spain, 
AND THE LATE Mb. ARCHIBALD P. HEAD, Member, of London. 



The development of the quick-firing gun has at once necessitated, 
and been rendered possible by, improvements in ammunition with a 
view to quick loading. Quick-firing guns differ from ordinary guns 
in having the propelling charge and the means of ignition contained 
in a metal case. The projectile may or may not be attached to the 
case, forming a complete cartridge, as this depends on the size of 
the gun. In ordinary guns the projectile, the propelling charge, and 
the primer or means of ignition, are all separate, the charge being 
usually contained in a combustible silk cloth or serge case. The 
advantages which metal cases present as compared with combustible 
cases are: (1) They are quicker in loading, since the primer forms an 
integral part. (2) The same reason reduces the probability of a 
miss-fire. (3) The sponging-out of the gun, to avoid the possibility 
of the burning remnant of a combustible case prematurely igniting 
the next charge, is avoided. (4) The expansion of a brass case under 
fire enables it to act as a gas-check, rendering the use of an obturator 
unnecessary. 



792 CARTRIDGE-CASE MANUFACTURE. Oct. 1905. 

Simultaneous loading with " fixed " ammunition, in which the 
projectile is attached to the cartridge-case, is practised with quick- 
firing guns up to about 3 inches diameter, above which size complete 
cartridges would be too unwieldy. Between 3 inches and 6 inches 
therefore, separate loading is the rule, with the projectile separate 
from metallic cartridge-case. Above 6 inches the gun ceases to be 
called quick-firing, and combustible cases with separate loading are 
used. Separate loading for larger quick-firing guns is desirable, 
not only because of the excessive weight of a complete cartridge, but 
also because of the danger of storing loaded and fused shell in the 
same magazine with loaded cases. With separate loading, the 
projectile may be placed near the gun at leisure, the cartridge-cases 
not being taken from store until the last moment. Such different 
conditions govern the storage, transport, and use of projectiles and of 
cartridge-cases that it is undesirable to attach them together. 

The object of this Paper is to describe the new plant recently 
completed at the Royal Spanish Arsenal at Trubia, near Oviedo, 
Spain, for the manufacture of brass cartridge-cases from 3 inches to 
6 inches diameter inclusive, the machinery for which was acquired 
in 1900 under Ithe direction of the first named author and to the 
designs and under inspection of the second named author. Cartridge- 
cases for quick-firing guns are universally made of brass, this 
material having been found to possess the qualities best suited for 
this exacting service. Some detailed considerations of the history 
and mechanical properties of brass will therefore not be out of 
place. 

Of all the numerous alloys that of copper and zinc, commonly 
called brass, ranks as one of the most important. At one period 
the generic name of bronze was given to this alloy as well as to 
that of copper and tin, to which it is now applied. The two alloys, 
copper-zinc and copper-tin, are each characterised by well-defined 
properties, and each should retain its proper name of brass and 
bronze respectively. Brass was known to the Greeks and Romans, 
although they were unacquainted with zinc in its pure state, and in 
the manufacture of brass they only used the compounds of zinc. 
The ancients used bronze and brass in the manufacture, of coins, 



Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 793 

arms, tools, works of art, and ornaments. Professor Thurston, an 
authority on bronze and brass, says that the latter alloy may be 
rendered hard or soft, brittle or ductile, strong or weak, elastic or 
inelastic, dull or as lustrous as a mirror, friable or almost as ductile 
as lead, merely by varying the proportion of the two constituents. 
No other metal or alloy, not even excepting iron, presents such 
widely varying qualities, or so great a field of application. 
Commercial brass consists of two parts of copper and one of zinc, 
and is used, with certain exceptions, in all countries for cartridge- 
cases, not only for rifles but for quick-firing guns. The exact 
composition is 67 per cent, of copper and 33 per cent, of zinc, with a 
margin of * 5 per cent, above or below for either metal. The French 
artillery department, which is noted for the care with which its 
specifications for cartridge-case metal are drawn up, not only specifies 
the above-named proportion and variation, but requires that the 
constituent metals shall be of accepted brands, and of known origin, 
the sources of supply of copper being limited to the following : — 
" Calumet and Hecla," " Tamerack," " Ovscila," " Atlanta," 
" Franklin," " Quincy," " Wallaroo," and that manufactured by 
electrolytic deposition. The brands of zinc specified are: — " Yieille 
Montague," known as " Extra Pure Fonte d'Art," " Oeschger 
Mesdach," " O.M. Art Zinc," and that of the Royal Asturian 
Company of Spain, known as the "R. C. A. Refinado." Subject 
to certain limitations, the use of scrap brass is also allowed. 

Commander Pralon, of the French Artillery, who has had 
great experience in the manufacture of cartridge-cases, published at 
Bourges, in 1892, the result of a remarkable research, dealing with 
the mechanical properties of brass for cartridge-cases, as ascertained 
by some 158,000 tensile tests carried out under his supervision. A 
question which has not yet been satisfactorily settled, is whether, 
generally speaking, the alloys of copper and zinc, and in particular 
that of 67 per cent, of copper and 33 per cent, of zinc, form a 
perfectly definite chemical compound. This is answered in the 
affirmative by Commander Pralon, who bases his opinion chiefly on 
chemical analysis. The researches of Calvert, Johnson, Mathiessen, 
and Muspratt, have left the problem unsolved. Calvert and Johnson 



794 CARTRIDGE-CASE MANUFACTURE. Oct. 1905. 

maintain that some metallic alloys, especially those of copper and 
tin, are true chemical compounds, but Mathiessen holds that they are 
solidified solutions. It is probable that metallic alloys are not true 
chemical compounds, but solid solutions of one metal in another, or 
in several. In order that two bodies shall form a new one, it is an 
essential condition, first, that they combine in a fixed ratio, and 
second, that the original characters of the elementary substances be 
lost in those of the compound. A compound cannot be separated by 
mechanical, but only by chemical agency. Metals which form alloys 
do not as a matter of fact mix in a definite ratio. This constitutes 
the chief evidence that alloys are not true chemical compounds, since 
the loss of original characteristics and assumption of others does 
not always occur in alloys. Thus the mechanical properties of 
alloys differ only in degree from those of the elements composing 
them. Metallic alloys have been happily compared with such 
substances as crystals or obsidians, termed solid solutions on account 
of their similarity to true solutions. Liquid solutions possess the 
characteristics that the component elements cannot be detected by 
the microscope, and that they are mixed in indefinite proportions. 

Dr. H. Gautier states that the alloys are crystalline substances, 
which may generally, although not always, be verified by a 
microscopic examination of the fracture. This view is confirmed by 
the fact that they exhibit one of the characteristic properties of 
crystalline substances, namely, sudden solidification accompanied by 
the abstraction of latent heat. From the fact that the alloys are 
crystalline, it results that the action of heat thereon is the same as 
on crystalline mixtures, mixtures obtained by fusion, and ordinary 
solutions. The examination of the physical properties of alloys, such 
as conductivity or melting-point, may indicate their composition. 
At Trubia Arsenal, the behaviour during cooling of an alloy of 
67 per cent, of copper and 33 per cent, of zinc, being that from which 
cartridge-cases are manufactured, has been determined with great 
accuracy. The instrument used was Le Chatelier's thermo-electric 
pyrometer, which was carefully calibrated experimentally, taking as 
fixed points the melting temperatures of copper 1,080" C. (1,976° F.), 
of aluminium 654" C. (1,209° F.), and of lead 332" C. (629" F.). A 



Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 705 

weight of 300 grammes of brass was melted in a furnace, precautions 
being observed to ensure slow cooling, and to ensure tbat the 
temperature of the melted mixture did not rise to more than 50 or 
60 degrees above the melting-point, in order that the quantity of 
zinc volatilised should be as small as possible. 

In Fig. 1 (page 796) A shows the resulting cooling-curve, in which 
the abscissae represent time, and the ordinates temperature. An 
examination of this shows that there is only one critical point, 
corresponding to the point of solidification of the metal, where 
the temperature remains practically constant with a variation 
of only 2° C, for 150 seconds. Cooling is then resumed, 
becoming slower as the temperature falls. This critical point 
corresponds to the evolution of the latent heat absorbed during 
the reverse process of melting. The fact that no other critical 
point occurs, indicates that in this alloy neither of the two 
metals is present in excess. No rise of temperature occurs which 
might be caused by allotropic change. The alloy of 67 per cent, 
of copper and 33 per cent, of zinc is therefore a eutectic alloy. 
Sir William Koberts- Austen, in his comments upon the researches 
of M. Charpy on alloys of copper and zinc, states that in a mixture 
of less than 30 per cent, of zinc, there exists only one point of 
solidification somewhat higher than that of pure zinc. In a former 
investigation of the cooling-curve at Trubia, starting from a 
temperature lower than the melting-point, another critical point 
was found approximately coinciding with that of the melting-point 
of zinc ; but on repeating the experiment several times it failed to 
reappear. The alloy of 67 per cent, of copper and 33 per cent, of 
zinc, has a melting-point of 930° C. (1,706® F.j, which agrees with 
that obtained by M. Charpy, and also with that found empirically 
by Mr. Mallet. In order to verify the statement by Roberts- Austen, 
the cooling-curve of an alloy of 80 per cent, of copper and 20 per 
cent, of zinc was determined, as shown by the curve B in Fig. 1 
(page 796). To verify the melting-point of the two metals, they were 
melted separately and then mixed. From the curve it is clear that 
the mixture has only one point of solidification at 990° C, from 
which it appears that this is also a eutectic alloy. 



796 



CARTRIDGE-CASE MANUFACTURE. 



Oct. 1905. 











j 




■ 










/ 






•« 








/ / 






^ 








/ / 






•o» 














O 








/ / 






f^ 








/ / 




■ 


?3S 








/ / 






s 








' 1 






•(S- 










































« 






/ 


/ 




■ 


Qi 






/ 


1 






K> 






/ 


j 






o 






/ 


1 






rO 








f 






C3 






®x/ 








fi 






^/ 


/ 






^ 














o 






.tj / 


f 






< 






/ 


' 






CO 

CO 






^/ / 








« 
^ 






w / 








e^ 






^^ 








^ 






/ 






■ 


80 






/ / 








:;j 






/ / 








«. 






/ /^ 








d 






/ k 








s^ 




/ 


M 






















o 




7 


/^ 








1 




/ 










r-< 




/ 










.; 




/ / 










H-l 




/ / 










b 


/ 


A 










/ 


/ 
















Fig. 4. 
Diagram of Stress. 




a. 



0^ 



^ 



Temperature Centigrade 



m 



stress, Tons per sq. in. 



Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 797 

Mattliiessen divided the metals constituting alloys into two 
classes : — 

(a) Those which impart to the alloys their physical properties 
in proportion to their presence in the alloy. 

(h) Those which do not. 

The metals of the first class are lead, tin, zinc, and cadmium, 
the remainder being probably in the second category. The physical 
properties may be divided into three classes : — 

(1) Those which are exhibited by the alloy in proportion to the 
presence therein of the respective metals. 

(2) Those which are not so exhibited. 

(3) Those which sometimes are and sometimes are not so 
exhibited. 

The physical properties typical of the first category are density, 
specific heat, and coefficient of expansion. Those typical of the 
second category are the melting-point and the crystalline form, 
while those typical of the third category are the thermal and electric 
conductivity, the ring, the tenacity, and the elasticity. To the 
manufacturer of metallic cartridge-cases, the tensile and compressive 
strength of brass is the most important property to enable it to 
sustain the stress of general permanent deformation without 
exceeding the breaking stress. 

The copper-zinc alloy, containing 67 per cent, of copper and 
33 per cent, of zinc, of which cartridge-cases are made, and which 
is also employed very widely for other purposes, possesses 
remarkable mechanical properties. Among those are high tensile 
strength and large percentage of elongation when annealed. 
Metallic cartridge-cases, however, cannot retain their annealed 
condition because they require a certain amount of hardness to 
enable them to withstand, without permanent deformation, the 
enormous stresses to which they are subjected at the moment of 
firing. Tensile strength and ductility are physical properties of 
zinc and copper which are imparted to the alloy in a much higher 
degree than that possessed by the component metals. For example, 
annealed copper has a tensile strength of about 13 tons per square 
inch and an elongation of 45 per cent., whilst zinc has a tensile 



798 CARTRIDGE-CASE MANUFACTURE. Oct. 1905. 

strengtli of 1*336 tons per square inch. But the alloy of 67 per 
cent, of copper and 33 per cent, of zinc has a tensile strength in the 
annealed state of 19 J tons per square inch, an elongation of 68*9 per 
cent, and a contraction of area of 29*4 per cent. (See curve A 
in Fig. 2, page 796.) The stress-strain diagrams published by 
M. Charpy, for various compositions of brass, indicate that the 
alloy containing 67 per cent, of copper and 33 per cent, of 
zinc possesses the greatest tenacity and ductility, and requires 
in the annealed state a greater expenditure of work for its 
rupture. The elastic limit is only 23*8 per cent, of the 
breaking stress and is reached without appreciable deformation, 
this being characteristic of annealed metals, but more pronounced 
in the case of this metal than in other metals with the same 
ultimate strength, such as wrought-iron. Professor Le Chatelier 
defines the perfectly annealed metal as that of which the elastic 
limit is zero, or in other words, as that in which the metal is 
completely plastic, and in which any force, however small, would 
cause permanent deformation. Such bodies do not exist in nature 
any more than their antitheses, the perfectly elastic bodies. In the 
authors' opinion, the state of perfect annealing may be defined 
as that corresponding to the highest degree of ductility which it 
is possible for the metal to acquire with the lowest tensile strength. 
Malleable metals, such as iron and copper, are usually worked hot, 
but brass, on account of its extraordinary ductility, is capable of 
undergoing in the cold state great deformation with the expenditure 
of comparative little force. The working of metals under these 
conditions, however, tends to produce hardness and brittleness. If 
the elastic limit is exceeded, as is essential, the substance acquires 
a new and higher limit of elasticity, the measure of which is the 
force which has caused the permanent deformation. By successive 
deformations it is possible for the elastic limit to approach the 
breaking stress. Test-bars cut from metal drawn into cartridge- 
cases exhibit a much higher tensile strength than annealed metal, 
and if the deformations have been excessive, the tensile strength 
may be more than doubled. The percentage of elongation is 
diminished to a remarkable degree, falling to 2 per cent, when the 



Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 799 

tensile strength is double that of the annealed metal. This may 
be explained by regarding the stresses to which the metal is 
subjected during drawing as combined tensile stress and lateral 
compression. Curves B^, and B.^, Fig. 2 (page 796), show stress-strain 
diagrams of brass in an incomplete state of brittleness, the elastic 
limit differing little from the tensile strength, which is 39 per 
cent, greater than in the annealed state, while the elongation is 
only 29*6 per cent, of that of the annealed metal. A state of 
perfect brittleness would be that in which the elastic limit coincided 
with the tensile strength ; there would be no permanent elongation, 
and the body would therefore be a perfectly elastic one. 
M. Le Chatelier has found that pure metals, when at a maximum 
degree of brittleness, have tensile strength approximately double 
that of the perfectly annealed metal. This law was arrived at from 
experiments in wire-drawing. When the wire attained double the 
original strength, it was found useless to draw it through the plate 
again as the strength could not be further increased. 

Before describing the manufacture of cartridge-cases, reference 
may be made to the method of annealing, an operation of the highest 
importance, and which requires to be repeated many times throughout 
the process. The manufacture of cartridge-cases consists of a series 
of cold drawings which impart brittleness, the more pronounced the 
greater the deformation. The ductility must be restored after each 
drawing by heat treatment carried out between certain limits of 
temperature, and followed by either a sudden or slow cooling. 
Experiments have been carried out by MM. Charpy and Pralon, 
which have been confirmed by other results at Trubia, and which 
show that such limits of temperature vary very widely, the low limit 
being lower the greater the degree of brittleness of the metal. The 
high limit is not far removed from the melting-point, and is 
determined by whether it is desired to diminish the ultimate strength 
only of the metal, or the elongation also. At the temperature 
required for the latter, the metal begins to become burnt, M. Pralon 
being of the opinion that cartridge-case brass begins to be burnt at 
600° C. (1,112° F.), but neither the experiments of M. Charpy nor 
those at Trubia bear out this opinion. M. Charpy states that the 



800 CARTRIDGE-CASE MANUFACTURE. Oct. 1905. 

metal may be annealed at a temperature of 700° C. to 730° C. 
(1,292° F, to 1,346° F.), and at Trubia blanks have been annealed 
with good results between 670° C. and 740° C. (1,058° F. and 1,364° F.) 
according to the thickness of the pieces. The most suitable 
temperature is 620° C. to 650° C. (1,148° F. to 1,202° F.). On 
reaching the required temperature, the metal may be cooled either 
suddenly or slowly, since the speed of cooling does not affect its 
physical qualities, as has been verified not only by tensile tests, but 
also in the working of the process. 

M. Le Chatelier draws attention to " spontaneous annealing," 
which can be effected by leaving a piece of metal, which has been 
rendered brittle by mechanical treatment, to itself at ordinary 
temperatures, when it will be found to have annealed itself slowly by 
the action of time, with consequent reduction in tensile strength and 
increase in elongation. This appears to be the most satisfactory 
explanation of the phenomenon that simple metals are capable of an 
indefinite amount of deformation. The degree of brittleness becomes 
constant after a certain amount of deformation, due to the fact that 
every increase of brittleness produced by new deformation is neutralised 
by a corresponding annealing effect. This phenomenon of spontaneous 
annealing was known to the ancients. Many castings with abnormal 
internal stresses, and other articles subjected to excessive and long 
continued strains, such as chains, acquire a crystalline texture, but 
recover their normal condition either by the action of heat or by the 
much more prolonged action of time. Many years ago, when cannon 
were constructed exclusively of cast-iron, an idea was current among 
artillerists that it was injurious to fire them immediately after 
completion, and that sufficient time should be permitted to elapse to 
allow the molecular state of the metal to reach the point of perfect 
equilibrium. 

Microstructure of Brass. — The micrography of metals is now 
regarded as an important branch of metallurgy. The microscope is, 
in fact, the most powerful assistant to chemical analysis, and by its 
means the metallurgist ascertains the distribution of the various 
constituents of metals and metallic alloys. Micrography also shows 



Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 801 

the relationship between the microstructure of the metal and its 

mechanical properties, chemical composition, and thermal treatment. 

Five photo-micrographs of brass for cartridge-cases have been 

prepared, and accompany this Paper. The method adopted in their 

preparation, especially the etching by electrolysis, is that recommended 

by M. Charpy. It is not necessary to polish the surface of brass to 

such a high degree of perfection as in the case of steel. The sample 

is first filed smooth, afterwards polished with emery cloth of 

successive degrees of fineness, and finally rubbed by doe skin with 

rouge and fatty substances. After removing the grease with benzine 

and alcohol applied with a brush, it is dipped for a few minutes in a 

hot concentrated solution of caustic soda, and finally washed in pure 

water. It is then etched by the electrolytic process, being placed on 

a platinum support immersed in a vessel containing water acidulated 

with sulphuric acid in the proportion of 1 to 10. In this vessel is 

another porous vessel containing a saturated solution of sulphate of 

copper, and a small strip of copper which is in electrical contact with 

the platinum support. By electrolytic action, the alloy is slowly 

dissolved, the etching process taking from a quarter to half-an-hour. 

A, Plate 38, represents the metal taken from an annealed disc 

X 150 diameters. 

B is the same metal which has passed through the first drawing of 
cupping process, and has been annealed at 650° C. x 50 diameters. 
C has passed through the fourth drawing operation and has 
been annealed at 630° C. X 50 diameters. 

In these three photo-micrographs of annealed brass, the 
crystalline structure of the metal is clearly seen, especially in A, 
which shows some fine and well-developed octahedral crystals. This 
might be expected, seeing that the thermal treatment by annealing 
favours the formation of the crystals, the state of which is greater 
the higher the annealing temperature and the slower the rate of 
cooling. In B and C the crystalline structure can be seen to have 
rearranged itself, in consequence of having undergone mechanical 
operations. 

D and E represent metal embrittled by the operation of 
cold-drawing. The structure differs widely from that of the 

3 I 



802 CARTRIDGE-CASE MANUFACTURE. Oct. 1905. 

annealed metal, the crystals losing their form, and finally disappearing 
in a confused and apparently homogeneous mass. This deformation 
of the crystals is more pronounced the more the cold working of the 
metal is prolonged, until the metal attains a perfect, or almost perfect 
state of brittleness. 

The Manufacture of Cartridge-Cases. — The entire manufacture of 
metallic cartridge-cases involves a series of operations which, with 
the exception of two or three, consist in cold-drawing. The brass 
used is capable of extreme deformation when cold, but cannot be 
worked hot. After being formed into a cup-shaped disc, the metal 
is subjected to successive drawings, the object of which is to 
diminish the diameter and thickness and increase the length, the 
volume undergoing no sensible alteration. At each drawing the 
metal is deformed to a point short of the breaking point, every 
drawing operation being followed by annealing until the desired 
form is obtained, namely, a long cylinder with thin walls, and 
closed at one end. 

The earlier operations, while the cartridge-case is still short, 
are carried out in a vertical press, but when the length is such that 
the manipulation and the withdrawal of the punch become difficult, 
the operation is continued in horizontal presses. The two most 
important tools are the punch and the die. The punch is carried 
upon the extremity of the ram of the press and transmits the power, 
acting upon the bottom of the cartridge-case, which is inserted in 
the larger end of the die, the latter being strongly secured to the 
head of the press opposite the hydraulic cylinder. The die consists 
of a ring of hardened and tempered steel, the interior having the 
shape of a truncated cone, the axis of which is in a straight line 
with that of the punch. The operation of drawing is performed by 
placing a partly-drawn case properly centred in the larger end of 
the die, and advancing the punch until it touches the bottom of the 
cup. The pressure then comes into play, forcing the cup through 
the small end of the die, thereby reducing the diameter of the cup 
and the thickness of the walls and increasing the length, a process 
which involves considerable flow of metal. 




Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 803 

During the process of drawing, the cartridge-case is subjected to 
stresses in general oblique to the surface, represented by P (see Fig. 4, 
page 796). This stress may be resolved into two, one of which is 
normal and the other tangential to the surface of contact between 
the brass and the die, called respectively N and T. If E represents 
the total pressure exerted by the punch upon the bottom of the 
cartridge-case, it is clear that equilibrium will exist when the vertical 
components of the normal and tangential forces are together equal and 
opposite to the force E. If a is the angle formed by the wall of the 
die with the axis of the punch, the equation of equilibrium will be : — 
E = T cos a — N sin a = 0. Thus it will be seen that the forces 
N and T vary with the magnitude of the angle a. As this increases 
sin a increases and cos a diminishes, and consequently the values of 
the components T and N also decrease and increase respectively. 
When a = 90° sin a = 1, and cos a = 0, and E is then equal to N. 
When a = then sin a = and cos a = 1, E being equal to T. 
In the first case this stress would be entirely normal, and in the 
second case entirely tangential and tensile. The two extreme cases, 
however, never occur. In practice the flow of the metal is never 
achieved by simple tensile forces, and compressive forces are always 
present. 

The brass for cartridge-cases, from 3 inches to 6 inches diameter, 
Fig. 5 (page 804), used at Trubia, is purchased in the form of discs, and 
contracts for it contain the provision that if more than 10 per cent, 
show cracks or other defects when first dished or cupped, the whole 
parcel is rejected. If the faulty discs do not amount to 10 per cent, 
the contractor is required to replace the defective ones. If they do 
not exceed 3 per cent., all are accepted. Tensile tests are also made 
by taking 10 strips, 160 mm. by 28 mm. (6*3 inches by 1-1 inch), 
from each 4 tons of brass, from which test-pieces are prepared. The 
tensile strength of the annealed metal must be at least 42,700 lbs. 
per square inch., with an elongation of 57 per cent. 

The manufacture of a 6-inch cartridge-case will now be described 
in detail. 



3 I 2 



804 



CARTRIDGE-CASE MANUFACTURE. 



Oct. 1905. 



3 



s 

I 

CO 












lO ^ 



^ »« 
fe > 



« 
> 









■ \^'.sv\'.\ss' 



.•,\svs\ .^ 



k^ \\ ^ ■■■.^W'.' 



.^^^^^^ v ■. v v \ v v ■■■.<.vvv 



£ 



4-8 ->• ^ 

^^^^^^^■.^^^^^k^^^■.».'■'■^^^^^y \ 



Lv ^^^^l.».^^'.^^^ ■■'.VV ^v; 



l>Tm>^^^^k^^^^^^<.■■^^^^^^■.v■^ ^ 



v^vvvv'-'.SSV\^'v'>'>\SS\S\\'; 



UlltkkVVV^V^ "-"^^ 



I t ^ K ^ ^ i.\ ^ ■■ ^ >, ■. >. I. ^^ ' 



L \ \ \ V s ■■ V ^ ^ ^ ^ V ^ vv 



U^ ..k ^^ ^^^ k v >. k i. k >.-T<r< 



^1 



LSV',\ ^ ■. ■. V ■.\'-\'-^ 



■ii > > V < 1 ■■ ■■^ k I. k ^-TV / 



1 \ \ ^ \ \ ^ ^ V v V ^ V 



.»^v».kki.k kkkC 



'•5 

i 



^\^^^■■ ^svvsk 



^ 



■.>>m^».i.\«.-.v'.6 ' 






3SSXSSSSS^ 



^ 



rr 



•5 



&s 



33sss3s:^ 






t. 



esssjn 



I 
I 



ssss^ 



sssss' 




hH 



&02" ^ 




"150>L^I 
" 5-'9 " 



II 

CD 

o 



158-8^ / 'r^ 



^\\\\Vv\^^^;\i\\\\v 



\± 



164-^ 
645' 



1 





(j5 








PM 




V 


«) 


S 


«0 




« 


<D 


o 


S 


«> 


-tJ 


e^ 


'^ 




05 

pi 


i 






1H 


CO 


o 




o 


■v 


bo 


<» 


rA 



0) 



o 



Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 805 

Cupping. — The brass discs for the 6-inch cartridge-cases, 
Fig. 6, Plate 39, are 361 mm. (14*2 inch) in diameter and 17 mm. 
(0*67 inch) thick, with an allowable variation of thickness of 
• 5 mm. (0 • 02 inch) above or below. They weigh 33 lbs. each, and 
should have a perfectly smooth surface with clean cut edges. 

The cutting of blanks from sheet brass is intended to be^ 
conducted in the 1,000-ton press with the cutting tools shown in 
Fig. 6, Plate 39, but hitherto the Trubia authorities have purchased 
blanks ready cut. 

The cupping is divided for convenience into two stages, the first 
being done with the punch and die illustrated in Fig. 6. Before 
commencing it is necessary to centre the die relatively to the punch, 
the breadth of the annulus being measured at three points. The 
stroke of the punch is then adjusted, so that at the end of each 
stroke it does not exceed what is necessary to thrust the disc clear 
through the small end of the die, and so avoid waste of time and 
power. At the commencement of the stroke an extra length of 
stroke of from 4 inches to 8 inches is given in addition to the amount 
actually necessary to clear the die, in order to give the operator time 
to place the discs upon the die. The die, punch, and disc are then 
well greased, and the latter is placed upon the upper surface of the 
die. Water is admitted to the cylinder and the punch advances, 
driving the disc through the die and out at the smaller end, whence 
it falls in the form of a cup into a receptacle placed below the press. 
The maximum pressure attained is 1,000 lbs. per square inch, as 
shown by the gauge attached to the hydraulic cylinder. The cup is 
then annealed for about 28 minutes at 740° C. (1,364° F.), having 
a steel clip placed round it. The scale which forms on the surface 
of the cup is subsequently removed by pickling in lead-lined wooden 
troughs containing dilute sulphuric acid, of a strength of 1 to 4, 
for a period varying from 8 to 15 minutes according to the strength 
of the bath. The cups are then washed by immersion in 
lead-lined wooden troughs, through which runs a stream of water, 
every trace of acid being quickly removed. The second cupping 
operation is made in exactly the same manner as the first, 
except that the punch and die shown in Fig. 6, Plate 39, 



806 CARTRIPGE-CASE MANUFACTURE. OOT. 1905. 

are substituted for those previously used, the same precautions 
being observed for centering and lubricating. The maximum 
hydraulic pressure indicated by the gauge is 1,150 lbs. per 
square inch, while the subsequent annealing lasts 20 minutes 
at a temperature of 650° C. (1,202° F.). The pickling and 
washing processes which follow this and all other annealings 
are as before described. The behaviour of the metal during cupping 
is an efficient test of its quality. The presence of impurities or 
improper annealing are quickly shown by cracks or a roughened 
surface. 

First Drawing. — Fig. 6, Plate 39, shows the punch and die 
used in this operation, also the resulting piece. The maximum 
hydraulic pressure is 1,300 lbs. per square inch. The pieces are 
then annealed at 650° C. (1,202° F.) for 28 minutes. 

Second Drawing. — This is performed with the tools shown in 
Plate 39 with a maximum hydraulic pressure of 1,350 lbs. per square 
inch. The subsequent annealing is at 650° C. (1,202° F.) for 
26 minutes. 

Third Drawing. — This is performed with the tools shown 
with a maximum hydraulic pressure of 1,320 lbs. per square 
inch. The subsequent annealing is at 640° C. (1,184° F.) for 
25 minutes. 

Fourth Drawing. — Before drawing, the bottom of the piece is 
flattened preparatory to indenting, which takes place after the fifth 
drawing, and is necessary for the formation of the primer hole. 
Flattening is accomplished by pressing the piece between the punch 
and a flat steel disc supported on the die. The disc is then 
withdrawn and the drawing proceeds as usual. The tools are shown. 
The maximum hydraulic pressure is 1,000 lbs. per square inch, 
and the subsequent annealing is at 630° C. (1,166° F.) for 
22 minutes. 



Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 807 

FiftJi Drawing. — This is the last operation performed in the 
vertical press. The tools used are shown in Fig. 7, Plate 40, the 
maximum hydraulic pressure being 700 lbs. per square inch. The 
subsequent annealing is at 630° C. (1,166° F.) for 20 minutes. 

Indenting for Primer. — This operation is performed in the vertical 
1,000-ton press. Upon the ram which moves upwards is placed a 
pressure plate, Plate 40, to which is hinged a steel punch-shaped 
piece having the same external form as the interior of the cartridge- 
case as it leaves the fifth drawing, and with an indentation at the 
top. This can be hinged to one side to facilitate the insertion and 
withdrawal of a cartridge-case. Upon the under side of the upper 
head of the press is a fixed holder, into which is screwed a flat piece 
of tempered steel having a small projection in the centre. The 
object of this is to form, in conjunction with the recess in the punch, 
the metal boss on the inside of the case, for the primer. The 
cartridge-case is subjected to a pressure of about 314 tons between 
the two surfaces, with a hydraulic pressure of 2,500 lbs. per square 
inch. No annealing is required after indenting. 

Sixth Drawing. — From this operation onwards the two larger of 
the three horizontal presses are used, because the length which the 
cartridge-cases have now reached does not permit of their 
manipulation in the shorter-stroke vertical press. The tools used 
for the sixth drawing are shown. Up to this point the cartridge- 
cases have been able to strip themselves from the punches by 
catching on the underside of the dies. Their expansion at the 
moment when drawing is complete, and when they are relieved from 
the considerable lateral pressure, prevents their being again drawn 
up through the die by the retreating punch. But from the sixth 
drawing onwards, the lateral pressure is less, and other means are 
adopted. Under each die is an attachment containing eight 
fingers pressed inwards towards the axis by springs. During 
drawing, they give way before the advancing case, retiring into 
recesses. But when the end of the case has passed them they spring 
out and keep the case from following the punch back, the inclination 



808 CARTRIDGE-CASE MANUFACTURE. Oct. 1905. 

of the recesses in whicli they move assisting this action. The sixth 
drawing may be performed either on the 18-inch or the 16-inch 
horizontal press, either having sufficient power. The subsequent 
annealing is at 630° C. (1,166° F.) for 18 minutes. 

Seventh Drawing. — The tools used for this operation are shown. 
The subsequent annealing is at 630° C. (1,166° F.) for 15 minutes. 

Eighth Drawing. — The tools used for this operation are shown. 
The subsequent annealing is at 600° C. (1,112° F.) for 14 minutes. 

Ninth Drawing. — The tools used for this operation are shown 
in Fig. 8, Plate 41. The subsequent annealing is at 570° C. 
(1,058° F.) for 14 minutes. 

Tenth Drawing. — The tools used are shown. This is the last 
drawing operation, and the blanks undergo no annealing upon its 
completion. 

Heading. — The formation of the head of the cartridge-case is one 
of the most interesting of the operations in the process of manufacture. 
The total pressure which the head of the cartridge is called upon to 
stand under fire is enormous. With the 6-inch quick-firing gun 
used in the Spanish service, for which these cartridge-cases are 
intended, the pressure caused by the explosion is about 17 tons per 
square inch. Even this pressure is exceeded when testing the guns, 
which is done with three discharges at a pressure of 20 tons per 
square inch. When the area of the cartridge-case head is considered, 
some idea may be formed of the enormous aggregate pressure to which 
it is subjected. It is essential for the satisfactory working of the 
guns that no deformation should take place under fire, and it is 
therefore important that during manufacture the head should be 
subjected to a pressure two or three times that likely to be experienced 
in practice. The operation of forming the head is made in the 
vertical 2,500-ton press in three stages. The tools used for the 
first stage are shown in Fig. 8, Plate 41. An iron casting, A — A, 
termed a bolster, is placed upon the ram of the press and serves to 



Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 809 

support the die-holder and die B — B, which latter imparts the form 
of the flange to the head. Inside the bolster is fixed a steel stem C, 
over which the cartridge-case is slipped in the condition in which it 
leaves the tenth drawing. This stem, which must be capable of 
withstanding an aggregate pressure of 1,650 tons, is of the best 
hardened steel, and is without question the most delicate of all the 
tools employed in the process of manufacture. The first heading 
operation is performed by inserting the cartridge-case between the 
stem and the bolster. Upon the top is also placed the punch D of 
hard steel, provided with a central depression, the object of which is 
to reduce the area of contact over which pressure is exerted on the 
head of the cartridge-case. The total pressure is 1,600 tons, which 
leaves the head with a central internal and external projection, and 
forces the metal outwards to form a flange. 

On Plate 42 is shown the second heading operation. This is 
performed with the same tools as the first, except that a smaller 
punch, 3 inches diameter, is placed over the cartridge-case, instead 
of the punch D previously used. A total pressure of 600 tons is 
exerted, with the result that the outside projection is flattened, 
and all the metal is driven into the internal boss, thus allowing 
sufficient metal for the primer holes. Finally, the third heading 
operation is performed with the tools shown, a total pressure of 
1,650 tons being applied, with the result that the head is rendered 
flat and shapely. 

Tapering. — This operation is for the purpose of giving to the 
cartridge-case its final external form, enabling it to fit the chamber 
of the gun, and to be easily inserted and withdrawn. It is performed 
in one or the other of the horizontal presses, in order to take 
advantage of their longer stroke. To the fixed head HH of the 
press, Plate 42, is bolted the cast-iron bolster A — A, inside 
which are placed seven rings of tempered steel BBB, the 
internal length of which when thus assembled is exactly equal 
to that of the gun chamber. The cartridge-case is driven into this 
space by the press, but as it is necessary forcibly to extract it after 
the operation, the special apparatus shown is made use of. The 



810 CARTRIDGE-CASE MANUFACTURE. Oct. 1905. 

cylindrical extractor C, having a head shaped to fit the inside of the 
headed cartridge-case, is connected rigidly with the ram of the press 
through the crossheads D and F and the tie-rods E E, and moves 
therewith, its position being kept central by the guide I. The 
punch G, bolted to the ram of the press, forces the cartridge-case in 
during the forward stroke, while the extractor C forces it out during 
the return stroke. At Trubia the tapering is divided into two 
operations with annealing between, to avoid risk of cracking. Before 
the first tapering the cartridge-case is annealed at 560° C. (1,040° F.) 
in a small vertical furnace, Fig. 28, Plate 47, care being taken 
to allow the head to remain outside the furnace in the air. It 
is then placed in the press and forced about half its length into the 
chamber, the precaution being taken to adjust the stop of the press 
so as to limit the stroke to half its usual length. On the return 
stroke, by the aid of a wooden distance-piece inserted between the 
extractor and the head of the cartridge-case, the latter is forced out. 
The case is then returned to the vertical annealing furnace, where it 
is exposed to a temperature of 600° C. (932° F.), care being taken 
as before not to anneal the head. Tapering is then completed in 
the press, the cartridge-case being driven completely home into the 
die chamber. 

Other Mechanical Operations. — The remainder of the operations 
are of a mechanical nature, such as turning the end, the head, the 
steps in the chamber, the attachment for the primer, cutting to the 
exact length, etc., none of which involve any features of special 
technical interest. It may, however, be mentioned that throughout 
the whole course of manufacture the thickness and diameter of the 
cartridge-cases are carefully checked with callipers and gauges, and 
particularly for the first two or three cases in each lot, in order to 
verify the accuracy of the dies and the setting of the tools. The 
ends of the cases are frequently turned to length between the various 
drawing operations, since there is a tendency, due either to the 
irregularities of metal or uneven annealing, to stretch unequally, 
leaving ragged edges. It is also of great importance that the 
thickness of the end of the cartridge-case should be closely checked, 



Oct. 1906. 



CARTRIDGE-CASE MANUFACTURE. 



811 



Fig. 10. 
Floor Plan of Cartridge-Case Factory. 



//y'//////^/////////yyy///////y^ 




1000 -7^/? 



Hyd. Heading Presses 
and Piimps 




2500-7b/^ 
AZx8-6ffor. 



\6K\00ffor. 



\8xl0-0 /for. 



22x4-71: l/pri^^ 



Compound Superposed 

Condensing Engine 



Boiler 



Furnace for _ 
Annealing Ends 




Double EnxLed 
Annealing Furnare 




'5> 



End Trimming 



^ 



PickUnj^ Trough 



V /// //////■////////// y /////////////// / i 




812 CARTRIDGE-CASE MANUFACTURE. Oct. 1905. 

and this is performed by limit gauges. Lubrication of the punches 
and dies is effected by olive oil or soapy water, according to the 
stage in the process. 

Plant. — Having now outlined the various processes involved in 
the manufacture of a 6-inch cartridge-case, the machinery by which 
such processes are effected will be described. The plant which, with 
the exception of the engine and boiler, was made in the United States, 
is intended to make cases from 3 inches to 6 inches inclusive, and 
has already turned out a considerable number. As no machinery 
has yet been acquired for casting or rolling brass sheets, discs are 
purchased abroad, but it is probable that in due course they also will 
be made at Trubia, so as to make the whole manufacture self- 
contained. In Fig. 10 (page 811) is shown a ground plan of the 
factory, which contains boiler, engine, four hydraulic pumps, 
accumulator, one vertical and three horizontal drawing-presses, 
one 1,000-ton and one 2,500-ton vertical presses, Plates 43-45, 
for indenting, heading, and cutting, annealing furnaces, pickling 
troughs, and trimming machinery. 

The boiler is of the Lancashire type, with a working pressure of 
125 lbs. per square inch, 30 feet long, and 8 feet diameter. The 
engine is of the Galloway superposed, compound condensing type of 
350 I.H.P., having cylinders 16 inches and 30 inches diameter and 
46 inches stroke, the high -pressure cylinder being on the top of the 
low-pressure cylinder, and inclined thereto at an angle, working on 
a common crank-pin. The air-pump is worked from a prolongation 
of the low-pressure piston-rod. The engine, which works at 70 
revolutions per minute, has a fly-wheel 15 feet diameter and 32 inches 
broad, and drives a shaft 8J inches diameter, which in turn operates 
four duplex-hydraulic pumps. Jaw-clutches are provided so that 
any one pump may be instantly disconnected from the shaft. Each 
duplex pump comprises four rams 4| inches diameter by 12 inches 
stroke, driven in pairs from two cranks set at an angle of 90^, and 
keyed to a spur-wheel, which is in turn driven by a pinion on the 
main shaft with a reduction of 1*7 : 1. Of each pair of rams one 
performs the working-stroke while the other draws in water, so that 



Oct. 1905. 



CARTRIDGE-CASE MANUFACTURE. 



813 



the relative position of the four rams produces a very even turning- 
moment. The pumps draw from a tank, into which the waste-pipes 
from all the presses are led, and deliver into a common pressure-pipe 
leading to the accumulator. The delivery is automatically cut 
off successively from one pnmp after the other as the accumulator 
rises by devices shown in Figs. 16 and 17 (pages 813-814). Fig. 16 



Fig. 16. 
Bye-Pass Valves for the 4f ins. x 12 ins. Duplex Pumps. 









(ns. 






ft. 


IZ 

I 1 1 


6 




1 1 


1 
1 



illustrates the bye-pass valve and connections, one being provided 
for each pair of rams worked from the same crank. The flanges 
a and b are connected one to the delivery of each ram. These pipes 
unite into a common delivery-pipe c so long as the valve d is closed. 
When d is opened the water escapes by the pipe e to the tank, and 
the pumping absorbs little or no energy. The opening and closing 



814 



OARTRIDGE-CASE MANUFACTURE. 



Oct. 1905. 



Fig. 17. 

Cam and Slide- Valve operated by Accumulator 
and controlling Bye-Pa<^ Valve, Fig. 16. 



Z' -K^^" -~~\rTo Rye-l'ass Valve 

i-4'0>,; O 

From Accumulator \ _I -" /<a^>v_^ 



^ To Tank 




TO 'f O 



Valve Base 



^ 




-n 




■9'^ 






s/ns. 



\ I \ 1- 



Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 815 

of the valve d is effected by a hydraulic cylinder in which works a 
piston/. High-pressure water is, under normal working, admitted 
to both sides of this cylinder by the pipes g and A, resulting in a 
greater pressure downwards than upwards, the difference being due 
to the sectional area of the piston-rod k. The valve is therefore 
kept normally shut. When the accumulator reaches the top of its 
stroke it operates successively one of eight small slide-valves, each 
of which is connected to a similar hydraulic cylinder. The movement 
of the slide-valve allows the water at the top end of the cylinder to 
escape to the tank, thus enabling the pressure underneath to open 
the valve d. A hand-lever I is also provided for effecting the 
movement by hand if necessary. To the flange m is attached a 
reliefrvalve to enable the water to escape if the normal working- 
pressure of 1,000 lbs. per square inch is exceeded largely. 

Fig. 17 (page 814) illustrates the mechanism of the small slide- 
valves. The shaft a is rotated slowly by the movement of the 
accumulator through a chain and sprocket wheel, not shown. On 
this shaft are keyed eight cam-wheels 6, each provided with a cam- 
recess c, in which works the cam-wheel d. The motion of the cam- 
wheel operates the slide-valve e through the lever / and valve-rod g. 
The valve as shown is in the normal position when the pump is 
delivering to the accumulator, and the bye-pass valve d. Fig. 16 
(page 813), of the pump is closed. High-pressure water is at all times 
admitted to the valve chamber by the pipe A, and in the position of the 
valve as shown is in communication with the upper end of the bye-pass 
valve by the port /. When, by the rise of the accumulator, the cam- 
wheel reaches the portion k I m of the groove, the valve moves to the 
left, closing the port j to pressure, and opening it to the port n, leading 
to the waste- water tank, thus allowing the bye-pass valve to rise, 
forced up by the constant pressure below the piston/. Fig; 16. The 
eight cam-wheels b are keyed to the shaft at different angles, so that 
the shutting off and opening of the different pumps as the accumulator 
rises and falls respectively is as gradual as may be desired. 

The accumulator is of the ordinary type, having a ram 14 inches 
diameter and 20 feet stroke. It is loaded with 54 tons of iron 
castings, in addition to the weight, about 14 J tons of the moving 



816 



CARTRIDGE-CASE MANUFACTURE. 



Oct. 1906. 



Fig. 18. 
22" X 47|" Vertical Hydraulic Drawing-Press 



haust 




Exhaust 



JOfteC 



Oct. 1905. 



CARTRIDGE-CASE MANUFACTURE. 



817 



parts. It is provided with a safety check-valve to prevent a too 
sudden descent in case of the failure of any pressure pipe, by gradually 
closing the exit pipe during the last 12 inches of the descent. 

The vertical drawing-press in which the cupping and first to fifth 
drawings are effected, is shown in Fig. 12, Plate 44, and Fig. 18 
(page 816). The hydraulic cylinder is 22 inches diameter by 
47 J inches stroke. Instead of a ram, a piston is used with a 
piston-rod 14 inches diameter. The annular space beneath the piston 

Fig. 19. 
Reversing- Valve for the 22" x 47|" Vertical Hydraulic Draiving- Press. 




Vciive 



o 

1.1.1 L. 



6 

-I L. 



IZlns. 

-i ^ I 



Development of 
Valve Ports 



is sufficient to provide upward force for the return stroke. The lower 
end of the piston-rod is guided by a cross-head a working between two 
vertical guides or columns h. The reversal is automatic, and the stroke 
can be set to any desired length. The reversing gear cc is described 
more in detail in the case of the 18-inch press. Hydraulic pressure 
is admitted to one end or the other of the hydraulic cylinder d, of 
which the piston-rod e is extended through both cylinder-covers and 
operates two reversing-valves. One of these is shown in detail in 

3 K 



818 CARTRIDGE-CASE MANUFACTURE. Oct. 1905. 

Fig. 19 (page 817). The opening a is attached to one end of the 
cylinder, h admits high-pressure water, and c leads to the exhaust. 
The rod d is operated by the auxiliary cylinder already described. 
The valves e are so formed as to stoj) gradually the flow of water 
either inwards or outwards. The base of the press, Fig. 18 (page 816), 
consists of a massive iron casting g on the top of which is placed the 
die, not shown, while the screwed recess li in the end of the piston- 
rod, holds a socket into which is screwed one of the punches, shown 
on Fig. 6, Plate 39. 

The horizontal drawing-presses, Fig. 10 (page 811), and Figs. 13 

and 14, Plates 44 and 45, which are used for the smaller cases, or for 

those which have been sufficiently elongated to need a longer stroke, 

are three in number and are 18 inches diameter by 10 feet stroke, 

16 inches diameter by 10 feet stroke, and 12 inches diameter by 8 feet 

stroke respectively. All have a working pressure of 1,000 lbs. per 

square inch. The first-mentioned, namely, 18 inches diameter by 

10 feet stroke, is selected for description, and is more fully illustrated 

in Figs. 20 and 21 (pages 819 and 820). It consists of a horizontal 

cylinder v, having a piston, piston-rod s, and cross-head t, guided by 

two guide-bars u, extending from the cylinder v to the die-head a. 

There are two tie-rods h uniting the cylinder v with the die-head, 

placed diagonally, one passing through one of the guide-bars u. The 

valve arrangement is similar to that described in the case of the 

22-inch vertical press, while the valve reversing-gear is shown in detail 

in Fig. 22 (page 821). The pin a on the cross-head strikes at one end or 

the other of its stroke the finger h or c, which, through bevel gear d and 

e respectively, rotate the weigh-shaft / through a small angle in one 

direction or the other. This motion is transmitted through bevel 

gearing </ to a shaft h at right angles to /, which in turn operates the 

pilot valve, admitting water to one end or the other of the auxiliary 

hydraulic cylinder, operating the inlet and exhaust valves of the main 

cylinder. A hand-lever h permits the operator to reverse when necessary. 

The die-head a. Fig. 23 (page 823), which forms part of the 

main structure is connected by the tie-rods h to the cylinder of the 

press. Within the die-head is inserted either a small chuck 

containing a die, shown in Fig. 6, Plate 39, or a larger chuck c, 



Oct. 1906. 



CARTRIDGE-CASE MANUFACTURE. 



819 




3 K 2 



820 



CARTRIDGE-CASE MANUFACTURE. 



Oct. 1905. 



Fig. 23. The latter serves the same purpose as the smaller chuck, 
but contains additional space for the insertion of a stripper plate d 
and a die-holder e, die /, and screwed ring or nut g, all of which 
are held in place by suitable clips h and set screws Jc. The stripper 
plate d, which has already been described, is only necessary when, 
as in the case of the 6th to 10th draws inclusive, there is a tendency 
for the cartridge-case to follow the punch on the return stroke, and 
is only used on the horizontal presses. 

The three horizontal presses are provided with a feed-table and 
attachment by which the cartridge-case can be held, and made to 
centre exactly with the die and punch. This feed mechanism is, 



Fig. 21. — 18" x 10' 0" Horizontal Hydraulic Drawing-Press. 



Section at BB (Fig. 20) looking 
towards Die Head. 



Feed 
AtZacfiment 



Section at AA (Fig. 20) looking 
towards Cylinder. 





10 Feet 



however, not absolutely necessary, the case being frequently pushed on 
to the punch by hand and driven home with a wooden mallet. The 
feed mechanism is shown more in detail in Fig. 24 (page 824), 
in which a is the lower of the two guide-bars w, shown on 
Fig. 20 (page 819), h is the hole for one of the two tie-rods, 
and c the feed-table securely bolted to a. Upon c is mounted 
a sliding bracket d carrying a grip mechanism. Four bent 
levers e are attached in pairs to hard-wood grip pieces / 
and hinged upon rods </, which are in turn held by the nuts A. 
The rods g are adjustable, and can be moved together or apart by 
a right- and left-hand screw It. The other ends of the four bent 



Oct. 1905: 



CARTRIDGE-CASE MANUFACTURE. 



821 



levers e are also adjustable along four vertical rods /, to wliich they 
can be locked by nuts m. Each pair of rods I are pivoted at one 
end eccentrically to a horizontal shaft n, which can be rotated by a 
hand-lever p, serving not only to put a final grip on or to release 
the cartridge-case r, but also to push forwards or backwards the 
whole mechanism along the table c into or out of position. By 
means of the movable fulcrums and adjustable lever ends, the 
mechanism can be altered to receive any size of cartridge-case. 



Fig. 22. 
Reversing- Gear for Valves, 
1 8-inch Dr arcing- Press. 




^ . \^Fin on Crosshead 



Plate 46 shows a side elevation of the 2,500-ton heading press, 
also a cross-section at AA, sections through CC of the pump, 
showing both the high- and low-pressure cylinders and valves, and 
an enlarged view of the main cylinder packing. A photograph of 
the heading press and pump is shown in Fig. 15, Plate 45. The 
1,000-ton heading press and pump is of similar but smaller design, 
and will therefore not bo described. The heading punch «, Plate 46, 
(of which there are three forms, as shown on Plates 41 and 42) is 



822 CARTRIDGE-CASE MANUFACTURE. Oct. 1905. 

securely held against a fixed pressure platen h, whicli is bolted to the 
press head c attached to the cylinder by four vertical tie-rods d. The 
punch a has no vertical movement. The tie-rods d at their lower 
extremities pass through a base plate e, on which is mounted a 
hydraulic cylinder/, having an internal diameter of 32 inches. On the 
top of the piston or ram g, which has a stroke of about 9 inches, is 
bolted a pressure platen h. Both cylinder and piston are fitted with 
leather packing-rings of L shape as shown in the enlarged view. The 
interior stem i is held vertically within the bolster /, in the upper 
end of which is the die j^. The bolster is bolted at its base to a 
carriage running on four wheels on a track, and capable of being 
quickly run into and out of position by a hydraulic feed cylinder Jc, 
5-inches diameter. Minute regulation of the travel of the piston is 
obtained by a stop screw Z, thus assuring exact centering between 
the cartridge-case and the punch a. 

The heading process is as follows : — The cartridge-case is first 
introduced between the bolster j and the stem i, and nearly driven 
home with a wooden mallet. The carriage is then run into position 
over the lower pressure platen h, and held there by the insertion of 
removable dowel pins. Both high- and low-pressure water is admitted 
to the cylinder by the pipe w, causing the ram ^, guided by the 
four vertical tie-rods d, to rise, carrying with it the whole carriage 
and all situated thereon. When the cartridge-case touches the top 
punch a, the low-pressure water is automatically shut off, and the 
high-pressure water completes the heading process. On the return 
stroke the carriage drops again on to its track, and is then pushed 
back by the feed cylinder until the base of the vertical stem i 
centres with the top of the ejecting punch n, for which a removable 
stop screw o is provided. Hydraulic pressure being admitted to the 
ejecting cylinder, the piston rises about 6 inches, carrying with it 
the stem i and the cartridge-case. On the return stroke the stem 
i follows by gravity, leaving the cartridge-case projecting, whence it 
is removed by a small hand-crane, not shown. When it is required to 
remove the stem i, the whole carriage is pushed along until the stem 
is over plate p, whereupon the latch q is released by the withdrawal 
of the hinged eye r, allowing the plate p to fall to the ground, 



Oct. 1905. 



CARTRIDGE-CASE MANUFACTURE. 



823 



Fig. 23. 
JDraioing and Stripping Device for the Three Horizontal Presses. 



Dte-Hcaxi a 



Stripper d 




824 



CARTRIDGE-CASE MANUFACTURE. 



Oct. 1905. 



Fig. 24. 
Feed Mechanism for the IS-inch x 10-/^. Press. 



d 



t^ r^S— f-Tl -K 



K 



a. 




/ 



h 



f 



tir-^- w I Li u I 

m \ 




Oct. 1905. CARTKIDGE-CASE MANUFACTURE. 825 

followed by the stem. The process of heading the 6-inch cases 
requires in practice a total pressure of 1,700 tons. 

The geared compound pump, Plate 46, contains three high- and 
three low-pressure plungers, all six being fitted with separate valves. 
The high-pressure plungers are J-inch diameter, and the low- 
pressure 4 inches diameter. The stroke in each case is 6 inches. 
All six plungers are connected with cross-heads which work between 
guides on the main frame-work, and are driven by eccentrics keyed 
on to two shafts s, each of which is provided with a spur wheel t, 
driven by a common pinion w on a central shaft v. This is fitted 
with a friction clutch pulley v^ 36 inches diameter, running at 
200 revolutions per minute, and driven by an electric motor. The 
hydraulic supply is obtained from the tank z by two suction pipes, 
w w, running through the bed plate of the pump. The delivery 
pipes X y, lead to three automatic valves shown on Fig. 26 (page 826). 
There is shown in section the bye-pass valve C, the check valve B, 
the main valve A, and a diagrammatic general arrangement in 
position for starting the working stroke. Delivery water from the 
low-pressure pumps arrives at a, passes round the valve &, which 
is held down on its seat by spring y, and leaves the bye-pabS valve 
by the pipe c on its way to the check valve. Here it pushes up 
the valve d, which is otherwise held on its seat by the spring d}, 
aided by high-pressure water which enters the valve by passages d"^. 
The raising of the valve d permits the low-pressure water to join 
the high-pressure which enters by the pipe e direct from the delivery 
of the high-pressure pump. The combined deliveries enter the main 
valve at /, leave by the pipe </, and are led thence to the cylinder of 
the heading press. The ram rises, causing the cartridge-case to 
come into contact with the punch of the press. The pressure then 
increases, and when it reaches 300 lbs. per square inch, the ram h, 
which is in constant communication with the high-pressure 
delivery, rises, compressing the spiral spring / and opening the 
valve 6, which is held in this position by the lever h interlocking the 
lever I. The opening of the valve h cuts off the low-pressure supply 
to the check valve, by allowing the water to flow back to the tank. 
The check valve d being now relieved from pressure below, is closed 



826 



CARTRIDGE-CASE MANUFACTURE. 



Oct. 1905. 




Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 827 

by the water-pressuro above, thus severing the connection between 
the liigh- and the low-pressure supplies. The high-pressure supply is 
now in direct communication with the cylinder of the heading press 
through the check valve and main valve by the pipe g^ and by its 
increasing pressure completes the heading process. This is stopped 
automatically by a safety device when the pressure reaches a given 
amount. The plunger m, which is in constant communication 
with the high-pressure delivery, is forced up against a spring n 
loaded to a definite pressure. This disengages the trip lever o from 
the hand lever jj, allowing the spring q to lift the rod r, which in 
turn raises the valve s sufficient to give a maximum opening of 
gJ^-inch. This opens a way for the high -pressure water through the 
passages t and u to escape to the tank. The sectional area of the 
passages being more than enough for the incoming delivery from the 
high-pressure pumps, allows the water to escape gradually from the 
heading press. When the pressure is reduced to almost zero, the 
springs q are able to raise the valve v, thus allowing free egress for 
the pump delivery to the tank. During the working stroke, the valve v 
is tightly held on its seat by high-pressure water, which is admitted 
to the upper surface through grooves lu. These grooves are closed 
at their bottom ends by the valve-seating x. The operator then 
depresses the treadle y, which locks the lever p and releases 
the lever k, thus replacing the various valves in their original 
positions ready for the next working stroke. A handle on the main 
valve gives the operator greater latitude in working the press. 

Fig. 27, Plate 47, illustrates in longitudinal and cross-section, a 
double-ended annealing furnace, for annealing the cases at the 
various stages of manufacture. The cases are placed therein from 
either end, packed in iron trays. The floor or bed of the furnace is 
25 feet 6 inches long and 6 feet 6 inches wide, and is covered with 
cast-iron pieces a to facilitate the insertion and removal of the trays. 
The furnace is designed for either coal or oil firing, special attention 
being paid to securing uniformity of temperature. The course of the 
gases can be readily followed by reference to the drawing. 

Fig. 28, Plate 47, shows longitudinal and cross-sections of an 
end-annealing furnace used only prior to the final tapering 



828 CARTRIDGE-CASE MANUFACTURE. Oct. 1905. 

processes. Oil fuel is used. A row of eight cast-iron annealing pots 
h, 3 feet 6 inches long and about 7J inches internal diameter, are 
suspended from the top of the furnace, and are provided with eye- 
bolts c for lifting purposes. Each pot is provided with suitable 
covers d, having holes of various diameters through which is passed 
the body of the case e to be annealed, and supporting the head by its 
flange. The head must remain hard to withstand the pressure of 
tapering, and is therefore kept in contact with the cool air and not 
annealed. The body and end alone are annealed to give the 
required ductility for the tapering process. The pot chamber or 
furnace proper is 7 feet 8 inches long by 11 J inches wide by 
3 feet 8 inches deep, the gases entering through port holes in the 
walls and leaving at the base as indicated by arrows. 

The makers of the various portions of the plant are as follows : — 
Drawing and heading presses, pumps, accumulator, and trimming 
lathes. The Waterbury Farrell Foundry and Machine Co., Waterbury, 
Connecticut, U.S.A. ; punches, dies, &c.,The Ferrocute Machine Co., 
Bridgeton, New Jersey, U.S.A. ; steam engine, Messrs. Galloways, 
Manchester ; boiler, Messrs. Daniel Adamson and Co., Hyde ; 
annealing furnaces, The Rockwell Engineering Co., New York. 

The thanks of the authors are due to Messrs. Deming, Oberlin 
Smith, Lamb, and Eockwell, for their help and information in the 
compiling of this Paper. 

The Paper is illustrated by Plates 38 to 47 and 15 Figs, in 
letterpress. 



Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 829 



Discussion. 

The Chairman (Mr. Jolin A. F. Aspinall), in according the 
thanks of the Institution to Colonel Cubillo, regretted that ho was 
unable to be present that evening, and also expressed regret at the 
loss the Institution had suffered by the death of the other author, 
Mr. Archibald P. Head. Mr. Head was a hard-working, studious, 
industrious young man, and an ornament to the Institution, and 
there were many members in the room who not only regarded his 
loss from the point of view of a, member, but also from that of a 
personal friend. 

Mr. B. W. Head said that Colonel Cubillo had asked him to 
represent him that evening, as lie was detained in Madrid in 
connection with President Loubet's visit to Spain. Unfortunately, 
he himself was not well acquainted with the subject to be discussed, 
but any questions arising in the course of the discussion would be 
submitted to Colonel Cubillo, who would then furnish a written 
reply. 

Mr. H. F. Donaldson, Member of Council, said the Paper was 
really divided into three parts — the material, the method, and the 
machines. To the majority of engineers he thought the actual 
manufacture of quick-firing ammunition did not convey very much 
information, but the subject was one which furnished food for study 
beyond the ordinary lines of engineering. Taking the material first, 
the Paper contained some photo-micrographs, and it had struck him 
that it might be instructive and useful to illustrate a few more 
examples of cartridge brass, and lie had therefore arranged to show 
some of these slides. The magnification of the first six was 40 
diameters on the Plate, and the remaining views were 2 diameters. 

Mr. Donaldson then exhibited a series of slides, Plates 48 
and 49, six of which (Plate 48) showed annealing at 600°, 700°, 
800°, and 900° C. (1,112°, 1,292°, 1,472°, and 1,652° F.), the sixth 
specimen, Fig. 34, being that of brass containing 1 per cent, of 
tin. Fig. 35, Plate 49, was a piece of cartridge case of normal 



830 CARTRIDGE-CASE MANUFACTURE. Oct. 1905. 

(Mr. H. F. Donaldson.) 

metal, and Fig. 36, a similar metal, but of different manufacture. 
Figs. 37 and 38 were taken from cases which failed, one by- 
spontaneous cracking in the wall and the other in the rim on 
firing. The natural assumption was that the metal was too hard, 
and had been overworked. The slides did not show the place of 
failure, being made for the purpose of examining the structure. 

There were one or two points with regard to the material with 
which he wished to deal. The authors spoke in several cases of the 
hardness and brittleness produced by working, and working cold, 
and they referred to the ductility being restored by annealing. 
According to his own experience and ideas, if a piece of brass or 
copper was overworked cold, however much it was annealed it would 
never be brought back to the same state it was in before the excess 
work was put into it. The authors had referred to the high limit 
as not being far removed from the melting-point, by which they 
probably meant the same thing. With regard to spontaneous 
cracking, it had been a cause of great trouble to a good many 
makers. If the metal was overworked, a condition of internal 
strain was left which did not appear perhaps for months afterwards, 
and then the case " went " of its own accord. 

Dealing with the subject of methods, he pointed out that the 
temperatures and pressures mentioned in the Paper were not quite 
the same, although very nearly, as those in this country. The 
number of operations used in Trubia in drawing were apparently 
two more than were found necessary in this country, but the 
pressures varied considerably. The first pressure mentioned by 
the authors was 1,000 lbs., whereas he found the pressure at 
Woolwich was about 1,550 lbs. per square inch. To give some idea 
of what that meant, it worked out to about 158 tons on the head. 
The second drawing in Trubia was 1,350 lbs. per square inch, 
whereas in this country it was run up to 1,400 lbs. The next 
drawing was about the same, but in the fourth the pressure was 
in this country, 1,120 lbs. against 1,000 lbs. in Trubia. The next 
drawing was similar in both countries, the English being about 
40 lbs. higher. But in the indenting for the primer the authors 
mentioned 2,500 lbs. per square inch, whereas in his own manufacture 



Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 831 

3,500 lbs. was used. He presumed everyone had to take his own 
experience to guide him, but it would appear that the authors had 
two more drawings and one more annealing than was found to be 
necessary in the manufacture here. 

Turning to the method, mention was made of the work always 
being done cold. It was within the knowledge of a good many 
members that large numbers of cartridge-cases were made in the 
initial stages from hot ingots, instead of from cold plates ; they 
were punched out from a square or hexagonal ingot from the end, 
and the amount of drawing was thus very much less. It might be also 
within the knowledge of some of the members that another method 
recently adopted, he believed with a certain amount of success, 
avoided drawing altogether in the proper sense, and the whole work 
was done by rolling. Theoretically that seemed to him a sounder 
process than the straight drawing, because if there was any 
" laying " of the metal at all, the metal would be laid helically, 
and anyone would imagine that was the best direction for restraining 
the bursting strain which must arise as one of the strains on the 
shock of firing. The length of the 6-inch case referred to was much 
longer than anything at Woolwich, and therefore perhaps the 
stripping arrangement referred to (page 809) might be necessary at 
Trubia, but was not found necessary with the process in this country. 

With regard to the tools or machines, he thought the interesting 
features were always those which gave the most trouble, and the 
question of valves on the presses probably gave more trouble than 
anything else. The cutting of the valves in both of the presses 
shown, he thought, would be very considerable. An automatic 
valve was shown (page 813), and a controlling valve in Fig. 17 
(page 814), and on page 817 again, and he thought the one on 
page 817 with the high pressures spoken of must have a very short 
life indeed. It seemed to him there was everything in favour of 
rapid cutting away when the pressure was wire-drawn. The same 
thing. to a certain extent would apply to the design shown in Fig. 17 
(page 814), but the design, and also that of Fig. 16, seemed to 
give a very pretty general arrangement, and an interesting one to 
those who had looked into the valve question for heavy presses. 



832 OARTRIDGE-OASE MANUFAOTURB. OcT. 1905. 

(Mr. H. F. Donaldson.) 

The pressures which came upon the work were not stated in the 
Paper. Of course it was a mere matter of calculation according to 
the area size of the surface taking the pressure ; but it might be 
interesting to know that with the system at Woolwich the cupping 
was performed at 158 tons, the first drawing 140 tons, the second 
drawing about 135 tons, the third drawing about 112 tons, the 
fourth drawing 80 tons, the fifth drawing 75 tons, and the sixth 
drawing 50 tons all on the work. Then came the indenting, which 
jumped up to 1,350 tons ; the seventh drawing came to 45 tons, 
the second indenting 1,350 tons, and then the eighth and last 
drawing, 40 tons. He was not quite clear about the meaning of the 
authors when they said that the process of heading in the 6-inch 
cases required in practice a total pressure of 1,650 tons. Whether 
that pressure was intended to be applied once or more than once he 
did not know. In his own case three pressures were used, 1,350, 
1,150, and 1,350 tons again. The annealing was uniform at about 
650° C. (1,202° F.) all through. In conclusion, he expressed his 
own feelings of regret that one of the authors had passed away, and 
the other was prevented from being present. 

Mr. Henry Lea, Member of Council, said that the subject of 
cartridge-making was not one with which he was very familiar, 
except in a general way, but the dealing with metal by what was 
known as the raising and drawing processes he had been acquainted 
with in work other than cartridge-making for a very long time, and 
he always regarded the process as being a very interesting one. He 
thought his first acquaintance with it must date back some 40 years, 
and what he was going to say related to the raising of vessels from 
very much thinner metal than was used in making cartridges — 
vessels of almost any size, not restricted by any means to 6 inches 
diameter. He had forgotten the name of the man who originally 
devised and introduced into Birmingham what was called the 
pressure-plate, but that plate was the basis of success in performing 
the first operation of converting a flat disc into a cup-shaped vessel. 
Fig. 39. He took a die, on which he laid a disc of metal. The die 
had a raised edge, and the depth of the recess was a very little more 



Oct. 1905. 



CARTRIDGE-CASE MANUFACTURE. 



833 



than the thickness of the blank. He then brought down upon it a 
stout plate, held down in some cases by cams, in other cases by 
springs, and sometimes by hydraulic pressure, so that the plate bore 
very hard on the raised rim of the die. Then the punch came down 
through the pressure-plate. If a pressure-plate were not used, the 
first thing the blank did, when it was depressed into the die, was to 
pucker up all round the edges, but by being confined between the die 
and the pressure-plate the puckers were smoothed out as fast as they 
were formed, by being drawn over the curved edge of the die. That 
was the solution of the dlfiiculty of forming deep vessels out of thin 
metal. Another device was subsequently resorted to. The idea was, 
that by holding the pressure-plate very hard down, and constantly 



Fig. 39. 



Fig. 40. 




down, the metal blank was subjected to too much stress in turning 
over the inner edge of the die, and another inventor therefore 
affixed the pressure plate to a steam-hammer, and hammered the 
metal disc at about 200 blows a minute, with the idea that as 
the plunger came slowly down, the plate should have the opportunity 
of puckering when not under pressure-plate stress, but immediately 
after the puckers had formed, the pressure-plate should come down 
with a blow and knock them all out again. That was remarkably 
successful. He once saw a blank of German silver, which was not a 
very easy metal to deal with, about 18 inches in diameter and ■^^^ inch 
thick, put into one of the hammering machines and drawn down into 
an article of about the size of a man's hat, 7 inches in diameter, and 
7 inches deep, at one operation without any annealing whatever. As 

3 L 



834 CARTRIDGE-CASE MANUFACTURE. OCT. 1905. 

(Mr. Henry Lea.) 

miglit be imagined, the delight of the inventor was very great. But 
then it was desired to do something more to the article formed, for 
which in its then state it was too hard, and so it was put into 
an annealing oven. The moment the vessel felt the heat it split up 
into about 50 ribbons, opening out in bold curves, making an object 
such as shown in Fig. 40 (page 833). Therefore the advantage 
which the hammering process gave by enabling a great deal of work 
to be done on a blank at one operation, was lost from the fact that it 
could not be annealed afterwards. 

There was a very interesting example of the way in which 
metals behaved under certain conditions. Fig. 41. It was a kind 
of stud or hollow rivet, with a bottom formed to it. Over the stud 

Fig. 41. Fig. 42. 



was slipped a washer with a conical mouth and a cylindrical body 
below. Pressure was brought to bear on the top of the stud, 
generally by a hammer and a punch, the punch having a little cup at 
the end. The first thing the punch did was to expand the top of the 
stud into the counter-sink of the washer ; then it carried the top 
down, and the washer with it, until the bottom of the cylindrical 
portion of the washer began to indent the curved shoulder of the 
stud. This went on until the metal of the stud began to fold 
downwards so as to enclose the cylindrical portion of the washer. 
Further blows caused the cylindrical part of the washer to come 
almost down to the base of the stud, Fig. 42. The stud could thus 
be reduced to about one-fourth of its normal height, or to any less 
extent, as the material to be fastened together might require. When 
one of these studs was cut down the middle and examined, it was 



Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 835 

an extremely pretty example of the way in wliicli tliin metal 
could be folded about upon itself, if one only went the riglit way 
to work. 

Mr. J. C. Aylan was greatly interested in the Paper, because 
the manufacture of cartridge-cases happened to be part of his work 
at Woolwich Arsenal, and he was naturally desirous of hearing 
or seeing anything of the latest processes. He did not wish in 
any way to repeat what had been already said by Mr. Donaldson, 
but would just say that the principle in the work of the development 
of cartridge-cases was probably on what might be called the 
physical side of engineering, and was distinctly a very interesting 
work, and one quite capable of eliciting a great deal of 
enthusiasm. There were just one or two points he desired to refer 
to in connection with the tools and the plant. Of course it would 
not be right to say that the plant as shown was the very latest, or 
that it combined the latest advantages, and he might illustrate that 
remark by one particular example. He noticed that they removed 
the die, which was a very massive part of the plant, for every 
particular heading operation. At Woolwich they had happily 
learned to keep the die stationary, and play about with the distance 
piece, which was very readily and very rapidly done by means of 
guides and runners, as easily as if it weighed only a twentieth part 
of the die. 

One thing that had struck him was the elaborate method adopted 
to extricate the case from the bolster. At Woolwich they certainly 
had not such long cases, but they had thin cases, probably as thin as 
those in the final draw of the case mentioned in the Paper ; but he 
had always been much impressed with the fact that it was very 
essential that the bolster should be highly polished, in order to 
eliminate entirely the possibility of pinning from the outside of the 
bolster on to the inside of the drawn case. At Woolwich they had 
no difficulty in extracting from the draw, and thought it was largely 
due to this very fine finish of the surface of the bolster. That, 
however, was an expression of opinion which he would leave for the 
consideration of the members. 

3 L 2 



836 CARTRIDGE-CASE MANUFACTURE. Oct. 1905. 

(Mr. J. C. Aylan.) 

He also noticed that in the tapering referred to in the Paper 
there were two operations. That might perhaps be associated in 
some measure with the great length of the case, but he did not think 
wholly so. At Woolwich they tapered entirely on one operation ; 
also the die for tapering was decidedly a much more economical one 
than the " built-up " die referred to in the Paper. It was made out 
of good cast-iron, bored out carefully to the exact curvature, and was 
quite a cheap die compared with the built-up one of best steel 
mentioned in the Paper. Those who had been associated with the 
development of small or large cartridges knew that there were so 
many intangible causes that brought about trouble that they were 
thankful always when things were going right. When a man was 
making a thing to dimensions he had something tangible to bring 
about, but in the manufacture of cartridge-cases the work was very 
trying sometimes, as the conditions to be met were very severe. It 
was evidently a class of work that was going to call upon the general 
engineer very much for good design of plant, and was also going to 
call upon the metallurgist and physicist to put their best into it. 
The possibilities of dealing with metals were intensely interesting, 
when it was realised what latitude there was for playing up and 
down with annealing and curvatures and so on, so as not to check 
the metal in its flow round the curves. It was in fact a constant 
and deeply interesting study, and at Woolwich they were rather 
enthusiastic about that branch of their work. 

Mr. S. W. Challen had been very much interested in what Jie had 
heard, and it had occurred to him that those who paid attention tj 
the manipulation of sheet metals were rather in want of some 
distinctive terms. The manufacture of cartridge-cases was hardly 
the same thing as drawing from thin sheet metal, mentioned by 
Mr. Lea. In connection with cartridge-cases, it was desirable to 
have some term accurately describing the manufacture, in order to 
show the difference between the two systems of drawing. He 
suggested that cartridge-drawing was more a reduction of thickness 
of material, and also an entirely different process from the class 
of work which Mr. Lea had referred to, and which, for want of a 



Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 837 

better term, lie should call pressure-plate drawing. In cartridge- 
making no pressure-plate was required ; the machines were of a 
simple character, single-acting machines. With the pressure-plate 
drawing the action was doubled ; there was one action to apply the 
pressure-plate, and another action for attacking the metal and 
changing its form. The subject was really such a large one that ho 
felt at a loss what point to speak upon. 

Mr. Lea had spoken about what had been done forty years ago, 
and he himself was hardly desirous of going one better. He was 
very young in those days, and could remember that when he was a 
pupil they had a machine for making door-knobs out of sheet metal, 
and it would be interesting to note that that machine was made in 
the year 1858. Another machine of a similar character was used 
for making the shells for cloth-covered buttons. Very great things 
were attempted. They tried to make five objects by each stroke of 
the machine, but the difficulty was to get the pressure uniformly 
on the five discs, and also to get the pressure around the 
periphery sufficiently equal to prevent one side going down more 
rapidly or more deeply than the other. He just mentioned that to 
show that pressure-plate drawing had been worked at for a great 
many years. He felt constrained to say another thing with 
reference to the remarks of Mr. Lea, who gave Birmingham the 
credit for having invented the pressure-plate. He would not 
dispute it ; but he himself had been told tuat it was first used in 
Paris, and from thence brought into Birmingham at a very early 
date, and used at the works of Messrs. Griffiths and Browett in 
Bradford Street, and Messrs. Hopkins in Granville Street. There were 
in those days some curious methods of working. One was that in which 
the pressure-plate resembling a pipe-flange was applied by hand and 
screwed down by a number of bolts and nuts, and the formation was 
done in a large screw-press with wheels that ran up to 8 or 9 feet in 
diameter, several men pulling at the wheel and walking round and 
producing a basin or tin bowl entirely by manual labour. It was 
a very expensive method compared with that by which such articles 
were produced in these days. When he mentioned that now tin 
basins and bowls could be produced in a modern press at the rate of 



CARTRIDGE-CASE MANUFACTURE. 



Oct. 1905. 



838 

(Mr. S. W. Challen.) 

70 or 80 gross in a day, by one man laying the blanks on tbe die, and 
at a cost in labour that was merely fractional, it would be seen that 
some advance had been made since those days. The subject of 
drawing sheet metals was such a large one that it deserved quite 
lengthy treatment, far more than he was able to give that evening. 



Mr. William Schonhetder said that on page 809 a word was used 
about which he had had something to say several times, and he 
had to refer to it again. The word was " tempered," and he thought 
it ought to be " hardened and tempered." It was used in speaking 
about the seven steel rings B, Plate 42. It was impossible to temper 

Fig. 43. 



Ra^nv or 
Packing Plate, 




steel, to do any good, unless it was hardened first. Since he made 
his first remarks against its use, the employment of the objectionable 
word had been discouraged, and " heat treatment " had been used. 
Either "heat treatment" or "hardened and tempered" was equally 
good. 

In the cylinder packing shown on Plate 46, there was a mistake 
in the design of the packing for the plunger. The plate 
forming the backing of the hydraulic leather was shown as being 
quite flat, thereby leaving a ring-formed space of triangular section 
between the inner wall of the cylinder, the said plate and the back 
of the leather. In this way the leather would be without any 
support at all on its back (the water in this space of course slowly 
leaking away past the ram and the pressure on it disappearing), 



Oct. 1005. CARTRIDGE-CASE MANUFACTURE. 839 

while the full hydraulic pressure would be on the inside of the 
leather. He regretted to say he had seen a good many of this kind of 
faultily constructed packing, and they generally burst after quite a 
short usage, sometimes almost directly. The packing plate or the 
ram should obviously be formed so as not to leave any such empty 
space as in Fig. 43 (page 838). 

Mr. J. M. Ledingham observed that much had been already 
said which it was quite unnecessary to repeat. There was one 
thing, however, on which he should like to say a word, namely, the 
valve arrangements. No doubt it occurred to many that the valve 
arrangements in the presses were very complicated, much more so 
than was really necessary, and he was sure that those present 
working with hydraulic pressures would agree that when dealing 
with 2J tons per square inch, it was^ most difficult to control 
cylindrical and automatic slide-valves and keep them tight. With 
regard to the bye-pass valve shown in Fig. 16 (page 813), which 
looked very pretty, it was said that when d was opened, the water 
escape by pipe e to the tank, and the pumping absorbed little or no 
energy. Apparently there were a number of those valves, and there 
would be very considerable leakage, resulting in a very great loss of 
energy, because a great deal of the water that was being pumped 
under pressure was passing into the waste-pipe. A much more 
satisfactory and simple arrangement was to have a connection from 
the accumulator which automatically worked the throttle-valve on 
the steam-pumps, so that when the accumulator was at the top the 
steam-pumps were stopped. Therefore he thought that the statement 
as to economy and saving energy was perhaps to be doubted. 
The valve arrangements altogether seemed to him to be quite 
unnecessarily complicated. 

Mr. J. J. Edwards had been intensely interested in the Paper, 
and supposed that, in common with other cartridge-case makers, he 
would be slow to admit that the process had very much in it that 
was superior to that with which he might be concerned, because when 
two or three persons set about making any particular article each 



840 OARTKIDGE-CASE MANUFACTURE. Oct. 1905. 

(Mr. J. J. Edwards.) 

one would attain perhaps the end in view by different methods. 
The gentlemen who had elaborated the method under discussion had 
attained their end on different lines from those of the shop in which 
he had been engaged, although in that shop they had reached the 
same end. The differences had been mentioned in several particulars, 
and he would only refer to one in connection with an important 
point on the subject of brass, namely that of annealing. It was said 
that they varied the temperature of annealing according to the 
thickness of the disc or cup to be dealt vdth. To his mind that was 
not sound. Every member had not, perhaps, had much dealing with 
brass, but he presumed that everyone had dealt with steel, and when 
dealing with a piece of steel it was known that to anneal a small 
piece, or a large piece, it was necessary to raise the temperature to a 
certain given point. The difference was, that it took longer to raise 
the larger piece to the desired temperature than it did to raise the 
thinner piece. His opinion was, that in connection with annealing 
brass, whether the brass was thin or thick, whether it was light or 
heavy, it required the same furnace temperature, but a longer time 
to get the larger piece of brass to that temperature than was 
required for the thinner. In his own practice he adopted one 
definite temperature, but varied the time according to the size of 
the piece of brass to be dealt with. He thought that was sound 
practice, and one that lent itself to facility in general work. It 
was possible to take large cups out of the furnace and follow 
with thinner material, or take thin material out and follow with 
thicker, using the same temperature but varying the time, as it 
took a longer time for the heat to permeate the larger masses 
than smaller. In general practice he found that was a correct 
principle. 

It was said that brass could be annealed by lapse of time. That 
statement was, he thought, open to some doubt. At any rate he 
found that, if he allowed a cartridge-case to go away from his works 
not properly annealed, it came back later on, perhaps after many 
years ; and no matter how long it had been on service, it never 
became properly annealed. The annealing must be done in 
connection with the work, and he believed no annealing would take 



Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 811 

place afterwards. Sometliing would take place, if the annealing had 
not been properly done. If the head was too hard, it would be found 
to come back, perhaps in two pieces or badly damaged, as the result 
of not being properly annealed before going away. It might crack 
spontaneously in the base, and one would bo fortunate that it did 
crack spontaneously before it was put into the gun. He hoped 
that some explanation would be given of the question of brass being 
annealed by lapse of time. It was a point there might be something 
in, although it seemed quite opposite to the general experience of 
people who had had to do with brass. 

There was a point in reference to the horizontal and vertical 
presses to which he wished to refer ; he understood that the earlier 
draws were done in the vertical press, and the later draws in the 
horizontal press. He did not know really why, because as far as 
he was concerned he should prefer the vertical right through. As 
a rule, three vertical presses could be put where it was possible to 
put only one horizontal. The work could be equally well done on 
either, as far as the drawing was concerned, but one was much more 
certain of getting concentricity in a vertical press than in a 
horizontal. The reasons were obvious. 

Captain E. W. Davies said the manufacture of brass cartridge- 
cases was an interesting subject, esj)ecially so at the present time, 
owing to the re-armament of the Horse and Field Artillery with 
quick-firing guns. It was impossible to obtain quick-firing unless 
the whole round could be loaded together, as otherwise there must be 
the separate operations of taking a tube from the pouch and inserting 
in the vent ; and in the case where there was an immovable carriage, 
and consequently very little alteration in laying was necessary for 
the subsequent round, there was no time for the operations to be done 
except at the expense of rapidity of fire. In future wars, he thought 
that side which had in reserve the power to pour in for a short time 
an overwhelming fire would have a tremendous advantage, as it would 
enable them to prevent the enemy's artillery from taking up more 
favourable positions during the battle, and to destroy them in detail 
if they attempted to do so. 



842 CARTRIDGE-CASE MANUFACTURE. Oct. 1905. 

(Captain E. W. Davies.) 

With regard to the heavier guns, where the projectile had to 
be loaded separately, and the whole surface of the gun was so 
much slower, the advantage of the brass case was not so marked, 
although here also it assisted in increasing the rate of fire chiefly 
through doing away with the necessity of sponging out after each 
round. Against that there was, however, the disadvantage that when 
the charge was assembled with the means of ignition, if an accidental 
explosion occurred outside the gun, the consequences were much 
more serious to the gun detachment than in the case of the smaller 
cartridges. 

The next point was, what was to be the material for the quick- 
firing case ? He thought it was generally agreed that on the whole 
brass was most suitable for the purpose, although the metal was not 
altogether in favour with certain experts, owing to the fact that 
occasionally the cases burst on firing, and caused serious injury to 
the gun. In order to get over that difficulty, experiments had been 
carried out with the view of obtaining more suitable material, but so 
far, he believed, they had only led to a slight alteration in the 
proportions of the existing ingredients, or possibly the addition of 
small percentages of other metals ; and he believed an alloy of copper 
would provide in the future, as it had in the past, the material from 
which the case was made. He thought it would be agreed that it 
must be a metal in order to get sufficient strength to secure the 
shell, and so make the whole round complete. The number of cases 
that had burst on firing was only a small fractional percentage of 
those used, but the consequences were so serious when they did 
occur that the attention of gunners all over the world had been 
specially called to it. 

Bursts might be of two kinds, either those at the base or those 
in the forward part of the case. The latter were not so serious as 
the former, because there was no escape of gas to the rear, and 
nothing more occurred than a scoring of the chamber of the gun, 
which however was very objectionable, in that it might cause other 
cases to fail at the same point in a similar manner, or make them 
difficult to extract after firing. Base splits were much more 
serious, as in such instances the gases came out to the rear round 



Oct. 1905. CARTRIDGE-CASE MANUFACTURE. 843 

the breecli-block, washing away the metal of the gun, and generally 
putting it out of action. The fracture usually occurred ia the 
angle at the bottom of the case, and in very bad instances might 
extend right through the rim as far as the j)rimer hole. The 
question arose as to what caused the metal to give in that manner, 
and he thought the answer was that it was generally due to defective 
manufacture. As far as the chemical purity was concerned, it was 
easy by frequent analysis to ensure that the specified formula was 
being worked to, but that would not help the manufacturer very 
much if his treatment of the metal in the various operations was 
wrong. 

He believed that the whole secret of success in the manufacture 
of brass cases was in the treatment of the metal, and the use of 
proper tools. As stated in the Paper, the head of the case must be 
as hard as possible, in order to withstand the high pressures on firing 
without deformation. From that point up to the mouth of the case 
the metal should gradually get softer, so that at the latter point the 
metal was fairly soft. That was necessary in order that the case 
itself might act as an efficient gas-check on firing. The important 
point was, that the change must be a gradual one ; any sudden 
change in the size of the crystals making a position of weakness 
where fracture was likely to occur on firing. On the other hand, 
the front part of the case must not be too soft, or there would be a 
difficulty in extracting the fired case, and if such occurred, then the 
term quick-firing as applied to the gun became a misnomer. In order 
to get the head hard, considerable work had to be put into it, and 
unless this was properly done it was liable to leave certain portions 
of the metal in a state of unequal strain, and in such cases sooner or 
later a burst was bound to occur at the point where such conditions 
existed. The inspection of quick-firing cases was therefore a 
difficult one to carry out. It was easy enough to test the chemical 
composition, and to ensure the correctness of the dimensions. The 
case could be examined for spontaneous cracks, although such were 
not likely to develop in the short time allowed for inspection, and 
such cases would not be used. But as regards testing them for the 
defects he had just indicated, and the condition of the molecules of 



844 CARTRIDGE-CASE MANUFACTURE. OCT. 1905. 

(Captain E. W. Davies.) 

the metal itself, that could only be done in a percentage of each 
batch made, and that necessarily a very small one. There were 
certain rough and ready methods that might be employed which 
would assist in the inspection, but the best one was by an examination 
under the microscope of specimens prepared in the manner described 
in the Paper, in which one could see all the pitfalls that a cartridge- 
case maker was likely to fall into — first, the metal in a state of 
unequal strain; secondly, the sudden change in the size of the 
crystals; thirdly, the metal being too hard; fourthly, the metal 
being too soft ; and fifthly, the fold in the angle near the base, 
which had frequently proved to be a cause of failure. In the Morris 
Tube Co., which he had recently had the honour of joining, they did 
not make cases as large as those mentioned in the Paper, although they 
made large numbers of 1-inch for their practice ammunition ; and as 
the manufacture of all solid-drawn brass cases was carried on under 
the same geijeral principles, he was very glad to attend the meeting, 
and take part in the discussion. 

The Chairman (Mr. Aspinall), in closing the discussion, said 
that a complete report of the discussion would be sent to Colonel 
Cubillo for his reply. The Paper had been extremely interesting, 
and he knew of no subject more fascinating to the engineer than the 
designing of machinery for the purpose of drawing metal. The 
pleasure it gave to anybody who had had to deal with the subject 
must be great, as he succeeded in producing the result he required, 
and observed the metal passing from one die to another, and being 
dealt with in that persuasive way which dies had so as to produce 
the final form. 

Another point in the Paper was that it showed the great 
advantage of annealing, which was not confined by any means to the 
class of metal dealt with. If one had to deal with such an ordinary 
matter as a steel boiler-plate under the testing machine, it was 
quite easy to show the value of annealing by stretching that piece of 
boiler plate by 10 per cent., then annealing, then stretching it 
another 10 per cent., and so on, so that one ultimately obtained 
without the slightest difficulty 85 to 90 per cent, elongation. 



Oct. 1905. CARTRIDGE-CASK MANUFACTURE. 815 

Perhaps he might point out that the Paper was another illustratioa, 
in these days when engineers were dealing so extensively with 
electrical machinery, of the fact that there were certain processes ia 
which the use of hydraulic machinery was absolutely essential, and 
that there was not the least likelihood of displacing that very useful 
and powerful agent. 



Communications. 



Colonel R. H. Mahon wrote that there were a few points on which 
remarks might be useful. His experience in the matter of temperatures 
of annealing referred to (pages 799 and 800) agreed with that of the 
authors, namely, that the temperature might reach at least 730^ C. 
(1,346° F.) before any of the indications of burnt metal occurred ; 
it was probable, however, that when the temperatures were approaching 
the maximum, the metal could not be left so long under the heat 
influence as when lower temperatures were used. 

On the question of spontaneous annealing a good deal might be 
said ; there was no doubt of the fact that brass finished in a state of 
internal stress, which expression might be held to include " hard 
rolled " or " hard drawn " metal, had a tendency to assume the 
original state, and this fact had an important bearing on the 
manufacture of cartridge-cases and fuses. It had also been noted 
that this effort to resume the normal state of " no stress " appeared 
to be facilitated in hot climates. The excellent photo-micrographs 
which accompanied the Paper, Plate 38, showed clearly the crystalline 
structure of the metal, and it was stated that the "embrittling" 
which followed successive operations of drawing or extending must 
be removed by the process of annealing. He thought, therefore, 
that it needed no apology to express an opinion against the 
maltreatment of the metal described under the operation of 
" heading " (page 808). He thought that there could be no doubt 
that the tendency to " reassertion " or " resumption of original 



I 



S-iQ CARTRIDGE-CASE MANUFACTURE. - OCT. 1905. 

(Colonel R. H. Mahon.) 

state " already referred to must be rendered active by such treatment, 
the more so that it was not usually followed by any annealing ; and 
it was a fact that, though large numbers of excellent cartridge-cases 
had been and were made by the method described, there had 
nevertheless been failures which were certainly attributable to some 
extent to the crushing to which the cases were subjected. To 
those who wished to go further into this matter, he recommended 
a study of the rolling method by which the base was rolled out 
with but slight pressure. The whole of this rolling process was 
worthy of attention, and there was little doubt that it left the metal 
in better condition to resist "time alteration of structure" which 
was an important point. 

One other point might be mentioned, namely, the question of 
annealing ; the class of furnace advocated exposed the brasswork to 
contact with the hot gases, rendering the use of the pickling baths 
necessary. An excellent class of annealing furnace had been 
designed in England, in which the metal was annealed out of all 
contact with air. and the necessity for acids was obviated. He 
thought that the abolition of this operation of pickling was in many 
ways an advantage : it was difficult to remove the acid entirely by 
subsequent washings, and it seemed probable that galvanic action 
subsequently slowly influenced the alloy. 

Mr. James Tennant wrote that it seemed to him that a very 
great deal of attention and thought had been given to the cupping, 
drawing, and heading, but that the annealing was still being 
followed on the old and wasteful lines. It was a well-known fact 
that all non-ferrous metals, excepting the rarer ones, such as gold, 
platinum, etc., had when heated a great afBnity for absorbing the 
oxygen contained in the atmosphere, and this amalgamation of 
oxygen and metal created a film more or less thick on the metal 
undergoing the annealing process. "When the metal was cooled 
down, it was found that its outside skin had been converted into an 
oxide of the metal which naturally must be removed before further 
progress could be made. This was subsequently removed by 
pickling and washing, as described by the authors (page 805). As 



Oct. 1905. CARTRIDGE-OASE MANUFACTURE. 847 

far as could be seen, eleven annealing operations were gone through 
with the usual pickling process after each heating, and it seemed 
that the loss due to this method of annealing must be very- 
considerable and the cost of handling very great. Experiments 
made showed a loss of 4 per cent, for each annealing and pickling. 

Quite recently a British-made annealing furnace or machine had 
been brought out. The writer had seen several of these furnaces at 
work, and could not be otherwise than impressed with their efficiency. 
It was called the Bates and Peard process.* The main object was 
to obtain the perfectly bright annealing of all non-ferrous metals, 
free from scale or discoloration, and thus to avoid the pickling and 
washing ; consequently there was a great saving in metal compared 
with the open-hearth annealing furnace process as on Plate 47. This 
method also provided an automatic means of charging and discharging 
the metal into and from the annealing chamber which reduced the 
handling to a minimum. This object was effected by the use of an 
air-tight chamber or tube, the ends of which descended and were 
sealed by or terminated in water. This chamber was suitably 
disposed in a brick or other heating furnace, and could be fired by 
gas, oil fuel, or coal. The metal to be annealed was conveyed 
through the chamber on a mechanical conveyor in the form of an 
endless-chain belt. The unannealed cartridge cuppings and drawings 
were placed in iron trays, and the trays were then stood on the 
conveyor at one end, and were then drawn nrst through one water- 
seal, then through the heated portion of the tube and thence out 
through the other water-seal by which it was cooled, the water-seals 
thus forming entrance and exit doors to the furnace. The British 
Government had adopted this process at the Eoyal Mint for 
annealing coin blanks and sheets from which the blanks were 
stamped, and experiments were about to take place also at the 
Eoyal Arsenal, Woolwich, for small-arms ammunition. 

There had long been a prejudice amongst workers in brass 
that a sudden cooling by immersion in water was detrimental to 
the article being annealed, it being assumed that the sudden chill 

* " Electrical Magazine," August 1905, page 105. 



848 CARTRIDGE-CASE MANUFACTURE. Oct. 1905. 

(Mr. James Teunant.) 

was liable to develop cracks in the metal, but be noted that mention 
was made (page 800) that the rate of cooling bad no effect upon the 
quality of tbe annealing, and tbis confirmed tbe opinion of tbe 
inventors of tbis process. Tbe annealing, wbicb took place in a 
non-oxidising medium, prevented tbe burning of tbe metal ; tbus any 
delay in removing wben fully annealed, wbicb by otber processes 
would damage tbe article being annealed, was of no consequence, 
and tbe evenly distributed incandescent beat tbrougbout tbe full 
lengtb of tbe annealing cbamber gave a very uniform annealing, 
wbicb would allow of more work being done on tbe metal between 
tbe various annealings tban was possible by tbe present means of 
softening. A means of varying tbe speed of tbe conveyor, so as to 
give longer or sborter time in tbe beat, was attached to these 
machines wbicb were thus capable of dealing with each of tbe 
successive operations in the manufacture of small-arms and quick- 
firing ammunition almost automatically. 

Mr. John H. E. Whinfield wrote that tbe subject of tbe Paper, 
in so far as it applied to cartridge-cases for large guns, that is, over tbe 
calibre required for field artillery, was rather belated, tbe days of tbe 
large brass cartridge-case being past as regards any new manufacture. 
As soon as tbe weight and size of the projectile required 
separate loading, tbe advantage of tbe brass case began to disappear 
and disadvantages arose. For example, the weight of an empty case 
for the 6-incb gun was about 14 lbs. Wben the breech was opened, it 
would not do to eject tbe cartridge bodily on to tbe ground or tbe 
deck of tbe vessel, as it might injure tbe men serving tbe gun, or it 
would be in the way of service and be deformed by such treatment. 
It was, therefore, necessary, after tbe breech had been opened and the 
cartridge loosened in its seat by the extractor, that the man should 
withdraw the case and carefully deposit it out of tbe way. The 
case was too hot to handle and be bad to use a special tool, so that 
after discbarge the case was rather an encumbrance. Before it could 
be used again, it would require to be carefully cleaned and if 
necessary reformed, an operation for wbicb there was no time in 
action. There was no more difficulty in obturating tbe breech of a 



Oct. 1905. CARTRIDGE-OASE MANUFACTURE. 849 

1'2 or 15 cm. (4*7 or 5*9 inches) gun than of guns of larger 
calibres, and, therefore, the combustible cartridge (by avoiding the 
disadvantages due to the disposition of the brass cartridge after 
use) seemed to be the likely future demand. Even with combined 
loading in smaller guns, such as field artillery, it was possible that 
the brass case would be found an inconvenience in the battery if firing 
continued for any time. 

Colonel CuBiLLo wrote thanking the Members for the kind 
reception they had given to the Paper. He was very pleased with 
the remarks of Mr. Donaldson, who had made such an excellent 
addition to the Paper by the photo-micrographs he had shown, 
Plates 48 and 49, especially Figs. 37 and 38, Plate 49, which were 
most interesting. With regard to the brittleness and hardness 
produced by cold-working, these were not special properties of 
brass, but were common to all metals in general, and it was also a 
general property of metals that the ductility lost by cold-working 
was restored by annealing. Cold-working could be assimilated, 
in its effects, to the hardening of steel, because the tenacity or 
breaking strain became higher and the ductility lower, and that both 
processes could be assimilated was shown by the fact that annealing 
restored the mechanical properties to their former values before 
hardening or cold-working. With regard to spontaneous cracking, 
he was entirely in accordance with Mr. Donaldson's views on the 
matter, and was of opinion that it was due to the metal not being 
properly annealed. 

It was quite natural that the pressures and the number of 
drawings might not be the same at Woolwich as at Trubia, the cases 
being different both in length and in diameter, but he noted with 
great pleasure that the general process was quite the same. With 
regard to Mr. Donaldson's remark about the mention made in the 
Paper of the work always being done cold, he (Colonel Cubillo) 
knew very well that, in the initial stages of making cartridge-cases, 
hot ingots were employed, the cases being punched out trom square 
or hexagonal ingots, just as steel was drawn, and at Trubia this 
process had been tried for making linings for the recoil cylinders of 

3 M 



850 CARTRIDGE-CASE MANUFACTURE. OCT. 1905. 

(Colonel Cubillo.) 

the mountings. The other method of rolling mentioned by 
Mr. Donaldson (page 831), and another followed at Madgeburg, in 
Germany, lengthen ing the cases by balls rolling on the metal, were 
not superior, in his opinion, to cold-drawing. This was the best for 
brass, because its high ductility at ordinary temperatures placed it 
in the same condition as iron and steel at the highest, and it was 
possible to have at the end of the whole process the cases with the 
metal in the best possible condition, if the annealings had been 
properly conducted. At Trubia 3-inch cartridge-cases had been made 
for mountain guns, which had fired forty to fifty rounds. 

With regard to the valves at the presses, they had given very 
little trouble, if any, thus proving the excellent design of the makers. 
It was interesting to know the pressures reached at Woolwich, which, 
as mentioned before, were in accordance with the difference of the 
cases. For the heading process the pressure at Trubia was applied 
three times, the first one of 1,600 tons, the second of 600 tons, and 
the third of 1,650 tons. 

Mr. Henry Lea's remarks (page 832) were very interesting, 
explaining the first device for converting a flat disc into a cup-shaped 
form. He noticed that the results obtained by the pressure-plate 
process without annealing were exceedingly good. 

With regard to Mr. Aylan's observations (page 835), he would 
say that the plant at Trubia for making cartridge-cases was one of 
the best in the world ; it was on the same lines as that erected at the 
Washington Navy Yard. Certainly the chucks and dies were very 
massive pieces, but in every operation only the die was removed ; 
the chuck was the same for the first five draws of the vertical press, 
and also the same for the remaining five draws in the horizontal press. 
The dies were very light pieces and easily removed. Mr. Ay Ian 
explained the necessity for the elaborate nature of the process for 
extracting the case from the bolster, due to the great length of the 
cartridge-case. It was certain that the diflaculties of this extraction 
increased greatly with the length of the case. It was indeed 
interesting to know the simple and economical device employed at 
Woolwich. As Mr. W. B. Challen had dealt in his remarks 
(page 836) with the statement made by Mr. Henry Lea on the 



Oct. 1905. CARTRIDGE-OASE MANUFACTURE. 851 

pressure-plate process, and not witli the cartridge-case drawing 
process, lie had to thank him for his interesting explanation, showing 
the wonderful advances made in mechanical devices for making the 
work more accurate, easy and cheap, with less fatigue to the men. 

The remark of Mr. Schonheyder about the word " tempered " was 
perfectly correct (page 838). It is unquestioned that " tempering " 
was the heat treatment to which the high carbon steels were regularly 
subjected after hardening, and was a different process from that of 
annealing. With regard to the observations about the packing of 
the cylinder as shown in Plate 46, he thought this illustration 
was correct. 

He had read with much pleasure Mr. Ledingham's remarks 
(page 839) about the valve arrangement as illustrated in the Paper, 
which struck him as being extremely complicated, but this very 
pretty arrangement had worked admirably, and had given cartridge- 
case makers entire satisfaction. In reply to Mr. J. J. Edwards, he 
would say that the processes as described in the Paper were not 
superior to any other practised by other cartridge-case makers, but 
he thought that these processes were conducted in accordance with 
the proper treatment required by the metal both in the cold and hot 
states, and that the machinery was of the best description, strong, 
well arranged, and economical. 

With regard to Mr. J. J. Edwards's remarks about annealing steel 
(page 840), the temperature must be raised to a certain given point. But 
in annealing, as in hardening, there was a certain range of temperature 
at which both processes could be properly conducted. In the Paper 
it was said that in nine out of twelve of the drawing and cupping 
operations the range of annealing was only from 650° C. to 600° C. 
(1,202° to 1,112° F.), a very small range, taking into consideration 
that, as ascertained by experiments at Trubia and at another 
works, the brass could be properly annealed at temperatures 
varying from 740° C. to 540° C. (1,364° to 1,004° F.). There 
was no doubt that not only brass, but any other metal, strained 
beyond its elastic limit could be restored to its primitive condition 
before being strained by the slow action of time, and in the 
manufacture of cartridge-cases, this result was attained by the 

3 M 2 



852 CARTRIDGE-CASE MANUFACTURE. Oct. 1905. 

(Colouel Cublllo.) 

quickest action of the annealing process. Perhaps tliis could bo 
explained by saying that the molecular state of metal, wlien 
strained beyond its elastic limit, was not stable at ordinary 
temperature, and then it tended to recover its primitive state and 
therefore its primitive tenacity and elongation. There was also a 
scientific reason for not annealing the cases at the same temperature 
in the diflerent drawings. The strain of the metal was not the same 
in each drawing ; it diminished in every successive operation, and 
the metal being less strained, it was not necessary to give it so much 
heat for the recovery of the ductility. With regard to the presses, 
the horizontal ones had worked very successfully. 

The author completely shared the views of Captain Davies referring 
to the loading and firing of the guns, and he also thought that, 
on the whole, brass was the most suitable metal for cartridge-cases. 
The alloy had, as stated in the Paper, many remarkable properties, 
and one of the best was certainly its little afiinity for oxygen. At 
Trubia some experiments had been carried out with short steel cases 
for modern howitzers, and the results obtained had been extremely 
good ; these short cases, of course, formed one of the best obturators 
it was possible to find for guns. 

With regard to the bursting of cartridge-cases (page 842), he 
thought Captain Davies' remarks were quite right ; the bursts at the 
forward part were the commonest and least serious ; those at the base 
were very rarely seen at Trubia, and it never had happened that the 
burst had extended right through the rim. The author was very much 
inclined to think that this accident was due to defective annealing. 
The most ordinary way of testing the cartridge-cases was to choose a 
certain number out of a batch, and to fire with them a stij^ulated 
number of rounds. With regard to chemical and mechanical 
properties, those must be ascertained before commencing the 
manufacture. What must be very useful, as Captain Davies pointed 
out, was the examination of the micro-structure, on the different 
parts of the case, but this test, as that of firing a certain number 
of cartridges, could only give an idea of the cartridge tested, not of 
the whole batch. He quite agreed with Mr. Aspinall's observations 
regardiug the hydraulic machinery and the cold-drawing process. 



Oct. 1005. CARTRIDGE-CASE MANUFACTURE. 853 

He entirely agreed with Colonel Malion's remarks (page 845) 
regarding the temperatures of annealing and also with Lis 
observations about the tendency of the hard-drawn metal to 
assume the original or annealed state. With regard to the heading 
operation, he thought it was absolutely necessary to submit the 
brass to a very high pressure, in order to prevent that of the 
powder-gases from deforming the metal. He thought it was a good 
idea to anneal the cases out of contact with the furnace gases, which 
obviated the use of the pickling bath. 

With regard to Mr. Tennant's remarks (page 847), he was very 
pleased to know the process described, the main object of which was 
to obtain a perfectly bright annealing of all non-ferrous metals free 
from scale or discoloration, without the necessity of pickling and 
washing. Undoubtedly that process must be adopted where the 
annealing of brass pieces was wanted in great quantities. He was 
glad that Mr. Tennant concurred in the author's opinion that the 
rate of cooling had no effect upon the quality of annealing. 

He thought that Mr. Whinfield's remarks (page 848) hardly came 
within the scope of the subject of the Paper which dealt with the 
Manufacture of Cartridge-Cases, but rather to the convenience of 
their use in the service of 6-inch guns. With some of these points, 
however, he was in agreement. 



Nov. 1905. 855 



ii^t ^UBiiktm d ^Ttcjanical Engineers. 



PROCEEDINGS. 



November 1905. 



An Ordinary General Meeting was held at the Institution on 
Friday, 17th November 1905, at Eight o'clock p.m. ; Edward P. 
Martin, Esq., President, in the chair. 

The Minutes of the previous Meeting were read and con^rmed. 

The President announced that the following five Transferences 
had been made by the Council since the last Meeting : — 

Associate Memhers to 3Iemhers. 

Cooper, Arthur Thomas, . . . Reading. 

HoPKiNSON, Allen Haigh, . . . Huddersfield. 

Lea, Frederick Mackenzie, . . . Birmingham. 

Marshall, Launcelot Paul, . , . London. 

Graduate to Associate Member. 
Stokes, Frank Torrens, . . . Johannesburg. 



856 BUSINESS. Nov. 1905^ 

The following Paper was read and partly discussed : — 

" Seventh Report to the Alloys Eesearch Committee : On 
the properties of a series of Iron-Nickel-Manganese- 
Carbon Alloys " ; by Dr. H. C. H. Carpenter, Mr. R. A. 
Hadfield, Member, and Mr. Percy Longmuir. 



The Meeting terminated shortly before Ten o'clock. The 
attendance was 131 Members and 64 Visitors. 



2sov. 1905. 85' 



SEVENTH REPORT* TO THE 

ALLOYS RESEARCH COMMITTEE: 

ON THE PROPERTIES OF A SERIES OF 

IRON-NICKEL-MANGANESE-CARBON ALLOYS. 



By Dk. n. C. H. CARPENTER, Mr. R. A. HADFIELD, Memher, 
AND Mr. PERCY LONGMUIR. 



CONTENTS OF REPORT. 

PAGE 

Introductory Note .......... 859 

General Introduction .......... 859 

Preparation and Meclianical and Heat Treatment of the Alloys at the 

Ilecla Works, Sheffield 8G1 

Heat Treatment of the Alloys at the National Physical Laboratory . . 863 
Table of Chemical Analyses ......... 864 

The Mechanical Properties of the Alloys — 

Introduction .......... 865 

Heat Treatment of Test-Pieces ........ 873 

Tests with forged Material — 

Bending Tests ......... 874 

Tensile Tests (Yield Point, Maximum Stress, Elongation and 

Reduction of Area) ....... 875 

Torsion Tests (Yield Point, Twisting Moment, Angle of Twist) 884 

Compression Tests ........ 885 

Modulus of Elasticity 886 

* For the First, Second, Third, Fourth, Fifth, and Sixth Reports," see 
Proceedings 1891, page 543 ; 1893, page 102; 1895, page 238 ; 1897, page 31 ; 
1899, page 35 ; and 1904, pages 7, 859 and 1319. 



858 



ALLOYS RESEARCH. 



Nov. 1905. 



CONTENTS — (continued). 

PAGE 

Shock Tests (Energy absorbed, Bending Angle) . . . 8S6 

Hardness Tests ......... 891 

General Summary of the Meclianical Tests with Key Table of 

Properties 894 

Alternating-Stress Tests ........ 896 

Alternating-Stress Testing Machine ...... 897 

Mechanical Properties of the Cast Alloys — 

Tensile Tests (Yield Point, Maximum Stress, Elongation, 

Eeduction of Area) 899 

The Physical, Chemical, and Metallographical properties, including — 

The Kesistivities 900 

. 902 

. 903 

. 905 

. 908 

. 908 

. 914 

. 914 



The Permeabilities 
The Specific Gravities 
The Dilatations . • 
The Corrodibilities — in water 

— in sulphuric acid 
The Eanges of Solidification — of the alloys . 

— of pure nickel 



The Critical Eanges on Cooling — 

The Cast Alloys 917 

Pure Nickel • ... 924 

The Forged Alloys 927 

The Critical Eanges on Heating — 

Cast Alloys 931 

Pure Nickel 935 

The Eeversibility of the Critical Eanges on Heating and Cooling . . 936 

Some Properties of Nickel ......... 938 

The Metallography of the Alloys — 

Introduction .......... 939 

A. The cast alloys, cooled from 900° C. (1,652^ F.) . . . 944 

The forged alloys 946 

The forged alloys, cooled from 800° C. (1,472° F.) . . . 947 

The cast alloys cooled to -100° C. (-148° F.) . . . 952 
The Structure of Alloy K upon which mechanical work of 

various kinds has been done ...... 953 

Subsequent Heat Treatment of K . . . . . . 956 



a 

B. 
E. 

F. 



Concludin": Eemarks .......... 958 



Nov. 1905, ALLOYS RESEARCH. 859 

Introductory Note. — The original scheme of the following 
Eesearch was drawn up by Dr. H. C. H. Carpenter and Mr. 
B. Keeling. 

In seeking a subject of research which appeared likely to yield 
results of both industrial and more purely scientific importance, and 
while paying due regard to the recommendations of the National 
Physical Laboratory sub-committees, it appeared to the authors 
that Mr. E. A. Hadfield's Research on the Alloys of Nickel and 
Iron * opened up more than one field of profitable enquiry. In 
particular, his opinion as to the function of nickel in triplex 
alloys of iron, carbon and nickel, put forward in the research 
mentioned, suggested the desirability of experimenting on a similar 
series of alloys containing more notable amounts of carbon. Mr, 
Hadfield agreed to join in the Eesearch and to prepare a series of 
alloys of the desired composition. The alloys investigated are thus 
in every respect comparable with his previous ones. A study of 
M. H. Le Chatelier's article,f " Les Aciers speciaux industriels,'* 
shows that the whole range of steels of the type about to be described 
are being studied in French Steel Works at the present time. 

Under pressure of other work in the National Physical Laboratory 
Mr. Keeling had to relinquish his share in the research before the 
alloys arrived. His place has been taken by Mr. Longmuir. The 
research has been carried out by the authors jointly, with the 
assistance of various members of the stafi^, to whom a detailed 
reference is made in the concluding section of the Eeport. 



General Introduction. — Although the literature of the manifold 
alloys of iron and nickel is considerable, the number of experimental 
researclies bearing on the investigations about to be described is by 
no means large. The authors have endeavoured to correlate their 
experimental results with those obtained by other workers in the 
same field, but, speaking broadly, only two other series of researches 
come under this head. 

* Proceedings, Institution of Civil Engineers, 1898-99, vol. cxxxviii, page 1. 
t "Eevue de Metallurgie," 1904, pages 574-590. 



860 ALLOYS EESEARCH. Nov. 1905. 

(1) The Prussian Society for tlie Encouragement of Industry, * 
as far back as 1892, inaugurated a series of investigations on nickel- 
iron alloys. In the course of a decade six reports were issued, of 
which practically only the last overlaps, and then only in a few 
places, th^ ground covered in the present research. These reports 
deal only with the mechanical and working properties of the alloys. 
(2) In 1903, M. Guillet f published the results of experiments, in 
which the mechanical and working properties were correlated to the 
structures, with three series of nickel steels, in each of which 
the nickel varied from 2 to 30 per cent., and the carbon was 
0*12, 0"22, and 0*82 per cent, on an average per series. These 
series, although low in manganese, are in other respects comparable 
with the authors' series and offer points both of comparison and 
contrast. 

The resources of the National Physical Laboratory have enabled 
the authors to investigate an unusually wide range of the properties 
of their alloys, which may be grouped under (a) mechanical, (b) 
physical (including metallographical), and (c) chemical heads ; 
and with respect more particularly to physical qualities it 
has been possible to compare and contrast results with those 
obtained, notably by M. Osmond and M. Guillaume in the same 
field. 

The literature of the so-called " Theory of the Nickel Steels," 
which is almost exclusively a product of modern French thought, 
and which at the present time endeavours to account for the 
properties of these steels on the basis of allotropic modifications 
both of iron and nickel, hardly comes in the same category 
as the work described by the authors. Nevertheless they venture 
to think that their experimental results, more especially in the 
physical section of the Report, may have interesting theoretical 
consequences. 



* Berichte des Sonderausschusses fiir Eisen-Nickel-Legieruugen (1892- 
1902). 

t Bulletin de la Societe d'Enoouragement pour I'lndustrie Nationale, 
May 1903. 



Nov. 1905. 



ALLOYS RESEARCH. 



8G1 



Preparation of the Alloys hy one of the authors — at the Hecla Works, 

Sheffield, 

The series of Iron-Nickel-Manganese-Carbon Alloys, A-K, whose 
properties have formed the subject of investigation of this research, 
has been prepared in the following manner : — 

The base of the material used was Swedish Charcoal Iron of 
special purity. This was melted along with the necessary 
ingredients. Nickel and Swedish White Iron. By these means 
approximately uniform percentages of carbon and manganese were 
obtained, the nickel varying from nil to 20 per cent, as aimed at. 
The series may be taken as showing satisfactory products, as the 
carbon percentages were on the whole uniform. They vary from 
0*40 to 0*52, 0*45 being the figure aimed at. The nickel 
percentages showed very little loss. The manganese percentages 
show a variation from 0*79 to 1-03. This element is notoriously 
difficult to obtain constant in any series of alloys. The exact loss on 
melting is shown in Table 1, col. 7. 



TABLE 1. 



] 

■1 


Hecla 


N.P.L. 

No. 


Nickel 
Alloy. 


Percentages. 


Percentage 

Loss of 
Manganese 
in ntjelting. 


VV UIKS 

Mark. 


Nickel. 


Carbon. 


Manganese. 


1 


2 


3 


4 


6 


6 


7 


1798 A/s 


39 


A 


Nil 


0-47 





95 


24 




, B/3 


40 


B 


1 


20 


0-48 





79 


29 




, c/, 


41 


C 


2 


15 


0-44 





83 


34 




, D2 


42 


D 


4 


25 


0-40 





82 


40 




, I 


43 


E 


4 


95 


0-42 


1 


03 


31 




, J 


44 


F 


6 


42 


0-52 





92 


36 




, E 


45 


G 


7 


95 


0-43 





79 


42 




, FA 


4G 


H 


12 


22 


0-41 





85 


45 




> G/4 


47 


J 


15 


98 


0-45 





83 


46 




, H/3 


48 


K 


19 


91 


0-41 





96 


38 



862 ALLOYS RESEARCH. Nov. 1905. 

Tlie original series did not include alloys E and F, col. 3. 
These were prepared some montlis after tlie research had been in 
progress, when it had become evident that more alloys were 
necessary to fill in the gap between 4 and 8 per cent, nickel, in 
which range the properties of the series underwent a profound 
change. 

The reference marks in Table 1 (col. 1) are those of the Hecla 
Works, in col. 2 those of the National Physical Laboratory. 

The reference letters in col. 3 will be used throughout the 
Report. 

The materials prepared, together with the thermal and mechanical 
treatment they received at the Hecla Works, were as follows: — 
Ingots 24 inches by 2-f inches by 2J inches were cast. 

Half the cast ingot was forged down to a round bar, IJ inch 
diameter, and from 4 to 5 feet long. A 10-inch length of this was 
rolled down to ^-inch diameter, giving about a 4-foot length.* 
(This was not done in the case of A, the nickel-free alloy.) The 
two series of bars were sent in this condition to the National Physical 
Laboratory. From the 1^-inch diameter bars nearly the whole of 
the materials used in the various tests — mechanical, physical, 
chemical, and metallographical — were machined, after the resistivities 
had been measured. The J-inch diameter rods were prepared 
primarily for the measurements of coefficients of dilatation. 

The other half of the cast ingot was machined (by the use 
of slotting tools in the case of the harder alloys) so as to give 
eight pieces Ij inch square and 5 inches long. Of these the first 
four alloys, A-D, were sent to the National Physical Laboratory, 
four each unannealed and four each annealed at 750^ to 770° C. 
(1,382° F. to 1,418° F.). The remaining six alloys could not be 
machined without a preliminary softening process, which is 
summarised in the following Table 2. 

* G and H cracked slightly under this treatment. 



Nov. 1905. 



ALLOYS RESEARCH. 



863 



TABLE 2. 



Ni Alloy. 


Ni 
Per cent. 


Softening Treatment. 


E 
F 
G 

H 
J 
K 


4-95 
6-42 
7-95 

12-22 
lo-98 
19-91 


Annealed at 750^-770° C. (1,382°-1,418° F.) 

Annealed at 750°-770° C. (1,382°-1,4I8° F.) 

Annealed at 750°-770° C. (1,3S2°-1,418° F.), then 
reheated and cooled slowly Irom 700° C. 
(1,292° F.) 

Annealed at 750°-770° C. (1,382°-1,4I8° F.), then 
at 700° C. (1,292° F.), afterwards at 650° (1,202° F.), 
and finally very slow treatment at 560° 0. 
(1,040° F.) 

Annealed at 750°-770° (1,382°-1,418° F.), then at 
620° C. (1,148° F.), afterwards slow treatment at 
590° C. (1,094° F.), and finally treatment at 
450° C. (812° F.) 

Annealed at 750°-770° C. (1,3S2°-1,418° F.) 



With reference to the special heat treatment necessary in the case 
of G, H and J, the temperatures which gave the best results Lre in 
the neighbourhood of the critical ranges of the alloys on heating. 

The softening treatment given the forged bars at the National 
Physical Laboratory was based in the first instance on the experience 
gained with the cast ingot at the Hecla Wcrks. Later, when the 
critical ranges on heating and cooling, and the types of structure of 
the series had been thoroughly investigated, some further experiments 
were made. 

The four hardest alloys F, G, H and J have a martensitic structure ; 
on the one side they are bounded by alloys, viz A-D, with a pearlitic 
structure, which are quite easy to machine ; on the other side by 
alloys with a polyhedral structure, of which K is an instance, which 
are also soft, and up to a certain point easily machinable. Attempts 
were accordingly made to so heat-treat the bars that : — 

(a) either they passed into the pearlitic class by a very slow 
cooling ; 

(6) or they were arrested in the polyhedral stage by a very quick 
cooling. 



864 



ALLOYS RESEARCH. 



Nov. 1905. 



CO 



CO 
00 






03 






















3 






















f-i 
























cq 


(M 


fM -H 


1— < 


•—1 


T— ( 


(M 


^^ 


CM 


I— ( 


&< 


o 


O 


O O 


o 


o 


O 


O 


o 


O 


o 


8 


o 


o 


O O 


o 


o 


o 


o 


o 


o 


o 


^ 






' V ' 
















P^ 






















»h' 






















-a 


-M 


(M 


CO CO 


CO 


CO 


CM 


<M 


.—1 


CM 


CM 


o 


O 


o o 


o 


o 


O 


O 


o 


o 


o 


's 


o 


O 


o o 


o 


o 


o 


o 


o 


o 


o 


m 






















§ 


r> 


OC 


^ CO 


^ 


o 


o 


t^ 


t- 


00 


CO 


a 


?— ^ 


J—* 


1—* r— 1 


I— t 


— H 


I— 1 


T-H 


1 — 1 


o 


1 — ( 
























r^ 


o 


o 


o o 


o 


o 


o 


o 


o 


o 


o 


a; 






















d 






















GG 






















O 






















P 


lO 


o 


CO o 


(M 


CO 


C<J 


o 


lO 


CO 


g 


e3 

to 


o 


L- 


GO CO 


00 


o 


o 


l^ 


00 


00 


o 


O 


o o 


o 


I— 1 


o 


o 


o 


o 


o 


c3 






^-^v— ^ 
















krH 






















f^ 






















o 


t- 


CO 


'tl t— 


o 


(M 


(M 


CO 


1—1 


"O 


l-H 


XJ 


--H 


-^ 


Ttt rH 


Tl 


'tl 


<o 


^ 


■+1 


■^ 


tH 


;-< 


• 






• 


• 


• 










e3 


o 


o 


O O 


o 


o 


o 


o 


o 


o 


o 
























'^^ 


• 


o 


lO l^ 


lO 


lO 


(N 


»o 


oq 


CO 


I— c 


• r-H 


CM 


1— ( r-H 


(M 


C5 


Tfl 


Ci 


CM 


o 


C5 




^ 


I— < 


(>l CQ 


tH 


tH 


i> 


t> 


CM 


lO 


C5 
















I— 1 


1— 1 


I— ( 


§ 






.'^ 












n5 




-rJ 


-+^ 


-4-3 


-M u 


-)^ 


-*J 


•fa 


-4^ 


-•-3 


o 


•4-a 


2 




OQ bo 


OQ 


OQ 


03 


CO 


03 


to 


QQ 


1— < 


a 


c3 


•^ S 


73 


e3 


e3 


cS 


a 


a 


o 


O 


o 


^^ 


o 


CJ 


O 


O 


U 


o 


Q 


O 






















>, 






















o 












































<1 


<i 


n 


Q 


p 


H 


f^ 


O 


W 


t-5 


w 


,^ 






















^ 























a -2 



C/2 =3 



CJ 


© 






0) 


-(-> 


w 


-1-3 

e3 


<X) 


>i 








i=l 


«•-( 


o 


o 






ri 



Si • — 






o 
XI 

•"" o 

I— I f-t 

a o 

o bo 

^ a 

fl a 

• pH "^ 

«> 9 

rO ^ 

O Oh 

0) '^ 

M 

o 



M 



a 
o 



3 § 

O c3 

S ^ 

to t" 

a o 

'^ o 

c ^ 

o 



m <0 

E-i ^ 



o 

■*^ 
o 
1=1 

C3 
O 

P 

o 



tS3 



o 



a 2 

o -^ 



O 

o 
bO 



n3 
cJ 



c3 

o 
o 






O 

o 



bi o 



c3 

I— < 

P4 



Xov. 1905. ALLOYS RESEARCH. SO 5 

One experiment to realize the condition described in (a) was 
'7ery successful. After tlie treatment at 450^ C. (842^ F.) a foot 
length of alloy J was kept for fifty hours at temperatures varying 
between 150° C. (302'' F.) and 130° C. (266° F.) [which are very 
near the end of the critical range on cooling (see Table 21, 
page 921)], and then very slowly cooled to the ordinary temperature. 
After this treatment it was machined with great ease. No separation 
<j>f carbon in the form of graphite or temper carbon was thus caused. 
G and H were accorded a similar low temperature annealing, the 
former being kept for 120 hours between 160° C. (320° F.) and 
120° C. (248° F.), the latter for 100 hours at 90° C. (194° F.)- 
Y5° C. (167° F.), but the bars, although machinable, were still quite 
tough. 

No success was reached in realizing the condition described in 
(h). A 5-inch length of ingot strip H was heated to rather above 
1,000° C. (1,832° F.), and then quenched in water at 4° C. (3*92° F.). 
No difference in either the hardness or the structure of the alloys 
could be detected. This result was confirmed in a second experiment. 



The chemical compositions of the alloys are shown in Table 3. 

The Mechanical Properties of Nickel Steel. 

Introduction. — The present Report embodies an exhaustive 
examination of the mechanical properties of a series of nickel steels 
containing an average of • 44 per cent, carbon and • 88 per cent, 
manganese. Before detailing the results obtained it may be well to 
note briefly some of the work already done in this field. 

That certain of the nickel steels are of high value is a fact well 
known to all engineers. Commercially the presence of nickel in, say, 
a structural steel is regarded as tending to raise the elastic limit and 
ultimate stress without seriously impairing the ductility. As types 
of commercial steels the following Table 4 (page 866) maybe quoted 
from a recent Paper * read before this Institution : — 



* Heat-Treatment Experiments with Chrome-Vanadium Steels ; by Captain 
H. Riall Sankey and Mr. J. Kent Smith, December 1904, page 1276. 

3 N 



866 



ALLOYS RESEARCH. 



Nov. 1905. 



TABLE 4. 



No. 


Analyses. 

1 


Yield 
Point. 


Maximum 

Stress. 


Elongation. 


Ni. 


C. 

0-320 
0-280 
0-310 


Mn. 


Si. 


46 

47 
48 


2-950 
3-010 
4-175 


0-512 
0-516 
0-625 


0-052 
0-123 
0-112 


Tons per 

sq. in. 

21-7 

21-9 

33-7 


Tons per 

sq. in. 

39-3 

.39-2 

50-4 


Per cent. 

on 2 ins. 

34-0 

32-5 

21-5 



No. 46. Annealed a* 600° C. (1,112° F.) to 650° C. (1,202° F.). 
No. 47. SheflBeld nickel-steel, probably annealed. 
No. 48. German nickel-steel oil-tempered as supplied. 

Other results could be quoted ; the foregoing are, however, 
typical of commercial products and are of recent date. It will be 
noted that in two of them the content of nickel is approximately 
o per cent., the contents of carbon are fairly similar, and in each case 
manganese is present in decisive quantity. 

Although these steels are a regular commercial product of fairly 
well-defined properties, the influence of varying contents of nickel is 
still a subject offering wide opportunity for research. 

In an earlier Paper * one of the authors published an examination 
of the effect upon iron of gradually increasing amounts of nickel, the 
latter element ranging from^ 0*27 per cent, to 49*65 per cent.; 
the average content of manganese being about * 81 per cent., and 
that of carbon 0*17 per cent. The necessity for the presence of 
manganese was indicated in the Paper, and its effect evidenced on the 
steels by the fact that [they were all sound and readily forgeable. 
The " Key Table 6 " (pages 868 and 869), extracted from the Paper 
quoted, gives at a glance the properties of the various alloys. 

These results show that an increase in the content of nickel raises 
the maximum stress and lowers the extension. Thus • 27 per cent. 



* " Alloys of Iron and Nickel." Proceedings, Institution of Civil Engineers,. 
1898--99, vol. cixxviii, page 1. 



ov. 1905. 



ALLOYS RESEARCH. 



867 



nickel gives an ultimate stress of 31 tons per square inch and an 
elongation of 35 per cent.; whilst 15-4:8 per cent, nickel gives 94 
tons and 3 per cent, elongation, both alloys being in the forged 
unannealed condition. Further increments of nickel tend to lower 
the maximum stress and increase the ductility. Thus 29 per cent, 
nickel yields a maximum stress of 38 tons and an elongation of 33 
per cent. The most characteristic feature of these results is found 
in the presence of a " brittle zone," members of which possess high 
tensile strength and low ductility. Alloys on either side of this zone 
are comparatively ductile. 

Though these results are familiar, their inclusion in this Report 
is necessary for comparison with others subsequently described. 

Herr Rudeloff in 1896 * published the results of an examination 
of the properties of cast nickel-iron alloys, a summary of which 
appeared in the earlier Paper on " Alloys of Iron and Nickel." Later 
Herr Rudeloff f has examined the properties of nickel-iron-carbon 
alloys in the cast, hammered, and rolled conditions. Two of these 
alloys bear somewhat on the present work, in that whilst low in 
manganese they are otherwise fairly comparable with two of the steels 
included in the series subsequently described. The analyses are as 
follows : — 



TABLE 5. 



No. 

3 
21 


C. 


Mn. ! ^j.^^^ Si. ' P. S. 

1 \ '< 


Al. Cu. 


1 
Per cent. Per cent. Per cent. 

0-41 0-06 3-07 

0-48 1 0-03 8-13 

1 


Per cent. 
0-05 

0-04 


Per cent. 
Trace. 

Trace. 


Per cent. 
0-01 

001 


Per cent. 
0-07 

005 


Per cent. 
0-03 

0-06 

1 



It will be noted that aluminium has been used as a deoxidiser, 
and it is of interest to note that some of the steels were " red short." 
The mechanical properties of Nos. 3 and 21 in the rolled and 
hammered condition are as in Table 7 (page 870). 



* Verhandlungen des Vereins zur Beforderung des Gewerbfleisses. 
t Sechster Bericht des Sonderausschusses fiir Eisen-Nickel-Legierungen. 

3 N 2 



808 



ALLOYS RESEARCH. 



Nov. 1905. 





■^ 




^«j> 


^ 


<» 


<ai 




'Oi 


« 




tq 


^ 


• 


•"*^ 


^ 


CO 






fin 


i^i 




o 


S:> 




fO 


a 




o 


* 






'^ 


<» 


<» 


r5i 


-o 


<U 




S 


t«a 




O 


'^ 


« 




^T^ 


« 


1 


rA 






CO 


o 






PQ 


OD 


< 


^ 


H 


"-<. 
<-<. 



[3 

'Eh 
O 
-4-> 

eS 

o 
E-i 

<D 

H 


CD 
03 

1 


Reduction 

of area 

per cent. 




Elongation 
per cent. 

on 
2 inches. 


1-1— 1,— lCDlOC0C0C<IC<lrHlO'+IQ0O5 
CO Tin Ttl 0:1 CO CO (M r-1 rH TtH -^ 


Maximum 

Stress. 

Tons. 


(rqC^C^lCOCOCOrJ^lOOOCDGOt-COCO 


Yield 
point. 

Tons. 


0— 'O(Mi000O<Nif5 liOJOCOiO 
(M(M(MCM(M(MCOCO-<tt |-<*HC<li— li-< 





Reduction 

of area 

per cent. 




Elongation 
per cent. 

on 
2 inches. 


ioi:D^cooi>?oc:5(Mcoi>cocOrrt 

CO CO CO CO CO (N <>4 r-H i-iCO^ 

a 


Maximum 

Stress. 

Tons. 




COCOCOCOCO'-tlTtlGOaiOSOit-CO 


Yield 
point. 

Tons. 


rHCM<N(MC^i:MCO-<Ti;OlOMHWiM 








rt 


C5iOC<l(MlOlOQOeOCOCOCOOCD(X) 

t>i>t>t>cococooociOia:oQOO 


(^ 


OOOOOOOOOOOi-lOrH 


□0 

o 


ci 


a5Tt^COTHOQOt:--i:CQOC0050-<tlCO 
1— li— li— li— (1— li— Ir-Hi— (i-HC-li— IrHi— li-H 

000000 c; 0000000 


i^ 


(MtOOiOJOOOOOifJCOTTliXltOOO 


oooi-(CO»oc-Ci.-Hioa:-^Oi05 

l-H I-H 1-1 (M (>J •<*( 




-^ p^ q &q fc^ ;■: 5J N.; K, k ).ci ^ l^; 



Nov. 1905. 



ALLOYS RESEARCH. 



869 









O 



l-H lO 

05 I 00 
CD I t^ 



t^ I -H 






I— I lO 

00 o 

t-00 



CO 
>4 






t^ 

L^ 



o 

CO 



CO t> 

co -^ 



to 

cS 
Cm 



o 






g 
p 



> 



to 

O 
P 






o 



a 

a 



p4 

a 

o 



ID 

a 


r:S-es* 

fcc — 

r «i 

a 
a 

a 



pq 



D O <I) O ® O 

;3 P P 3 3 1=1 ^ '^'^ 

o o o o o o 
QPPQPQ 



GO 



O '—' © o fl^ 



a 



© o © © © 

3333o3oo 

i-lr-(r-|-jOaC_)CO 



3 13 S 

o o o o 
PPQO 



o 

p 






© 

O 

© 

r-H 
© 

!=l 



(=1 

© 

pq 



oooSoooo 

^OlO'^Oi— Ir-OO 



^^ 



O t> rH 



o o o o o 

I— I CO O Ttl CD 
I— t I— I CO I— • 



© 



13 
© 







o3 



^ TO OS 

.5 • a 

-^ © in 



000000^.^0 

Q0O5 — O'^rH-j-^CM 
r-H (M Oi lO (M r-( ^ 



o o o o o 

G^ (N 7^1 CO .-O 

r-i CO «:> 



13 -M 



S "S "» 






d 
d 

a 

© 

be 

o 



d ^ 02 

d fl o 

o Jh o 

^ © 1— t 

GO A 



CO r^ 

CO lO 



oooii- ii>co:o 

lO ir; Tfi •»t^ rfi CO CO 



CO CO ^ . 
ri Ttl © 

a 



-S.S d 
•d o o 






(M ^I O t^ 00 O O O 



03 



o o o o o o 
o 00 00 o tN ;?; 



-:^=qOC)Eqt,^tej H ^!^^^^ 



870 



ALLOYS RESEARCH. 



Nov. 1905. 



TABLE 7. 



No. 


Condition. 


Yield 
Point. 


Maximum 

Stress. 


Elongation. 


Modulus 

of 
Elasticity. 


Compression 

at 71 -8 tons 

per sq. in. 


3 

21 


rRolled and| 
\ hammered J 

Annealed . 


Tons per 
sq. in. 

21-9 
29-5 


Tons per 
sq. in. 

31-0 

50-5 


Per cent. 
19-9 
14-4 


Lbs. per 

sq. in. 

28,945,840 
27,807,920 


Per cent. 

30-3 

81 



No. 3 lies between C and D of the authors' series, whilst No. 21 is 
near to G. It may be remarked in passing that the tensile results 
obtained by Eudeloff from his No. 21 are, in comparison with the 
authors' G, high in ductility and low in ultimate stress. 

M. Guillet in his recently published " Les Aciers speciaux " gives 
results obtained from three series of steels, containing respectively 
0*12, 0*22, and 0*82 per cent, carbon, nickel in each series varying 
from 2 to 30 per cent. Table 8 (page 871) embodies these results, 
and it may be noted that the carbon of each series represents 
an average, whilst the content of nickel is expressed to the nearest 
whole number. The three series are however very comparable, and 
show clearly the influence of nickel on steels of varying carbon 
content and containing only traces of manganese. 

These results are very instructive. However, for the present 
attention may be confined to the following features : — 

a. Each series contains a brittle zone. 

h. In the first series, 0*12 per cent, carbon, the lowest ductility is 
found at a content of 15 per cent, nickel. 

c. In the second series, * 22 per cent, carbon, the lowest ductility 

occurs at 10 per cent, nickel. 

d. In the third series, 0-82 per cent, carbon, the lowest ductility 

is found at a content of 7 per cent, nickel. 



In concluding this introduction, reference must be made to an 
iron-manganese-nickel alloy previously described by one of the 



Nov. 1905. 



ALLOYS RESEARCH. 



871 



00 







1' 


-^ 
























too 




r> 


O 




o 


"O 


o 


o 


o 


>o 






o 


to 


o 


o 


C>0 


o 


■* 


C5 


CM 


(N 






c 


s^ 


I— ( 


r-i 












CM 


CO 






_o 
























O o 


s 






















a 
























8?^ 


3 no 

a g 


09 


o 


<M 


o 


'tl 


t> 


CS 


O 


t> 


CD 


H 


*S J:J 


o 


t- 


4i 


t> 


l> 


CM 


00 


O 


b- 


o 




C5 II 




H 


lO 


CD 


iO 


CD 


iO 


CM 


CO 


'^ 


»o 




S 








































































fH 


























c; 


























o 




72 


05 


O 


o 


"*• 


t^ 


o 


CO 


CM 


»o 






T'-' O 


o 


00 


t- 


r> 


^* 


CM 


T~> 


CO 


4i 


o 






>Kfg 


H 


(M 


CO 


o 


CD 


o 


CM 


CM 


CM 


CO 






d 


























o 


-4^ 
























• I-I 


CI 








lO 


o 


lO 


lO 










S?„ 


1-H 


o 


o 




• 






Oi 


CM 






iJD 




(M 


C^l 


CM 


CO 


TfH 


iO 


rH 


CM 


CO 




-4J 


a 


^ 






















fl 


o 


OJ 






















ti_i gj 


i—H 


Ph 






















O « 


H 






















S O) 














































i-i 


O) o, 
























1— 1 




a 

.S '^ 






















no 

W 

M 

P3 


8^ 


00 

a 

o 




o 


CO 

CO 


CO 


00 
00 


rK 
(N 


CO 
O 




CO 
lO 




^11 


H 


C<1 


CO 


CO 


CD 


I> 


I> 


00 


CO 


CO 




























ci 


























o 


2 "S 
:~" o 


no 

i=i 
O 




CI 


o 

CD 


CO 
CM 


lO 
lO 


1-H 

lO 


CM 

CD 


CM 

OO 


I-H 
1-H 






>Hfg 


H 


I— 1 


CM 


<M 


iO 


»o 


rH 


iO 


l-H 


CM 






d 




















i 






.2 


-4^3 
























Vs 












lO 


»o 


O 


o 


»o . 






?« 


CO 


lO 


i-^ 


00 
















• to 




CM 


C^l 


CM 




CO 


o 


»o 


»o 


C5 




-fj 


a 


;h 














l-H 




CM 




fl 


o 


Q 
























H 


Ph 




































































M 


i^ e. 


a 
























fl (M 






















CO 


8r 




32 


CO 


o 


CD 


■^ 


o 


CM 


o 


CD 


CD 


Ph 


*i? _fe 


o 


C<l 


ZD 


OO 


1— ' 


y—i 


CM 


CD 


i-H 


00 




cS II 


eiCC 


H 


Csl 


(M 


CM 


■^ 


CD 


CD 


t> 


CD 


CM 




























o3 


























U 


1^ -♦J 

'o .2 


oj 


00 


CO 


CM 


^ 


i-H 


CD 


lO 


O 


t^ 






:"" o 


o 


00 


05 


l-H 


CO 


Ci 


4i 


(>1 


o 


00 






>^iS 


H 


r— 1 


i-H 


CM 


CO 


CO 


lO 


lO 


CO 


l-H 


-(^ 






















1 




fl T 






















1 




?1 "!■ 

6 5i 


) 




(M 


»o 


t> 


o 


cq 


o 


o 


lO 


o 




! 










l-H 


l-H 


I— 1 


CM 


CM 


CO . 



872 



ALLOYS RESEARCH. 



Nov. 1905. 



autliors.* In Table 6 (pages 868 and 869) the liigh tensile strength 
and low ductility of alloy J" (15 • 48 per cent. Ni) will have been noted. 
A similar brittle product, but of much lower tensile strength, is 
obtained from iron alloyed with manganese to the extent of about 
7 per cent., as sho^vn in the following Table 9 : — 



TABLE 9. 
Manganese-iron alloy containing nearly 7 per cent, of manganeoe. 



Mark. 


Analysis. 


Treatment. 


Maximum 

Stress. 


Elongation 
on 8 ins. 


Reduction 
of Aiiea. 


c. 


Mn. 


4D 

i 


Per 
cent. 

0-52 


Per 

cent. 

6-95 


Forged into test 
bar and no further 
treatment. 


Tons per 
sq. in. 

25-43 


Per cent. 
1-5 


Per cent. 



A most singular feature, discovered by the author quoted, lies in- 
the fact that the simultaneous presence of manganese and nickel in 
the foregoing amounts confers extremely high ductility, and the 
results of Table 10 are of much interest. 

These results indicate some of the possibilities of manganese- 
nickel-iron alloys, an ultimate stress of 50 tons and an elongation of 
75 i^er cent, forming a distinct record. 

The series of alloys investigated in the research about to be 
described are practically alloys of iron, carbon, nickel and manganese,, 
the latter element being however comparatively low and varying 
within only small limits throughout the series. The results are^ 
therefore not comparable with the foregoing ones ; still some fairly 
exceptional properties are recorded in one instance, as reference to- 
the various tests of K will show. 



* Engineering Conference, 1903. Proceedings, Institution of Civil'. 
Engineers. Supplement to vol. cliVj'page 118. 



Nov. 1905. 



ALLOYS RESEARCH. 



873 



TABLE 10. 

Manganese-NicJcel-Iron Alloy. 



Mark. 



Analyses. 



Ni. 



Mn. 



Treatment, 



a 
*o 






I Per 
; cent. 

I 
1109 D' 14-55 



J' 


>5 


>> ' 


J> 


>> 


>) 



3417 Ft 



3418 F 



Per ! Per 

cent, i cent. I 

j 

0-60 ' 5-01 ! Bar as rolled 



JUnannealed, notl 
\ toughened J 

[ Quenched in 

I water from 

I 1,100° C. 

I. (2,012° F.) 

As forged . 

Quenched in 

water from 

1,100° C. 

(2,012° F.) 



Tons 

per 

sq. in. 

27 



18 



27 



20 



Tons 

per 

sq. in. 

54 

56 



50 



5G 



52 



-5 a 



Per 

cent. 

45 

68 

75 




57 



^^. 



:j o 



Per 

cent. 

38 
44 

47 
42 
52 



Mechanical Properties. — The tests outlined in the Annual Eeport 
of the National Physical Laboratory for 1903 were selected with a 
view to determine a wdde range of properties. Owing to the many 
test-pieces required to cover this range, duj)licate determinations 
could not be made in the majority of cases. But although individual 
tests were not repeated, the concordant results obtained from different 
methods of testing enhance the value of any one result. For instance, 
if the results obtained from bending, tensile, torsion, and elasticity tests 
are compared, coincident properties will be noted. This agreement 
carries a far higher value wdien obtained by four different methods of 
testing, and is certainly more convincing than quadruple tests by one 
method. 

In order to obtain strictly comparative conditions, all the tests 
have been conducted on "material" that has been heated to 800^ C. 



874 



ALLOYS KESEARCH. 



Nov. 1905. 



(1,472° ¥.). Eeference to the heating curves will show that a 
temperature of 800° C. safely clears the critical ranges, and this 
temperature was therefore selected as most suitable for normalising a 
series of steels of very dissimilar critical ranges. A similar " annealing 
temperature " was selected by Herr Rudeloff.* In the present 
work the above treatment is differentiated from annealing in that the 
bars were simply heated to the required temperature, the source of 
heat shut off and the muffle allowed to cool down to atmospheric 
temperature. The machined test-pieces were packed with bone ash 
in wrought-iron tubes, the ends being closed with a plug of asbestos 
and magnesia. Two rows of tubes, placed in an 18 inches by 14 
inches by 8 inches Fletcher muffle, were treated at one time, a 
thermo-couple placed in the centre of the top row indicating the 
temperature. The time taken in falling from 800° C. (1,472° F.) 
to atmospheric temperature was usually about sixteen hours. Each 
set of test-pieces was treated at one time in order to ensure identical 
conditions in the series. 



Bending Tests of Forged Steels. — The test bars employed were 
J inch in diameter by 8 inches in length. About ^ of the bar fitted 
into a y^y-inch hole drilled in a solid anvil, and the free end was 
turned over by blows from a " striking " hammer. The results are 
given in Table 11 (page 878), col. 5. 

The first four members of the series all bent parallel, and in 
doing so developed no apparent flaws. The fifth steel marks a 
pronounced change, and it will be noted that E breaks at the 
comparatively small angle of 30°, Yet the content of nickel in E is 
only • 7 per cent, greater than in D, a steel which bent parallel 
without sign of flaw. There is a difference of 0*21 per cent, in the 
content of manganese, and a very slight one of 0*02 per cent, in 
carbon ; such differences however can hardly affect the issue, and tho 
change in properties must be due to nickel. Further increments of 
nickel decrease the bending angle until a minimum is reached at a 
content of 7*95 per cent. Still further increments of nickel mark 
the gradual return of ductility as shown by angles 10°, 60°, and 

* Sechster Bericht des Sonderausschusses fiir Eisen-Nickel-Lesrierungen. 



Nov. 1905. ALLOYS RESEARCH. 875 

180° for steels containing respectively 12*22, 15*98 and 19*91 per 
cent, of nickel. The last steel, containing nearly 20 per cent, nickel, 
not only bent parallel, but was afterwards closed up, a procedure 
which developed no visible flaw. 

In the diagram, Fig. 1 (page 876), bending angles are plotted 
against the content of nickel. The resulting curve, whilst showing 
no difference between five of the steels, does however indicate features 
worthy of early recognition in this report, namely, the abrupt change 
in properties at E, the improved ductility of H and J, and the high 
ductility of K. 

Tensile Tests of the Forged Steels. — The average dimensions of the 
test-pieces employed were f inch in diameter by 2 inches parallel. 
This figure represents the largest diameter suitable for testing a 
series of bars, which of necessity had to be broken in a 10-ton 
testing machine. Actually in one case this diameter proved too 
large for the machine, and the piece had to be reduced before it could 
be broken. With this exception, mentioned hereafter, all the pieces 
w^ere of comparative size. 

In Table 11 (page 878), col. 6, " yield point " represents the first 
appreciable permanent set as observed by means of fine dividers. 
In the case of A to D the yield points noted in this way were 
coincident with the " drop of the beam." With the exception of F 
and G all the yield points were well marked, the two exceptions 
showed no yield point, or at any rate none that could be detected. 

Maximum stress (col. 7) increases fairly gradually up to D, and 
at the same time ductility does not seriously decrease. For some 
reason not yet apparent the ductility of B is less than that of C. The 
abrupt change between D and E noted in the bending is further 
emphasised in the tensile tests, and it will be noted that an increase 
in nickel of * 7 per cent, raises the maximum stress some 12 * 23 
tons per square inch, at the same time decreasing the elongation by 
18 per cent. Ductility has disappeared entirely in F, which gives 
the maximum tensile strength of the series. When first tested this 
bar, I inch diameter, was loaded to a weight equivalent to 107 tons per 
square inch without effecting fracture. The diameter of the piece 



876 



ALLOTS RESEARCH. 



Nov. 1905. 



Fig. 1. — Bending Tests. 




10 15 

JVickel per cent. 



Fig. 2.— Tensile Tests. 
Elongation, Eeduction of Area, and Twisting Angle. 



700° 



600 




WO 



^ 



200 



Nickel per cent 



Nov. 1905. 



ALLOYS RESEARCH. 



877 



Tensile Tests. 
Fig. 3. — Yield Points. 




Nickel per cent. 



Fig. 4. — Maximum Stresses and Twisting Moments. 

125 T — 1 1 , ■ — rlO.OOO 




8 000 



•« 



i 



6,000 <; 



4.000 



5 
I 



2,000 



■^ 10 15 

Nickel per cent . 



ALLOYS RESEARCH. 



Nov. 1905. 



TABLE 11 (continued to page 881), 



Properties of the Series of Alloys. Condition of Material, 
Forged and cooled from 800° C. (1,472° F.), 



Ni 

Alloy. 


Analyses 




Bending 

Test. 


Tensile Tests. 








Yield 
Point. 


Maximum 

Stress. 


Elastic 
Eatio.* 




Ni. 


C. 


Mn. 






Per 

cent. 


Per 

cent. 


Per 

cent. 


Bent to 


Tons per 
sq. inch. 


Tons per 
sq. inch. 






A 


— 


0-47 0-95 


180° U. 


21-00 


38-19 


0-55 




B 


1-20 


0-48 


0-75 


180° U. 


23-93 


40-93 


0-58 




C 


2-15 


0-47 


0-86 


180° U. 


23-67 


41-52 


0-57 




D 


4-25 


0-40 


0-82 


180° U. 


29-16 


47-86 


0-61 




E 


4-95 


0-42 


1-03 


30° B. 


33-95 


60-09 


0-56 




F 


6-42 


0-52 


0-92 


10° B. 


None detected 


110-57 


— 




G 


7-95 


0-43 


0-79 


5° B. 


None detected 


77-38 


— 




H 


12-22 


0-41 


0»85 


10° B. . 


34-56 


80-24 


0-43 




J 


15-98 


0-45 


0-83 


60° B. 


28-53 


80-24 


0-35 




K 


19-91 


0-41 


0-96 


180° U. 


15-33 


43-92 


0-35 




1 


2 


3 


4 


5 


6 


7 


7a 





In column 5, U. signifies Unbroken. B. Broken. 
* Added after Discussion. 



Nov. 1905. 



ALLOYS RESEARCH. 



870 



(^continued on next page) TABLE 11. 



Properties of the Series of Alloys. Condition of Material. 
Forged and cooled from 800^ C. (1,472^ F.). 



Tensile Tests. 



Elongationj Keduction 

on of 

1 2 inches. Area. 



Per cent. 
25-0 

21-0 

24-5 



2-0 



Nil 



Nil 



1-0 



5-5 



55-0 



Per cent. 
51-73 



51-83 



20-0 33-06 



3-71 



Nil 



Nil 



1-63 



7-33 



Torsion Tests. 



Twisting 
Moment. 



Angle of 
twist at 
fracture 



Compression 
at 100 tons 
per sq. in. 



Modulus I 

of I Ni 

Elasticity. Alloy, 



Inch -lbs. Degrees. 
4,277 



42-80 i 5,077 



5,609 



5,071 



6,429 



7,497 



7,938 



8,621 



63-11 5,662 



10 



405 



621 



Per cent. Lbs. per sq. in. 



177 



20-1 



31-5 



7,329 118-5 



690 



11 



36-98 
36-21 
37-54 



468 31-13 



42-6 5-73 



3-57 



7-52 



9-31 



29-94 



12 



32,100,000 A 

30,700,000 ! B 

30,500,000 C 

29,900,000 j D 



6-64 29,500,000 i E 



28,000,000 ! F 



27,300,000 G 



27,500,000 



27,400,000 



29,600,000 



16 



H 



K 



880 



ALLOYS RESEARCH. 



Nov. 1905. 



TABLE 11 (^continued on next page). 



Properties of the Series of Alloys. Condition of Material. 
Forged and cooled from 800° C. (1,472° F.). 



Ni 
Alloy. 


Shock Tests. 


Hardness Tests. 




Fall of 
46-7 lbs. 
Hammer. 


Energy 
absorbed. 


Bending 
Angle. 


Indentations in yJfjjj 
inch. 


Brinell 

Ball Test. 

Hardness 

No. 

(Normalised 

specimens.) 




Load in 
tons 1*5. 


Load in 
tons 2*5. 






Inches. 


Inch 
Pounds. 


Degrees. 










^ 


13-23 


451 


18-0 


7-2 


15-0 


2U2 




B 


13-05 


428 


17-0 


6-4 


14-5 


207 




C 


13-67 


454 


16-5 


7-0 


19-5 


212 




D 


13-92 


460 


15-5 


6-0 


12-3 


217 




E 


13-67 


217 


Broken 


4-2 


8-7 


321 




F 


13-67 


105 


Broken 


2-5 


5-5 


532 




G 


14 15 


230 


Broken 


2-5 


5-7 


578 




H 


1 14-17 


436 


7-5 


3-2 


6-2 


555 




J 


13-33 


432 


14-5 


50 


10-3 


293 




K 


13-77 


452 


28-0 


16 

1 


40 


131 




, 


14 


i 1^ 


16 


17 


18 


19 





Nov. 190;"). 



ALLOYS RESEARCH. 



881 



(concluded from page 878) TABLE 11. 



Properties of the Series of Alloys. Condition of Material. 
Forged and cooled from 800° C. (1,472° J^.). 



Specific 
Gravities 

at 

17-18° C. 

(62-6- 

64-4° F.) 



7-880 
7-890 
7-884 
7-867 
7-876 
7-885 
7-883 
7-904 
8-026 



Specific 

Volume Dilatations 
at 17-1 8° C. Coefficients 

(62-6- X 10' 

64-4° F.). 



20 



0-1269 



0-1267 



0*1268 



21 



11-22 



11-12 24-8 



0-1271 11-36 



0-1270 12-07 



8-122 0-1231 19-65 



22 



Eesis- 
tivities 
Microhm 
Cm. at 
17° C. 
(62-6° 
F.). 



Corrosion Tests. 
Percentage Loss 



After 

32 days 

in 

water. 



20-3 



22-6 



29-1 



39-3 



0-1268 12-23 j 42-8 

0-1269 12-13 1 43-9 

0-1265 . 13-28 50-5 

0-1246 17-54 63-3 



75-4 



After 
33 days 
in 
sea- 
water. 



0-15 
0-11 
0-10 
0-11 
0-12 
0-12 
0-12 
0-10 
0-09 
0-09 



After 

17 days in 

50 percent. 

Sulphuric 

acid. 



23 24 



0-23 
0-26 
0-26 
0-24 
0-23 
0-23 
0-23 
0-18 



25 



0-80 
1-10 
0-72 
0-71 
0-66 
0-58 
0-54 
0-44 
0-28 



Ni 
Alloy. 



0-14 0-28 



26 



B 
C 
D 
E 
F 
G 
H 



K 



27 



3 o 



882 ALLOYS RESEARCH. Nov. 190h. 

was then reduced, and on re-testing fracture occurred at 110 "57 tons 
per square inch. G, 7*95 per cent, nickel, marks a distinct fall in 
tenacity, and shows no tendency towards regaining ductility. This 
steel in the bending tests also gave the lowest bending angle of the 
series. H, 12*22 per cent, nickel, showed a distinct yield point at 
34J tons, fractured at a higher load than G, and marks the return of 
measurable ductility. J, 15 • 89 per cent, nickel, whilst giving the same 
tensile strength as H, shows a better elongation (Table 11, page 879, 
col. 8). K, 19 • 91 per cent, nickel, is characterised by the lowest yield 
point, the highest elongation and reduction of area of the series 
(Table 11, col. 9). This steel, remarkable in many respects, derives 
its chief interest here from the fact that an ultimate stress of 43 * 9 
tons per square inch is associated with an elongation of 55 per cent, 
on 2 inches. 

The foregoing tensile results are plotted on three separate curves, 
see Figs. 2, 3, and 4 (pages 876 and 877). Reference to Fig. 2 shows 
at a glance the influence of an ascending content of nickel on the 
ductility of the steels. The full and dotted curves represent 
respectively reduction of area and elongation. The general similarity 
of these two curves with the one plotted from the bending tests will 
be noted, though the greater sensitiveness of the tensile test 
differentiates between the first four steels which gave equal results 
on bending. 

The yield points noted in tension are plotted on Fig. 3 (page 877). 
In order to complete this series of curves the authors had hoped to 
include the yield points noted in torsion and compression, a hope 
which unfortunately could not be realised. 

The maximum stress curve plotted on Fig. 4 (page 877) specially 
emphasises steel F, and at the same time shows the influence of an 
ascending content of nickel. Finally both tensile and bending tests 
show that the series of steels may be roughly divided into two groups, 
one including ductile and the other brittle steels. Up to i\ per cent, 
nickel, ductility is fairly high; 4*95 per cent, nickel marks an 
accession of brittleness, whilst 15-98 per cent, nickel denotes the 
eturn of ductility. This return is complete at 19*91 per cent, 
nickel. The properties of intermediate steels would naturally be of 



Nov. lf)05. 



ALLOYS RESEARCH. 



883 



much interest, especially as the range in passing from the ductile to 
the brittle zone is represented by 7 per cent, nickel. However, the 
narrow extent of this range indicates that the change from ductile to 
brittle is, as already described, " abrupt." 



Tensile Tests of the Forged Steels at the temperature of Liquid 
Air. — In continuation of the research work of one of the authors 
regarding " The effect on mechanical and other properties of Iron 
and its Alloys produced by Liquid Air Temperatures," * similar 
experiments were also made with the same series of steels dealt with 
in the present research. The test-bars were of standard type, as 
used in the research above mentioned. The following Table 12 is 
a list of the tests : — 

TABLE 12. 







Tests at -182° 


C. (-295°F.). 


Ni 
Alloy. 


Percentage 
Nickel. 






Tensile Tons 


Elongation 


! ^ 




per sq. in. 


per cent. 


1-20 


75-4 


7-5 


C 


2-15 


95-1 


12-7 


D 


4-25 


75-4 


10-0 


E 


4-95 


88-0 


nil. 


F 


6-42 


142-2 


2-5 


G 


7-95 


91-1 


0-5 


: H 


12-22 


87-2 


— 


J 


15-98 


144-2 


2-5 


K 


19-91 


157-2 


15-5 

1 



All the above bars were magnetic before and after testing with 
the exception of K, which was only very slightly magnetic before 
testing in liquid air. After the test it was strongly magnetic. 

* Journal, Iron and Steel Institute, 1905, 1, page 147. 

3 O 2 



884 ALLOYS RESEARCH. Nov. 1905. 

Maximum Stress. — Comparing the figures with those obtained at 
ordinary temperatures there is in every case an increase of tenacity. 
Alloy H shows the least increase, namely, from 80*2 to 87*2 tons per 
square inch. With Alloy K, however, the value 43 • 9 tons obtained 
at the ordinary temperature rises at — 182° C. ( — 295° F.) to the very 
high figure of 157-2 tons maximum stress. The value obtained for 
C. at —182° C. seems somewhat high. The sharp rise of stress 
between E and F, and the almost as rapid fall between F and G, are 
equally well marked in both series. 

Elongation. — The alloys maintain the same general character 
at —182° C. ( — 295° F.) as at the ordinary temperature. Up to 
4 per cent, nickel they are appreciably ductile, although to a 
diminished extent. Between 5 and 16 per cent, nickel they undergo 
fracture with almost no elongation. Finally, at 20 per cent, nickel 
there is a reversion to ductility to the extent of 15*5 per cent, 
elongation. 

Thus, considering the two series of tests as a whole, the general 
character of the alloys is unaltered by cooling them to — 182° C. 
( — 295° F.), although the tenacity is greater, and the ductility less, 
than at ordinary temperatures. The most marked difference is 
found with K, where the tenacity is 3*58 times greater, and the 
elongation 3 • 55 times less than at ordinary temperatures. 

Torsion Tests. — The test-pieces employed were 0*625 inch in 
diameter by 2 inches parallel. The results obtained include the 
twisting moment at fracture, expressed in inch-pounds, and the amount 
of torsion expressed in degrees. Eeliable records of elastic strains 
were not obtained, but from curves subsequently plotted a series of 
approximate yield points was obtained. These curves were plotted 
from observations of twisting angle taken at each increment of • 2 inch- 
ton, and from them the figures in Table 13 (page 885) were obtained. 

Definite results obtained from the torsion tests are given in Table 1 1 
(page 879). cols, 10 and 11. 

The twisting angle of A cannot be accurately given owing to the 
fact that an autographic recorder attached to the specimen failed to 



Nov. 1905. ALLOYS RESEARCH. 885 

give a record. However as estimated by the flow lines on the broken 
piece, the actual twist is greatly in excess of that undergone by B, and 
probably exceeds that of C. Subsequent tests from A onwards were 
made by actual observation of the twisting angle at each increment 
of 445 inch-pounds. 

The characteristic features of the tensile tests are reproduced in 
the torsion results, namely, an increase of rigidity is associated with 
a rising content of nickel until a maximum is attained ; from thence 

TABLE 13. 



Ni Alloy. 


Apparent Yield Point. 


A 
B 

C 
D 
E 
F, G, H 
J 
K 


No record. 
2,205 inch-pounds. 
2,646 „ 
2,205 „ 
3,969 „ 

None detected. 
2,764 incli-pounds. 
1,32.3 „ 



rigidity falls and ductility increases. The maximum in this case 
occurs with H, and not with F as in the tensile tests. The resulting 
curve is plotted on Fig. 4 (page 877), for comparison with that of 
maximum stress, and it will be noted that this curve is of a more 
regular character than the companion one of tensile strength. 
Although the two maxima occur at different contents of nickel the 
general character of the two curves is one of fair agreement. This 
agreement is rendered more evident by a comparison of the twisting 
angle with the elongation and reduction of area yielded by the 
tensile tests. Reference to Fig. 2 (page 876) will show this 
comparison, and it will be noted that the ductility of the series in 
tension or torsion is practically similar. 

Torsion results of K are again worthy of a special note, namely, 
a twisting moment of 5,662 inch-pounds associated with a twisting 
angle of 090°. 

Compression Tests. — The compression cylinders were • 35 inch in 
diameter by 0*56 inch high, these dimensions being selected in order 



886 



ALLOYS RESEARCH. 



Nov. 1905. 



to attain a maximum load equal to 100 tons per square inch. See 
Table 11 (page 879), col. 12. The compressibility of the series 
decreases with an increase of nickel until a minimum is reached at G, 

7 • 95 per cent, nickel. The compression curve, Fig. 5 (page 887), has 
a strong resemblance to preceding curves representing different 
aspects of ductility. It will, however, be noted that, in the return of 
the curve, the right-hand branch does not attain its initial height. 

Modulus of Elasticity. — Elastic measurements were taken by means 
of Professor Ewing's extensometer from bars • 5 inch in diameter by 

8 inches parallel. Duplicate determinations gave a series of very 
concordant results, which, on the whole, form one of the most valuable 
sets of the series. 

The Modulus results are given in Table 11 (page 879), col. 13, and 
are plotted on Fig. 6 (page 887). Attention is directed to the fact 
that the base line of the curve does not correspond to zero. The curve 
bears a marked resemblance to those in Figs. 1 and 2 (page 876), 
except in respect of A and B. These exceptions impart a more 
regular contour to the curve. 

It will be noticed from these results that the modulus of elasticity 
falls at first as the percentage of nickel is increased. Now 
corresponding to a small modulus a great elastic compression is 
produced by a given force. Thus at first the elastic compression or 
extension due to the action of pressure or tension within the elastic 
limit rises as the nickel content is increased. The rise is at its 
maximum at about 8 per cent, nickel. From 8 to 16 per cent, nickel 
the modulus is approximately constant. But it has been seen that 
the elongation at break decreases very markedly with the rise of 
nickel until a content of about 6 J per cent, is reached. From 6 J to 
12 per cent, it is practically nil. 

Thus it follows that the specimens of very small ductility yield 
most to pressure or tension within the elastic limit. 



Shock Tests. — The following description of the Impact Testing 
Machine used for Shock Tests, together with Fig. 7 (page 888), has 
been furnished the authors by Dr. Stanton and Mr. Jakeman. 



Nov. 1905. 



ALLOYS RESEARCH. 



887 



Fig. 5. 
Compression Tests. 




Zol 



Nichel pe/r cent^ 



888 



ALLOYS RESEARCH. 



Nov. 1905. 



" The machine consists of a cast-iron anvil A and a tup 
T, each supported by four pieces of steel strip J inch wide 
by ^Q inch thick and about 12 feet long. The anvil A has two 
heavy bosses on the sides, through which pass two pieces of round 



Fig. 7. — Impact Testing Machine, at National Physical Laboratory. 




J /■;„>/ 



steel bar, BB. These can be adjusted to protrude any distance towards 
the middle of the anvil, and arc locked by means of the set-screws 
CC. The ends of these bars are cut away to hold the knife-edges K, 
against which rests the specimen S, kept at the right height by 



Nov. 1905. ALLOYS RESEARCH. 889 

adjustable supports E. The tup T is provided with a steel knife- 
edge E, adjustable outwards so as to just touch the specimen when 
the tup and the anvil are at rest. 

" From the back of both tup and anvil a string is carried over a 
pulley near the roof with a small weight attached (just sufficient to 
keep the string taut). The rise of these weights is a measure of the 
height through which the tup or anvil is raised. The actual heights 
of the tup and anvil corresponding to the observed motions of the 
small weights on the strings were obtained by separate experiment. 

" The anvil weighs about 60 lbs. and the tup about 47 lbs. The 
specimens used in these tests were 6 inches long by f inch square, 
and were notched on the tension side with a small V groove. The 
knife-edges were placed 4 J inches apart. 

" The method of test was as follows : — The specimen was placed 
in the anvil and the tup tied back at the desired height by a piece of 
thin string. The tup was then released by severing the string with 
a sharp knife, and an observer noted the height to which the anvil 
was forced, while a second observer noted the height to which the 
tup swung after the blow. 

" The work given as that required to deform or break the 
specimen is the difference between the kinetic energy of the system 
before and after the blow, calculated from the heights to which the 
masses were raised." 

Table 11 (page 880), cols. 14, 15 and 16, give the results of the 
shock tests, fi*om which it will be seen that only three of the steels 
fractured, the remaining ones bending to a greater or less extent. 
The weight of striking hammer was 46 * 7 lbs. ; distance apart of knife- 
edges, 4*6 inches; dimensions of specimens, 5 inches X 0*375 inch 
X • 375 inch ; depth of notch (V groove), • 04 inch. The fact that 
the whole of the steels have not broken probably imparts to the results 
a greater value than if fracture had occurred in each case, for by 
selective action the dangerously brittle steels are at once picked out. 
Generally speaking the differences between the first four steels,* 
nickel from to 4*25 per cent., are not great ; they were all bent and 



* It must be remembered that these four steels did not break. 



890 



ALLOYS RESEARCH. 



Nov. 1905. 



shock developed no visible flaw in any one. When however the 
nickel content is raised from 4-25 to 4*95 per cent, then decisive 
brittleness under shock is shown ; this increment of only • 7 per cent, 
results in the steel fracturing with comparative ease. The minimum 
is found with F, a result which coincides with that of the same steel 
under tension. The behaviour of H is noteworthy, a steel which in 
the preliminary bending tests fractured on reaching an angle of 10°. 

Fig. 8. — Impact Tests. 




Nichcl per cenl-. 



Under shock this steel absorbs 436 inch-pounds, and bends to an 
angle of 7J° without developing any apparent flaw. The expenditure 
of an equal amount of work in the case of A and K shows the higher 
ductility of the latter even when that work is applied as shock. 

By taking the bending angle in conjunction with the energy 
absorbed, a very fair idea of the properties of the steels under 
impact is obtained. These two sets of observations are plotted on 



Nov. 1905. ALLOYS RESEARCH. 891 

Fig. 8 (page 890), where it will be noted that the series ol steels 
follow generally the order observed in the earlier tests. 

Hardness. — Considerable difficulty was encountered in machining 
some of the steels. In fact E to J were almost beyond the limit of 
ordinary tools, and the various test-pieces of these steels were more 
conveniently ground to size. Types of hardness are shown in the 
foregoing tests ; it was however desired to ascertain some measure of 
the resistance offered by the series to the penetration of a tool. 

In the first place two series of scratch tests were made with (1) a 
hardened steel point such as is used for marking scales, and (2) a 
diamond, both sets of scratches, representing standard conditions, 
being made under a uniform load in a dividing engine. On 
magnifying the resulting scratches, in order to measure their width, 
the following difficulties were met with : — 

1. The edges of all the scratches were torn and ragged. 

2. The form of the scratches was an approach to a V. Hence 
a combination of rough edge and sharply receding width of scratch 
rendered accurate measurements of width difficult if not impossible. 

The second method followed was based on a relative penetration 
of a hardened steel point. Taking Swedish iron as unity the 
following figures were obtained : — 

TABLE 14. 



Material. 


Relative hardness. 


Swedish Iron 


10 


Ni Alloy A 


1-6 


B 


1-8 


C 


1-6 


D 


1-5 


G 


2-2 


H 


2-2 


J 


20 


K 


1-2 



These values, determined before E and F were added to the 
series, are relative only and represent material that has not been 
heated to 800° C. (1,472° F.). 



892 



ALLOYS RESEARCH. 



Nov. 1905. 



Finally, a series of heat-treated specimens, f inch diameter by 
about f inch high, and polished on one face, was tested for hardness 
by means of Unwin's indentation test. This method is familiar ; it 
may however be well to recall the fact that the indenting tool is a 
straight knife-edge formed by a square bar of hardened steel ; and 



Tons 
5-5 



Fig. 9. — Indentation Tests for Hardness. 




20 

lOOO 



that the indentation is produced by a steady load. The indentations 
were measured in one-thousandths of an inch, and the penetration at 
two loads is shown in Table 11 (page 880), cols. 17 and 18. 

In each case a load of • 5 ton represents the zero reading. With 
C and K loading could not be carried further without danger of 
deformation. 

The results (including load in tons of 3*5, 4*5 and 5*5), 
plotted on Fig. 9, yield a series of curves which show at a glance 



Nov. 1905. 



ALLOYS RESEARCH. 



893 



the relative resistance offered to the penetration of a hardened 
knife-edge. Plotted in another form, as on Fig. 10, the effect of 
an increasing content of nickel on the relative hardness of the series 
is shown. The two curves, plotted from readings 1 * 5 and 2 • 5 tons 
respectively, are again similar to preceding curves in which one or 
more aspects of ductility figure. The high position of K on the return 
part is of much interest. This steel, though the softest of the series, 
is difficult to machine owing to the fact that work has an appreciable 

Fig. 10. — Indentation Tests for Hardness, arranged in another form. 

40i 




' 20Z 



Nickel per cent 



hardening effect. Whilst the steel cuts freely at first, as the work 
progresses the hardening effect is shown by the diminished cutting 
power and blunting of the tool. Thus in sawing small sections of 
the steel by means of a power-driven hack-saw, when a notch of about 
\ inch in depth is obtained, the saw absolutely refuses to go further, 
and if the action is continued the teeth of the blade are stripped. 
The authors fully recognise that an apparent contradiction between 
working properties and hardness determinations is here shown. 



894 ALLOYS RESEARCH. Nov. 1905. 

They however venture to think that an explanation will be found in 
the section dealing with the influence of work on the structure 
(page 953), in which it is shown that work creates a fresh structural 
material which is hard. 

Hardness Tes/.— Brinell Ball test carried out at the Hecla Works, 
Table 11 (page 880), col. 19. For purposes of comparison it may be 
stated that Swedish charcoal iron has a hardness number of 122 as 
forged, and 90 as normalised or annealed. Forged nickel (99*8 per 
cent, nickel) has a hardness number of 106. 

The results in Table 11 place the alloys in the same order as 
those obtained by the knife-edge method. The first four alloys give 
nearly the same values. At 5 per cent, nickel a marked increase in 
hardness is seen. The hardness reaches a maximum value at about 
8 per cent, nickel, the values over the range 6*5 to 12 per cent, 
being however very similar. At 16 per cent, a very marked drop 
has occurred, and at 20 per cent, the softest alloy of the series is 
seen. 

General Summary of the Mechanical Tests. — The results of the 
mechanical tests are summarised in Table 11 (pages 878-880), cols. 
6 to 19. Dealing with this series of tests of the different mechanical 
properties, it will be seen that they all give the same kind of result. 
With the introduction and increase of nickel content up to 4 per 
cent., the change in the properties is gradual ; the material when under 
elastic stress yields more to the stress ; at the same time after the 
apparent yield-point is passed the maximum stress increases. 
Although the change of properties is gradual there is in nearly every 
case a more or less pronounced kick in the curves between and 4 
per cent, nickel. At some point between the percentages of 4*25 
and 4*95 nickel there is a very sudden change in nearly all the 
properties, evidenced by a rap ^'^^ increase of maximum stress, 
which reaches the highest value at 6-42 per cent, nickel, a fall of 
ductility, and an increase of brittleness as shown by the bending, 
tension, torsion, and shock tests. Thus, as far as industrial products 
are concerned, a danger limit for nickel content is found at 4:^ per cent., 



Nov. 1905. ALLOYS RESEARCH. 895 

when carbon and manganese are present to tJie extent 0/ • 44 per cent, 
and • 88 per cent, respectively. 

After this sudden break in the curve the various properties alter 
more slowly again, until a percentage somewhere in the neighbourhood 
of 16 is reached, that is, the brittle zone extends from about 5 to 16 
per cent. From this point on the change is more rapid and in the 
reverse direction to the original rapid change. 

Fig. 1 (page 876) is thus roughly typical of any of the curves. 

The nickel-iron alloys quoted on Table 6 (pages 868 and 869) 
show a sharp decrease in ductility at 9*51 per cent, nickel, a slight 
improvement is manifested at 11*39 per cent., whilst from 15*48 
per cent, to 19*64 per cent, the product is distinctly brittle. At a 
content of 24*61 per cent, nickel an improved ductility is shown 
which is rendered more complete at 29 per cent. The alloys 
included on Table 4 (page 866) are directly comparable with the 
authors' present series, in that they contain similar contents of 
manganese and were produced under similar conditions. Therefore 
the two series show that the influence of carbon, when present to the 
extent of 0*44 per cent., is to lower the beginning of the brittle zone 
to between 4*25 per cent, and 4*95 per cent, nickel. Not oiAj is the 
beginning lowered by the presence of carbon, but also the end of the 
brittle zone, as evidenced in the various tests of K. This steel may 
be compared with alloys M and N on Table 6 (pages 868 and 869), 
and such comparison will show that ductility is regained at a much 
lower content of nickel, when the alloys contain carbon. 

M. Guillet's results quoted in the introduction have shown the 
lowest ductility in the • 82 per cent, carbon series at a content of 
nickel in the vicinity of 7 per cent. The authors' various tests of a 
* 44 per cent, carbon series containing manganese give the lowest 
ductility over a range of from 5 to 8 per cent, nickel. Further, Guillet's 
brittle zone, 0*82 per cent, carbon series, extends from 7 to 15 per 
cent, nickel ; the 5 per cent, steel giving an eloDgation of 10 per cent., 
whilst the 20 per cent, nickel steel gives an elongation of only 9J 
per cent. It will be remembered that K, 0*41 per cent, carbon and 
19-91 nickel, of the authors' series gave an elongation of 55 per cent. 



896 



ALLOYS RESEARCH. 



Nov. 1905. 



Alternating Stress Tests. — The construction of the Alternating 
Stress Machine in the Engineering Department * (a brief description 
of which is given at the end of this paragraph, also an illustration 
on Plate 59) was completed while the authors' experiments were in 
progress, and a series of pieces of the heat-treated forged bars was 
put through the test. The diameter of the test-piece was 0*251 
inch. The results are given in the following Table 15 : — 



TABLE 15. 



Ni 
Alloy. 

1 


No. of 

Reversals 

for fracture. 

2 


Reversals 
per minute. 1 

8 


Stress in Tons per square inch. 


Tension. 
4 


Compression. 
5 


Range. 
6 


A 

B 
C 
D 

E 
F 
G 
H 
J 

K 


69,000 
177,000 

99,000 

111,000 

Unbroken 

after 
1,070,000 

102,000 

266,000 

70,000 

238,000 

Unbroken 

after 
1,295,000 


805 
810 
807 
845 

851 
825 
855 
842 
851 

851 


18-2 
18-5 
18-3 
20-1 

20-3 
191 

20-5 
19-9 
20-3 

20-3 


130 
13-2 
13-0 
14-4 

14-5 
13-7 
14-7 
14-3 
14-5 

14-5 


31-2 
31-7 
31-3 
34-5 

34-8 
32-8 
35-2 
34-2 
34-8 

34-8 



* See also " Engineering," 17th February 1905, page 201. 



Nov. 1905. 



ALLOYS RESEARCH. 



897 



Fresh specimens of E and K were turned to 0*200 inch 
diameter and gave the following results on testing : — 



12 3 

1 


4 5 


6 


E 45,480 
K 14,600 


758 
816 


25 1 
29-2 


18-0 
20-8 


48-1 

50-0 



Beyond directing attention to the high Reversals Nos. for Alloys 
E and K, in comparison with the remaining eight alloys, the authors 
prefer to give the Table 15 with no comment other than that, 
except in the case of K, the stress was in each case less than the 
apparent yield point. Conclusions with reference to the behaviour 
of the alloys under a test of this kind can only be drawn from a 
large number of experiments, and it has not been possible to attempt 
this. Accordingly these results have been intentionally omitted from 
the preceding section, and no attempt has been made to correlate 
them with the results of the other mechanical tests. 



* Alternating- Stress Testing Machine. — " The principle of this 
machine is that of employing a rotating crank to cause periodic 
motion of a reciprocating mass by means of a connecting-rod, the 
specimen under test forming the connection between the reciprocating 
mass and the crosshead. This device has been employed by Professor 
Osborne Reynolds in the testing machine at Owens College, which 
is of the vertical type with a single balanced crank. In the National 
Physical Laboratory machine there are four cranks operating four 
specimens, the motion of the specimens being in a horizontal plane. 
By this means the balancing of the machine is made independent of 
the ratio of the crank-arm to the connecting-rod, so that a length of 
crank-arm has been adopted which enables experiments to be made 
at moderately low speeds, that is, from 600 to 1,000 revolutions 
per minute. Although this arrangement causes the motion of the 
specimens to deviate from the simple harmonic law, the effects on the 



* This description has been furnished by Dr. T. E. Stanton. 

3 p 



898 



ALLOYS RESEARCH. 



Nov. 1905. 



stresses set up in the specimens are suflBciently small in value to be 
neglected, so that the maximum tensile force on the specimens may 
be taken as 

W r 

— (oV (1 + i) pounds, 

and the maximum compressive force on the specimens as 



W 



T 

V (1 — -j) pounds; 



where W = weight of mass attached to end of specimen in pounds ; 

r = radius of crank-pin ; 

o) = mean angular velocity of crank-shaft ; 

I = length of the connecting-rod. 

" It will be observed that the maximum tensile stress is 1 • 4 

times the value of the maximum compressive stress, which is 

approximately the ratio of the stresses in the piston-rod of an 

ordinary reciprocating steam-engine. The form of specimen adopted 

TABLE 16. 



Ni 
Alloy. 


Condition. 


Yield Point. 


Maximum 

Stress. 


Elongation 
on 1-75 in. 


Reduction of 
Area. 


B 
B 


As cast. 

/Ca8theated\ 
\ to 800° 0. / 


Tons per 
sq. in. 

27-47 
22-07 


Tons per 
sq. in. 

49-54 
43-23 


per cent. 
6-28 

14-28 


per cent. 

5-86 

14-88 



is the same as in Reynolds and Smith's experiments, consisting of a 
5-inch bar screw cut 2T~i^ch Whitworth, and turned down in the 
centre to a diameter of ^ inch for a length of half-an-inch. Great 
care has to be taken in the preparation of the specimens to ensure a 
gradual change of section in the turned-down part, as the effect of a 
change of section on the resistance of the specimen is much more 
marked in the case of tests under alternating stresses than in 
statical tests. For the tests described in the present Report, as the 



Nov. 1905. 



ALLOYS RESEARCH. 



899 



material was too hard to be screw-cut, special holders wore made to 
which the specimens were attached by means of set pins." 

A photograph of the machine is given on Fig. 71, Plate 59. 



Mechanical Properties of the Cast Material. — As an example of the 
influence of heating the cast material to 800° C. (1,472° F.) the 
tensile results obtained from B may be quoted. {See Table 16.) 

The following Table 17 gives the tensile results obtained from 
the cast steels after cooling from 800° C. 

TABLE 17. 



Ni 
Alloy. 


Content of 


Yield 
Point. 


Maximum 

Stress. 


Elongation 
on 

1-75 in. 


Reduction 
of Area. 


Ni 
per cent. 


C 
per cent. 


Mn 
per cent. 


A 
B 
C 
D 
E 
F 
G 
H 
J 
K 


1-20 

2-15 

4-25 

4-95 

6-42 

7-95 

12-22 

15-98 

19-91 


0-47 
0-48 
0-47 
0-40 
0-42 
0-52 
0-43 
0-41 
0-45 
0-41 


0-95 
0-79 
0-86 
0-82 
1-03 
0-92 
0-79 
0-85 
0-83 
0-96 


Tons per 
sq. in. 

19-24 
22-07 
25-79 
28-11 
36-00 
37-13 
41-47 
39-85 
31-59 
17-94 


Tons per 
sq. in. 

38-49 
43-23 
42-84 
43-49 
56-09 
57-51 
74-03 
71-19 
76-55 
29-69 


per cent. 
15-42 
14-28 
17-70 
13-10 
14-28 

6-20 

4-50 

6-20 

4-00 
14-30 


per cent. 
16-26 
14-88 
23-51 
17-94 
24-43 

8-42 

5-68 

6-88 

4-21 
19-05 



These results follow generally the same order as those of the 
forged material, maximum tensile strength, however, being represented 
at a content of 7*95 per cent, nickel. Singularly enough this 
maximum is associated with an elongation of 4J per cent., and the 
three steels, which in the forged condition are distinctly brittle, show 

3 p 2 



900 ALLOYS RESEARCH. Nov. 1905. 

in the cast normalised state elongations of 6-2 per cent., 4*5 per 
cent, and 6*2 per cent. These results are remarkable and unusual, 
but some analogy will be found on reference to Table 6 (pages 868 
and 869). Thus alloy H; 9*51 per cent, nickel, in the forged annealed 
condition fractures at an angle of 5° ; the same steel in the cast annealed 
state bends to an angle of 8°. Somewhat similar results were found in 
the case of J, and it will be noted that, in bending, this steel in the 
unforged condition reaches an angle of 9°, whilst in the forged state 
fracture occurs at an angle of 3°. The compression results of I 
are also of interest, the cast unannealed alloy shortening some 
5 per cent., whilst the forged unannealed shortens only 1 per cent, 
at 100 tons per square inch. 

Physical, Chemical and Metallographical Properties. 

Besistivities. — The Mass Resistivities were determined by the 
Thomson double-bridge method. Measurements were made on both 
the 1^-inch and the J-inch bars, except in the case of Alloys A and 
H, where only the J-inch bars were used. Very closely agreeing 
values were obtained in the two series. In the case of the 1^-inch 
bars the length tested was 60 centimetres ; in that of the J-inch bars 
100 centimetres, except in C where a 50-centimetre length was tested. 
From the mass and length measurements of the bar under test, the 
mass resistivity was calculated. The quotient of this by the density 
gives the resistivity. 

The results are given in Table 11 (page 881), col. 23, in the form 
of Microhm Centimetre Resistivity at 17° C. (62*6° F.), (the 
temperature at which the tests were made). To enable a comparison 
to be made with copper, the value of the standard soft copper of the 
Electrical Standards Committee is 1*702 microhm centimetre. The 
results are also plotted in Fig. 11 (page 901), the co-ordinates being 
resistivity and percentage nickel. The resistivity rises with ascending 
nickel, the values for alloys A-D, H, J, and K lying practically on 
a smooth curve. Alloys E, F, and G give values above the smooth 
curves. These values have been doubly checked. The sudden 
change at E is coincident with a structural change in the series, and 



I 



Nov. 1905. 



ALLOYS RESEARCH. 



901 



constitutes additional evidence of the sharp break in the continuity 
of the properties of the series which is manifested in the great 
majority of tests made. 

Increase of resistivity with rise of nickel was found by Barrett in 
his measurements of five alloys, D,E,I, K, L, in the low carbon 
nickel iron series.* There was nothing examined between 3 • 82 and 

Fig. 1 1 . — Besistivities. 




Nickel per cent. 



11*39 per cent, nickel, and the curve between these percentages has 
been drawn smoothly. It is precisely in this region that the 
irregularities in the authors' series have occurred. The authors' 
first set contained no nickel between 4*25 and 7*95 per cent, and 
the latter alloy was the only one that lay off the smooth curve. 
When alloys E and F were inserted, it was found that the slope was 
more irregular than had been previously indicated. This is a 
striking instance of the caution that is necessary in interpreting 



* Journal, Institution of Electrical Engineers, vol. 31, Table 1, pages 680-1 
also Fig. 2. 



902 ALLOYS RESEARCH. Nov. 190 . 

from curves the variations of properties of a series of alloys containing 
comparatively few members. The resistivities of alloys H to K 
come out high, but they do not approach the figure 97 * 5 found by 
Barrett with an alloy containing 25 per cent, nickel, 5 per cent, 
manganese, and 1 • 18 per cent, carbon. The temperature coefficients 
of the alloys were not determined. 

Magnetic Tests.* — " A series of magnetic tests was made on rods 
of the alloys by means of a Ewing permeability bridge. The rods 
were all cylindrical and of about 20 centimetres length and 0*713 
centimetres diameter. It was found when using the permeability 
bridge that it was important that the bar under test should be very 
similar in magnetic quality to the standard bar against which the 
comparison was made. Thus the ideal standard in each case would 
be a rod whose permeability curve would approximate closely at all 
points to the curve for the rod under test. As the samples varied 
very widely in magnetic quality, it would not have been easy to 
construct a set of standard rods agreeing all along their curves with 
the rods tested. The bridge, however, may be used with sufficient 
accuracy with rods whose curves may be widely different, provided 
the comparison is made only at one point, namely, the point of 
intersection of the two curves. Thus, if a number of different 
standard bars are available, the points at which their curves are 
crossed by that of the bar under test can be determined and thus a 
more or less complete curve obtained. 

In order to obtain a series of standard rods whose magnetic 
qualities should be sufficiently spread over the wide range required, 
several expedients were adopted. The standards were all made in 
pairs, great care being taken to get the rods in each pair as uniform 
as possible ; they were tested for equality in the Ewing bridge, and 
if sufficiently near had their approximate permeability curve 
determined by means of Ewing's small yoke method. 

The materials for the standard bars comprised mild steel, tool 
steel, grey and white cast-iron, and one of Heusler's copper- 
manganese-aluminium alloys, the three last-named alloys being 

* The Report on the Magnetic Tests has been submitted by Mr. Campbell. 



iJov. 1905. ALLOVS RESEARCH. 903 

prepared in the metallurgical department. In addition to standard 
bars, tubes of mild steel were also used and proved efficient ; they 
were all of the normal outside diameter, but of various inside 
dianaeters, thus giving a variety of ap2)arcut permeability curves. 

Tests were also made by the ordinary ballistic method on a 
series of small rings of all the alloys except H ; the outside diameters 
of these rings were of the order of 1 centimetre. The rings were all 
admitted to preliminary heat treatment as similar as possible to that 
to which the bars had been subjected. 

For alloy B, the curve obtained by the ring method agreed at the 
lower values with the curve for the rod compared against cast-iron, 
and at the higher values with that obtained against wrought-iron. 
This agreement made it possible to take B as a standard, and 
accordingly A, C, and D were tested against it, their curves being 
not far apart. The remaining alloys were tested against the cast-iron 
and other standards already described. The agreements between 
curves obtained against different standards made it permissible to 
consider that the fiual curves given in Fig. 12 (page 904) represent 
the actual properties of the rods to a sufficient accuracy." 

Table 18 (page 905) gives for each rod a statement of the 
standards against which it was tested. Some comparisons were also 
made between the rods themselves ; these are included in the Table. 

The results of the permeability tests show how intimately these 
are bound up with the structures of the alloys. In Fig. 12 the 
structural characteristics are tabulated alongside of the curves. The 
four pearlitic steels show the highest permeabilities. The martensitic 
steels give very much lower values. K, the polyhedral steel, is the 
lowest of the series. 

Specific Gravities. — These were determined on small cylinders of 
the forged bars. The latter were weighed in air and afterwards in 
water, being suspended in the latter case by " hair " platinum wire. 
During the determinations the temperature varied between 17° and 
18° C. (62-6° F. and 64-4° F.). The results are given in Table 11 
(page 881), cols. 20 and 21, in which the specific volumes have been 
also included. 



904 



ALLOYS RESEARCH. 



Nov. 1905, 



Fig. 12. — Permeabilities. 



20^00 




Character of 
Striictur& 



Pearliiic, 



Miirtensitlc 



Martensitic 
and^ Polyhedral 



Polyhedral 



Magnetising Force 
Fig. 13. — S'pecific Gravities and Specific Volumes. 




Nickel per cent. 



Nov. 1905. 



ALLOYS RESEARCH. 



905 



TABLE 18. 





Ni Alloy. 


Standards used. 




A 


B. 




B 


Soft Iron and Tool Steel. 




C 


B. 




D 


B. 




E 


Mild Steel Tube and Cast Iron. 




F 


Mild Steel Tube and G. 




G 


Cast Iron and Heusler's Alloy. 




H 


Heusler's Alloy. 




J 


Tool Steel and Mild Steel Tube. 




K 


Mild Steel Tube and G. 



The values are also plotted in Fig. 13 (page 904), specific 
gravities and volumes being on the vertical ordinate, nickel 
percentages on the horizontal ordinate. The specific gravities of the 
first eight alloys do not differ markedly from one another. From 
12*22 per cent, nickel upwards the specific gravity rises sharply. 
Thus the pearlitic and martensitic steels of the series have similar 
densities, and the polyhedral steels are characterized by considerably 
higher values. 



Dilatations. — The dilatations of bars B to K were determined in 
the comparator of the Metrological Department. Bars of J inch 
diameter and about 40 inches long were used. The average range 
of temperature was from 1° to 29° C. (33-8° F. to 84-2° F.). The 
results are contained in Table 11 (page 881), col. 22. 

When these values are plotted against nickel percentages (Fig. 14, 
page 907), it is seen that they do not lie on a smooth curve. 
Neglecting the slight drop between B and D, there is (a) a sharp 
rise at E (4 • 95 per cent, nickel) ; the curve returning to the 



906 



ALLOYS RESEARCH. 



Nov. 1905. 



original slope at G. There is also (6) a distinct rise at J (16 per cent, 
nickel), followed by a return at K. These results may be correlated 
with structural changes. E is on the threshold of the change from 
pearlitic to martensitic structure, while in J the polyhedral type has 
made its appearance. (The slight rise at B may be due to the 
introduction of nickel.) 

An instructive instance is thus furnished of the care that is 
necessary in drawing conclusions, from the behaviour of a small 
number of alloys spread over a wide field, as to the properties of 
alloys of intermediate composition. In the authors' original series, 
with no alloys between D and G (a gap of 3 • 7 per cent, nickel), the 
results up to 12 per cent, nickel would have given a practically 
smooth curve. There was nothing to suggest that in the range from 
4*25 to 7*95 per cent, nickel, the values would lie considerably off 
the smooth curve. A similar case is instanced by the resistivities 
(Fig. 11, page 901). 

Increase of nickel causes an increment in the coeflSicient of 
expansion in a somewhat irregular manner as already noticed. 
There is nothing approaching the very small coefficients exhibited 
by high nickel steels of the " Invar " type, but these, it must be 
remembered, contain only minute amounts of carbon, manganese, 
silicon, &c. In the authors' series the high carbon and manganese 
are quite sufficient to account for the absence of small coefficients.* 

Corrosion Tests. — These were carried out on pieces of the J -inch 
diameter forged bars in two ways. 

(a) Immersion in well-aerated fresh water ) , , . 

,, ^ ^ . . ^ , , . . T ? at ordmary temperatures. 

(o) Immersion in 50 per cent, sulphuric acid^ 

The tests under (a) were carried out to get some idea of the 
extent of fresh water corrosion under " natural " conditions ; those 
under (h) to enable a comparison to be made with the behaviour of 
low carbon-nickel-iron alloys.f 



* Guillaume. " Nouvelles Recherches sur les Aciers au Nickel." Comptes 
Rendus, 2nd February, 1903. 

t Proceedings, Institution of Civil Engineers, 1898-99, vol. cxxxviii, page 43. 



Nov. 1905. 



ALLOYS RESEARCH. 



907 



2()i 



Fig. 14. 
Dilatations. 




10 15 

Nickel per cent . 



10 






s 




^0-5 

^ A 



c 

-^^ 



C D 



Fig. 15. 
Corrosion Tests. 







Nickel per cent . 



908 ALLOYS RESEARCH. Nov. 1905. 

Fresh- Water Corrosion Tests. — The specimens weighed about 
70-80 grammes each. They were suspended by string in separate 
glass pots filled with water. The latter was renovated each day, and 
the film of brown hydrate gently brushed off. The pots stood 
beside an open window, night and day. 

Duration of Test. 32 days. 

At the completion of the test, the brown scale was removed as far 
as possible by careful rubbing with fine emery paper. The bars were 
heated for IJ hour at 110° C. to 120° C. (230° F. to 248° F.) allowed 
to cool in a desiccator, and weighed. The losses in weight ranged 
from • 07 to • 1 of a gramme. Table 11 (page 881), col. 24, gives the 
percentage losses. The latter, which range from 0*15 to 0*09 per 
cent., are plotted against nickel content in Fig. 15 (page 907). The 
differences found are very slight, but they tend to show that from 
12 per cent, nickel upwards the tendency to corrode under these 
conditions diminishes. 

Sea-Water Corrosion Tests. — These were carried out under 
conditions similar to those described for the fresh-water corrosion 
tests. The duration of the test was thirty-three days, and the sea- 
water was changed once a week. The losses in weight varied from 
0*11 to • 22 of a gramme. The percentage losses are given in Table 
11 (page 881), col. 25, and are plotted against nickel content in Fig. 
15 (page 907). They are about twice as great as in the case of the 
fresh-water tests, and they agree with the latter in indicating that up 
to 12 per cent, nickel no marked difference of behaviour is exhibited 
by the alloys, and that with higher percentages of nickel the 
tendency to corrode is less marked. 

Acid Corrosion Tests. — The bars used in the previous experiments 
were afterwards immersed, each in a separate trough, and supported at 
two points on thin glass rods, to enable the liquid to circulate round 
the bar, in acid (at the ordinary temperatures) made by mixing equal 
volumes of distilled water and concentrated sulphuric acid. More 
than sufficient acid was present to dissolve the whole of each bar. A 
vigorous action soon set in with the early members of the series, the 



Nov. 1905. ALLOYS RESEARCH. 909 

gas evolution being less marked among the later members. As time 
went on, bars A-H gradually became coated with a deposit of white 
salt. (This was probably a mixture of anhydrous sulphates insoluble 
in sulphuric acid.) After a fortnight the powders were broken off, 
and the liquids well stirred. After seventeen days the reaction had 
practically ceased among the early members. The bars were then 
taken out, washed, dried at 120° C. (248° F.), cooled in a desiccator 
and weighed. The losses in weight ranged from about 0*21 to 0*88 
grammes. The percentage losses are given in Table 11 (page 881), 
col. 26. They are also plotted against ascending nickel in Fig. 15 
(page 907). 

Except for a rather high corrosion figure for B, the values 
decrease fairly uniformly with ascending nickel. The irregularity 
at B will have been noticed in various mechanical and physical 
tests. 

The alloys with pearlitic structure are attacked most readily. 
„ polyhedral „ „ „ least „ 

„ martensitic „ occupy an intermediate position. 

[The same order was found in etching the alloys with picric or 
nitric acids for microscopic examination.] The differences in the 
structural characters of the alloys render these differences of 
behaviour towards attacks by acids quite intelligible. In the solution 
in acids of the pearlitic alloys with duplex structure, differences of 
potential are probably set up, which doubtless facilitate the attack ; 
in the polyhedral alloys, on the other hand, where the structure 
appears to be homogeneous, such differences of potential cannot 
occur, and solution will be more difficult. 

The figures given for the percentage losses of low carbon nickel- 
iron manganese alloys by one of the authors run in the same direction, 
but the absolute values are considerably higher. Thus the percentage 
loss of the alloy with 7 • 62 per cent, nickel was 2 • 77 * as against 
0-54 for the alloy with 7*95 per cent. (G), but the former has a 
pearlitic, the latter a martensitic structure. 

* Proceedings, Institution of Civil Engineers, 1898-99, vol. cxxxviii, page 43. 



910 ALLOYS RESEARCH. Nov. 1905. 

Summary of Physical and Chemical Tests. — These properties are 
summarised in Table 11 (page 881) columns 20-26. Except the 
corrosion tests, which are not very sensitive, the sharp change in 
properties at some point between 4*25 and 4-95 per cent, nickel 
is well brought out, especially by the Eesistivity and Dilatation 
Tests. 

The concluding sections of this research, dealing with the critical 
ranges and structures of the alloys, will show how intimately the 
properties of the various members are associated with definite 
positions of the critical ranges and with definite types of structure. 

Up to 4J^ per cent, nickel the positions of the critical ranges on 
cooling and the types of structure are similar to those of iron carbon 
steels cooled normally. 

From 5 to somewhere between 12 and 16 per cent, nickel the 
critical ranges on cooling are both lower and very much wider, and 
the structures are similar to those of quenched iron carbon steels 
(more or less markedly martensitic). 

Between 16 and 20 per cent, nickel a range is entered in which 
the critical change that takes place on cooling appears to be due to 
nickel, and the structure is polyhedral. The alloy is non-magnetic 
at the ordinary temperature. 



Determination of the Bange of Solidification.* — The method used 
has been that described by one of the authors in a series of similar 
determinations in connection with iron carbon alloys.f The melting 
of about 3 lbs. of each alloy occupied about one hour, at the end of 
which the temperature of the molten alloy was about 1,500° C. 
(2,732° F.). Cooling was followed by taking simultaneous 
observations of time and temperature. By plotting these as 
co-ordinates the curves on Fig. 16 (page 911) were obtained, the dots 
representing observed points. 

* The present state of knowledge as to the range of solidification of these 
alloys is known to be incomplete, and investigations are already in hand to 
throw more light upon this subject (January 1906). 

t Journal, Iron and Steel Institute, 1904, 1, pages 227-284. 



Nov. 1905. 



ALLOYS RESEARCH. 



911 



Trniperaturr Fahrenheit 




P3 

d 
•i-i 

a 



d 



d 

&: 
-M 

<X) 

.£> 

O 

o 

d 

03 



4) 



Tc/ri/yern/ure Cen/igrad^ 



912 



ALLOYS RESEARCH. 



Nov. 1905. 



Beginning of Solidification. — The sudden change of slope which 
marks this point may be caused in one of three ways : — 

(a) There is an actual recalesence. In this case the temperature 
of incipient solidification has been passed, and there is a rise of 
temperature to the point. The curves of A, B, C, E, G, and of pure 
nickel, show this. 

(b) The temperature becomes steady for a definite interval of 
time. Alloys F and H showed this. 

(c) The rate of cooling is very much retarded. This occurred 
with alloys D, J, and K. 



Fig. 17. — Temperature Ikinges of Solidification. 



■I 

c3 



14-00 



5 

^ 1300 

1^ 



1200- 



F. 
175Z° 



Pure Ni. 



J552 



„"^ 



10 15 

Nichei per cent. 



20 



'S 






. 12192 

100? 



End of Solidification. — This temperature cannot be given with the 
same precision. It has been estimated in the following way. 

Before solidification sets in the time-temperature curve is 
practically a straight line. The beginning of solidification causes a 
marked change of slope, as stated in the preceding paragraph. As 
solidification proceeds, the curve, which is concave upwards, gradually 
approaches the same slope that it had before solidification began. 
(It never quite reaches this, for the furnace as a whole is cooling 
more slowly at the lower temperature.) Finally a temperature is 
reached at which the rate of cooling of the alloy reaches a maximum ; 



Nov. 1906. ALLOYS RESEARCH. 913 

it remains constant for an interval of about two to three minutes, and 
then begins to diminish, and the slope of the curve becomes convex 
upwards. The temperature at which the steady rate begins, 
indicating that the temperature of the furnace and the alloy are 
about the same, has been taken as representing the end of 
solidification. 

The ranges of solidification are given in Table 19 (page 914). 
Further, the ranges have been plotted against nickel content in Fig. 
17 (page 912). Besides the ten members proper of this series, three 
other alloys have been included. Nos. 2 and 9 * at the beginning, and 
pure nickel at the end. 

Alloy No. 2 is an iron containing 0*02 per cent, carbon, traces 
of manganese, and no nickel. It begins to solidify at 1,504° C. 
(2,739° F.) and finishes at about 1,470° C. (2,678° F.) ; the range of 
solidification is thus about 34° C. (61° F.). The introduction of 
• 45 per cent, carbon lowers the temperature of incipient solidification 
some 22° C. (39° F.) and widens the range to about 66° C. (119° F.). 
This latter is not an experimental figure, which is wanting for this 
alloy, but has been estimated by taking the mean of alloys with • 38 
and 0*63 per cent, carbon whose ranges are 63° C. (113° F.) and 
69° C. (124° F.) respectively. Where 0*95 per cent, manganese is 
added, the beginning of solidification is lowered some 62° C. (112° F.) 
and the range is widened to about 180° C. (324° F.). It will be 
noticed that the upper limit of this range, namely, 1,420° C. (2,588° F.) 
is almost identical with that of pure nickel, determined by one of the 
authors as 1,427° C. (2,600° F.) f The introduction and successive 
additions of nickel up to 20 per cent, lower neither the upper nor the 
lower limits of the range to any great extent ; the curves have the same 
character throughout. 

For the upper limit the extreme values are : — 

1,451° C. (2,644° F.) in the case of D, and 1,383° C. (2,522° F.) 
in the case of K. The former is 32° C. (57° F.) above, the latter 
87° C. (66° F.) below, 1,420° C. (2,588° F.), the value found for 

* Journal, Iron and Steel Institute, 1904, 1, Table II, pages 232-233. 
t This value has been exactly confirmed by Dr. Harker of the National 
Physical I^aboratory, using four diflferent thermo-junctions. 

3 Q 






914 



ALLOYS RESEARCH. 



Nov. 1905. 



as 
1—1 

pq 
< 









CV. 


























p 


. o 


o 


O 


o 


c 


o o 





o 


o 


o 


o 


O 


o 




tt^ o 


Pm5 


05 


(M 


CO 


o 


O 'CD 


CM 


r-H 


oo 




CO 


CD 


—< 




O Ti 




(M 


1—4 


Ol 


CO CO 


't" 


00 


00 


CD 


Oi 


Ci 


lO 




Range i 
Solidifical 




1-H 


CO 


CO 


CO 


CM CM 


CO 


CM 


(M 


CM 


CM 


CM 


I— t 


I- 


II 


II 


11 


II 


II 


il II 


II 


II 


II 


II 


II 


II 


II 




o 


^~^ 


o 


o 


o 


o c 


o 


O 


o 


O 


o 


O 


o 




,-< "*• 


c>-. 


c:) 


rJH 


00 


CM r-< 


o 


•-D 


o 


o 


■^ 


'^ 


•rh 




O CO 


o 


i> 


t-- 


t^ 


CM CO 


a 


i-O 


CD 


"<*• 


CO 


CO 


00 






o 


•-H 


'■^ 


r^ 


T^ — ^ 


1—^ 


""" 


1-H 


1—1 


rH 


r— ( 










CO 






























/'^ 




























o 


CV. 


o 


o 


o 


o o 


o 


o 


o 


o 


O 


O 


o 






. 00 


^^^ 


CD 


CD 


'CD 


05 O) 


M 


l^ 


l^ 


t> 


l> 


CD 


Oi 




d 


^^. 


o 


«o 


CD 


'&< 


CM O 


Oi 


o 


o 


o 


CO 


CM 


■* 




o 


o 


CM 


<N 


C^ 


■* •* 


I-H 


CO 


CM 


CO 


CM 


(M 


"^ 




«« '^ 




CO 


•- 


•^ 


•^ 


•> •-, 


.% 


•^ 


^ 


^ 




^ 








rN 


lO 


iM 


<M 


c^> 


CM CM 


CM 


CM 


CM 


CM 


cm" 


CM 


CM^ 


CO 


H:2 


II 


(>) 


II 


II 


II 


II II 


II 


II 


II 


II 


II 


II 


II 




o 

o 


II 


O 

r— 1 


O 


o 

o 


o o 

CM O 


O 

o 


o 


o 

CM 


o 


o 

CM 


O 

Oi 


o 
eo 




1—^ 


d^^ 


o 


^ 


•^ 


CO 


CO CM 


o 


CD 


■<tl 


CD 


'^l 


1—1 


■*( 






CD 
I— 1 


G<l 


(M 


(M 


CO CO 


CM 


CM 


CM 


CM 


CM 


CM 


CO 






T-H 




1— ( 


t— 1 


1—1 


l-H 1-H 


1—1 


r-( 


1-H 


1— ( 


1— 1 


1— ( 


1— H 






o 


O 


O 


O 


o 


O O 


o 


O 


O 


O 


O 


o 


o 




f. . • 


. <^ 


Ci 


GO 


a 


CD 


Oi '# 


t)H 


00 


lO 


00 


CO 


CM 


o 




^ fl 


fef2^ 


as 


CO 


i> 


CD 


^ -^ 


CO 


CO 


lO 


CD 


CD 


CM 


o 




^ o 


to 


Ui 


40 


lO 


CD CD 


»o 


o 


lO 


lO 


o 


O 


CO 




fcfrj 






.N 


•^ 


*^ 




















9 =3 


c<r 


oT 


(M 


<M 


(M 


CM CM 


<M 


cm" 


cm" 


c<r 


cm" 


cm" 


(m" 


U9 




II 


II 


II 


II 


II 


II II 


II 


II 


II 


II 


II 


II 


II 




o 


o 


o 


O 


O 


o o 


o 


o 


o 


o 


o 


o 


o 




■^ 


-M 


o 


<o 


00 


rfl i-H 


o 


o 


CM 


Ci 


CD 


CO 


t^ 




«l 


1—1 


CO 


(M 


1-H 


o 


lO irj 


Oi 


CM 


o 


o 


o 


00 


CM 




T— I 


1— 1 




1— 1 


1-H l-H 


CO 
1—1 


1— ( 


T-H 


1— ( 


tH 


CO 

1—1 


1— ( 






C? 
































OQ 


































O 




lO 


Oi 


CO 


CM 


CO 


(M 


Oi 


o 


CO 


CO 




Ti< 




be 


Oi 


t^ 


CO 


00 


o 


Oi 


t> 


CO 


00 


Oi 


1 






fl 


H 


H 


o 


o 


o 


o 


I-H 


o 


o 


o 


o 


o 








03 
































S 






























OQ 
































































£3 
O 

-s 

OS 


CN 


t- 


t^ 


X 


tH 


o 


CM 


CM 


eo 


1— 1 


»o 


1—1 




eo 


o 
o 


o 


o 


o 


o 




o 


o 

o 


o 


o 


o 


o 


1 




Si 


Q 






























O; 
































Px 


































.2 








o 


•o 


lo 


lo 


CM 


lO 


(M 


CO 




o 






'^ 


173 


•^ 


<M 




(M 


Oi 


Ttl 


Oi 


CM 


Oi 


Oi 


CO 


e« 




J2; 


;2; 


:zj 


1— 1 


G<l 


4h 


tH 


CD 


t- 


<N 


1/5 


Oi 


Oi ■ 






^ 




















1-H 


1— 1 


1—1 


Oi 




1 
























































'o 






























M 




X 


























o 


1^ 


N 


* 


05 


< 


A* 


o 


P 


H 


fe 


G5 


w 


•-5 


^ 


2 

s 



eo 

CO 

CM 

CM 

CO 
CM 

09 

(D 

to 

03 






CD 
Oi 



CO 

d 

03 

d 
o 



^1 

d 
o 



Nov. 1905. 



ALLOYS RESEARCH. 



915 



alloy A, the " blank " of the series. Alloy D, as being exceptionally 
high, was repeated on a fresh sample. The two closely agreeing 
values are given in Table 19 (page 914). 

The variations of carbon and manganese are certainly partly 
responsible for the irregularities found. The former range from 
0-40 to 0-52 per cent., the latter from 0*79 to 1-03 per cent. If 
the lowering of the upper limit is proportional to the quantity, 

0*1 per cent, carbon will cause a drop of about 5° to 6 
(9^ to 10-8° F.). 

TABLE 20. 



C. 



Ni 
Alloy. 


Carbon Change in previously 
melted alloy by direct cooling. 


Carbon Change in unmelted 
alloy by differential cooling. 




C. F. 


C. F. 


A 


692° = 1,277° 


665° = 1,229° 


B 


658° = 1,216° 


649° = 1,200° 





618° = 1,144° 


616° = 1,141° 


D 


.557° = 1,034° 


558° = 1,036° 
(mean of two determinations.) 


E 


500° = 932° 


513° = 955° 



• 1 per cent, manganese will cause a drop of about 6° to 7° C. 
(10-8° to 12-6° F.). 

From the carbon and manganese contents alone, alloy E might 
be expected to begin to solidify about 15° C. (27° F.) lower than D. 
But the beginning of solidification is not related simply either to — 

(a) the carbon content 

(b) the manganese content 

(c) the sum of these. 



Change of Composition on Melting, — A thin film of slag is left on 
the solidified alloy. The following evidence, however, tends to show 
that the changes of composition of the alloys are only slight. 

3 Q 2 



916 ALLOYS RESEARCH. Nov. 1905. 

The time and temperature observations of alloys A-E were 
continued below the temperature of the change from hardening to 
pearlite carbon. In Table 20 the steady temperature maintained by 
this change in the alloy cooling from the molten state is compared 
with the steady or quicken ing-up temperature observed in the 
differential method of detecting this change in the unmelted alloy, 
(cf. Critical Eanges. Table 21, pages 920 and 921.) 

There is perfect agreement in the two cases for alloys C and D. 
The differences for B and E are 9^ C. (16° F.) and 13° C. (23° F.) 
respectively, but these are not greater than are sometimes found in 
duplicate determinations by the differential method. A, however, 
shows a difference of 27° C. (48° F.). As it seemed probable that 
in this case the discrepancy was to be accounted for by the passage 
of manganese into the slag, a chemical analysis was made for 
manganese. The value 0-77 per cent., as compared with the original 
0*95 per cent., was obtained. 

In the case of nickel the composition of the metal was 99-6 per 
cent, before, 99*3 per cent, after, melting. 



The Critical Ranges of the Alloys. 

These have been determined on rising temperatures (heating 
curves) and falling temperatures (cooling curves). A differential 
method of taking heating and cooling curves (the latter introduced 
by the late Sir William Roberts- Austen in the Fifth Alloys Research 
Report) has been used. A detailed description of the method has 
been given by one of the authors,* so that it is unnecessary to give a 
full account here. The authors would only say that the photographic 
method of recording the curves has not been used by them, but an 
open scale method, in which observations are taken and the curves 
plotted from them. Much greater sensitivity in reading the 
temperatures is thus attained.f 



♦ Journal, Iron and Steel Institute, 1904, I, pages 235-237. 

t The Committee are of opinion that further research is necessary, and a 
comparison of the " differential " with the " inverse-rate " method ;of taking- 
heating and cooling curves will be made. (January 1906.) 



Nov. 100'). ALLOYS RESEARCH. 917 

Critical Ranges on Cooling.— Cooling curves of the cast material 
were taken, cylinders ;!-incli long and -j-inch diameter being machined 
from the ingot. Cooling curves of four alloys in the forged condition 
were also taken. The results were so similar to those obtained with 
the cast alloys that it was deemed unnecessary to experiment with 
the remaining six forged alloys. The temperature from which the 
cooling cui'ves were taken was 900° C. (1,652° F.), except in the 
cases of Nos. 1 and 9, where it was 1,000° C. (1,832° F.). This 
temperature is well above the critical ranges on heating of any of 
the alloys. The alloys with 6 • 4 per cent, nickel and upwards have 
critical ranges which are considerably lower than those of pure iron 
carbon steels, and it was necessary to take cooling curves down to 
the ordinary temperatures. A special tube furnace was constructed 
for these in which the winding of the wire carrying the heating 
current was different from that described in the Journal already 
referred to (rather less compensation for the cooling effect of the 
ends being resorted to). Only a thin layer of quartz for insulating 
purposes was used, and the outside tube was made of brass to promote 
a more even distribution of heat in the furnace than a fire-clay tube 
would give rise to. This type of furnace renders it possible to take 
a cooling curve from 900° C. (1,652° F.) to 50° C. (122° F.), in 
about 2 J hours. 

The cooling curves of the cast alloys have been plotted on Fig. 18 
(pages 918 and 919). The vertical oidinate represents absolute 
temperatures of the alloy during cooling. The horizontal ordinate 
represents differences of temperature between the alloy and a cylinder 
of platinum cooling under the same conditions, as registered by 
movements of the differential galvanometer. The deflections have 
been reduced in plotting. The distance between two vertical lines 
represents about 9J° C. The dots represent observed points, and 
each curve has been obtained by drawing a line through these points. 
A summary of the temperatures over which the critical ranges 
extend is given in Table 21, col. 5 (page 921). Alloy No. 9* has been 
included in the curves and the Table, as it contains the same carbon 



* Journal, Iron and Steel Institute, li)04, I, pages 232-233 



918 



ALLOYS RESEARCH. 



Nov. 1905. 



Temperature fahrenheLt 



o 

(M 

lO 
CO 



o 

o 
o 






a 



"T^^ 



^ 

3 
? 







Tentperalure Centi^rad^ 



Nov. 1905. 



ALLOYS RESEARCH. 



9iy 



TerrvptrcLticre Fahrenheit 




Tt'tnpcrai 14 re 



Centi^ra de 



920 



ALLOYS RESEARCH. 



Nov. 1905. 



TABLE 21 — (continued ow op]^osite page). 



1 


2 


8 


4 






Percentages. 






Alloys. 










Nickel. 


Carbon. 


Manganese. 




1 * 


Nil. 


0-01 


Trace. 




9* 


Nil. 


0-47 


Trace. 




A 


Nil. 


0-47 


0-95 




B 


1-20 


0-48 


0-79 




C 


215 


0-44 


0-83 




D 


4-25 


0-40 


0-82 


' 


E 


4-95 


0-42 


1-03 




F 


6-42 


0-52 


0-92 




G 


7-95 


0-43 


0-79 




H 


12-22 


0-41 


0-85 




J 


15-98 


0-45 


0-83 




K 


19-91 


0-41 


0-96 




Nickel. 


99-6 


— 


— 





* Journal of the Iron and Steel Institute, 1904, No. 1, Table II, pages 232-233. 



Nov. 190.5. 



ALLOYS RESEARCH. 



921 



TABLE 21 — (concluded from opposite page). 



Critical Kanges on Cooling. 



900^-754° C. (1,652°-1,389° F.) 

770°-688° C. (1,418°-1,270° F.) 
(Temperature steady at 690° C.) 1 ,274° F. 

706°-658° C. (1,:]03°-1,216° F.) 
(Temperature steady at 665° C.) 1 ,229° F. 

687°-646° C. (1,268°-1,195° F.) 
(Temperature steady at 649° C.) 1,200° F. 

661°-604° C. (1,222°-1,119° F.) 
(Temperature steady at 616° C.) 1,141° F. 

646°-544° C. (1,195°-1,011° F.) 
(quickening at 561° C.) 1 ,042° F. 

644°-422° (1,191°-791° F.) 
(quickening at 513° C.) 955° F.) 

644°-125° C. (l,191°-257° F.) 
(quickening at 171° C. (340° F.) 

500°-123° C. (932°-253° F.) 
(quickening at 156° C.) 313° F. 

513°-78° C. (955°-172° F.) 
(quickening at 100° C.) 212° F. 

623°-149° C. (l,153°-300° F.) 
(quickening at 246° C.) 475° F. 

600°-189° C. (l,112°-372° F.) 
(with no marked quickening) 

630°-280° (1 , 166°-536° F.) 
(quickening at 350° C.) 662° F. 



Critical Ranges on Heating. 



729°-755° C. (1,344°-1,891° F.) 
700°-721° C. (1,292°-1,330° F.) 
664°-721° C. (1,227°-1.330° F.) 
634°-705° C. (1,173°-1,301° F.) 
634°-693° C. (1,173°-1,279° F.) 
622°-684° C. (1,151°-1,263° F.) 
586°-G74° C. (1,087°-1,245° F.) 
586°-660° C. (1,087°-1,220° F.) 

586°-616° C. (1,087°-1,141° F.) 

Nothing found between 400° & 
800° C. (752° & 1,472° F.) 

342° C to [?] (647° F. to [?J) 



i^22 ALLOYS RESEARCH. Nov. 1905. 

percentage as the alloys of this series and renders it possible to see 
the influence of 0*8 to 0*9 per cent, of manganese on the position 
of the critical ranges. A comparison of alloy No. 1 (same Table) 
with alloy No. 9 shows the effect of • 47 per cent, of carbon on the 
position of the critical ranges. 

Two cooling curves of each alloy were taken, the positions of 
the cylinders being altered between the two determinations. In some 
cases three and even four curves have been taken. A movement of 
the curve to the right. Fig. 18 (pages 918 and 919), indicates an 
evolution of heat in the alloys. Accordingly the cylinders were 
placed in such a position in the furnace that the slope of the curve 
during cooling was from right to left. [For reasons given on page 237 
of the Pa2)er referred to, the cooling curve is not a vertical straight 
line, even in the absence of critical changes.] 

The curves on Fig. 18 will now be discussed : — 

The critical ranges of No. 1 begin at 900° C. (1,652° F.). The 
change at this temperature is rapid, and at about 892° C. (1,637° F.). 
the curve is beginning to return to its normal slope. At about 770° C. 
(1,418° F.), another change sets in, and at 754° C. (1,389° F.), the 
curve again begins to return to the normal slope. The lower limits 
of the two critical ranges cannot be given, for they are doubtless a 
fraction of the rate at which cooling is taking place. The lower limit 
given in Table 21 (pages 920 and 921) is the temperature at which the 
curve is furthest from its normal slope. The curves have been taken 
under similar conditions, and the results are therefore comparative. 
Further, alloy G was tested in two furnaces, cooling at different rates, 
when the lower limit found was the same in both cases. 

The introduction of 0*46 per cent, carbon, alloy No. 9 (the 
average carbon percentage is 0*44 in the series A-K), lowers the 
appearance of the first critical range from 900° to 770° C. (1,652° 
to 1,418° F.), namely, about 130° C. (234° F.), and quite alters the 
character of the curve. At this temperature a gradual change 
begins, extending to about 720° C. (1,328° F.), when the curve 
begins to come back ; at 690° C. (1,274° F.), a sudden change sets in 
and at 688° C. (1,270° F.), the lower limit is reached. This latter 
is the change from hardening carbon to pearlite carbon. The 



Nov. 1905. ALLOYS RESEARCH. 923 

introduction of 0*95 per cent, manganese (alloy A) (the average 
manganese percentage is 0*88 in the series A— K) lowers the 
temperature of the first critical range still further, namely, from 
770° to 706^ C. (1,418° to 1,303° F.), this is, 64° C. (115° F.). The 
character of the curve is similar to that of No. 9 except that the 
second change follows more closely on the first ; but the two 
changes are quite distinct.* 

With the introduction of nickel the critical ranges are lowered, 
the lowering being fairly uniform up to about 4 per cent, nickel, 
(alloys B, C, and D). Further, up to this point the curves maintain 
the same general character. 

The next alloy E (nickel percentage 4*95) represents a critical 
percentage of nickel in this series. Although it contains only * 7 
per cent, nickel more than D, the character of the curve is altered 
and the lowering of the critical ranges considerable. There is a 
small thermal change beginning at 644° C. (1,191° F.), the curve 
resuming its normal slope at 560° C. (1,040° F.). A second change 
sets in at about 549° C. (1,020° F.), and instead of being a sudden 
one (as in the case of the preceding five alloys), it extends over about 
127° C. (228° F.). 

Alloy F (nickel percentage 6*42) shows a further and very 
considerable lowering and alteration of the character of the curves. 
A very slow change sets in at 644° C. (1,191° F.), and continues 
apparently without intermission to 125° C. (257° F.), there being a 
more marked evolution of heat at 171° C. (340° F.). The critical 
ranges of this alloy thus extend over about 520° C. (936° F.). 

Alloys G and H show similar curves to alloy F. Attention is 
directed to the fact that in alloy H, the lower limit is 78° C. (172° F.). 

Alloy J gives a somewhat similar curve (critical ranges 623° 
to 149° C), 1,153° to 300° F., with a slight quickening in the heat 
evolution at 246° C. (475° F.). 

Alloy K gives a curve in which a slow change between 600° and 
189° C. (l,112°-372° F.) is evident, but in this case there is no 

* A cooling curve taken from 1,250° C. (2,282° F.) does not displace the 
position of the critical ranges appreciably, although the upper limit is not so 
well marked (see Fig. 18, pages 918 and 919). 



924 



ALLOYS RESEARCH. 



Nov. 1905. 



quickening of the heat evolution at any point. The polyhedral 
structure of this alloy, which microscopic examination had previously 
revealed, and the fact that it is non-magnetic at the ordinary 
temperature, led the authors to suppose that the slope of the curve was 
due to a change in the nickel (19-91 per cent.) not associated with a 
change of physical structure. The correctness of this supposition 
was established by the results of cooling curves that have been taken 



TABLE 22. 

Critical Ranges on Cooling. 



Ni Alloy. 


Cast. 


B 


687^-646° C. (1,268°-1,195^ F.) (steady at 649° C.) 1,200° F. 


C 


661°-604° C. (1,222°-1,119° F.) (steady at 61G° C.) 1,141° F. 


G 


500°-123° C. (932°-253° F.) (quickening at 156° C.) 813° F. 


K 

B 


600°-189° C. (l,112°-372° F.) (no quickening). 


Forged. 


687°-642° C. (1,268°-1,187° F.) (steady at 649° C.) 1,200° F. 


C 


679°-(;08° C. (l,2o4°-l,126° F.) (steady at 619° C.) 1,146° F. 


G 


500°-128° C. (932°-262° F.) (quickening at 160° C.) 320° F. 


K 


600°-200° C. (l,n2°-392° F.) (no quickening). 



with pure nickel (99 • 6 per cent.). Fig. 19 (page 925). Three cooling 
curves of the nickel are given. No. 1 was obtained in a furnace 
cooling very slowly ; Nos. 2 and 3 in a furnace cooling considerably 
more quickly. Between these latter the specimens were shifted 1 
inch in the furnace. A curve of K taken under the same conditions 
as 2 and 3 is also given on this Fig. There is the same slow 
change between about 630° and 280° C. (1,126° to 536° F.), with a 
perceptible quickening at 350° C. (662° F.). The rather smaller 



Nov. 1905. 



ALLOYS RESEARCH. 



925 



Tempera-ture FaJireafveLt 



fee 



a 

r 






rJS 



2 !^ 










i=l 
o 

a 

A 
Eh 



O CO 

s w 



o 






s 



Temperature 



Centigrade 



Note. — The important questions raised by the above curves are being made 
the subject of further investigation. (January 1906.) 



926 



ALLOYS RESEARCH. 



Nov. 1906. 



Tempcrcvtwre Fa^hrervheit 




Tenvperaytiore Cervtv^rcud'e 



Xov. 1905. ALLOYS RESEARCH. 927 

range 350° C. (630" F.), is quite intelligible in view of the fact that 
in this case a pure metal is being dealt with, whereas in the case of 
K, about 80 per cent, of foreign matter is present. The inspection 
of the curves F, G, H and J shows that the same slow change is 
taking place in these. And the conclusion suggests itself that in 
these alloys nickel is actually present as such. At any rate if it is 
dissolved in the iron, in some such form as mixed crystals, a change 
very similar to that taking place in pure nickel occurs on cooling. 

Cooling curves of alloys B, C, G and K in the forged condition 
were taken. Cylinders §-inch long were cut from the i-inch 
diameter forged bars. The curves are given on Fig. 20 (page 926), 
and a comparison of the results with those obtained with the cast 
alloys is to be found in Table 22 (page 924). 

The agreement between the results is complete. Unless forging 
produces an actual chemical change in the alloys, which is unaltered 
by heating to 900° C. (1,652° F.) no difference in the results, beyond 
that caused by a slight alteration in the composition of the materials 
during forging, was to be expected. But the experiments have been 
carried out in order to test the validity of the objection raised during 
the discussion of the Paper by one of the authors,* that the results 
obtained with forged materials might be different from those obtained 
with cast materials. 

Five alloys of series A-K (medium carbon) form a suitable basis 
of comparison with five alloys of a series of low-carbon nickel iron 
alloys made by one of the authors some years ago.f The critical ranges 
of the latter were determined by M. Osmond. { Table 23 (page 928) 
renders it possible to trace the influence of increasing percentages of 
nickel on the critical ranges of the two series (the carbon and 
manganese percentages being approximately constant in each series). 
M. Osmond's method of determining the critical ranges (inverse- 
rate curve) is different from that employed by the authors.§ Some of 



* Journal, Iron and Steel Institute, 1904, 1, page 251. 

t Proceedings, Institution of Civil Engineers, 1898-99, vol. cxxxviii, pp. 38, 39. 
X Proceedings, Institution of Civil Engineers, 1898-99, vol. cxxxviii, pp. 312- 
317. Also Plate 2. Experiments on Alloys of Iron and Nickel ; Floris Osmond. 
§ Journal of the Iron and Steel Institute, 1902, Part I, page 230. 



928 








ALLOYS 


RESEARCH. 




Nov. 1905 




I 

1 






/-N 


^ 


/-V 


^-s 










OS 

to 




Q O^ Q fe 


•fe* 

^ o 


o^ 










O Ci 


O r-l O O, 


o o 


O (M 










93 


yD p-H 


-544 
-1,0 

-123 
253- 


05 O 


SS t^ 










P^ 


■* - 


-*! CO 


00 CO 










"-2 


1 , 


--H 1 

1 o 


rH 1 

1 o 










.2 


o' o' 


o o o _i 


O CO 


O (N 


©5 








1^ GO 




CO lO 


O T-4 


o 








'-3 


OO -o 


(M rH 


O r-l 


-*J 








I— I 


CO - 

rH 


r—t 


Pi 






. o 




v^ 


•v^ 












C3 O 













i 




















^ 










"i 






oQr^ 


O) 










d3 






^*-^ 


fl 


C5 


(M Ci 


CO 


O 


m 






1^ 


QJD 


l> 


CO t^ 


00 


05 








-go 


(2 


O 


o o 


o 


O 


rH 






ego 


CS 










a 






la 


§ 










1 






r^ 


GO 


O CO 


lO 


i—i 


H 






'^'- 


S 


■*! 


-^ -*t 


•^ 


■* 


Tf 






s-S 


^ 


O 


o o 


o 


o 


o: 








Q 










a 






1 












s 












/- 






Q 


. 










OS 










^^ 


lO lO 


GO 


rH 






, 




(N 


(M 05 


05 


Ci 






55* 




o 


rH 


■* t- 


>o 


05 


'> 






15 






^^ 


rH 




















^ 














^ 


CO 






si 


w 


P O 


1-5 


« 


"c 


<1 


o 

Oh 




<5 










P 




as 




• &I 'fa 

O Q 
^ o ^^ o 


dp4 




o 

a 
o 

Qt 

r- 


H 




iJD 


O r-( 


O ?0 O T-i 


o o 


o o 




r-o 




c 


O r- 


cr- 00 O lo 


O T« 


OO 







e 




OS 


O - 


CO OS O 00 


'ii 00 


T—l CO 


»> 




« 
•^>i> 




M 


1 1 


to 1 -* 1 
1 O 1 o 


7<^ 


7^^ 






>fcS 




) 


o o 


O O O (M 


o o 


o o 


c 




6 




03 


lO D- 


O (M O (M 


O ^ 


Ot)H 


•r 






.2 


t^(M 


«D cq >0 O 


05 t^ 


'^ GO 


C 




/~\ 


U^ 


t>Tt^^ 


CD ^ lO - 


rH CO 


r-l iM 


C 












r-l T— I 


V.^^ 


V— ' 


g 






o 


g 


^C_ 


s~/ v~^ 


, ^ 


^ 








(3) CO 




Normal Cooling. 


Hastened 


Cooling. 


C 






•S^-- 




1 








c 






©v^ 


OJ 
















cc . 


cc 


i 








C 






fl^ 


ID 


1 '^^ 


lO 00 


r*^ 


CO 


^C 






Oo 


c3 


1 t-- 


?p -o 


OS 


OS 


^ 






"IS 




o 


o o 


o 


o 


r 
■41 






OS 










a 






|a 

































^2 


1 


CO 


05 t- 


CO 


OS 


S 






* ^ 


1— t 


i-H r— 1 


<M 


?-H 








% 


03 


o 


O O 


o 


o 


T 
C 






1 


O 










< 














< 
t 








1 


lO 


(N lO 


00 


Tji 


P 








05 


GO <X> 


^ 


«D 


« 








o 


o 


CO t> 


ifO 


OS 










S 






rH 


rH 
























1 S 




















o 


&q ^ 


S 


^ 










w^ 













Nov. 1905. ALLOYS RESEARCH. 929 

liis curves were taken from 875° C. (1,607^ F.) (a temperature near 
theirs of 900° C. (1,G52" F.), others from 1,030^ C. (1,886° F.), 
;ancl others from 1,270' C. (2,318° F.). But he says (loc. cit.) : 
" The quickness of the cooling and its initial temperature . . .only i^lay 
here a secondary part, of a similar order to that of iron containing 
similar percentages of carbon, and this notwithstanding the large 
percentages of nickel present." He says further : " The samples A 
to G inclusive were left to cool in a porcelain tube loosely closed at 
the ends with stoppers of asbestos ; the others were placed in the 
'Open air, cooling in the tube occurring too slowly at the critical and 
interesting point." 

Inspection of Table 23 shows that the lowering of the critical 
ranges is not markedly different in the two series up to 4 per cent, 
nickel. At 8 per cent, nickel, however, there is a sharp drop in the 
medium, but not in the low carbon series. [Reference to Table 21 
(pages 920 and 921) shows that the drop occurs at nickel percentage 
6 '42.] In the alloy with about 16 per cent, nickel the lower limit is 
nearly the same in the two series (the minimum for the medium 
carbon series having already been passed at alloy H). At 20 per 
cent, the position of the critical range is raised in the medium carbon 
series, but is still falling in the low carbon series. 

The changes'of position of the lower limits of the critical ranges 
are thus of the same kind in the two series, being somewhat more 
rapid in the medium carbon alloys for a given nickel percentage. 
One difference between the characters of the curves of the two series 
must, however, be noted. M. Osmond found that, although increasing 
amounts of nickel lower the position of the critical range, they do 
not actually widen the range itself. It will be seen from Table 23 
that the ranges are really squeezed together. 

For C the range is about 175° C. = 315° F. 
„ E „ „ „ „ 130° C. = 234° F. 
„ G „ „ „ „ 95° C. == 171° F. 
„ J „ „ „ „ 50° C. = 90° F. 
„ K „ „ „ „ 30- C. =: 54° F. 

On the other hand the authors' results show that, although the 
" lower limit " falls with increase of nickel, the upper limit is not 

3 R 



930 



ALLOYS RESEARCH. 



Nov. 1905. 



lowered below 500° C. (932^ F.) (G), and that in J it has risen to 
above 600° C. (1,112° F.) ; and that from 6*4 per cent, nickel 
upwards the range is considerable, varying between dOO° to 500° C. 
(752° to 932° F.) 

For F the range is about 519° C. = 934° F. 
„ G „ „ „ „ 377° C. = G78° F. 
„ H „ „ „ „ 435° C. = 783° F. 
„ J „ „ „ „ 474° C. = 853° F. 
„ K „ „ „ „ 411° C. = 739° F. 

In discussing these ranges it has been stated (page 923) that the 
change which begins at from 500° to 600° C. (932° to 1,112° F.),is a 
small slow thermal change, and that a considerable quickening up 
occurs at temperatures between 100° and 200° C. (212° to 392° F.). 
If this slow change be disregarded, and the range measured from the 
quickening-up temperature, the following values are obtained : — 



E . . . 


. 91° C. = 164° F. 


F . . . 


. 46° C. = 83° F. 


G . . . 


. 33° C. = 59° F. 


H . . . 


22° C. = 39° F. 


J . . . 


. 97° C. = 174° F. 


K . . . 


no quickening-up temperature 



E to H then show a progressive squeezing together of the ranges. 
In fact the numerical values for E, F and G are almost the same as 
those for G^ J and K. The following quotation from M. Osmond's 
Paper (loc. cit.) appears to supply the key to the position (bottom of 
Xmge 315) : — 

..." Besides the critical points above described and identified, 
some of the cooling curves show small variations of which the author 
cannot suggest a satisfactory explanation. Several times during the 
experiments a sudden fall of temperature was observed, followed by 
a retardation. This phenomenon for the same steel was never 
repeated twice at the same temperature, and consequently tends to 
disappear on the average. But there are other retardations of a more 
regular character lohich may have a real existence, although from their 
smallness they cannot he distinguished from experimental errors. In 
these doubtful cases the method recently described by Sir W. Boberts- 



Nov. 1905. ALLOYS RESEARCH. 931 

Austen * could he applied with great advantage." ..." The method of 
heating and cooling as applied hy the author is not sufficiently delicate 
for alloys containing more than 25 per cent. nicJcel, these alloys having 
their transformations very feebly marked or spread over too large an 
interval of temperature to he clearly identified.^' 

It would appear then that M. Osmond's speculations with regard to 
slow thermal changes in nickel-iron alloys have been confirmed by 
the results obtained by the authors, using a method which is in 
principle the one referred to by him, but which is more sensitive than 
that actually employed by Sir William Eoberts- Austen. 

Osmond examining a specimen of alloy N (49*65 per cent, 
nickel) found a slightly marked retardation between 370° C. (698° F.) 
and 340° C. (644° F.). The mean of this, 355° 0. (671° F.), is 
the temperature at which the authors have found a marked quickening 
to occur in the cooling curve of pure nickel. They think this tends 
to bear out their suggestion (page 927) that in the high nickel alloys 
nickel is actually present as such. 

Critical Ranges on Heating. — A knowledge of the position of the 
critical ranges of an alloy on heating appears to the authors quite as 
important as that of the position of the ranges on cooling, seeing that a 
good deal of heat treatment industrially of alloys is done on a rising 
temperature. Accordingly the cast alloys (A-K) after being cooled 
from 900° C. (1,652° F.) had heating curves taken. The curves are 
plotted on Fig. 21 (page 932), and the results are tabulated in Table 
21 (page 921), col. 6. The curves were obtained by the differential 
method of heating, and the co-ordinates have the same meaning as in 
Figs. 18 to 20. In these curves a movement of the differential 
galvanometer to the left indicates a critical change in the alloy, the 
change being attended with absorption of heat. 

An inspection of the curves shows that, except in the case of A, 
they are not of a duplex character. In this they differ from the 
cooling curves. Further they do not show the remarkable change of 
character which is so evident in the cooling curves in passing from D 



* Fifth Keport to the Institution of Mechanical Engineers. 

3 R 2 



932 



ALLOYS RESEARCH. 



Nov. 190"). 



Temperature Fahrcnfu'Lt 



Temperature FahreaheU 




Temperature Centigrade 



Temperature Centigrade- 



Nov. 1905. ALLOYS RESEARCH. 933 

tlirougli E to F. They are of the same type throughout except K, 
which shows no departure from the normal slope of the curve. With 
the introduction and increase of nickel the temperature at which the 
critical range is entered upon is lowered gradually from 729° C. 
(1,344° F.) to 586° C. (1,08G° F.). As regards the upper limit of 
the critical range on heating, the same convention will be observed 
as for the lower limit of the critical range on cooling, namely, the limit 
given will be the temperature at which the curve is furthest from its 
normal slope. The widest range is that of G, which extends over 
about 88° C. (158° F.). 

It was not to be expected that the critical ranges on heating 
would be as wide as those on cooling. The reason for this has been 
pointed out lately by Professor H. Le Chatelier,* in an article on rapid 
tool steels. With reference to the carbon change at about 700° C. 
(1,292° F.), in a steel of this kind he says : — (Translation) . . . 
" The speed with which this change occurs at a given temperature is 
regulated by a law which governs chemical phenomena generally. 
The speed is greater 

" (1) The higher the absolute temperature. 
" (2) The further it is removed from the transformation point. 
Above the change point the two factors act in the same direction, 
and it is impossible to heat the metal notably above it without its 
taking place." The range on heating varies from about 21° C. 
(38 F.) to 88° C. (158 F.), in the series A-K. ..." Below the change 
point the two factors act oppositely in cooling. The speed of change, 
which is nil at the transformation point, increases, passes a maximum, 
then decreases and finally becomes nil at low temperatures. The 
size and position of the maximum vary besides with a number of 
circumstances." The range on cooling varies from about 40^ C. 
(72° F.) to 520° (936° F.) in the series A-K. 

These results are thus in general agreement with Professor H. 
Le Chatelier's remarks. 

M. Osmond has likewise published heating curves of the 5 low 
carbon nickel iron alloys whose cooling curves have been given. He 
describes his method thus f : — " A Mermet furnace heated by gas is 

* "Revue de Metallurgie," June 1904, pages 334-347 
t loc. cit., page 313. 



934 



ALLOYS RESEARCH. 



Nov. 1905. 





?5s 




s 




to 




« 


^ 




(M 


s 




o 


Pd 


«0 


i-:i 




W 


s^ 


< 





nri 


Ph 


Ol 


a 


f^ 


<u 


<v 


-*J 


m 


t>. 


d 
I 


OS 


03 


% 


O 


o 


3 


j3 


P 


-«^ 






'?3 



K 



iS 



Ph 
eS O 

^a 



w 



© 

P5 



O 



be 

p 

OS 



p^ 


[^ 


f=; 


f^ 


n3 o o 

fl '=' ^ ^ 


^% 


<-^i^ 


^i^ 


^iin 


o cc 


o O 


O '^ 


O tJH 


.- cc 


»0 CO 


-* (M 


CD i-H 


Qeq 


<M - 


o - 


t^ - 


1-H »- 


_ o l> 


^T7 


^T7 


^7 


CD .-H 

1 1 


bfigo^^ 


o o 


o o 


o o 


o o 


a i>S r-^ 


O (M 


^ CO 


CD l> 


CD C^ 


rJ^ -Jj ^d "TJ 


O Ci 


CO c- 


00 00 


CO 00 


I-H 


CO T-t 

I— • 


to o 

1-H 


lO o 

t-H 


o S ^ S 


^^ 


^w^ 


N—/ 


^-^ 


c:) 


(M 


05 


CO 


CO 


i^ 


0& 


t> 


GO 


05 



o 
O 



oo 



o 



o 



iO 



o 



© 

4sl 



o 






OS 



00 



— o 



05 

CI 



M 



o 



H^ 



w 



be 

n 



Q 



O ■*! 

o to 

oo I— I 

o'o' 
lO t~ 

CO <N 



O^ 
o (M 

lO •*! 

o o 

lO CO 

CD I— I 



^ 



^ 



^ 



O IM 

O CO 
(M - 

I I 

o o 

o o 

.—I CO 

CD I-H 



O lO 

JO f— I 
(M ^ 

CD r-i 
I I 

o o 

lO Ol 
CD 'tl 
lO O 



Q 



© 
p 

cS 
blD 

a 

03 



o 






CO 
I— I 

o 






CO 

CD 



CO 
05 



CO 
05 



O (M 


H 


o - 


, , 


CD I-H 


• r-l 


1 1 


>■ 


o o 




lOfH 


o 


lO CO 
iO o 


ft-) 
o 



CO 

© 

C3 

H 
"*" 

1-H 

CC 

© 
bo 

03 



O 

> 

Oi 

I 
oo 

05 
00 



bo 



o> 
o 



I-H 

o 



CO 



CO 



a 
•i-i 

© 
© 
© 
o 
u 

p-l 



© 



to 


(M 


to 


CO 


-H 


o 


00 


CD 


-tl 


CD 


o 


CO 


t> 


lO 

1-H 


C5 


O 


K? 


e 


s 


^ 



Nov. 1905. ALLOYS RESEARCH. 935 

lit, and a temperature of about 1,000° C. (1,832° F.) obtained. 
When the temperature becomes stationary a small tube containing 
the specimens and wrapped in asbestos cloth is introduced into the 
furnace tube. The metal is then heated rapidly and regularly." In 
the authors' measurements the cylinders were introduced into the 
furnace cold, and the heating current was afterwards switched on. 
Table 24 (page 934) enables a comparison to be made between the 
critical ranges on heating of corresponding members of the two series. 

In the low carbon series with increasing nickel the critical range 
is gradually lowered, and also squeezed together. In the medium 
carbon series the critical range is lowered, but the narrowing is not 
noticeable except in passing from G to /. Broadly speaking the 
ranges of the medium carbon series, as was to be expected, are 
lower throughout; but for nickel content 15* 5-16*0 per cent, the 
values are nearly the same. 

The heating curve of nickel given on Fig. 19 (page 925) indicates 
a critical change at about 340° C. (644° F.). This is not far 
removed from the temperature of about 350° C. (662° F.), at which 
the critical change on cooling quickens up. It is also in the 
neighbourhood of the temperature, 340° C. (644° F.),* at which 
nickel loses its magnetism on heating. 

Although alloy K gives no evidence of a critical range on a 
heating curve, either in the cast or forged condition, when it has 
been previously cooled from 900° C. (1,652° F.), it is proper 
to state that the material when it comes from the forge does 
give evidence of a critical range beginning at about 670° C. 
(1,238° F.) on heating. Mechanical work of various kinds — rolling, 
hammering, bending, stretching, compressing, twisting — causes this 
alloy to undergo a structural and magnetic and chemical change. A 
fresh structural constituent appears which has much the same 
appearance however it is produced ; the material has also become 
magnetic. This condition is removed by heating to 800° C. 
(1,472° F.). The whole matter is dealt with in a special section of 
this Eeport (pages 953 to 958). 

* Guillaume. " Nouvelles Keclierches sur les Aciers au Nickel." Comptcs 
Rendus, 1903, page 10. 



dSO 



ALLOYS KESEARCH. 



Nov. 1905. 



The Meversibility of the Critical Manges on Heating and Cooling. — 
In the foregoing paragraphs the term critical range, and not critical 
point has been used, the thermal changes extending in some cases 
over a considerable number of degrees. Accordingly in this 
concluding paragraph, in which the reversibility of the changes will 
be considered, the upper limit of the range on heating will be 
compared with the upper limit on cooling — the lower with the lower ; 



Fig. 22. — Critical Banges or Heating and Cooling. 




NiX'kcl per cent 



the terms upper and lower limit having the following signification 
(which has been used throughout). 

The \ jj \ limit on a \ ^^ -.. ^ \ curve is the temperature at which the- 
curve begins to bend from the normal slope. 

Thtl JJ \ limit on a \ jt )■ \ curve is the temperature at which the 

curve is furthest from its normal blope. 

Fig. 22 and the left-hand division of Table 25 (page 937) give 
a comparison of the upper limits of the critical ranges on heating 
and cooling. It will be seen that the differences for alloys A-F are 



Xov. 1905. 



ALLOYS RESEARCH. 



937 








1 

OQ 






















j 




a 


• 


o 


o 


o 


o 


o 


o 


O 


o 


o 






o 


h~H 


CO 


I- 


CO 


<M 


c^ 


■^ 


-« 


»o 


t^ 




.2 


j-t 




(r-i 


o 


o 


CO 


oo 


o 


CO 




CO 




to 

•T3 




rH 




t— 1 


I— 1 


CO 


c» 


GO 


o 


L- 




^ 




II 


II 


11 


11 


11 


11 


11 


11 


11 


1 




'S 


. 


O 


o 


o 


o 


o 


o 


o 


o 







O 


<— ( 


tH 


o 


o 


CM 


1- 


CO 


oo 


l^ 




2 






L^ 


o 


CO 


Ci 


^1^ 


C5 


CO 


o 


CO 






o 












<M 


rft 


tH 


lO 


'n 




^ fcb 
























e*-( a 


























O ;i; 


















































OQ O 


























-*^ o 


























a "" 






o 


o 


o 


o 


o 


O 


o 


o 


o 


O 


.StU 






O 


iO 


o 


r-H 




1^ 


CO 


CM 


o 


(M 


'"' fl 




• 


-~^ 


o 




l-H 


Ci 


»o 


to 


I- 




l- 


M CO 


fcjo 


fiH 


<M 


I— 1 


r— 1 


O 


I- 


(M 


M 


r-l 


CO 


CO 


<o . 


a 




*^ 


^ 


•^ 


,^ 














O .3 






I— I 


J—^ 


j^^ 


,-^ 














^ 
























<— 1 ^ 


o 




O 


o 


o 


o • 


o 


o 


o 


O 





o 


,a 2 


• 


CO 


CD 


— H 


'tl 


G^l 


lO 


CO 


OO 


o 


o 




Q 


lO 




o 


Ti^ 


(M 


(M 


(M 


L- 


-tH 


oo 


-t-> r-H 






CO 


CO 


CO 


o 


-^l 


t— 1 


1— ( 




rH 


rH 


fl d 


























o o 












































































-w fcc 


























"^ != 
























1 


-=2 S 






o 


o 


o 


O 


O 


O 


o 


O 


o 


1 


o 






TtH 


(M 


t^ 


CO 


CO 




1^ 


l^ 


1^ 




be 

.9 


• 


'tH 


o 


C<J 


r^ 


I- 


»o 


GO 


00 


CO 




Pm 


CO 


CN 


(M 


1— ' 






O 


c: 


o 




2i 


H^ 




!-H 


r-i 


I— 1 


1-H 


t— 1 


1—1 


r-i 


T—l 


rl 


C^« 


,<jj 


eS 
























5tt 






o 


O 


o 


O 


o 


o 


O 


C 


c 




ft 




o 


O 


■^H 


-H 


-tl 


(M 


CO 


CO 


CO 






O 


(M 


O 


CO 


CO 


CO 


(M 


CO 


oo 


CO 










l^ 


L^ 


CO 


CO 


CO 


CO 


iO 


*-J 


*o 


i 




02 




o 


o 


o 


o 


o 


o 


o 


o 






, 


<y 


I^* 


CO 


1— ( 


c» 


CO 


GO 


(M 


CO 


-tH 






03 


JH 


ou 


CO 


o 


o 


00 


L^ 


1-H 


CO 






o 


be 








r— 1 


I— ( 






CO 


(M 








© 
























-tJ 


r3 




II 


11 


II 


11 


11 


11 


11 


11 


1 


1 




O 




o 


o 


o 


o 


o 


o 


o 


o 






© 


Q 


o 


■r*^ 


o 


o 


<Zi 


o 


^+1 


l^ 






d 




tH 


CO 


CO 


»o 


tH 


TtH 


l> 


'tl 






















1— ( 


1—1 






-^ fcc 


^ 
























e« C 


























°;^ 
























[ 


M O 


























.t: o 


























aJ^ 






o 


o 


o 


o 


c 


o 


o 


o 


r, 





•S -C 






CO 


oo 


(M 


'O 


1—1 


I— 1 


(M 


o 


CO 


(M 


" a 


, 


• 


o 


CO 


(M 


o 


o 


C5 


CO 


o 


iO 




a '^ 


be 


P^ 


CO 


(M 


CJ 


I— 1 


r— 1 


I— 1 


C5 


o 




r- 


p,&c 


_H 




I-H 


t— 1 


I— < 


r— 1 


1— ( 


, 






j_^ 


1-^ 


2 S 


'o 
o 




o 


o 


o 


O 


O 


o 


o 


o 


o 


o 


• 


«3 


1^ 


I— 1 


CO 


■^ 


^+1 


o 


CO 


CO 


o 


a-S 




O 


o 


CO 


CO 


T+H 


•^ 


-+^ 


o 




<M 


o 






r> 


<^ 


CO 


CO 


CO 


CO 


to 


i-"^ 


■CD 


CD 


a S 


























o o 


























P cc 


























j? 0) 


























-tJ tUO 


























O fl 


























-^ s 






o 


o 


o 


o 


o 


o 


o 


o 


o 




o 

a 






1—1 


o 


o 


^H 


CI 


CO 


lO 


o 


T—t 




bb 




Ci 


CO 


CO 


o 


1^* 


CO 


""tl 


(M 


tH 




f^* 


CO 


CO 


CO 


CO 


CM 


(M 


(M 


(M 


1—1 




QJ 






,_, 


rH 


,__, 


,_, 


,_i 


1-H 


rH 


T-^ 


,_l 




,03 


"S 






















CW 


ia 


0/ 




o 


o 


o 


o 


o 


o 


O 


c 


o 




ft 


• 


o 


*-i^ 


r—t 


lO 


CO 


■rH 


^+1 


o 


CD 






o 


'O 


(M 


-M 


o 


o 


CO 


1^ 


CO 












t- 


L-- 


L^ 


L^ 


CO 


CO 


CO 


o 


S 




> 


% 
























< 






<i 


ft 


O 


ft 


H 


r^ 


o 


K 


•-5 


1 



938 



ALLOYS RESEARCH. 



Nov. 1905. 



fairly constant, varying between 34° to 60° C. (61° to 108° F.), or from 
17° to 30° C. (30° to 54° F.) on either side of a mean temperature. 
Alloys G and H show much greater differences, namely, 174° and 
147° C. (313° to 264° F.) respectively, or 87° and 73° C. (156° to 
131° F.) on either side of a mean temperature. In alloy J there is 
practically no difference between the temperatures of the two limits, 
and in alloy K the difference cannot be estimated. 

Fig. 22 and the right-hand division of Table 25 (page 937) enable 
the same comparison to be made for the lower limits of the ranges 
on heating and cooling. 

Inthisseries alloys A-D show differences of 54°to90°C. (97° to 162° 
F.),or 27° to 45° C. (48° to 81°F.) on either side of a mean temperature. 
A sharp break occurs at E, where the difference is 106° C. (191° F.) 
on either side of a mean temperature, and in the case of the alloys 
F-J the differences are very large, namely, of the order of from 220° 
to 250^ C. (396° to 450° F.) on either side of a mean temperature. 

Thus, whichever limit of the ranges be considered, alloys A-D 
may be classed as at any rate approximating to reversible alloys, 
while alloys G and H are seen to be irreversible. E and F occupy 
an intermediate position, their upper limits being reversible, their 
lower limits irreversible. J also shows the same property. For K 
no classification is possible. 



The Bange of Solidification and the Critical Banges of Nickel on 
Heating and Cooling. — The authors have had occasion to determine 
these properties during the present research, and they think it may 
be useful if the results are summarised here. 



TABLE 26. 




Nov. 1905. ALLOYS RESEARCH. 939 

The nickel used for solidification was in the form of " berries " 
made by the Mond Nickel Process. That used for critical ranges 
was a cylinder cut from a nickel bar, belonging to the Meteorological 
Department. The values of the melting point of nickel to be found 
in the literature are very divergent, for example, the following are 
given in Castell Evans' Physico-Chemical Tables. 

1,371° C. (2,500° F.), highest forge heat, 1,450° C. (2,642° F.), 
1,600° C. (2,912° F.), 1,392° to 1,420° C. (2,537° to 2,588° F.), with a 
probable value of 1,450° C. (2,642° F.). 

Holborn and Wien found it to be 1,484° C. (2,703° F.) * 

The authors' value of 1,427° C. (2,600° F.) for the beginning of 
solidification, has been exactly confirmed by Dr. Harker working with 
four different thermo-junctions under quite different conditions. The 
values all refer to nickel in an oxidising atmosphere. 

The very sensitive method of detecting critical ranges used by the 
authors has enabled them to detect and follow a slow thermal change 
in pure nickel on cooling. The range is from about 630° to 280° 0. 
(1,166° to536°F.), with a marked acceleration at about 350°C.(662°F.), 
This latter temperature is quite near that found recently by Nagaoka 
and Kusakabe for the beginning of magnetism in nickel on cooling.'j' 

A critical range beginning at about 342° C. (647° F.) on heating 
is also indicated in the authors' curves, and a reference to Nagaoka 
and Kusakabe's results shows that at about this temperature 
magnetism is disappearing. 

Peculiarities in the thermo-electric and resistance properties in 
the neighbourhood of 300° C. (572° F.) have long been known. The 
authors' results are thus in general accord mth the change of 
properties of nickel previously known. 

The Metallography of Nickel Steels, B-K. 

The metallography of the series of low carbon-nickel-iron alloys 
(A to N) (average • 17 per cent, carbon), prepared by one of the 

* Ann. der Physic und Chemie, vol. Ivi, page 360. 1896. 
t Journal of the College of Science; Imperial University of Tokyo, 
vol. xix, 1904. 



940 



ALLOYS RESEARCH. 



Nov. 1905. 



authors, was studied by M. Osmond.* He divides the alloys into 
three groups : — 

Group 1. (Up to 7-65 per cent, nickel.) The structure 
resembles that of low carbon steels containing no nickel. (Ferritic -f- 
Pearlitic structure.) 

Group 2 extends to steels with about 25 per cent, nickel. 
(Martensitic or acicular structure, similar to that of hardened carbon 
steels.) 

Grouj) 3 comprises the alloys still richer in nickel. They are 
non-magnetic at ordinary temperatures. (Polyhedral structure, 
similar to that of pure iron above A3.) 

In 1903| M. Guillet, to whom the foregoing classification of nickel 

iron alloys was unknown, published the results of a very thorough 

study of the metallography of three series of alloys and adopted a 

similar method of grouping. 

Series 1 contained about 0*120 % carbon, and from 2 to 30 % nickel. 
'> •^ jj 0*220 ,, ,, ,, ,, 

jj 3 ,5 0*800 ,, ,, ,, 5, 

The three series contained only small amounts of the ordinary 
impurities, and were low in manganese. M. Guillet worked on materials 
(a) forged,! (h) quenched, (c) reheated, (d) cold worked, (e) cooled 
considerably below atmospheric temperatures. With one exception 
(500 diameters) all the photo-micrographs have a magnification of 300 
diameters, which is insufficient to characterise the pearlitic structure. 
The following is a tabulation of his results with the forged 
alloys : — 

TABLE 27. 



Group. 

1 
1 


Micro jrraiiluc ^ -.^ ■, ^ ^^ , 
characteristics. « ' ^^ carbon. * 22 carbon. 


0*80 carbon. 


1 

1 

2 
3 


per cent. I per cent, 
o iron + pearlitc. to 10 Ni. to 7 Ni. 
Martensite. 10 to 27 „ ' 7 to 25 „ 
Polyhcdric - y iron. over 27 „ | over 25 „ 


per cent. , 
*0 to 5 Ni. 1 
5 to 15 „ ; 
over 15 „ j 



Group 2 overlaps with Groups 1 and 3. 



* Proceedings of the Institution of Civil Engineers, 1898-99, vol. cxxxviii, 
pages 322, 323. 

t Bulletin dc la Socictc de rEncouragemcnt, 31 May, 1903. 

X " The Metallographist," 1903, pages 277-283, erroneously translates " brut 
de forge " as " cast steel." 



Xov. 190.'). 



ALLOYS RESEARCH. 



941 



lu each series the first steel with a polyhedric structure is 
nou-magnetic at ordinary temperatures. The greater the sum of the 
nickel and carbon contents, the smaller the percentage of nickel at 
which the structure changes. 

It will be noticed that M. Guillet fixed the same limits for the 
0"22 per cent, carbon series as M. Osmond for the 0*17 per cent, 
carbon series. The latter series, however, contained from 0*65 to 
1 • 08 per cent, manganese, which doubtless is sufficient to account for 
the diiference. 

Quenching Experiments. — Thermo-couple used. Quenching bath ; 
water at 15^ to 20" C. (50° to 68" F.). 

TABLE 28. 



i Group. 

i 

1 


Microstructure 
before quenching. 


Quenching temperature a little higher than 

the magnetic transformation i)oint 

on heating. 


! 1 

' 2 

1 

o 

O 


(a iron + pearl ite or) 
\cementite + pearlite. j 

Martens! te. 

Pure 7 iron (polybedric.) 

1 
1 


Same effect as with ordinary steels. ! 

j Tendency towards polyhedric structure but 
( predominance of martensite. ; 

1. For the first steels of each series with i 
polyhedric structure, acicular crystals 
and more finely divided crystals. 

2. No appreciable change in the case of 
feteel containing more nickel. 



Beheating Experiments. — Members of Group 1 are affected in the 
same way as pure carbon steels. With members of Groups 2 and 3 
the same results follow as on quenching. Certain of the steels with 
polyhedric structure, upon which quenching was without effect, gave 
martensitic structure on heating. M. Gaillet concludes that the changes 
of structure caused hy the quenching experiments are not due to 
quenching proper, hut merely to the reheating preceding it. 

Cold-WorJcing Experiments. — Steels with polyhedric structure, 
when pressed or hammered beyond the elastic limit, pass into 
martensitic structures. 



942 ALLOYS RESEARCH. Nov. 1905. 

Low-Temperature Experiments {at —78^ C.) ( —108 '4^ F.). — 
Steels of Groups 1 and 2 undergo no change. As regards Group 3, 
tlie tendency is to convert poljlaedral into acicular crystal. The 
latter appear sometimes white, sometimes black, after etching. In 
one specimen the authors found a crystal coloured white on one side, 
black on the other. When martensitic crystals began to he formed^ 
magnetism appeared. 

M. Guillet has recently published a reprint of the Paper, of which 
a very brief abstract has been given, under the title of " Les Aciers 
speciaux," and has added a section entitled " Nouvelles recherches 
sur les Aciers au Nickel." It is necessary to direct attention to three 
of his conclusions : — 

1. " The martensite of the nickel steels appears to be a special 
martensite, that is, different from that of carbon steels." 

2. " A steel which has once been changed from the polyhedric to 
the martensite condition, by whatever method, cannot be regenerated 
either by annealing or quenching," 

3. " Page 44 contains a diagram in which, from a knowledge of 
the carbon and nickel contents of a nickel steel, it is possible, for a 
given carbon percentage, to read off — 

(a) the nickel content at which tlie structure changes from 

pearlitic to martensitic, 
(6) the nickel content at which the structure changes from 
martensitic to polyhedric." 
The diagram of course only holds for steels in which the ordinary 
impurities are low. 

M. Guillet has correlated the three types of structure of nickel 
steels with their mechanical properties : — 

" Group 1 steels (Pearlitic) have properties similar to pure carbon 
steels ; the yield point and ultimate strength being a little higher. 

" Group 2 steels (Martensitic) are characterised by properties 
similar to those of carbon steels containing notable quantities of 
hardening carbon ; they are extremely hard, difficult to work and 
brittle. They are unaffected by quenching. 

" Group 3 steels (Polyhedric) have a low elastic limit, are not 
brittle, and can be worked with great ease." 



Nov. 1905. ALLOYS RESEARCH. 943 

The authors are glad to state that the results of their 
metallographic study of alloys B-K are in complete agreement 
'with M. Guillet's conclusions on many points. On one or two 
matters, however, they have formed a different opinion from his, 
and to these special attention will be directed. 

M. Guillet's first Paper appeared while their research was 
in progress, and in view of his results they decided to exclude 
quenching experiments. The scheme of work has included the 
microscopic examination of — 

A the cast alloys, cooled from 900° C. (1,652° F.) 

Ji the forged alloys. 

C the forged alloys cooled from 800° 0. (1,472° F.) 

D the cast alloys cooled to -100° C. (-148° F.) 

E the structure of K upon which mechanical work of various kinds has 

been done. 
F Subsequent heat treatment of K. 

A 5-per-cent. solution of picric acid in alcohol was used at the 
outset for developing the structures. With this reagent — 

Etching required about 5 minutes (ordinary temperatures) for alloys A to D. 
,, „ 10-15 „ „ 55 55 E to J. 

5 5 „ more than 4 hours ,, „ alloy K. 

A 1-per-cent. solution of nitric acid in alcohol gave the same 
results in a very much shorter time. 

For alloys A-D about 30 seconds. 
„ E-J „ 1 to 2 minutes. 

„ K „ 30 minutes. 

The latter reagent has been used throughout the research. 

It has been found convenient to work at three magnifications : — 

(a) In a few cases at about 10-12 diameters, in order to obtain 
a general survey of a comparatively large portion of the etched 
surface. A Zeiss planar of 50 mm. focal length proved suitable for 
this work. 

(6) At about 150 diameters. This is suitable in most cases for 
'• low-power " photo-micrography. 

(c) At 1,000 diameters, for the characterisation of pearlitic and 
martensitic structures. 



944 ALLOYS RESEARCH. Nov. 1905. 

Terminology. — Tlie types of structure which are characteristic 
of alloys A-K can be classified in the same three groups that have 
been adopted by MM. Osmond and Guillet, namely, (a) pearlitic, (6) 
martensitic, (c) polyhedral. Typical photo-micrographs of the 
three varieties of structure are given on Plate 50. The nature 
of the martensite of hardened steels is however still a matter of 
controversy. Further, M. Guillet's conclusion that " the martensite of 
the nickel steels appears to be a special martensite, different from 
that of carbon steels," is, regarded from a chemical standpoint, 
absolutely self-evident, seeing that the former variety contains 
nickel and the latter does not. For this reason, in Group 1 the 
word pearlitic is used as a type of structure, not as a constituent. 

A. Structures of the Cast Alloys cooled from 900° C. (1,652° F.). 
Plates 51 and 52. Group 1. Pearlitic Steels. — Alloys A to E 
(50 to 55 per cent, nickel) come under this head, the introduction of 
nickel and its rise to about 5 per cent, causing a fining of the 
structure. The photo-micrographs at 500 diameters show the 
pearlitic structures to be lamellar as well as granular, in spite of 
the presence of an average of • 88 per cent, manganese. 

Group 2. Martensitic Steels. — Alloys F to J (6*42 to 15*98 per 
cent, nickel). The authors have found, as Guillet did, that the 
martensitic structure is light and dark. In F the light variety is 
present to the almost entire exclusion of the dark ; in G the dark 
j^reponderates ; in H the two are about equally distributed, while in 
J the polyhedral structure appears side by side with the martensitic, 
the latter appearing regularly darker. It may be pointed out that 
the] dark type of martensitic structure recalls the so-called 
*' Troostitic " structure of carbon steels, the latter, however, not 
being markedly martensitic. 

Group 3. Polyhedral Steels. — Alloy K (19*91 per cent, nickel). 
This requires a more detailed description than has been accorded the 
members of Groups 1 and 2. As previously mentioned the etching 
properties of this alloy place it in a class separate from the others. 



I 



Nov. 1905. ALLOYS RESEARCH. 945 

After 5 hours' treatment with picric acid reagent, or 30 minutes 
with nitric acid reagent, it is seen by the naked eye to be covered 
with a black deposit. If this be rubbed off, the structure seen in 
photo-micrograph, Fig. 25, Plate 53, at 10 diameters is obtained. 

The photo-micrograph at 150 diameters, Plate 51, shows the 
structure to consist of large white polyhedra with ripple marks.* 

The photo-micrograph at 500 diameters, Plate 52, shows the 
etching pits or minute holes left by the escape of gas at one of the 
polyhedral boundaries. If the black deposit be partly rubbed off, 
the following appearances are obtained : — 

Fig. 26, Plate 53, at 5 diameters. 

Figs. 27 and 28, at 75 diameters, which show 3 types of colour. 

(a) White polyhedra. 

(h) Green and red polyhedra. 

(c) Striped (black and white) polyhedra, resembling zebra 
markings. 

(Fig. 30 gives the last-named at 1,000 diameters.) 

Figs. 27 and 28 are taken through green glass and red glass 
respectively in order to differentiate between the two coloui's. 
Fig. 27 taken through green glass shows the green markings light 
and the red markings dark. Fig. 28 shows the opposite. In order to 
try and see which colour was in the hollows this field was rubbed 
with * 000 ' emery, and it was found that the green was in the hollows, 
the red deposit being scratched off. (Fig. 29 shows this.) Direct 
focussing on green and red parts, though less definite, led to the 
same conclusion. When the red and green deposit had been 
rubbed off, it was seen that the red lines corresponded to the dark 
ripple marks, and the green areas to the white hollows. 

The deposit was chemically investigated. One of the strips cut 
from the ingot, measuring 5 inches in length, IJ inch square, was 
cleaned of rust, and was then immersed for 8 weeks at ordinary 
temperatures in 5 per cent, picric acid solution in absolute alcohol. 
From time to time the brownish black deposit was scraped off with a 
glass rod, and the flakes allowed to accumulate at the bottom of the 

* cf. M. Osmond's Fig. 6, Plate 6, Proceedings, Instilution of Civil Engineers, 
1898-99, vol. cxxxviii. 

3 s 



946 



ALLOYS RESEARCH. 



Nov. 1905. 



containing vessel. The level of tlie liquid was kept constant. 
Even after two months the deposit accumulated was very small. 
It was thoroughly washed with water, alcohol, and ether. It 
contained traces of iron, but no nickel or manganese, and had about 
the same density as water. It left a slight residue on ignition in 
^ platinum crucible. What was left over after the above i tests was 
estimated for carbon by the combustion method, but only a minute 
quantity of carbon dioxide was weighed. It appears probable 
therefore that the deposit consists of a highly bydrated, slightly 
carbonaceous matter. 

The series just described, containing an average of • 44 per cent, 
carbon and 0*88 per cent, manganese, is seen from the following 
Table to be almost identical in micrographic characteristics with 
M. Guillet's series, containing an average of 0*82 per cent, carbon 
and 0*08 per cent, manganese. 

TABLE 29. 



Group. 


Micrographic 
Characteristics. 


0*44 carbon 
0-88 manganese. 


0-82 carbon. 
• 08 manganese. 


1 
2 

3 


Pearlitic. 

1 Marten sitic "1 
\ (dark and light). / 

Polyhedral. 


Per cent. 
to 5 or 6 Ni. 

5 or 6 to about 16 Ni. 

over 16 Ni. 


Per cent. 

to 5 Ni. 

5 to 15 Ni. 
over 15 Ni. 



1 



For a given percentage of nickel, 0*38 per cent, carbon is 
equalled by 0*80 per cent, manganese in its influence on the 
structure of the alloy. Thus broadly speaking carbon is seen to be 
rather more than twice as powerful as manganese in its structural 
influence. 

B. Characteristics of the Forged Alloys. — (Sections cut from 
J inch diameter round bars.) (Plates 51 and 52.) 



Xov. 1905. 



ALLOYS RESEARCH. 



947 



Group 1 is narrower tlian in the cast alloys, the pearlitic structure 
ending at about 4 per cent, nickel (Alloys A-D). 

Group 2 contains the remaining members of the series. Alloy E 
is markedly martensitic (see photo-micrograph at 500 diameters). 
F contains much more of the dark etching material than the cast 
alloy. In G the dark etching material has taken a cellular structure. 
H shows no change. In J the clear martensitic structure is very much 
broken up and the alloy etches darkly. K has entered this group, 
forging having produced a dark etching material, and the structure 
resembling that of J. It now etches readily. It is doubtless this 
change which is responsible for the heating curves of the forged and 
cast alloys being dififerent. 

These results confirm Guillet's view that mechanical work causes 
the polyhedral to be replaced by the martensitic type of structure. 

The series investigated by the authors shows that Group 2 
absorbs members from its lower as well as its upper limit (Alloy E). 

TABLE 30. 



Group. 


Micrographic Characteristics. 


Forged Alloys. 
0*44 carbon. • 88 manganese. 


1 
2 


Pearlitic 
Martensitic (dark and light) 


Per cent. 
to 4 Ni. 
5 to 20 Ni. 



G. Structures of Forged Alloys, cooled from 800° C. (1,472° F.).— 
(Plates 51 and 52.) 

Group 1 has the same nickel limit as in the forged alloys 
(A-D). 

Group 2 includes alloys E-J. E is similar to that in B. F 
contains a good deal more of the dark etching material. G and H 
resemble one another strongly, but are different from the corresponding 
members in A and B. J has reverted to the structure of the cast 
alloy, except that the distribution of the dark and light martensitic 
structure is not so uniform. 

3 s 2 



948 



ALLOYS RESEARCH. 



Nov. 1905. 



Group 3. Alloy K returns to this. The meclianioal effects of 
forging are still evident, but the structure has reverted to the 
polyhedral. The authors wish to draw particular notice to this fact on 
account of M. Guillet's dictum that " a steel which has once been 
changed from the polyhedric to the martensitic condition, hy ivhatever 
method, cannot be regenerated either by annealing or quenching." 
This observation contradicts it. 

TABLE 31. 



Group. 


Alloys. 


Micrographic 
Characteristics. 


Forged Alloys 

cooled from 800° C. 

(1,472° F.) 

0*44 carbon. 

0*88 manganese. 


1 
2 
3 


A-D 

E-J 

K 


Pearlitic 

("Martensitic (dark and\ 
{ liglit) / 

Polyhedral 


Per cent. 
to 4 Ni. 

5 to 16 Ni. 

above 16 Ni. 



The mechanical properties of the alloys have been determined in 
the forged heat-treated condition, and it is interesting to correlate 
them with the three types of structure which have just been described. 
A system of classification based on mechanical properties would place 
them in three groups, in which the same members would be found in 
corresponding groups. 

Alloys A-D have properties similar to normalised carbon steels, 
the yield point and ultimate strength rising with increasing nickel. 

Alloys E-J approximate in their properties to hardened carbon 
steels, with high ultimate stress, and low ductility. They are hard 
and brittle. 

Alloy K is characterised by a low yield point and great ductility, 
and it combines the qualities of hardness and softness in a very 
remarkable way.* M. Guillet's contention that steels of this class " can 

* To use M. Osmond's description of steels of this class, it possesses " a 
characteristic combination of strength and plasticity." Proceedings, The 
Institution of Civil Engineers, 1898-99, vol. cxxxviii, page 163. 



Xov. 1905. ALLOYS RESEARCH. 949 

be worked with great ease " is not confirmed either in M. Osmond's * or 
in tlie authors' experience. The alloy is certainly not so hard to 
machine as alloys E-J (with martensitic structure), which have 
proved very difficult to manipulate. But it is by no means easy to 
machine, except at the very beginning. The explanation appears to be 
as follows : — The polyhedral structure is characterised by softness. 
The effect of mechanical work, of whatever kind, is to convert the 
polyhedral into the martensitic structure which is characterised by 
hardness. Therefore, after machining has been in progress a short 
while, it is being done on a hard material. 

The bearings of the results of the foregoing micrographic 
analysis on the question whether the influence of nickel on iron is 
direct or indirect in nickel steels can now be briefly discussed. 

Into the three series of alloys examined microscopically by 
M. Guillet, in which with varying percentages of nickel the carbon 
average was 0*12, 0*22 and 0*82 per cent., the authors' series with 
• 44 per cent, carbon exactly fits. M. Guillet's series contained traces 
of • 02, • 08 per cent, manganese respectively ; the authors' contains 
• 88 per cent. There is perfect agreement between M. Guillet's results 
and the authors' in respect of the types of structure produced. 
These types — the pearlitic, the martensitic (dark and light), the 
polyhedral — have their analogues in carbon steels, containing no 
nickel, in pearlite, troostite, martensite, and austenite respectively, 
the last three being the subject of controversy as to whether they 
are or are not to be regarded as constituents. The analogy between 
the dark martensitic structure in the nickel steels with the troostite 
of carbon steels is warranted from the " colour " standpoint, though 
it is open to doubt from the " morphology " standpoint. But with 
this exception and the fact that in pure carbon steels the polyhedral 
structure has not been obtained free from the martensitic structure, 
there is no difference in respect of structure between the series of 
nickel steels examined by M. Guillet and the authors on the one hand, 
and that obtained in nickel free carbon steels by suitable thermal 
treatment on the other hand. 

* Iron and Steel Metallurgist, vol. vii, page 18. 



950 ALLOYS RESEARCH. Nov. 1905. 

The next point to be emphasised is that the changes of structure 
occur in the same order ; in the nickel steels, with any given carbon 
content up to 0-8 per cent., by simply increasing the nickel 
percentage ; in tlie carbon steels by quenching from higher and 
higher temperatures. This suggests that the changes in the former 
series are very similar to those in the latter case ; that the changes 
•which in pure carbon steels require quenching to prevent their taking 
place are simply avoided in the nickel steels by the presence of 
nickel which acts as a brake. The authors have already shown 

that a comparison of their / l'^^ ^'' ^^^*- ^^^^«" | series with 

10 '88 per cent, manganese J 

M. Guillet's /- per cen ^cai^ on | ggj-jgg enables the conclusion to 

(0-08 per cent, manganese) 

to be drawn that the action of manganese in presence of carbon and 

nickel is similar to that of carbon, though less than half as powerful. 

Further, a comparison of M. Guillet's three series among themselves 

renders it possible to compare the action of carbon with that of 

nickel. 

Thus, considering the change from the martensitic to the 

polyhedral range, three comparisons are possible : — 

(0*22 — 0*12 carbon) = 0"10 carbon is equivalent to 2 per cent, nickel. 
(0-80 -0-22 „ )=0-58 „ „ 10 „ 

(0-80 -0-12 „ ) =0-G8 „ „ 12 

From which it follows that carbon is 20, 17 and 18 times 
respectively — an average of about 18 times as powerful as nickel, in 
the mechanism of this particular change. 

M. Guillet's "Manganese Iron Alloys" * enables a comparison to be 
made between carbon and manganese in the absence of nickel. With 
given carbon content the introduction and increase of manganese 
causes a series of structure changes similar to those in carbon steels 
containing only small quantities of manganese. The analogy is 
even more complete than in the case of the nickel steels, for the 
*' sorbitic " and " troostitic " structures, as well as the martensitic and 

* " Les Aciers spe'ciaux." 



i 



Nov. 1905. 



ALLOYS RESEARCH. 



951 



polyhedric types, pass in succession in the series containing 0*86 
per cent, carbon. 

TABLE 32. 



Group. 


Micrograpliic Characteristics. 


0-18 carbon. 


0-86 carbon. 


1 
2 
3 


Pearlitic (including sorbitic) 
Martensitic ( „ troostitic) . 
Polyhedral .... 


Per cent. 
to 5 ]\In. 

5 to 12 Mn. 

above 12 Mn. 


Per cent. 
to 3 Mn. 

3 to 5 Mn. 

above 5 Mn. 



In the change from Group 1 to Group 2, 

(0*86 to 0*18) = 0*G8 per cent, carbon is equivalent to 2 per cent, manganese; 

that is, for this change manganese is about one-third as powerful 
as carbon. And in passing from Group 2 to Group 3, 

0*68 percent, carbon is equivalent to about 7 per cent, manganese; 

that is, carbon is here about ten times as effective. 

The conception of the " metallurgical equivalency " of hardening 
carbon, manganese and nickel, is due to M. Osmond,* who states, 
" M. Guillet, who was even now at work to determine the numerical 
values of these equivalents with greater precision than had yet been 
attempted, had found that 1*65 parts of hardening carbon, or more 
exactly 1 • 65 parts of total carbon containing the maximum amount 
of hardening carbon, was equivalent to 12 parts of manganese and to 
29 of nickel." The authors' alloys, containing carbon, manganese 
and nickel, constitute, they think, additional evidence of the truth of 
M. Osmond's contention, that the action of these three elements upon 
iron is of the same kind, though not of the same strength. If the 
action of carbon on iron is described as " direct," so also must those 
of manganese and nickel be thus conceived. Thus the authors' 
conclusion is that, neither in the three series of nickel steels 



* The Iron and Steel Metallurgist and Metallographist, vol. vii, page 18. 



952 ALLOYS RESEARCH. Nov. 1905 

microscopically examined by M. Guillet, nor in the present series of 
nickel manganese steels similarly examined by them — the carbon 
ranging from 0*12 to 0*82 per cent. — is there any evidence of a 
special carbide of nickel. The possibility of the existence of a 
special carbide in presence of higher percentages of carbon is, however, 
not excluded. 

D. The Cast Alloys cooled /o-100° C. (-148° i^.).— (Photo- 
micrographs Figs. 31 to 38, Plate 54.) Small cylinders with polished 
surfaces were cooled to about — 1 00° C ( — 148° F.), by immersion 
in a freezing mixture produced by dissolving solid carbon dioxide in 
acetone. (The temperature was read by means of a toluene 
thermometer, kindly lent by Dr. Harker.) The cylinders were kept 
in the mixture for six hours, which gradually assumed the ordinary 
temperature. After this treatment the surfaces of cylinders A to D 
and G and H (alloys E and F had not at the time of these experiments 
been made) preserved their polished appearance. J and K had 
become dimmed with a crystalline pattern on the polished surface, 
the development in K being much the stronger of the two. Photo- 
micrographs at 100 and 700 diameters, both of the unetched and the 
etched surfaces, are appended (Figs. 31 to 38), scratches being made 
to enable the same field to be taken. 

J unetched presents a martensitic pattern, seen best at 1,000 
diameters. 

Etching enables the martensitic structure to be compared with that 
of the untreated alloy. 

K unetched has a strongly developed pattern. The surface was 
so uneven that at 1,000 diameters no field could be obtained entirely 
in focus. The speckled appearance characterises the whole surface. 

The alloy now etched as readily as J and had become magnetic. 
The micrograph at 1,000 diameters of the etched surface shows the 
dark etching martensitic pattern alongside a patch of unaltered 
structure. 

The authors' results are thus in complete agreement with those 
obtained in M. Quillet's experiments at - 78° C. (- 108-4° F.), 
namely : — 



Nov. 1905. ALLOYS RESEARCH. 953 

1. The members of Groups 1 and 2 (A to D and G, H) are 

unaffected by this treatment. 

2. The members of Group 3 (J on the border between Groups 2 

and 3 and K) show a partial conversion of the polyhedral 
into the martensiric type of structure, the change being 
accompanied by the appearance of magnetism. 

E. The Structures of K upon wliich mechanical work of various 
kinds has been done. — (Plates 55 to 58.) Preliminary investigations 
of the structures of the alloys which had been subjected to various 
mechanical tests showed that alloy K is better fitted for a complete 
study than any other members of the series. It has a characteristic 
homogeneous structure in the untreated state. It consists of large 
polyhedra which even after prolonged etching appear white (when 
the black film left by the reagent has been rubbed off). Mechanical 
work of several kinds examined by the authors, which stresses it 
beyond the elastic limit, causes an entirely new type of structure to 
appear, the latter taking various shapes, hut alivays etching dark. In 
other words, what may be termed the simplex structure in the 
unstressed state becomes a duplex structure in the plastic state. 
Further, magnetism appears icith the duplex structure. This well- 
marked structural change lends itself well to photo-micrographic 
reproduction. 

(1) Tensile Test. Forged Bar. — A transverse section as near the 
fracture as possible was chosen. (The section was of course polished 
and etched.) This test may be expected to influence the structure 
equally, the distortion being the same on all parts of a section cut 
at right angles to the direction of stress. Examination showed this 
to be the case, and the distortion, as evidenced by the appearance of 
the black areas, was fairly uniform in all parts of the section. A 
typical field is shown in photo-micrograph, Fig. 53, Plate 57 
However, in one place, near the edge, a larger collection of the black 
material was observed, and is shown in Fig. 54. 

The high-power photo-micrographs, Figs. 55, 56, and 57, show 



954 ALLOYS RESEARCH. Nov. 1905. 

that the black bars are grouped in triangles, in rhombi, as well as 
quite irregularly. 

A section cut from the cast tensile piece, which has a much 
lower ultimate strength than the forged material, and was not so 
long under stress in the plastic state, showed that although the 
distribution of the white and black areas was fairly uniform over the 
surface, yet the two types occur in very much larger patches, 
Fig. 64, Plate 58. A field is shown in which a large yellowish 
polyhedron occurs with only tiny splashes of black structure except 
for one black patch of moderate size, Fig. 6Q. 

2. Compression Test. Forged Bar. — One end of the cylinder 
(barrel-shaped) was filed and polished. Etching developed a series 
of concentric circles. Although the marks left by the turning 
tool were quite obliterated in polishing, the etching reagent 
developed these rings. Accordingly the end was ground down 
about i inch, polished, etched, and the structure examined. All 
traces of rings were absent. From the edge to a distance of about 
-^^ inch inwards the structure was unaltered. Fig. 58, Plate 57. But 
there the darkly etching patches appear and increase to a maximum 
at the centre. Fig. 59. A longitudinal section cut through the 
centre of the cylinder showed the same characteristics. Fig. 60 
is typical of the centre. The structural changes produced by 
compression stress increase to a maximum at the centre. 

3. Torsion Test. Forged Bar. Transverse section cut near the 
fracture. — The central area of the section, where the torsional stresses 
have produced the least strain, shows practically no alteration of 
structure. A few tiny black patches can however be made out, 
Fig. 39, Plate 55. These patches gradually increase from the 
centre to the edge, where they constitute almost the whole of the 
structure. Figs. 40, 41, and 42. The relative positions of these 
photo-micrographs are as shown on Plate 55. 

In positions. Figs. 40 and 41, several prong-like black lines will 
be noted. A high-power photo-micrograph of some of these is seen 
in Fig. 43. Other features are typified in Figs. 44 and 45. 



Nov. 1905. ALLOYS RESEARCH. 955 

4. Bending Test. Forged Bar. — A ticansverse section cut from 
the place where the bar had undergone on one side the greatest 
tension, on the other the greatest compression, was examined. (The 
bar had been bent round 180° and was squeezed together in a vice.) 
Travelling from the tension side of the section, where the black 
patches are by no means numerous, Fig. 46, Plate 56, the latter 
increase continuously until on the compression side they constitute 
the larger part of the structure. Fig. 48. There is no neutral axis 
of the bar where no change has occurred. 

Under the high-power objective white, yellow, and black tinted 
patches were found. The yellow patches do not show under low 
magnification. They probably appear grey under a low-power 
objective. Fig. 50 shows the three differently-coloured patches. 
Figs. 49 to 52 show the structures in various parts at 1,000 
diameters. 

5. Alternating Stress Test. Forged Bar. — A transverse section 
was examined as near the fracture as possible. Typical structures 
are seen in photo-micrographs. Figs. 62 and 63, Plate 58. One low- 
and one high-power photo-micrograph are sufficient to characterise 
the field. The black patches are seen in the form of parallel bands, 
the orientation differing from polyhedron to polyhedron. Under 
high-power magnification twinning is sometimes noticed. 

No cracks were observed. 

6. Forging Test. — That forging produces a darkly etching 
material has been already noticed. Plates 51 and 52. The irregularly- 
shaped black patches are quite uniformly distributed. 

It appears hardly necessary to anticipate a possible objection as 
to the nature of the black patches of various shapes. The authors 
think there is no room for doubt that these are indications of a new 
structural entity produced by mechanical work. It may, however, be 
alleged that the black patches are actual cracks. This objection can 
be met in two ways. Firstly, it is known for certain that the black 
patches cannot be seen before etching — which they could be if they 
were fissures — and that after etching they can be removed by rubbing 



956 



ALLOYS RESEARCH. 



Nov. 1905. 



with chamois leather moistened with benzine. They consist, therefore, 
of a black film. Secondly, a sure criterion of a crack is the 
possibility of focussing from the edge of the crack down the side. 
Nothing of the kind is possible with the black patches. 

That the darkly etching material is neither graphitic nor 
amorphous carbon was shown by taking turnings near the fracture 
of the piece broken by tensile stress and dissolving these in nitric 
acid (specific gravity 1*2). They dissolved without leaving a 
residue. It seems very probable that the new structural constituent 
is a hard amorphous substance produced by mechanical work on the 
soft crystalline material, and that the case is parallel to that of 
silver investigated by Mr. Beilby.* The fact that heat restores the 
white crystalline material from the amorphous is also in conformity 
with Mr. Beilby's results. 

The results of the microscopical examination are summarised in 
the following Table 33 :— 

TABLE 33. 



Test. 


Distribution of darkly etching patches in a 
transverse section. 


(a) Cold Worlc. 
Tensile . 

Compression . 

Torsion . 

Bending 

Alternating Stress . 
(b) Forging. 


Uniform. 

Minimum at Edge. Maximum at Centre. 

Minimum at Centre. Maximum at Edge. 

/Minimum at Tension side. Maximum at Compression 

\ Side. 

Uniform. 
Uniform. 



F. Restoration of the original structure K hy heating the mechanically 
stressed bars to 800° C. (1,472° F.).— That this treatment is effective 



♦ " The Hard and Soft States in Metals." Journal, Faraday Society, June 1904. 



Nov. 1905. 



ALLOYS RESEARCH. 



957 



in the case of the forged bar has already been shown, Plates 51 
and 52. It is also completely eflfective in the case of the bending bar, 
Fig. 67, Plate 59. On examining the "Torsion" section it was 
found that the distribution of the black and white areas was altered, 
the dark patches being at a maximum near the centre. Fig. 70, 
and almost entirely absent at the edge, Fig. 69. The forged 
" tensile section " was almost unaltered. Fig. 68. 

The following sections were next kept at 900° C. (1,652'' F.) for 
two hours, and were afterwards re-polished and etched, with the 



following results :- 



TABLE 34. 



Test. 



Compression. 

Forged Tensile. 

Cast Tsnsile. 
Alternating Stress. 

Torsion. 



Distribution of darkly etching patches. 



I Dark patches absent. 
\ restored. 



White polyhedral structure 



(White polyhedral structure in central area; black 
\ patches otherwise uniformly distributed. 

Distribution uniform. 

Distribution uniform. 

I White polyhedral structure in central area; distribution 
\ of dark and white otherwise uniform. 



Summarising the results, it is seen that the structures induced 
by (a) forging, (6) bending, and (c) compressing, which deformed the 
bars, however, without fracturing them, can be removed, and the 
original types reverted to, by a short heat treatment at 800° to 900° C. 
(1,472° to 1,652° F.) ; while the structural results produced by stresses 
which caused fracture are not so easy to remove. However in the 
case of the forged tensile and torsion sections there was an 
appreciable reversion in parts towards the original polyhedral 
structure. It seems probable therefore that a prolonged heat 
treatment at a temperature of about 900° C. (1,652° F.), would cause 
a complete restoration in every case. 



958 ALLOYS RESEARCH. Nov. 1905. 

The authors' results are thus not in agreement with those 
obtained by M. Guillet, who affirms (pages 942 and 948) that " a steel 
which has once been changed from the polyhedric to the martensitie 
condition, by whatever method, cannot be regenerated either by 
annealing or quenching." 



Concluding Eemarks. 

In writing this Eeport the authors have endeavoured to present 
the results in such a form that any particular property of any one 
alloy can be referred to with as little loss of time as possible. To 
this end, the detailed index, given at the beginning, has been 
compiled, the properties being classified under broad headings. 
Further, in any section dealing with any one series of tests, the 
results are illustrated as fully as possible by Tables, figures, curves, 
or photo-micrographs as the case may be, and where possible, a short 
summary is given. 

They venture to hope that the results contained in the 
" Mechanical " and " Heat Treatment " sections will be of some 
practical value, and that those in the " Physical " (including the 
" Metallographical ") section will have both a practical and 
theoretical interest and importance. 

In conclusion, they hope that the foregoing research will not be 
found unworthy to take a place among the previous researches 
presented to the Alloys Eesearch Committee. They acknowledge 
with much pleasure the interest taken by Dr. Glazebrook during its 
progress. They are indebted to several members of the National 
Physical Laboratory staff for much willingly-given advice and 
assistance. Special mention must be made of Dr. Stanton and Mr. 
Jakeman of the Engineering Department, who have assisted in the 
majority of the mechanical tests, who have had to bear the brunt of 



Nov. 1905. ALLOYS RESEARCH. 959 

the difficulties associated with the preparation of test- pieces of 
the hard alloys, and to whom the sole credit for the results of 
the Alternating Stress and Shock Tests is due. For the rest, the 
magnetic permeabilities were determined by Mr. Campbell ; the 
resistivities by Mr. Melsom ; and the coefficients of dilatation by 
Mr. Attwell. In the chemical investigations the authors have been 
assisted by Mr. Eichardson and Mr, Kobinson. To those gentlemen 
the authors' cordial acknowledgment of services rendered is due. 

The Paper is illustrated by Plates 50 to 59 and 22 Figs, in the 
letterpress. 



Discussion on Friday, 17th November 1905. 

The President remarked that, before opening the discussion on 
this most important Eeport which had just been read and which 
represented an enormous amount of work, it was his duty and great 
pleasure to propose a hearty vote of thanns to the authors, which 
he hoped would be carried by acclamation. 

The resolution was carried by acclamation. 

Dr. H. C. H. Carpenter stated that he would like to show a few 
slides connected with the Report, because the photo-micrographs 
in Plates 51 and 52 had been reduced and compressed in reproduction, 
and had lost some of their clearness. 

The structure seen in Fig. 23, Plate 51, B (forged normalised), 
was that referred to in the Eeport as pearlitic, the dark patches 
being pearlitic and the white patches ferritic. 

Fig. 24, Plate 52, B (forged normalised), was a photo-micrograph 
of a small portion of the field in Fig. 23 at a magnification of 



960 ALLOYS RESEARCH. Nov. 1905. 

(Dr. H. C. H. Carpenter.) 

600 diameters. The indistinctness in the banded structure was due 
to the presence of manganese in a notable amount. 

Fig. 23, Plate 51, F (forged), was typical of the steels referred to 
as martensitic, and the structure instead of being duplex was rather 
more what might be termed simplex. 

Fig. 24, Plate 52, F (forged), represented a piece of the field seen 
in the last Fig. at a magnification of 500 diameters. 

Fig. 24, Plate 52, J (cast normalised), was a photo-micrograph of 
the alloy containing about 16 per cent, of nickel, in which the light 
and dark types of martensitic structure were clearly seen. 

Fig. 25, Plate 53, represented the type of steel referred to as 
polyhedral. This structure was developed with some difficulty, the 
etching requiring something like 30 minutes, whereas in the other 
cases of the pearlitic and martensitic types it was a matter of only 
at most two minutes. The material was practically homogeneous ; 
the lines, which somewhat resembled ripple marks in sand, were due 
to unevennesses left on the surface by the prolonged action of the 
etching fluid. 

Fig. 23, Plate 51, K (cast normalised), was comparable with B on 
the same plate. This was the structure of the alloy containing 
20 per cent, of nickel in the cast state. 

Fig. 23, Plate 51, K (forged), showed the effect of mechanical 
work which stressed alloy K beyond the yield point. In this case 
it was the work due to forging, and the black structural constituent 
which made its appearance after the action was well seen. 

Fig. 23, Plate 51, K (forged normalised), represented the structure 
of the forged material as seen in K (forged) subjected to a 
temperature of 800° C. (1,472^ F.), which caused a complete 
reversion of the black amorphous into the white crystalline 
material, but it did not do away with the mechanical effects of 
forging in breaking up the large crystals of the cast alloy into 
comparatively small ones. 

In conclusion, he wished to say that the originals of the 
photographs and the specimens which had been treated in the 
tensile, torsion, impact, and compression tests, were on exhibition 
in the meeting hall. 



Nov. 1905. ALLOTS RESEARCH. 961 

Mr. R. A. Hadfield said that, whilst having been concerned in 
what might be termed the very important research under discussion, 
he would like to take the present opportunity of offering a tribute 
to the admirable work of Dr. Carpenter and Mr. Longmuir. The 
very clear manner in which they had presented the results made 
them of unusual value. Although their work was founded upon 
his own alloys, it was to them he must warmly offer his 
congratulations, and to them was the great credit of the research 
due. He did not accept all the theoretical conclusions stated in 
the Report, but he was sure it would be understood that that was 
quite natural, as in a research of that kind, which covered so much 
ground, there were necessarily very many conclusions to be drawn, 
and in some cases different interpretations of the same facts. As 
the Report mentioned, the experiments had been carried out at the 
National Physical Laboratory, whose very able head, Dr. Glazebrook, 
had assisted the authors in every possible way. He was sure all the 
members would agree that the work done was but another proof of 
the great value of the National Physical Laboratory, and in a sphere 
of work which was peculiarly its own. They had seen in certain 
papers what he thought were, to some extent, uncalled-for remarks 
regarding the National Physical Laboratory. He could personally 
bear his own testimony, from what he had seen going on there, that 
it was indeed an institution of which the country not only had the 
greatest possible need, but of which it ought to be proud ; and he 
was sure Dr. Glazebrook must have given an immense amount of 
time to organising such an admirable laboratory in so short a time. 

The research into the series of nickel alloys under discussion 
was of considerable interest, owing to the percentage of carbon 
present being somewhat a peculiar one. As Dr. Carpenter had 
pointed out, there had been previous researches, but none of them 
had contained that particular percentage of carbon as in the 
present series of specimens, which it would be noticed varied from 
0*40 to 0*52 per cent. He had the pleasure of a steel-maker 
sitting near to him, and he thought that gentleman would admit 
that the variation in carbon was comparatively small for such a 
considerable series of steels. There were something like eight or 

o T 



962 ALLOYS RESEARCH. Nov. 1905, 

(Mr. E. A. Hadfield.) 

ten different casts, and lie thought that percentage of carbon was a 
proof that the utmost had been done to keep this element as constant 
as possible. The percentage of carbon also differed from his own 
former series, of which he presented particulars to the Institution of 
Civil Engineers,* in which the carbon was about 0*2 per cent., 
and also from M. Guillet's nickel-iron series containing • 8 per cent, 
of carbon. Moreover, he was not aware of any previous series, 
undertaken either in this or any other country, which showed the 
remarkable test which had been obtained from material in its cast 
form. He referred particularly to Table 17 (page 899). Being 
specially interested in that particular form of steel — that is, the 
cast condition, he could only say that Table 17, from his own point 
of view, contained a really remarkable set of tests. He never 
remembered seeing before in any research the extraordinary 
tenacity of 75 tons obtained from a material which had never been 
forged or treated mechanically. It would be seen that specimen G 
had a tensile strength of 74*03 tons, and that specimen J had a 
tenacity of 76 • 55 tons. It was also somewhat remarkable to note, 
as Dr. Carpenter pointed out, that there was quite an appreciable 
percentage of elongation, varying from 4 to about 6 per cent., 
showing that steel even in this form was not entirely wanting in 
ductility. 

Dr. Carpenter referred (pages 872-873) to the peculiar manganese- 
nickel-iron alloy which he (Mr. Hadfield) had described more 
fully elsewhere ; f but in passing he might mention that one of the 
peculiarities of that alloy was tbat it offered, he believed, the 
highest known electrical resistance, namely, over 90 microhms, and, 
of course, it was equally as bad a conductor of heat. 

He understood that the statement on page 956 might be claimed 
as "supporting the allotropic theory, and if that was so he must 
dissociate himself from such an opinion, that is, he could not 






* ProceediDgs, Institution of Civil Engineers, 1898-99, vol. cxxxviii, 

page 1. 

t Proceedings, Institution of Civil Engineers. Engineering Conference, 
1903. Supplement to vol. cliv, page 118. 



Nov. 1905. ALLOYS RESEARCH. 963 

accept any tlieory whicli involved the acceptance of an adamantine 
form of pure iron. He did not think, however, that Dr. Carpenter 
for one moment meant the adjective " hard " in the tenth line on 
that page to bear such an interpretation, but if he did, he 
(Mr. Hadfield) did not quite agree with that point. He had 
received a letter from his friend, M. Osmond, with regard to that 
question, but did not think that Mr. Beilby's researches could be 
construed into the particular acceptance which M. Osmond claimed. 
There might be material alteration in the character of the particular 
form of metal present, in that case iron, by mechanical work, yet he 
could not think that it altogether bore the interpretation claimed. 

There was only one other matter to which he would like to refer, 
namely, that since the research was started he had had the pleasure 
of preparing tensile bars from all the specimens of the series, and 
he had had them tested at liquid-air temperatures in Sir James 
Dewar's laboratory. Some of the members were probably aware 
that he had presented a research on that question, but at the time he 
did so the particular specimens he mentioned were not available, and 
were now therefore included in this Report. The results he obtained 
from this series were so remarkable that he thought he might be 
pardoned for taking up the time of the meeting for a few moments 
in referring to them.* Taking specimen B, which contained 
the lowest percentage of nickel, under the National Physical 
Laboratory test, a tensile of 41 tons with an elongation of 21 per 
cent, was obtained. At liquid -air temperatures the very remarkable 
increase was obtained up to 75 tons per square inch, and, of course, 
with considerable diminution in elongation, 7J per cent. Specimen 
C rose to 95 tons with 13 per cent, elongation ; and specimen F, 
containing about 6*42 per cent, nickel, rose to the extraordinary 
tenacity of 142 tons per square inch with 2 J per cent, elongation. 
That, he thought, was a remarkable result, but still more so 
was specimen J containing 15*98 per cent, of nickel. The tenacity 
of that material at the ordinary temperature was 80 tons with 5 per 



* The results here referred to are recorded in Tables 11 and 12 (pages 878-881, 
and 883). 

3 T 2 



964 ALLOYS RESEARCH. Nov. 1905. 

(Mr. R. A. Hadfield.) 

cent, elongation ; but the remarkable rise was obtained, at liquid-air 
temperature to 144 tons with 2J per cent, elongation; and specimen 
K, with 19 • 9 per cent, nickel, rose from 44 tons to 157 tons per 
square inch ; that was exceedingly remarkable. There was an 
extraordinary diminution in the tenacity at ordinary temperatures^ 
which dropped from 80 tons in specimen J to 44 tons in specimen K^ 
while the elongation increased from 5 to 55 per cent. ; this was more 
ductile than iron. But at liquid-air temperatures specimen K rose 
from 44 tons tenacity to 157 tons per square inch, and the 
elongation, although dropping considerably, still remained the very 
high figure of 15 J per cent. He thought the President, from 
his knowledge of steel, would admit that this was indeed a 
remarkable result. 

Sir William H. White, K.C.B., Past-President, remarked that 
the Keport in its title was described as the Seventh presented to the 
Alloys Research Committee of the Institution ; and although the work 
had been done at the National Physical Laboratory, it was a great 
honour to the Institution to have been able to assist in its prosecution. 
When he became Chairman of the Committee, some years ago, he 
discussed with the late Sir William Eoberts- Austen the desirability 
of transferring the work from the Mint to the National Physical 
Laboratory, as soon as that Institution was on its feet. Sir William 
Roberts-Austen was very desirous of that transfer being made, but 
before it was actually accomplished Sir William's death occurred ► 
It was in Sir William Roberts-Austen's mind, however, that one 
important section of work at the National Physical Laboratory ought 
to be the further prosecution of the research that he had undertaken 
on behalf of the Institution ; and the Council of the Institution, 
recognising the desirability of this arrangement, had assisted the 
National Physical Laboratory by means of an annual grant for 
purposes of research, and so might claim a special interest in the 
very valuable results obtained. Whatever Mr. Hadfield might say,, 
the Institution did not intend to take him oif the list of those who 
had carried out the research. His previous work had indicated 
important lines which had been developed during the inquiry, and 



Nov. 1905. ALLOYS RESEARCH. 9G5 

his association with it, not merely in the provision of the alloys 
themselves, but in the conduct of the investigation, had been of 
immense value throughout — a view which Dr. Carpenter would 
willingly support. They all regretted Mr. Longmuir's absence. 
Mr. Longmuir was recommended to Dr. Glazebrook by Professor 
Arnold, so that they had the pleasure, in the later stages of the 
research on the alloys of steel, of having Professor Arnold heartily 
with them, and finding for them a man who had proved of very great 
value, as Dr. Glazebrook, Dr. Carpenter, and Mr. Hadfield all agreed 
that the choice of Mr. Longmuir was amply justified by his 
oontribution to the research. He was very sorry Mr. Longmuir 
was not present to give the members the benefit of his broad 
-conclusions. 

With regard to the nature of the research itself, it would ill 
•become him to attempt to say much ; but there were certain features 
which any outsider like himself, interested in all metallurgical 
processes, could not help remarking. The Eeport furnished a 
wonderful confirmation of many conclusions reached by previous 
investigators of nickel steel ; and although it might, and did greatly 
■extend what had been before available, and although it might 
correct in some respects certain details where previous investigations 
had been only partial, yet, in the main, the present Eeport 
substantiated most of the special features and characteristics of 
mckel steel to which Mr. Hadfield and others had drawn attention 
in previous researches. Some qualities of the material were more 
clearly defined and amplified in a way that could not fail to prove of 
great value to all makers and users of steel. He spoke as one who 
had been a considerable user of steel; and it was of the greatest 
interest to note the possibilities which the steel-maker might be able 
to offer to the constructive engineer in the future. Some of the 
alloys, he presumed, at present hardly came within the region of 
practice because of their cost. But there they were, and if special 
materials were required for special purposes, the user knew now 
where to look for them. In addition, there was the fact that the 
investigation had proceeded on definite lines. Mr. Hadfield had 
epoken of the special percentage of carbon that was employed, and 



966 ALLOYS RESEARCH. Nov. 1905. 

(Sir William H. White, K.C.B.) 

of the almost constant percentage of carbon tliat was maintained. 
In the Report itself it would be seen that allusion was made to the 
comparatively small variations in the percentage of manganese ; 
that is, in making the alloys it was endeavoured simply to vary the 
nickel and to keep the other constituents close to an average value. 
There could be no doubt that it was a true method of research not to 
vary many conditions simultaneously, otherwise it was impossible to- 
determine what was due to each individual change. If he might 
venture to say so, if any further investigation of the alloys of steel 
were undertaken by the Institution, he would like to see nickel 
steels tested with variations in the manganese content. He had 
seen a good deal of steel made and tested, and was of opinion that 
it might be possible to obtain important results at less cost for 
certain purposes. He said that with hesitation in the presence ot 
steel-makers like the President and Mr. Hadfield, but that was his 
conviction, and possibly it was not out of place to mention the 
matter. 

Another point to which he would refer was of great importance 
to the particular branch of engineering he had been connected withy 
namely, the question of corrosion. The tests recorded in the Report 
were made under certain defined conditions which were clearly 
stated ; everyone could therefore judge for himself how closely they 
might approach the conditions of practice. He ventured to suggest, 
however, that from the point of view of the user, and, in fact, from 
the practical point of view, no good purpose was served by expressing 
losses by corrosion as percentages of total weight. Corrosion was 
essentially a surface process, and what was required in any structure 
was to know how the structure would wear by corrosion from the 
surface. One could translate the results given in the Report into 
rates of surface wear, but if Dr. Carpenter would add to the Tables- 
comparative tests on similar-sized specimens of ordinary mild 
steel similarly treated, and give the relative losses of mild and 
nickel steels per unit of surface in a given time, he would confer 
further obligations on engineers. The claim to protection froni 
corrosion in nickel-steel had been made with great persistence. No 
doubt it was true in some degree, but what engineers wanted to know 



Nov. 1905. ALLOYS RESEARCH. 96/ 

for many purposes was what degree of advantage nickel-steel 
possessed. He remembered some years ago, in connection with tlie 
manufacture of water-tube boilers, there had been certain difficulties, 
and the question arose as to whether the use of nickel-steel tubes 
would overcome " pitting." Some nickel-steel tubes were ordered 
and delivered, and the steel-makers were confident that they would 
fulfil all requirements ; but as the tubes were rusty when they were 
unpacked the user did not think they were likely to be non-corrodible 
in the boilers. That was an illustration of how a man introducing a 
novelty might probably take a sanguine view of its advantages. An 
important and careful investigation of the relative rates of corrosion 
of nickel-steel and ordinary steel was still badly wanted, and if the 
facts could be added to the Eeport, the members of the Institution 
would be more deeply indebted to the authors. 

Lastly he would like to refer to the point Mr. Hadfield had 
raised as to heat treatment. For many years he had been looking 
forward to the time when steel-makers would produce, by chemical 
and physical means, results equal to those produced hitherto only or 
chiefly by means of mechanical work. He had every confidence that 
that would be accomplished eventually ; when it was done, there 
would be little short of a revolution in steel structures. No one 
could look through Table 12, to which Mr. Hadfield had referred, 
and compare the behaviour of cast specimens which had been treated 
in a very simple manner with the behaviour of specimens which had 
been forged and largely reduced in sectional area, without seeing 
how far steel-makers had already advanced towards realising that 
ideal. The Report was really a treasury of information. Not 
merely was it a credit to the authors and to the Laboratory where 
the work had been done, but it would long remain the standard 
authority on the alloys of nickel and steel, and be quoted in all 
discussions on the subject. 

Professor Thomas Turner said that the previous Eeports of the 
Alloys Research Committee had been so widely read throughout 
the world by those who were interested in metallurgical subjects, 
and had been so very much appreciated, that if the members said 



9G8 ALLOYS RESEARCH. Nov. 1905. 

(Professor Thomas Turner.) 

that the present Eeport maintained the standard which was set in 
previous Eepoits they paid it a very high meed of praise. He felt 
sure that all who had read the Eeport carefully could say that it 
was in no way behind those that had appeared previously in 
connection with the work of the Institution. The Eeport, as had 
been said, was very full of information. It was difficult to summarise 
it in a word or two, but there were three things which struck him 
as being very important in the conclusions that were brought 
forward ; and, although they were not altogether new, still their 
knowledge had been amplified and made more definite by the 
Eeport. The fact was well illustrated in the diagram that there was 
a brittle zone, with nickel alloys, the position of which depended 
upon the proportion of carbon, and that with each proportion of 
carbon there was a definite brittle zone. Corresponding to 
that, there were, as represented in the photo-micrographs. Plates 50, 
51, and 52, and beautifully illustrated in the slides thrown on the 
screen by Dr. Carpenter, the three characteristic structures that 
were plainly shown. Thirdly, there was the remarkable effect on 
the position of the arrest points in heating and cooling, as shown 
in Fig. 18 (pages 918-919) and Fig. 21 (page 932), that as the 
proportion of nickel increased so the point of arrest on cooling was 
lowered, until the important result was obtained that ultimately 
it was below 100° C. That, of course, had been taken advantage 
of by the authors in the treatment of those special steels which were 
so hard to cut and difficult to machine. It was noticed more than 
fifty years ago — he was not sure that it was not nearer a hundred 
years — that certain high carbon steels could be partly tempered by 
plunging them into boiling water ; they were therefore not unprepared 
to find that steels would be sensitive to relatively low temperatures. 
But he did not know that any accurate information was available, 
similar to that brought forward in the Eeport, showing so low an 
arrest point on cooling, and at the same time so wide a range 
between the points of arrest on heating and on cooling respectively. 
He was sure the members were very much indebted to the authors 
for I ringing that important information before them. 



Nov. 1905. ALLOYS RESEARCH. 960 

He hoped if future work of the kind were carried out a somewhat 
larger quantity of material would be used for the tests. The weight 
of the ingots employed was, he believed, about 45 lbs. ; and though 
the authors had shown that the experiments had not materially 
suffered from shortness of metal, still that quantity was small, and 
in some cases it would no doubt have been an advantage to have 
been able to duplicate the tests. But there was another aspect to 
which he wished to direct attention, namely, that a large amount of 
skilled work, analytical, microscopical and physical, had been 
devoted to that small quantity of metal. He thought it would be 
a great advantage if two or three times as much material were 
obtained, and if the Institution could retain some of the metal, so 
that at any future time, if it were requisite to develop the 
research and introduce new methods, there would be a supply of 
material, which had been investigated with the highest scientific 
knowledge of the present time, so that students ten, fifteen, or 
twenty years hence might be able to test the same material further. 
That was done some little time ago by an International Committee 
in connection with chemical tests of steel, and there were now 
available in this country samples of steel, of which he was the 
custodian, of known composition, with various proportions of 
carbon : and the British Association's standards had proved very 
useful in several cases in questions of appeal, or in working out 
new processes. The cost of the extra metal would be extremely 
small. 

He was interested in another matter mentioned in the Report, 
namely, the testing of the hardness of the samples. The first 
method employed for testing the hardness was by means of a scratch 
(page 891). As he had used a somewhat similar process for 
something like twenty years, with a special instrument called a 
sclerometer, he would like to say a word or two on the matter. 
The authors attempted to measure the width of the scratch by 
means of a microscope, after having made a scratch with a uniform 
load, using a dividing engine. He had never known any microscopic 
method of dete