T H E CYCLES OF GAS AND OIL ENGINES JAMES D, ROOTS, ;=©s=&:=85=8=: JFmnxng anb ITabor. LIBRARY Universityof Illinois. CLASS. BOOK. & 2. 1 .V "V\CA VOLUME. A Accession No. ... THE CYCLES OF HAS AND OIL ENGINES. BY JAMES D. ROOTS. ALL RIGHTS RESERVED. LONDON : OFFICE OF THE ENGINEER, 33, NORFOLK STREET, STRAND, W.C. 1899, m PREFACE. So far as I am aware, in no published work has any attempt been made, either in this country or abroad, to deal with the cycles of internal combus- tion engines, much less to deal with them in their relation to one another ; and only indefinite and perfunctory attempts at classification of these engines have been made in the various works on the subject. It is therefore hoped that this little work will be found useful by engineers. Considerable additions have been made since the matter appeared in the pages of The Engineer. JAS. D. ROOTS. London , November , 1898. 63532 Digitized by the Internet Archive in 2016 with funding from University of Illinois Urbana-Champaign Alternates https://archive.org/details/cyclesofgasoilenOOroot THE CYCLES OF GAS AND OIL ENGINES. CHAPTER I. Introductory. The words “internal combustion engines” contain a fairly approximate descriptive definition of gas, oil, and spirit engines ; and for the later period of these motors — for the whole may be conveniently divided into two periods, that before and that subsequent to the introduction of the Otto engines — this title is possibly the best that can be devised, yet as these words will necessarily include gunpowder, gun-cotton, and all similar fulminating engines, it is not sufficiently accurate for a work upon the cycles of these motors. The words “ explosive engines ” are excluded for the like reason, for in using gunpowder and gun-cotton the working pressure is produced by the decomposi- tion of the explosive into gases, and the presence of air is not essential to its operation. In gas and oil engines air is essential to the decomposition of the explosive. The old title, therefore, of the Patent-office, “ air and gas engines,” is a very correct one, so far as it goes. To include powder and gun-cotton engines in this category would be not only to widen the field of the subject considerably — and it is a sufficiently wide one as it is — but it would take us back to the year 1256, and ascribe to Roger Bacon, or else some ancient and unknown Chinese worthy, the invention of the gas engine. The words “ gas motor ” contain in themselves a sufficient definition to exclude powder engines. M. Witz has endeavoured to show in a manner characteristically French that the Abbe Hautfeuille was the inventor of the gas engine, because he published some ideas upon a gunpowder motor. Upon the same reasoning, the cannon used at Crecy were equally if not more practical and workable gas engines. On the other hand, the Barber patent, which is given as the earliest invention in the English works upon the subject, is only a fanciful, and as far as I can judge, as imprac- ticable a motor, from the modern standpoint, as the steam engine of Hero of Alexandria. I have therefore excluded from this book all powder and gun-cotton engines, although those in Class 3 sometimes approach closely to the hot-air engines, and some of them are decidedly on the borderland between the internal combustion engine and hot-air engine. I have also excluded from consideration all engines not possessing a cylinder and piston. The classification of the cycles of gas and oil motors necessitated the reading through carefully, in addition to the known works upon the subject, the patents dealing with the subject of cycle from the earliest time up to the end of the year 1894 — a task I should have recoiled at had I realised the necessity for it before having collected many notes and written considerable portions of the book. The inventive faculties of inventors have worked in such riotous profusion in the matter of cycles, the variations have been so many, so wide, and so ingenious, that it is often a matter of extreme difficulty to arrive at what the inventor means. I had thought that I might have adopted the classification in one of the known works upon this subject. In the “ Gas Engine,” by Mr. Dugald Clerk, and in his paper read before the Institution of Civil Engineers, upon “ Recent Developments in Gas Engines,” these engines are divided into three types, which are in effect somewhat similar as to the division of the classes to those I have adopted ; but on an examination of Mr. Clerk’s definition of type 1 — “ engines igniting at constant volume without compression ” — it will be seen that though this is true as a rough theoretical approximation, yet it is impossible in practice to ignite at constant volume. The flame takes an appreciable interval of time to pass through the explosive mixture of gas and air, during which interval the piston is moving at its greatest velocity ; the volume during ignition is therefore continually changing. If the ignition in this cycle were to take place at the dead point of the stroke, it might be possible to effect ignition at almost constant volume ; but taking place at from one-third to one half of the stroke, there must always be an appreciable relative interval of time between the commencement and the completion of the ignition. Flame passes through a mixture of gas and air at atmospheric pressure much less rapidly than through a compressed mixture. The diagram, Fig. 1, is an indicator card of this type of cycle, taken from aHugon engine, by Professor Tresca, of Paris. The time taken to complete the ignition is clearly shown by the rise of the ignition line. If the ignition had been at constant volume, the ignition line would have been a vertical one, as it is in the indicator diagram of the atmospheric engine. Hirn says, in “The Theory of the Lenoir Gas Engine,” “ the gas mixture not only requires a certain time in order to become ignited, but there must be upon the beginning of the working stroke a certain interval ere the flame enters.” With regard to type 3, the definition, “ engines igniting at constant volume ( 7 ) with previous compression,” is rather more accurate than that of type i, as the ignition is usually upon or near the dead point in this class. There are, how- ever, many engines that could not be included in either Mr. Clerk’s or M. Witz’s classification. It appears to me advisable to classify the engines by revolutions rather than by stroke, because this classification by stroke of the engine ignores the strokes of the pump, an essential part of the cycle, the so-called two-stroke engine being generally in reality a four-stroke engine. On the other hand, if it be called a four-stroke engine, it is placed in the same category as engines of the de Rochas cycle, which would be clearly wrong. The British Patent-office now classifies the compression engines by strokes, as also does M. Witz and M. Richard. Such well-known engines as the Stock- port, Clerk, Midland, and engines of this class, have in this classification their two pump strokes ignored, and these really, although the operations are carried out in two cylinders, are four-stroke engines. The Trent gas engine and engines of that construction are also of the same type, as they really possess two pistons in two cylinders of different diameter placed tandem. My first intention was upon these grounds to classify the cycles by the number of revolutions per cycle solely, but the former objection would apply equally unless some explanatory subdivision were made. The classification I have finally adopted is shown in the table, page io. It will be seen that so far as the first classification goes, the three main classes are in effect somewhat similar to those of Mr. Clerk and M. Witz. Mr. Clerk, however, places atmospheric engines as “ type ia,” while M. Witz places them in a separate fourth class, together with others which he describes as “ mixed.” In 1884 and 1885 the British Patent-office classified internal combustion engine cycles by revolutions, and although this is undoubtedly a good method, yet the classification was coupled with a system of abridgment that formed a very elusive paradise for the searcher, and the system was subsequently very properly discontinued. The Atkinson cycle engine would have presented a difficulty under the classification, as it is a four-stroke engine, yet is not upon the de Rochas cycle, as the strokes are of different length. The two extra or pumping strokes are obtained in one revolution and by the one piston by an ingenious mechanical device ; it is therefore classed with the one-revolution cycles. Of course it may be urged that an engine consists of mechanical devices, nevertheless it is clear — and this may be taken as a definition of cycle — our consideration in every case must be the disposition of the working fluid in the engine and the processes it is subjected to before and after combustion. I have endeavoured in classifying to follow the line of fewest objections, and the classification followed was adopted after having tested and rejected five other systems of classification which I had drawn up. r 2 Class i contains all non-compression engines. It is divided into two types, of which the first includes those engines in which “ power is developed directly by explosion,” and the second those in which “ power is produced indirectly by atmospheric pressure.” Class 2 contains all compression explosion engines. These are divided into six different types of cycle, three of which are completed in one revolution, the first being “ with the aid of a separate pump or pumps,” type 3 ; the second “ with the use of the opposite side of the working piston as a pump,” type 4; and the third “ without a pump,” type 5. There is a subdivision of cycles of two revolu- tions, the well-known de Rochas or Otto cycle being the first of the two columns, type 6 ; the second, type 7, a modified de Rochas cycle, having a greater expan- sion, ie., the working charge is expanded to a larger volume than it occupied at atmospheric pressure before ignition, and is so expanded in the same cylinder ; these form types 6 and 7. The cycles of three revolutions are included in type 8, but neither the engines nor patents having three revolutions are numerous. At present the patents coming under type 9 are also far from numerous, although in all probability the immediate future will see them enormously increased, when possibly some further subdivision will have to be made. Type 9 includes all inventions described in patents in which compounding is the leading idea of the specification. By compounding I mean the expanding the working charge in a second expanding and contracting chamber after ignition, and generally after it has done one or a portion of one working stroke. There is no word in the English language that expresses an expanding and contracting chamber simply ; the word “ cylinder,” unless it be expressly excluded, includes the combustion or clearance space also, which in some engines — the earlier Otto to wit — may include one-third more than the total cubic space swept through by the face of the piston in one stroke. It is obvious that the compounding or further expansion may take place either in another cylinder or in the same cylinder on the other side of the same piston, i.e., not in the same contracting and expanding chamber although in the same cylinder. This type only includes those engines in which a further expansion of the working charge after ignition takes place in a second “contracting and expanding chamber,” and not those in which a greater expansion, beyond that of the normal cycle, occurs in the same “ contracting and expanding chamber.” These latter are dealt with in type 7. Lastly, of Class 3 there is only one type or division. The engines of this class are of the continuous combustion type. There are many of them on the borderland between the internal combustion engine and the hot-air engine ; there is, in fact, no very sharp line of demarcation. All internal combustion-engine cycles are thus divided into ten different types — under one or the other of which types I believe any gas or oil engine that has yet been invented would be classed. In the columns of the Table are set down the chief patents of each year in any way relating to cycles since the commencement, 1794 up to the end of 1894. It must not be supposed that the lists contain all the patents of each year. They contain practically all the patents that relate to each type of cycle. The names marked with an (e) are not taken from British patents, but from French works on gas and oil motors — chiefly those of M. Witz and M. Richard, to whom I am indebted for most of these names. Those names marked (c) refer to specifications in which, a special effort is made by the inventor to clear out all the products of the previous combustion, so as to ignite a practically pure charge. The earlier specifications are frequently described as “ Gas and Oil Motors,” but those marked (b) refer to petroleum engines. Those marked (d) are those specifications in which a varying volume is compressed and ignited, or in which one of the chief points of the specification is an intention on the part of the inventor to vary the quantity of the charge compressed and ignited. Between the years 1830 and 1850 there are some patents not included in the Table, most of which would come under types 1 or 2, in which a charge of H 2 and O is exploded by an electric spark in a cylinder without compression. It is not to be understood that there are precisely the same cycles in each type of the Table ; indeed, the cycles often vary considerably in each type — particularly those of type 5, “one revolution without pump.” No doubt it will be found that I have made some omissions in so long a task, but I believe that I have in the Table a very large majority of the specifications and inventions referring to cycles of internal combustion engines between the year 1794 and the end of 1894. ( IO ) CLASSIFICATION OF OIL AND GAS ENGINE CYCLES INTO TEN TYPES— 1794 TO THE END OF 1894. (a) Are same cycle, (b) Petroleum engines, (c) Efforts to clear out all products, (d) Vary- ing volume ignited, (e) Names from foreign works, (f) First engine in which an ignition tube is used, (g) Coal dust or powdered fuel. CLASS 1. Non-compression Engines. Type i. — P ower developed directly by explosion. Street 1794 Wright 1833 Balestrino 1855 Torassa 1856 Lenoir i860 Hugon 1865 Kinder and Kinsey 1867 Errani and Anders 1872 Hock 1874 Bischopp 1875 Ravel (e) 1878 Benier and Lamart 1881 Ord 1881 Alcock 1881 Sambart 1881 Haigh and Nuttall 1882 Hutchinson 1882 Andrew . . 1883 Economic Motor Company . . . . 1883 Dawson 1885 Hartig 1885 Economic Motor Company . . . . 1885 Brine 1886 Dawson 1886 Taylor 1886 Dawson 1887 Laviornery (e) 1888 Wilkinson 1892 Schoenner 1894 Type 2. — -Power produced indirectly by atmospheric pressure. Brown 1823 Brown 1826 Johnson 1841 Bolton 1853 Schollick 1853 Barsanti and Matteucci pro vis. . . 1854 Hugon . . i860 Barsanti and Matteucci 1861 Hugon 1863 Wenham 1864 Otto 1863 Langen and Otto 1866 Gilles 1867 Fogarty 1873 Turner 1873 Soderstrom and Dick 1870 Gilles 1874 Kirkwood and others 1874 Daimler 1874 Dixon 1875 Crossley 1875 Daimler 1875 Hallewell 1875 Linford 1876 Hallewell 1876 Robson 1877 Crossley 1878 Turner 1879 Turner 1880 Francois (e) 1882 Robinson 1882 Schweizer 1883 Butterworth 1884 Dewhurst 1884 Hopkins 1884 Hill 1884 Lowne 1889 Vermand 1890 Terwhella 1891 '( II ) CLASS 2. Compression Engines. Cycles of One Revolution. Type 3. — With the aid of a separate pumping piston. Lebon 1801 Barnett 1838 Newton (Drake ?) (f) 1855 Million 1861 Simon 1877 Clerk 1878 Wittigand Hees . . . . 1878 Koerting Lieckfeld (e) . . 1879 Rider 1879 Linford 1879 Simon 1879 Rider 1880 Otto 1881 Clerk 1881 Crossley and Holt 1881 Atkinson 1881 Clerk 1882 Atkinson 1882 Clerk 1882 Andrew 1883 Niel 1883 Koerting and Lieckfeld 1883 Martini 1883 Skene 1884 Eteve and Brama 1884 Atkinson 1884 Clayton 1884 Weatherhogg 1884 Sombart 1884 Browett 1884 Atkinson 1885 Simon 1885 Andrews 1885 Butterworth 1886 Nash 1886 Welch and Rook 1886 Shaw 1886 Fielding 1886 Niel 1886 Gas Motoren Fabrik Deutz . . . . 1886 Nydpruck and Heyne 1886 Clerk 1886 Murray 1886 Robson 1886 Taylor 1886 Wordsworth and Wolstenholme . . 1886 ' Priestman (b) 1886 Niel (d) 1886 Wright and Charlton 1886 Davey 1887 Cycles of One Revolution. Type 3. — With the aid of a separate pumping piston. Chart and others 1887 Sturgeon * 1887 Wordsworth 1887 Williams 1887 Ravel and Breitmayer 1887 Sturgeon - . . . . 1887 List and Kosakoff 1887 Adams 1887 Beeckey 1887 Schmid and Beckfeld 1887 Gaze 1888 Campbell 1888 Purchas and Friend (b) 1888 Butler 1888 Royston 1888 Richards (d) 1888 Simon (Trent Gas) 1888 Williams 1889 Hoelljes 1889 Covert 1889 Taylor 1889 Theermann 1889 White and Middleton 1889 Clerk 1889 Bull 1889 McAllen 1889 Proell 1890 Beckfeld and Schmid 1890 Binns 1890 Williams 1891 Wertenbruch 1891 Miller (b) 1891 Barclay 1891 Higginson 1891 Connelly 1891 Higginson 1892 Robert 1892 Webb 1892 Clerk 1892 Gas Motoren Fabrik Deutz . . . . 1892 Von Oechelhausen and Junker . . . . 1892 Davy 1893 Martin 1893 Smethurst, Vickers, and Rogers . . 1893 Garner and Sherrin 1893 Dickinson 1894 Benier 1894 Maccallum 1894 Farmer 1894 ( 12 ) CLASS 2 ( continued ) Cycles of One Revolution. Type 4. — Use of the opposite face of the working piston as a pump. Robson 1877 Leichsenring 1878 Robson 1879 Robinson 1880 Williams and Malian 1880 Fielding 1881 Emmet 1882 Porteous 1882 Maxim 1883 Nash 1883 Shaw 1884 Steel and Whitehead 1884 Wright and Charlton (b) 1885 Hale 1885 Ravel (e) 1885 Royston 1885 Deacon 1886 Deacon 1886 Hosack (c) 1887 Ravel and Breitmayer 1887 Hale (d) 1887 Sington 1888 Nash 1888 Von Oechelhausen 1888 Pinkney 1888 Butler 1888 Pinkney (b) 1890 Baldwin 1890 Day 1891 Ridealgh 1891 Day 1891 Atkinson (b) 1892 Instone 1892 Sohnlein (b) 1892 Bilbault (b) 1892 Cock 1892 Roots (b) 1892 Okes (d c) 1893 Okes (c d) 1893 Roots (b) 1893 Sintz (b) 1893 List and Kosakoff (b) 1893 Roots (b) 1893 Maccallum (g) 1894 Harris 1894 Cycles of One Revolution. Type 5. — Without a pump. Capitaine (b) 1885 Atkinson (c) 1886 Atkinson (d c) 1887 Roots (a) 1889 Lentz Czermak and others. . . . 1890 Coffey (d c) 1891 Southall (a) 1892 Gas Motoren Fabrik Deutz (a) . . .. 1892 Cycles of Two Revolutions. Type 6. — Ordinary De Rochas cycle. De Rochas 1862 Otto 1876 Crossley 1877 Linford 1879 Linford 1880 Otto . . 1881 Holt and Crossley 1882 Martini 1883 Steel and Whitehead 1883 Crossley 1883 Pickering 1883 Spiel 1883 Picking and Hopkins 1883 Hale 1883 Serrell 1883 Warchalowski (e) 1884 Holt and Crossley 1884 Delamare and Malandin (e) . . . . 1884 Wirth (g) 1884 Wirth 1884 Crossley 1884 Magee 1884 Lenoir 1885 Sombart (e) 1885 Spiel 1885 Durand (e) 1885 Daimler 1885 Daimler (c b) 1885 McGhee and Magee 1885 Crossley 1885 Priestman (b) 1885 Koerting Boulet (e) 1885 Smyers 1885 Southall 1885 Rogers (c) 1885 Bickerton (c) 1885 ( i3 ) CLASS 2 ( continued ) Cycles of Two Revolutions. Type 6 .— Ordinary De Rochas cycle. Magee 1886 Roots (b) 1886 Capitaine and Brunler (b) 1886 Davy 1886 Magee (d c) 1886 Humes (b) 1886 Hutchinson 1886 McGhee 1886 Stuart and Binney (b) 1886 Southall (c) 1886 McGhee 1886 Adam (e) 1887 Priestman (b) 1887 Gas Motoren Fabrik Deutz . . . . 1887 Spiel (b) 1887 Wastfield (c) 1887 Davy (c) 1887 Roots (b) . . . . 1888 Johnston (c) 1888 Ragot (e) 1888 Kosztovits . . 1888 Purnell 1888 Lalbin 1888 Rowden (c) 1888 Royston (c) 1888 Quack (c) 1888 Schnell 1888 Noel (e) 1888 Stubbs 1888 Roots, gas 1888 Roots (b) 1888 Dougill 1888 Stuart and Binney (b) 1888 Lanchester 1889 Niel (e) 1889 Miller 1889 Lalbin (e) 1889 Daimler (c b) . . . . 1889 Rogers and Wharry 1889 Lindner 1890 Hamilton (c) 1890 Butler (b) 1890 Stuart and Binney (b) 1890 Higginson (c) 1890 Pinkney (b) 1890 Dheyne, Nydpruck, and Hoult (b) . . 1890 Robson 1890 Forest (e) 1890 Qvens .. .. .. . , .. .. 1890 Cycles of Two Revolutions. Type 6 — Ordinary De Rochas cycle. Jeanperrin (e) 1890 Pinkney (b) 1891 Wertenbruck (c) . . . ’. 1891 Pinkney 1891 Capitaine 1891 Poussant (e) 1891 Hornsby and Edwards (b) 1891 Roger (e) 1891 Roots (b) 1891 Van Rennes (b) 1891 Letombe (e) 1891 Crossley and Holt (b) 1891 Koerting and Boulet (e) 1891 Fiddes and Fiddes (c) 1891 Lacoin (e) 1891 W. and S. Priestman (b) 1891 Weiss (b) 1891 Gas Motoren Fabrik Deutz (bed).. 1891 Vanduzen 1891 Vanduzen 1891 Dawson 1891 Evans 1891 Seek (b) 1891 Anderson (c) 1892 Webb 1892 Held (b) 1892 Durr (b) 1892 Rankin and Rankin (b) 1892 Owen (b) 1892 Bengger (b) 1892 Charter (b) 1892 Gas Motoren Fabrik Deutz (b) . . . . 1892 Spiel (b) 1892 Lanchester (b) 1892 Hogg and Forbes (b) 1892 Berk 1893 Cronan 1893 Fiddes and Fiddes (c) 1893 Hamilton (c) 1893 Roots 1893 Crossley and Atkinson (c) 1893 Fielding 1893 Gas Motoren Fabrik Deutz . . . . 1893 Burt and McGhee 1893 Dixon 1893 Evans 1893 Hartley and Kerr 1893 Bellamy 1893 Hamilton (c) 1893 ( 14 ) CLASS 2 (continued) Cycles of Two Revolutions. Type 6. — Ordinary De Rochas cycle. Cycles of Three Revolutions. Type 8. Wattle 1893 Clerk and Lanchester (c) 1894 Skene (b) 1894 Griffin (b) 1894 Lindemann 1894 Foster (c) 1894 Low 1 894 Gibbon (b) 1894 Henriod-Schweizer (b) 1894 Hirsch (b) 1894 Bryant (b) 1894 Robinson 1894 Howard, Bousfield, and Bastin (b) . . 1894 Wiseman and Holroyd (b) 1894 Campbell (b) 1894 Withers (c) 1894 Linford . . 1880 Griffin 1883 Spence 1884 Rollason 1886 Rollason 1886 McGhee 1886 Griffin 1887 Griffin 1887 Lindley and Browett 1888 Wordsworth 1888 Roots (double expansion) 1888 Weatherhogg (b) 1889 Peebles 1889 Cycles of Compound Engines. Type 9. Cycles of Two Revolutions. Modified De Rochas cycle. Type 7. — Reducing the charge fired to increase' the expansion. Wordsworth and Lindley 1883 Crossley .. 1884 Nash 1886 Wordsworth and Wolstenholme . . 1886 Shaw 1888 Crossley 1888 Butler- 1888 Gas Motoren Fabrik Deutz . . . . 1888 Royston 1888 Fnrrp<;t arirl Gnllirp Skene (c) 1886 Milburn and Hannan (c) 1886 Gas Motoren Fabrik Deutz (c d) . . 1887 Sterry (c d) 1887 Griffin (c d) 1887 Davy (c) 1887 Charon (d) 1888 Browett and Lindley 1888 Rowden (d c) 1889 Green 1889 Southall (c) 1889 Fielding (d) 1890 Roots 1890 Brayton (b) 1890 Deboutteville and Malandin (d) . . 1890 Rouzay (deb) 1891 Shaw and Ashworth (c) 1891 Fernand and Gallice 1891 Bellamy 1893 Letombe (d) 1894 1 UiiV/Jl CU1U Cti.X-l.V_/V_/ •• •• .. .. -L UUM Niel 1889 Butler (c b) 1889 Melhuish 1890 Stuart (b) 1890 Beckfeld and Schmid 1890 Higginson 1892 Robert 1892 Anderson (c) 1892 Piers 1892 Hartley and Kerr 1892 Davy 1892 Durr (b) 1892 Hartley and Kerr 1893 Davy 1893 Berk 1893 Clerk and Lanchester (c) 1894 Davy 1894 Roots (b) 1894 Schimming 1894 Davy 189 ( *5 ) CLASS 3. Continuous Combustion Engines. Type 10. Wilcox i860 Siemens i860 Million 1861 Hock 1872 Brayton . . 1872 Jenkins 1874 Simon 1878 Foulis 1878 Simon 1879 Williams and Baron 1879 Williams 1879 Simon and Wertenbruch 1881 Watson 1881 Siemens 1881 Tonkin 1881 Maynes 1882 Odling 1883 Foulis 1883 Crowe 1883 Type 10. Livesey 1883 Wilcox ■ 1885 Babcock 1886 Hargreaves (b) 1888 Von Oechelhauser 1888 Gardie 1889 Hargreaves (b) 1889 Hoelljes 1889 Crowe 1889 Clerk 1889 Atkinson 1889 De la Touche 1890 Holst (c) 1890 Diesel (d) 1892 Societe Anonyme Gardie 1892 Brunler, three patents (b) 1893 Arschanloff 1894 Gas Motoren Fabrik Deutz . . . . 1894 CHAPTER II. Cycles of Types i and 2, Class t. Robert Street may fairly be described as the originator of the internal combustion engine, since in his patent we have the first description of a possible and workable engine. It is entitled “ Method to produce an inflammable vapour force by means of liquid air, fire, and flame, for communicating motion to engines, pumps, and machinery ; ” is dated May 7th, 1794, and numbered 1983. It is on similar lines to most of the following engines on this cycle, except that the piston is connected to a beam instead of to a crank. It is a non-compression engine of type 1. The illustration accompanying the specification is a mere sketch — Fig. 2— but the description not only of the construction of the engine, but also of its working, is very clear and concise. I cannot do better than quote this part of the specification : — “ A, an iron cylinder ; B, a solid iron piston made to fit cylinder ; D, a strong frame in which the cylinder is suspended ; E, a stove to keep the bottom of the cylinder hot ; F, a counter-sunk touch-hole, and near the bottom of the ( i6 ) cylinder. As soon as the bottom is sufficiently heated, pour a small quantity of spirits of tar or turpentine into the funnel, whicli falls on the hot part of the cylinder, and instantly the liquid is converted into an inflammable vapour ; at the same moment raise the piston by means of the lever G, which sucks in the external condensed air, and also raises a light to the touch-hole, the confined vapour takes fire similar to gun-powder, and by the combined power of inflam- mable and rarified air thus incorporated together, forces the piston B up the frame, and also raises with it the long shaft K, which descends with the piston to the bottom, and works the pump or other machinery at the opposite end at L. The two sides of the frame — Fig. 2 — are made hollow, like a groove, to guide the piston in its return into the cylinder; the same operation continued, a constant motion is communicated.” It will be seen that here are the main points as to construction and working of the subsequent engines upon this cycle that afterwards achieved success. The piston at the commencement of the cycle draws in a charge of air to mix with the vapour of the spirit previously injected for a portion of the stroke; flame is then drawn into the cylinder to ignite it, the piston is driven out by the resulting explosion, and during the return or in-stroke of the piston the products of the combustion are displaced from the cylinder, when the cycle commences again. It will be noted that this is the same process for vaporisation, though in a much cruder form, that has been adopted at the present day in some oil engines upon the two-revolution or de Rochas cycle, viz., the dropping the fuel fluid upon a hot surface of metal within the cylinder to vaporise and ignite it. In i860 the Lenoir engine was produced in France. On the 8th February of the same year Lenoir took out a patent in this country, numbered 335, and entitled “Improvements in obtaining motive power and in the machinery or apparatus employed therein.” The Lenoir engine was undoubtedly the first gas engine that was commercially a success ; it sold in considerable numbers. Fig. 3 is a sectional plan taken from the patent drawing ; it very much resembles the steam engine of that time. A is the cylinder, B the piston, G is the admission slide, and H is the exhaust slide. It is double acting, explosions taking place on ( *7 ) both sides of the piston. Assuming that it commences to move from the position shown, air is first drawn into the cylinder end Ai, and then air and gas are drawn through the slide G, and the port C (the gas from the port T and gas-cock R), to the same end Ai, the piston-rod end of the cylinder. This admission of the working charge is continued for about one half of the stroke, at which point the pressure in the cylinder is below the atmospheric — see Fig. 4. ( i8 ) The slide then closes the port or channel C, and the charge is ignited by an electric spark produced by a Rhumkorff coil at the igniter J. The explosion pressure rises to about 85 lb. per square inch ; the piston is driven to the end of the stroke, at which point the pressure has fallen to about 7 lb. per square inch above atmosphere. The exhaust slide H then opens the port D, and on the return stroke the products of combustion are displaced from the cylinder up to the dead point of the return stroke, whereupon the cycle commences again. The excentric Gi works the slide G, and the excentric Hi on the other side of the engine works the exhaust slide H. is the electric igniter at the other end of the cylinder. The same operations take place at the other end of the cylinder A2. The indicator card, Fig. 4, is from a Lenoir engine. It will be seen from the sloping ascent of the ignition lines that ignition did not take place at constant volume in this engine. The diagram, Fig. 5, further explains this cycle. We now come to Type 2 of Class 1. The earliest engine of this cycle is Samuel Brown’s. There are two patents, the one, No. 4874, dated December 4th, 1823, and the second, containing improvements upon the first, No. 5350, April 25th, 1826. Fig. 6 is taken from the earlier of the two patents. A A are the cylinders, B B the jackets open at the top between A and B to allow water to overflow into the working cylinder A from the jacket. C C are the pistons suspended by chains from the rocking beam D. EE are valves sliding upon the piston-rods, and suspended from the rocking beams by the chains F F, ( 19 ) passing over wheels attached to the spiral spring G. H is the connecting-rod transmitting the motion of the rocking beam to the crank-shaft I carrying the fly-wheel. J is a rod connected to the rocking beam D, tiansmitting motion to a second rocking beam K by means of the tappets J x and J 2 . The ends of the rocking beam K operate the slide valves L L, which open and close the ports MM. N is the gas reservoir or chamber supplying the nozzles O O inside the cylinder, and the pilot light burners P P outside the ports M M. The end of the rod J operates the pump Q in the water tank Q l5 which pumps water ( 20 ) into the jackets. E 2 are small mushroom relief valves opening upward and outward fitted in the piston covers or valves E ; there are two in each. Water stands in the cylinders to the level of the top of the pipes R R ; any excess flows back into the tank Q x by these pipes. The slide N x in the chamber N controls the supply to the gas pipes feeding the nozzles O O, inside the cylinders. The burners P P remain alight during the running of the engine. The move- ment of the slide N x is effected by the tube S, containing mercury to act as a governor ; the tube S is oscillated on the centre S x by the tappets T T on the rod J. As the action in both cylinders is alike, what follows applies to both equally. Assuming the piston to be at the top of the cylinder, as it is in the right-hand cylinder — we are considering the right-hand cylinder only — the burners P P are lighted, the fly-wheel is turned, and the piston on the right of Fig. 6 descends ; the tappet J\, when the piston is nearly at the bottom, strikes the beam K, moves the slide L, opens the port M, and lights the gas at the nozzle O inside the cylinder ; this continues to burn, the valve E being open. The piston ascends on the up-stroke, the combustion continues, and when the piston is near the top of the stroke the tappet J., strikes the beam K, which closes the valve M, and the pump Q causes the water to overflow from the jacket on to the ascending piston, and at the same time the piston-rod guide closes the valve E by stretching the spring G. The vacuum produced by the cooling of the products of combustion in the cylinder drives down the piston until the pressure within nearly equals the atmospheric, when the spring pulls up the valve E, and the continued descent of the piston forces out the remaining products. The slide L again opens the port M, the rocking tube S again opens the slide N 1} and the cycle recommences. Although Street was the first inventor and maker of a gas or spirit engine, yet this engine of S. Brown is the first gas engine, in spite of its complication, that was an industrial and continuous working mechanical success. A number were made and, it is stated, worked successfully for some years ; they were used chiefly for pumping, and propelling boats and barges. See Mechanics' Magazine , vol. ii., pages 360 and 386. Although I have placed Brown’s engine as the first in Class 1, type 2, its chief and leading idea being the working by the pressure of the atmosphere , owing to the vacuum created in the cylinder, yet a close examination of the construction of the engine shows that there was more or less continuous combustion during that part of the working stroke prior to the production of the vacuum, the flame burning from the jet or nozzle in the cylinder. It has not been placed under Class 3, type 10, for the reason stated. The best known and most successful engine of this class and type is that invented by Langen and Otto, the patent in this country being dated February 12th, 1866, and numbered 434. Fig. 7 illustrates this engine. It was exces- sively noisy, but it was economical ; it consumed far less gas than any gas engine preceding it. It was, however, of a very similar construction to Barsanti and Matteucci’s engine, which appeared in 1861 ;* the cycle was precisely the same, but it had an advantage in having flame instead of electric ignition ; they were both free piston engines. The Langen and Otto was commercially successful, while the Barsanti and Matteucci was not. The Langen and Otto is still occa- sionally to be found doing regular work even in this country. I know of one that does its daily work satisfactorily to its owner; and after many years of service the owner, when I last saw him, had no intention of replacing it by a compression engine, in spite of its noisiness. In the Langen and Otto engine the power of the explosion is not transmitted to the shaft. The pressure produced by the explosion throws upward in the cylinder a free piston, i.e., a piston not connected to the shaft during the time in C ( 2 2 ) which the explosion is acting upon the piston. The piston having completed the upward stroke due to the explosion is then automatically clutch-geared to the shaft, and the power stroke is effected by the atmospheric pressure upon the piston during the larger part of the return or instroke. The cycle is as follows : — When the piston B is at the lowest position in the cylinder A, the fly-wheel is turned by hand, lifting the piston by means of the toothed wheel C on the shaft D, which gears into the rack E fixed to the piston. The rack moves in the guide F. The raising of the piston draws in a charge of gas and air through the port G, the slide H being moved to allow it. The port I in the slide permits communication between the port G and the air inlet port J and the gas supply K. The charging continues for about one-tenth the whole stroke of the piston ; when the piston reaches the position shown the slide H opens the port G to the flame pocket L in the slide. The pocket is supplied with gas and air just before each ignition, and the mixture in the pocket is ignited by the burner flame M. The flame from the pocket is drawn into the port G, ignites the charge in the cylinder, and the piston is thrown up by the explosion. A free upward movement is allowed to the rack E by the toothed wheel C as it is connected to the crank shaft by a clutch, having an equivalent action to that of a ratchet and pawl. The quantity of charge drawn into the cylinder is adjusted to be only that amount necessary to throw the piston to the top of the cylinder ; any excess force there may be is taken up by a rubber buffer provided to receive the blow. A vacuum is formed in the cylinder by the prolonged expansion, and the pressure of the atmosphere on the outside of the piston impels it on the return stroke, upon which the clutch N connects it to the fly-wheel shaft. The clutch N contains wooden rollers, which allow free movement in one direction but have a frictional grip in the other. The piston continues to descend, giving power until the pressure within the cylinder reaches that of the atmosphere again, when the momentum of the fly-wheel continues its movement to nearly the bottom of the cylinder, expelling the products. The fly-wheel momentum commences the upstroke of the piston, drawing in the new charge of gas and air, which is ignited, and the cycle is repeated. ( 2 3 ) The indicator card at Fig. 8 — published by Mr, Crossley, for which I am indebted to Professor Robinson’s book, “ Gas and Petroleum Engines ” — shows clearly the action of this cycle. Fig. 9 shows the respective lengths of the pro- cesses of this cycle — not lengths of time, but of operations of this cycle, as represented by piston stroke. With regard to Type 1 of this class — non-compression engines deriving their power from the explosion — I do not think that any economy, or indeed any improvement, is likely to be effected in the future once more to. place it in the market as a competitor of compression engines. This cycle is, I believe, dead as far as motors are concerned ; but it may possibly be used again for some special purpose, such as a gas hammer. Although, however, the atmospheric cycle is practically dead, I think it is quite within the range of possibility that a competitor of existing engines, on the ground of economy — but not on any other ground, for such an engine will always be larger and heavier for its power than v a compression engine— might be devised, in the light of our present knowledge and of improvements effected in gas engine practice since engines of this type were designed. ( 2 4 ) CHAPTER III. Class 2. Compression Engines of Type 3. Lebon, a French engineer, was the first to suggest a compression gas engine. To him is therefore due the credit of the invention of the first engine under Class 2. Lebon used two pumps as Barnett did later. M. Witz says, in an addition to his first patent, he explained a project of a gas engine in these terms : “ I shall show the means of collecting this expansive force (of gas) and of mode- rating its energy, and of using it only in a measure and in proportion to the power required and to the strength of the machine which is permissible.” “ In a cylinder A the combustion of the inflammable gas takes place, it is introduced by the pipe B, while the atmospheric air necessary for combustion is supplied by the pipe C. The cylinder A receives gases produced by this com- bustion, its piston intercepts all communication between the parts E and F.” In a footnote referred to at this point, M. Witz says : — “ This description can be followed easily without an illustration.” “ The working piston also drives two pumps — the one a double-acting air pump, the other a similar double-acting gas pump. For the ignition an electric machine worked by the motor can be arranged.” Lebon, as far as I can discover, never made an engine, and probably did not attempt to do so ; his was the credit of the idea.* About four years after taking out his patent he was assassinated, just after the maniacal times of the French revolution. M. Witz continues : “ This document establishes in an indisputable manner the title of Lebon to the invention of gas engines. We find in the preceding lines the idea of the most notable improvement which has yet appeared in the construction of gas engines. I refer to the idea of the compression of the gas and air before explosion.” I have accepted the latter part of this statement, and placed Lebon as the leader of this column, but I do not find anything in M. Witz’s statements regarding Lebon to prove that he — Lebon — compressed the charge before ignition. In Wright’s patent in 1833, No. 6525, in which is given a full and clear description with drawings, the statement is made that he compresses the charge to 2 lb. per square inch ; he also used a double-acting air pump and a double- acting gas pump, but by closely following the specification it becomes manifest * Professor Robinson in his work, “ Gas and Petroleum Engines,” falls into a curious error. He says the Lebon engine was practically identical in principle and construction with the Lenoir, ( 25 ) ( 26 ) that the charge when in the combustion chamber immediately before ignition is necessarily not appreciably above atmospheric pressure. Wright is therefore placed in Type i of the Table, q.v. Although Lebon was the first, according to M. Witz, to suggest a compression gas engine, it was William Barnett who first made one. Barnett was an ironfounder of Brighton, Sussex. A patent was granted to him dated April 18th, 1838, and entitled “Certain improvements in the pro- duction of motive power.” The engine, or rather engines, described in this patent were, judged even by the modern standard, practicable, and well arranged engines ; but no water jacket is shown on this engine, although I have reason to believe that one was fitted in the engines ntade. Barnett’s engine is similar in principle— except in the fact of possessing separate pumps for gas and air — to the Clerk gas engine which appeared fifty years later. Barnett’s engine is of interest, too, as the first internal combustion engine in which, we have ample and indisputable evidence, compression was used ; it is peculiar also in that its constructional arrangements and cycle were equal to many of those that came after it, and were more successful — that is, so far as success may be determined by the number of engines sold. Barnett’s specification is so clear and concise that I cannot do better than quote from his patent as to the action of the cycle. The second engine differs from the first in being double-acting ; the third is also double-acting, but it is without the receiver. The two first are provided with receivers for the temporary reception of the gas and air. Fig. 10 shows a double-acting gas engine taken from Barnett’s patent specifi- cation. A is the cylinder, B the piston, C the slide-valve case, having the two slides upon the one rod. C x C x are the two receiving chambers, into which the air and gas are delivered by the air and gas pumps. D is a double-acting pump, which draws in air at the valves G and H, and forces it into the receivers through the valves I and K. There are also two single-acting pumps (not shown), the cylinders of which are open at one end, which draw in gas, and deliver it by their delivery valves into the receivers C x by the ports E and J. The gas and air pumps, by means of the spur wheels P and Q, make their ascending and descending strokes simultaneously with the ascending and descending strokes of the piston B. S is a pipe proceeding from the slide-valve case C, to convey the products of the explosion or any unconsumed portions of the explosive mixture from the cylinder, and having at its extremity a light valve opening outwards. T T are two igniting cocks. These cocks are placed over the gas burners V V in the position shown in the drawing ; and by turning the cock the flame communicates with the interior of the cylinder. The operation of this engine is as follows : — Eet it be supposed that the lower receiver is changed with the explosive mixture, and that the slide valve be then moved so as to open the communication between the lower receiver and the cylinder at the under side of the piston, while the cylinder upon the upper side of the piston is open to the exhaust pipe S, the mixture in the lower receiver will immediately flow into the cylinder, and upon turning the lower igniting cock the ( 2 7 ) mixture will ignite, and by exploding will expand and impel the piston to the top of its stroke, any air which may be above it being expelled into the exhaust pipe S. During the ascent of the piston the gas and air pumps are drawing in air and LJ gas respectively, and delivering them into the upper receiver C. Upon the piston reaching the top of its stroke the slide is reversed, the explosive mixture is admitted above the piston, and upon the explosion taking place the piston is ( 28 ) driven to the bottom of its stroke, the products of which explosion are then exhausted, the cycle continuing. The second engine selected to represent this cycle is the Clerk gas engine, it being probably the most widely known of the type. Fig. 1 1 shows a sectional plan of Mr. Clerk’s engine. The patent is numbered 1089, March 14th, 1881, and is entitled “ Improvements in motors worked by combustible gas or vapour.” A is the working cylinder, B the pump cylinder, C the working piston, D the pump piston, E a conical combustion chamber or clearance space. F is the bed of the engine. The working trunk piston C is connected to the crank in the usual manner, while the pump piston D is connected to a crank pin on the fly- wheel fixed about 90 deg. in advance of the working crank. G is the ignition and gas-controlling slide reciprocated by the bell crank lever H through the slide Dead Clerk Inner circle shows Pump actios) . Otter circle working Piston. Pump 95 m advance of working Piston Cylinder of Pump has double the cubic capacity of the working cylinder block I, which is connected by an excentric rod to an excentric on the crank shaft. The pump B has an inlet and delivery valve placed below the delivery port J. K is the gas supply pipe conveying gas through the slide port L to the port M, from which it passes through small orifices in the mushroom admission valve seating, together with air, to the pump B. N is the pipe through which gas and air are drawn into the pump, and by which the mixture is returned to the clearance space E through the port J by an ordinary mushroom lift valve immediately below it. O is the gas burner, P an ignition pocket supplied with gas in the slide G, Q is the water jacket, W is the ignition port through which ignition of the charge is effected by the slide ; R R are the exhaust ports uncovered by the piston at the end of the working stroke. The volume of the piston sweep of the pump is considerably larger than that of the working cylinder. The working operations are as follows : — The pump draws in its charge of air and gas, the gas being controlled by the slide port L until the explosive charge equals in volume the cylindrus * or the cubic capacity * Continuous difficulty arises in dealing with internal combustion engine cycles in that we have no word which expresses the cubic capacity of the space swept through by a piston in its stroke, i.e. , the area of the piston x the stroke. The word “cylinder,” in speakiug of volume, expresses a very different quantity, viz., the area, x the stroke + the combustion clearance space. I have used the wmrd “ cylindrus ” to express the space or volume, piston area x stroke, throughout the rest of this book. Brevity recommends it. ( 2 9 ) of the piston sweep of the working piston, the gas is then cut off and air only drawn in for the rest of the pump suction stroke. While the exhaust ports R R are fully open and the working piston is at the end of its stroke, as shown, the pump piston is delivering its charge into the working cylinder through the port J, driving out through the exhaust ports the waste gases of the previous explosions. As the contents of the pipe N, at least, are pure air without gas, this air is necessarily first delivered to the clearance space of the working cylinder. The inventor’s idea of the relative positions of the air and charge in the pump cylinder seems to partake very much of the Otto stratification theory. This he has since removed. The plate S and ring T were fitted in the positions shown, with the object of preventing the mixing of the charge first drawn in the pump with the air last drawn in, and of delivering the air first to the working cylinder before the charge. The working piston returns, compressing the charge delivered to the working cylinder ; meanwhile the pocket P of the slide is moving forward with its charge ignited by the burner O. At the dead point of the piston, or at the commence- ment of the return stroke, the ignition slide pocket is brought opposite to the port W, the working charge is ignited, and the piston is impelled outward by the explosion until the exhaust ports R are opened ; when the waste products are exhausted the new charge is delivered, and the cycle commences again. With regard to this cycle, one can only remark that it is difficult to see the advantage of using two pistons where one will do the work as well, if not better. This cycle, as compared with the de Rochas, has about double the piston friction and a much increased piston leakage, to say nothing of the increased first cost in making an engine with two cylinders instead of one. The advantage of an impulse every revolution is heavily outweighed by these disadvantages. Fig. 12 shows the cycle. Fig. 13 is an indicator diagram from a Clerk engine. This engine had other disadvantages which prevented successful competition with the Otto engine, and it soon dropped out of the market. There is one other arrangement of the series of operations properly coming under Type 3, “with the aid of a separate pumping piston,” which is briefly as follows : — The pump is used to draw the products of the previous combustion from the working cylinder, and thus, by lowering the pressure in the working cylinder below that of the atmosphere, to induce an inflow of the air and gas to the working cylinder to form the new charge. We might take the section of the Clerk gas engine, Fig. it, to illustrate oneway of carrying this into effect. It must first be assumed that the valves of the engine have been altered to suit this ( 3 ° ) arrangement of cycle. Let it also be assumed that the piston D is moving on the latter part of its outstroke, while the piston C has completed its outstroke — the relative positions as shown in Fig. 2. The outstroke of the piston D in the cylinder B creates a certain degree of vacuum in the cylinder B, the suction and delivery valves of the cylinder B being held closed to permit it. Then before the end of the stroke of the piston B, the working piston C having completed its out or working stroke, and by uncovering the exhaust ports R, has caused the pressure of the products of combustion in the cylinder A to fall to atmosphere. Then the piston C shall commence its return stroke and close the ports R, whereupon the cylinder A is placed in communication with the vacuum in the cylinder B, an intermediate valve being opened for the purpose. The products of combustion in the cylinder A immediately flow into the cylinder B, and the continuation of the stroke of the piston D increases this outflow. In this way the pressure in the cylinder A is kept below that of the atmosphere during a portion of the instroke of the piston C. Meanwhile the admission valve is held open, and the new charge flows into the cylinder A ; the admission valve is then closed, the charge compressed and fired, the working stroke takes place, and the series of operations in both cylinders commences anew. Some attempts have been made to carry into effect such a cycle as that sketched out above, but for various reasons, some of which are obvious, no success has been attained. It would appear that it must be always a more indefinite, uncertain, and less positive method of displacing the old charge and substituting the new, than that of direct delivery of the new charge by the pump, as the degree of vacuum, and, therefore, the quantity of new charge drawn in, would probably be an ever varying quantity. The above diagram, Fig. 13a, shows this modification of the cycle. ( 3i ) CHAPTER IV. Class 2. Compression Engines of Type 4. On the next column the earliest invention on this cycle that I can find is Robson’s. I had an impression that I had seen an earlier patent upon this cycle at the Patent-office during some long-forgotten search, but if it exists I have been unable to find it again. I have taken Robson’s second patent, No. 4501, November 4th, 1879 — “ Improvements in Gas Engines ” — as the first representative of this type, this specification being much clearer than the earlier one, and devoted exclusively to this cycle. In this type of cycle the front or piston-rod side of the piston in one cylinder is used as a pump— both ends of the cylinder being closed — to draw in the air or the charge, and deliver it to the other or working side of the piston, where it expels, or assists in expelling, the products of previous combustion. The piston then compresses the delivered charge, ignites it at the point of highest com- pression ; and during the next stroke the charge expands, doing a working stroke. At the end of the working stroke the exhaust valve or port opens, the new charge is again delivered from the opposite side of the piston, displacing the exhaust, and is again compressed and ignited, continuing the cycle. Many inventors, including Robson, added an intermediate receiver or reser- voir to receive the charge on its way from the pump end of the cylinder to the working end ; but I believe, from my own experience, that the so doing has no advantage, and indeed much reduces the power of the engine. Fig. 14 shows the Robson, 1879, method of carrying out this cycle. The cylinder is closed at both ends. A is the explosion or working end of the cylinder, Ai the pump end of the cylinder, B the piston, C the piston-rod, D the connecting-rod, E the crank, F the gland through which the piston-rod passes, G the gas cock controlled by the governor Gi, H the gas and air inlet, or suction valve to the pump end of the cylinder Ai, I the valve through which the charge drawn into the front or pump end of the cylinder is delivered by means of a pipe not shown in the drawing to the port J in the reservoir K, in which the charge is stored under pressure ; L is the valve, and M the pipe through which the charge is conveyed to the working or explosion end of the cylinder A ; N are deflecting plates arranged in the cylinder “ for the more uniform distribution of the charge in the cylinder.” T T are the exhaust ports controlled on the outside of the cylinder by a mushroom valve — not shown — opened by an excentric rod worked by the crank shaft, O is a valve opening outward, which during the maximum pressure of the explosion allows a certain quantity of the gases to escape, the gases are conveyed by the pipe P to the reservoir Q, to be there ( 3 2 ) retained and used for working a small engine similar to a steam engine for start- ing the gas engine. Such a device, however, has nothing to do with the cycle of the engine ; it would not, I think, tend to render any engine more economical. If the engine were starting from the position shown, and the piston were moving towards the back cover of the cylinder, it would draw in air and gas by the valve H to the front or pump end of the cylinder Ai ; the return or forward stroke— still considering the pump end of the cylinder— would deliver the charge under compression by the valve I and the port J to the reservoir K. When the piston in its forward or outward stroke has uncovered the ports T, and the exhaust ( 33 ) valve has been opened, the pressure in the cylinder falls nearly to the atmo- spheric ; the charge in the reservoir K being at a pressure above atmosphere, it opens the valve L and flows by the pipe M to the working end of cylinder A, displacing before it the products of the previous combustion through the ports T ; the piston then moves on its inward or backward stroke, the exhaust ports close, and the charge is compressed in the working end of the cylinder A, and at the dead point is ignited by a slide, not shown. The working stroke then takes place, at the end of which the exhaust ports are uncovered, the valve is opened, and the cycle recommences. This patent contains a modified form of this cycle, in which also the two sides of one piston are employed ; the one side as a pump, the other receiving the working pressure. The modification consists in using a smaller quantity of charge each working stroke at a higher pressure, thus effecting a greater expan- sion, but this arrangement has the great disadvantage of a much delayed ignition, and a corresponding reduction of the working stroke. The Robson engine was manufactured by Messrs. Tangyes, Birmingham. This was also the cycle of the Sunderland and the Day engines, which have been manufactured and placed on the market. The two chief difficulties the inventor selecting this cycle has to encounter are (i) the firing of the new charge as it enters the working end of the cylinder by delayed combustion of the previous charge, or sparks in the explosion space or ports ; and (2) the difficulty of displacing approximately all the products of com- bustion in the short time available ; or, if this difficulty be avoided, then there arises that of preventing a certain portion of the new charge from passing out of the exhaust port. If, however, the fuel is pumped in after the exhaust valve is closed, this last difficulty will necessarily be overcome. Fig. 15 shows the cycle. The outer circle shows the operations in the work- ing end of the cylinder, the inner circle those in the pumping end. The cycle, although it has the advantage of giving an impulse every revolution, which may outweigh all the disadvantages for certain purposes, has necessarily the disad- Robson Enytne Cycle ( 34 ) vantage of compressing the charge twice — once on each face of the piston for each working stroke. It follows, therefore, that it cannot be so economical a cycle as the de Rochas ; and, in all probability, it will not be quite so trust- worthy. Yet it is possible that given a new and perfect system of ignition, this cycle might displace the now supreme de Rochas cycle. The “ compressing the charge twice ” objection applies also to the former, or Lebon cycle, in which there is a separate working cylinder and pump cylinder, or else a separate piston, but this cycle, as compared with the Lebon, has the enormous advantage of the friction and leakage of one piston only. As the other representative of the cycle now under consideration, I have selected the engine described in my specification patent No. 23,571, dated December 7th, 1893, “ Improvements in internal combustion engines.” Fig. 16 is a horizontal section of the engine. A is the piston, provided with the valve Y ; B is the cylinder ; C is the combustion space ; E the exhaust valve ; F the exhaust valve excentric rod; Fi, its excentric ; G, the gas pump ; G2, the inlet valve to the pump from the gas pipe ; G3, the excentric operating the gas pump ; G4, the excentric rod ; H is the gas-delivering valve ; J is a per- forated nozzle for the purpose of spreading the gas throughout the air in the cylinder ; X is the ignition channel into which the ignition tube is screwed ; Z is a port in the side of the cylinder. When the piston A commences its working stroke after ignition, starting from the position shown, the front or piston-rod side of the piston displaces or ejects air previously drawn into this end of the cylinder through the port Z. When the port is covered by the piston, it — the piston — compresses the remaining air on the piston-rod side until at or near the end of the working stroke — the pres- sure having on the working side of the piston fallen to the atmospheric — the compressed air opens the mushroom valve Y, and flows through to the other side of the piston. The piston then begins its return stroke, and again draws in air into the pump or piston-rod end of the cylinder through a mushroom self- acting valve — not shown — placed at the top of the cylinder, until the port Z is uncovered, when the air enters through the port Z ; on the return of the piston, air is displaced through the port Z until it is again covered, when the air remain- ing in the cylinder is compressed and flows as before stated through the valve Y to the combustion or working side of the piston. The air in flowing into the working end of the cylinder displaces the products of the previous combustion through the exhaust valve E which has already been opened. The exhaust valve is kept open during half the return or instroke while products are being displaced from the combustion space C. When the exhaust valve E closes, the gas pump G, by means of the excentric rod G4, delivers its charge of gas through the valve H and nozzle J, which spreads it throughout the air remaining in the cylinder just before the dead point. At the dead point, as shown, the mixture is forced into the ignition tube — not shown — by the port X, and the ignited working charge expands on the next outstroke to nearly the full stroke of the piston Just before the outer dead point is reached at the end of the working stroke the exhaust valve E is opened, the pressure falls to atmo- ( 35 ) sphere, the valve Y opens by the air pressure on the other side of the piston, and the operations recommence. The objects sought to be obtained by this construction were (i) expanding the working charge to double its volume before compression when at atmospheric pressure; (2) avoiding the compression, and therefore the back pressure on the piston for half the working stroke. The intent of these, as of most improve- K 3 ( 3 ® ) ments attempted in gas engines, was economy, but the economy sought was not obtained. This is a most expensive cycle to experiment with. Nevertheless, I believe that this modification of this cycle will be much heard of in the future, and I think that ultimately economy will be obtained somewhat upon the lines laid down in this patent, care being taken to effect a complete mixture of the fuel with the air, and to use a high compression. The objections to this cycle before mentioned are therefore partly obviated by the described construction. The patent 9618, of 1893, is upon the same cycle. It will be observed that Types 3 and 4 are similar, so far as a part of the cycle is concerned, as in each there is a pump delivering its charge to the working cylinder, which charge is compressed in the working cylinder before ignition, but in Type 4 the charge is usually delivered at a different time, and in a different manner. It was therefore necessary to mark such a difference, and I thought that the best way to do this was by the leading and chief difference of construction, mentioned at the head of the column in the table. There is one other cycle arrangement, a modification of the foregoing, and therefore coming within the definition of Type 4. “ Use of the opposite face of the working piston as a pump.” In this modification, instead of the pump “ cylindrus ” delivering its charge of gas and air to the working “ cylindrus,” or working end of the cylinder, the pump “cylindrus” (1) may create a certain vacuum into which the products are permitted to flow after the pressure has been lowered by opening the exhaust ports, or (2) may simply pump out the products after the opening of the exhaust ports, without obtaining the vacuum prior to so doing ; (3) the combination of 1 and 2. Shaw’s specification 8579, June 4th, 1884, has been selected to represent this arrangement of Type 4. It is of that modification which is numbered 3 in the foregoing division of this modification. Fig. 1 6a is a sectional view of his engine. A is the cylinder, of which B is the combustion end, and C the pump end, D is the piston, E is the admission valve, and E 2 a pipe to convey air to the combustion chamber, and F the gas admission valve. The gas is delivered to the valve F by a separate pump — not shown — of ordinary construction, the piston-rod of which is attached to and operated by the pin Z. G is the ignition port, H is a self-starter somewhat similar in principle to those now in use in that the contents are first ignited to convey ignition to the working charge in the combustion chamber. I J are exhaust valves which open outwards, K is the exhaust valve leading to the con- denser — not shown — which valve is operated by the cam K 2 by means of the lever K 1 , which oscillates on the fulcrum K 3 . The pump end of the cylinder is connected to the other end of the condenser before mentioned at the port L below the valve J, having a self-closing valve fitted between. The action of this engine is intended by the inventor to be as follows : — Let it be assumed that a working charge has been ignited through the ignition port G. Then the piston performs its working outstroke until it un- covers the port leading to the valve I, when all excess pressure is intended to be relieved by means of the valve I, which action continues until the end of the ( 37 ) stroke represented by the dotted line, whereupon the valve K is opened by the lever K x and cam K 2 , and the products from the combustion end of the cylinder B flow into the condenser, which flow is accelerated and prolonged by the return movement of the piston pumping further products from the condenser. The pressure is now reduced below the atmospheric in the combustion chamber B, and air enters by the valve E and pipe E l5 and gas is delivered by the pump through the valve F to form the combustion mixture ; the continued inward movement of the piston after the closing of the valve K compresses the mixture, which is again fired through the port G, and the cycle in the combustion end of the cylinder recommences. On the pump end of the cylinder C, assuming the piston to be moving on its outstroke from the position shown, products are first forced through the valve I, and when that is covered, through the valve J until the dead point at the end of the stroke be reached. The inward stroke of the piston then tends to create a vacuum in the pump end C, and products therefore flow by the port L into the cylinder from the condenser, which are ejected by the piston from the cylinder through the valve I on the outstroke or working stroke, and the cycle continues in the manner described. The chief disadvantage of such an arrangement as that before described would be the uncertainty and unreliability resulting not only from the use of the condenser in this manner, but also from any endeavour to cause the inflow of the new charge by an attempt to create a vacuum which, in an internal combustion engine, will in all probability be in ever varying degree and quantity. In any case, to obtain a certain degree of success with such a modification of this cycle, simplicity should be the chief aim; Shaw’s engine was too complicated, and had too many valves to have attained success. CHAPTER V. Class 2. Compression Engines of Type 5. The Atkinson cycle engine is selected as the first representative of Type 5 — the third division of one-revolution cycle compression engines, those not having a separate pump. This is one of the most ingenious attempts to overcome the difficulties 01 increased expansion in an internal combustion engine. The results of tests of this the Atkinson cycle engine, Fig. 17, had not been previously approached for economy of gas consumption. The ingenuity of construction of this engine must always command the highest praise, and the inventor of such an engine deserved the best pecuniary reward that a power-using public could have bestowed. Indeed, this motor is, from the inventive point of view, not only a P ( 38 ) more ingenious, but a more novel one than its predecessor, the two-revolution Otto, as this was new as a combination of devices to produce a certain result ; whereas the Otto engine only embodied in practice the academic ideas of de Rochas. In so far that there were the same series of operations, viz., suction, com- pression, explosion, and exhaust, each occupying one stroke — the engine was similar to the two-revolution de Rochas, or Otto cycle. But there the resem- blance ended, for each of the four strokes was of different length, and all four strokes were effected in one revolution of the crank shaft. The proportional length of each stroke was approximately as follows : the suction stroke 6, the compression 5, the combustion and expansion n, and the exhaust 12. It is somewhat difficult to convey an idea of the working of this engine briefly, and at the same time lucidly. Imagine the crank pins of two cranks connected together by a link or connecting-rod. Now, if one crank were turned, assuming that they were of the same throw, say 12m., it is possible, though not very probable, that the one will rotate the other. But if the throw of the one crank were reduced to, say, ioin., and the ioin. crank were turned, it would compel the 12m. crank to describe a part of a circle only, and thus perform a kind of prolonged oscillation. The crank A, to the shaft of which the fly-wheel is keyed in the Atkinson cycle engine, is the one having the smaller throw, while the piston is connected to the longer arm B. A little reflection on this will now make it clear why the engine makes four strokes to one revolution of the crank. The dotted lines show the path of the crank-pin centres Y Z of the arms B and A respectively, and also how the strokes differ in length. On reference to the Fig. 17 it will be seen that the piston connecting-rod is not connected to the same centre on the oscillating lever as the link D, but to a centre a short distance from it. It is this difference in centres that causes the difference in the lengths of the strokes, that is, between 11 and 12 and 6 and 5 respectively — Fig. 17A. The connecting rod C, together with the link D, and the oscillating crank B, formed a kind of toggle joint. The proportionate length of this series of four strokes was, no doubt, arrived at by experiment, and served well in practice. First a short suction or out-stroke of the piston E in the cylinder F, proportion 6 ’3 ; a shorter return or in-stroke compressing the charge of gas and air, pro- portion 5 ; at the end of which compressing stroke the charge is ignited by a hot tube G ; then a long out-stroke during the combustion and expansion of the charge, proportion ii'i ; and lastly, a still longer return or in-stroke, proportion i2’4; expelling the products of combustion. It must be noted, too, that the piston at the end of the exhaust stroke — position of piston as shown — leaves so very little clearance space that practically all the products are cleared out of the cylinder, so that at the end of the next suction stroke, which commences the cycle again, the contents of the cylinder approach most nearly to a pure charge of gas and air. Nearly twice the expansion of the charge was effected in this engine as com- pared with that in the usual de Rochas or Otto cycle ; thus, when the exhaust valve opened, instead of about 38 lb. per square inch of pressure being wasted ? ( 39 ) with its corresponding ratio of heat, there was only about 15 lb. per square inch above atmospheric pressure. Although theoretically it would have been better to have expanded down to atmospheric pressure, in all probability no greater economy would have followed from the so doing. In comparison, if the length of stroke in an Otto engine were 6*3, the same as the length of the suction stroke in the Atkinson engine under consideration, then the working stroke length of the Qtto would be y6, while that of the Atkinson is ii*i, or, making a similar p ? ( 40 ) reduction for the exhaust valve lead, 10. The charge was ignited by a hot tube without a timing valve. The piston speed being much greater than had up to that time been generally adopted, the engine not only gave a greater range of expansion, and therefore used more of the heat units of the fuel in work upon the piston, but also, as the working stroke was completed more quickly, allowed less time for the heat of combustion to pass into the jacket water. The diagram Fig. 1 8 is taken from the judge’s report in the Society of Arts Journal In the Society of Arts trials, in which the Atkinson competed against the Crossley and the Griffin, it was shown that although rather more cooling water was used than in the Crossley or in the Griffin, a double-acting Otto or de Rochas, the temperature was not raised to such a degree as in the Otto and the Griffin. The Atkinson (9*48 brake horse-power) used 680 lb. of water per hour, the temperature of which was raised 50 deg. Fah., while the Crossley (1474 brake horse-power) used 713 lb. of water, the temperature of which was raised 128 deg. Fah. It is therefore manifest that less heat per brake horse-power passed through the walls of the cylinder to the jacket water in this cycle than in the Otto. In comparing, however, one thing should have been taken into considera- tion that is not mentioned in the report, viz., the respective thicknesses of the cylinder walls. Clearly, the thicker the wall the longer time will the heat take to pass through it. I have found that with two de Rochas cycle engines of the same size and construction, the one having a liner fin. thick and the other Jin., the engine having the fin. liner used less water than the engine having the ^in. liner ; but this was more than compensated for by the fact — they were both gas engines — that after an hour’s run the engine with the fin. liner would decrease slightly in power, while the power of the other would remain approximately the same. The Atkinson engine consumed 22*61 cubic feet per brake horse and hour, while the Otto consumed 24*ift. per brake horse and hour ; but as these figures in both cases include the gas for heating the ignition tube, and the Atkinson consumes one cubic foot more for this purpose than the Crossley, the difference of consumption, 2*49 cubic feet, owing to the difference of cycle, was rather greater than appeared without the ignition gas deducted. The Atkinson engine not only wasted less heat in the jacket water, but less also by the exhaust. It must be taken into consideration, in making a comparison of respective ( 4i ) expansions, that there is a much larger clearance space in the Otto cycle, and that, therefore, a larger admixture of products with the charge occurs every charging stroke than in the Atkinson cycle. Although the Society of Arts motor trials conclusively proved the greater economy of this cycle over any other competing, and by inference over any other then existing, yet the particular Atkinson engine competing in these trials did not exhibit that degree of economy that one might have been - led to expect from the theoretical advantages of the cycle, nor the comparative economy relatively to the other competing gas engines that it should have shown. This is doubt- less to a great degree accounted for by the fact that the mechanical efficiency of the Atkinson cycle was shown to be less than the Crossley engine. There was necessarily a larger friction, and therefore a greater difference in the ratios of the indicated and brake horse-powers, as there were two additional bearings to transmit the power to the crank shaft. On reference to the heat accounts to be found in the report of the Society of Arts trials it will be seen that in the Atkinson heat account this engine is credited with 2 2*8 per cent, of heat turned into work, while the Crossley — which, con- structed upon the Otto cycle, we may fairly regard as the standard cycle with which other cycles must be compared — is credited with 20*9 per cent., a differ- ence of only 1 - 9 per cent. The Atkinson lost 27 per cent, of the heat supplied to it in the jacket water, while the Crossley lost 43*2 per cent., a difference of i 6’2 per cent. The Atkinson lost in its exhaust products 37*9 per cent, of its heat, while the Crossley lost 35*5 per cent. It would appear in these last figures that there is a mistake somewhere. It has been shown that in the Atkinson engine the charge is expanded to almost double its original volume before com- pression, and when the exhaust valve is opened the pressure of the products of combustion is only about 15 lb. above atmosphere; on the other hand, in the Crossley engine the charge is expanded to an extent slightly less than the volume before compression, and immediately before the exhaust valve is opened and the waste products leave the cylinder the pressure is nearly 40 lb. I cannot see, therefore, how it is possible with those conditions for the Atkinson cycle engine ( 42 ) to waste in the exhaust products 37 *9 per cent, of the total heat supplied, while the Otto engine wastes only 35 ’5 per cent. No engine before this time had reached such a degree of economy, and the engines upon this cycle demonstrate that economy is to be found upon these lines and in those directions that theory would indicate, viz., greater expansion, higher piston speed, and reducing the quantity of products left in the cylinder to mingle with and dilute the next charge. There can be no doubt that a much higher compression would have increased the economy. Fig. 19 shows the action of the cycle and the differences in the operations between this and the de Rochas cycle. The second example of this type — Type 5 of the table — “ one revolution without a pump,” is taken from the Roots engine — patent No. 3972, March 6th, 1889, gas engines. In this engine all the four processes — suction, compression, explosion, exhaust — occur in one revolution, and on one side of the piston in the cylinder, that is, the one side is not used as a pump to force the charge to the other side on which the explosion takes place, nor to pump the products from the explosion chamber, nor is a mechanical device employed to obtain four strokes. Fig. 20 is the drawing from the patent specification of the engine having the complete cycle in one revolution, and with two strokes. A is the cylinder, B the piston, and P the red-hot tube igniter of ordinary arrangement. A small charge having been drawn in through the port E to start the engine, the piston is returned to compress the charge into the hot tube P, and the explosion occurs ; the piston B makes a working out-stroke until the exhaust port R in the cylinder A is uncovered, when the products pass through the port R, the exhaust valve S is opened by the pressure of the exhaust, and the excess products flow out. The valve is then closed by the spring S 1 and the piston B continues its outstroke^ during which time the products in the length of tubing forming the water-jacketed ( 43 ) chamber T are cooling, and in contracting cause more products to flow from the cylinder into T, thus assisting the lowering of the pressure in the cylinder below the atmospheric caused by the continued out-stroke of the piston B. A certain vacuum is produced, and the fresh charge of gas and air enters by the sloping admission port E and admission valve E 1 from the bell chamber F, which is open to the air and to the gas pipe G. By the heat of the cylinder A, and the coming in contact with the cylinder cover and walls, owing to the slope of the port E sending it in that direction, the new charge expands, displacing more of the exhaust through the port R. The piston now commences its in-stroke, forcing exhaust products through the port R until the piston covers it, when the com- pression commences ; and at the end of the instroke, at the highest compression, ignition occurs, when the cycle commences again. Thus, briefly, the operation is in the out-stroke, explosion, part of exhaust, and suction ; in the in-stroke, the remainder of exhaust and compression. There is only a short length of tubing shown for the cooling chamber T ; but it may vary in size, according to the length of the suction stroke and the degree of expansion required ; and the chamber is made with the jacket T 1 , through which the cold water passes on its way to the cylinder jacket. Very little will be done with this cycle unless the cooling or condensing surface in the chamber T is nine or ten times the size indicated in the figure. There may be three or more tubes in one jacket. There are only two ports in the cylinder, and only two valves — one inlet and one outlet — both automatic or lift valves. The condensing chamber T much increases the efficiency by permitting the suction portion of the stroke to be shorter, and therefore the working portion of the stroke longer, and this chamber can be made of such size and cooling surface as to do nearly the whole suction work of the engine. The dotted lines — i, 2, and 3 — show about the respective lengths of the stroke, 1 to 2 being about the length of tfie working portion of the stroke ; and 2 to 3 of the suction portion of the stroke, and, as shown, the proportion is about as 9 to 4; but it will vary according to the size and construction of the chamber T. When the com- pression commences, the cylinder contents are in about the same proportion — new charge 9, products 4. I spent much time during the year 1889 endeavouring to make this cycle work successfully. The chief, and finally the only difficulty was that one misfire would stop the engine. Until there was a misfire the engine worked perfectly, and, considering that it was the first engine and a small one, with a very low consumption of gas. If a misfire occurred, however, the next out stroke drew in a full cylindrus of charge, instead of the normal quantity, a portion of which was then discharged through the exhaust pipe, and the rest being compressed, the charge ignited considerably before the dead point and pulled the engine up. Had this cycle been successful, it will be seen the engine has a greater expansion than any other, including the Atkinson just described. It had also a high piston speed. Instead of the jacketed condenser T, the engine was really finally run with a condenser made very similar in action to the Korting con- denser. A spray of water was delivered just below the port R upon a series of gills ; the jet must be delivered intermittently and just after the opening of the port R. For some reason I was unable to discover, the engine always ran best with a low compression. Fig. 20A shows the operations of this cycle. In this column of the table, q.v., Type 5, although all the cycles are com- pleted in one revolution without a separate pump or “ cylindrus,” there are much greater differences to be found between cycles of the respective engines in this type than there are in either of the others. CHAPTER VI. The Two-Revolution, or De Rochas Cycle. In 1862, January 7th, Beau de Rochas took out a patent in France for a new cycle of gas engine working, and proceeded to explain it in a pamphlet published shortly afterwards, in which he not only sets forth with great clearness the two-revolution or four-stroke cycle, but also suggests points to be aimed at in construction to obtain the best results, and that most correctly, judged by our later knowledge. He describes the cycle thus: — (1) “Aspiration” of the explosive mixture during one complete stroke of the piston. (2) Compression of the mixture during the following stroke. (3) Ignition at the dead point and expansion during the third stroke. (4) Driving out of the burnt gases from the cylinder during the fourth or last stroke. The conditions he asserted to be necessary to get the best results were : — (1) That the cylinder should be as large as possible, while having that form which would give the least surface. (2) That the piston speed should be as high as possible. (3) That the gases are expanded as far as possible. (4) The charge should have a high compression. As far as we know, de Rochas did not put the invention he so concisely and lucidly described into practice. M. Witz says no motor was made, and the failure to pay the second tax caused the patent to lapse. On May 17th, 1876, was applied for in this country Nicolaus August Otto’s patent, entitled “ Gas Engines in which the motive agent burns gradually.” In this patent is described the same, or practically the same, series of operations as de Rochas described in his patent in 1862. Clearly, however, Otto did not regard the cycle as the most important part of his patent, as in the commence- ment of the specification we find the words : “ A combustible mixture of gas or vapour and air together with air or other gas, that may or may not support com- bustion, in such a manner that the particles of the combustible m'xture are in an isolated condition in the air or other gas, so that an ignition instead of an explosion ensuing, the flame will be communicated gradually from one com- bustible particle to another, effecting a gradual development of heat and expan- sion of the gases.” Then he proceeds to describe an engine upon the non- compression or Street cycle, which was first described. And lastly comes the description of the de Rochas cycle. But throughout, Otto lays much more stress upon, and evidently attaches more importance to, the stratification idea of his, than to anything else in the patent. This was a much vexed question about twelve years ago, and most men interested in internal combustion engines at that time debated the subject of stratification with some degree of heat. Do you believe in stratification ? was a question one frequently had to reply to at that time. The doctrine of stratifica- tion appeared to divide engineers interested in internal combustion engines into two camps, the one for and the other against the doctrine. Of course, stratification has long since “exploded,” and is going the way of other errors into oblivion. Every practical man knows now that, even at fairly slow speeds for an internal combustion engine, the new charge enters at such a velocity that the contents of the cylinder are all in one mighty swirl, which leaves no chance, unless it is in the ports or behind mechanical divisions in the clearance space, for any separation or division of the components of the charge, and even then it would not amount to stratification. Otto differs in one respect from de Rochas in that he — Otto — desired to obtain a slow combustion, a gradual development of the flame throughout the working charge. De Rochas, in spite of his in all probability possessing much less experience of gas motors than Otto, yet had a much clearer perception of the conditions required for obtaining the best result from the cycle upon which both were working, Otto practically, de Rochas as far as we know only academi- cally, theoretically. The whole spirit of de Rochas’ instructions implies rapid combustion and rapid expansion. ( 46 ) The Otto Gas Engine. Fig. 21 shows the Otto engine as originally made by Otto. A is the cylinder, B the piston, B 1 the connecting-rod, D the slide valve, E is the exhaust port provided with an ordinary mushroom outlet valve. F is a connecting-rod for reciprocating the slide valve by a crank pin G on the rotating side shaft H ; I is the crank shaft having the bevel wheel J keyed on it, gearing into another bevel wheel K of twice the diameter. The larger bevel wheel K is fixed to the end of the shaft H, and rotates it at half the speed of the crank shaft I. The slide D, therefore, makes one reciprocation while the piston B makes two. L is the slide cover, M is the ignition burner, N is a small gas slide controlling the supply of gas from the pipe N 1 to the pipe N 2 , from whence it passes to the port R in the slide D. The governor O slides the cam P along the shaft which operates the slide N ; N 3 is the gas port in the slide cover L. S is the air inlet port in the slide D, and T the port into the cylinder. U is an ignition channel in the cylinder end, and W is a port of a corresponding channel in the slide through which ignition from the flame M is effected. Air is first drawn into the cylinder though the port S in the slide and port T in the cylinder end to the clearance space A 1 for the first half of the suction stroke, until the piston has travelled from the dead point to the middle of the suction stroke ; during the remaining half of the stroke gas is drawn in through the port R with the air to form the explosive mixture. It was the inventor’s idea that the contents of the cylinder would remain in stratified order at the com- mencement of the in-stroke. The second or in-stroke compresses the charge, which is ignited by the flame at the burner M through the intermediary of the channels U and W in the cylinder and slide D, at or immediately after the dead ( 47 ) point. The explosion effects the working stroke, which is the second out-stroke, and the third stroke of the cycle. Near the end of the working stroke the exhaust valve beneath the port E is opened by the lever E 1 and the cam E 2 on the rotating side shaft H. The following instroke and last stroke of the cycle expels the products of combustion from the cylindrus, leaving products in the clearance space A 1 . At the next out-stroke the cycle re-commences. The Rochas or Otto cycle, therefore, consists of four strokes of the piston or two revolutions of the crank shaft. During the first stroke in a single-acting engine the piston moves outward, or away from the cover towards the crank, drawing in its charge, consisting of the fuel gas or vapour mixed with air. In some engines on this cycle the admission valve is opened by mechanism actuated by the crank shaft, and in others by the pressure of the atmosphere on the outside of the valve. This charge, consisting of the mixed fuel and air mingled with the products of the previous combustion, is then compressed by the piston to a pressure determined by the volume of the clearance or combustion space, the ignition is effected by one of the various devices in use for that purpose just upon or a little after the dead point, and the piston moves outward again on its explosion or working stroke, until a little before the end of the stroke the exhaust valve is opened, and from that point until the end of the working stroke, and also throughout the return or inward stroke, the exhaust valve is held open, and the products are being expelled through it by the piston. The indicator diagram, Fig. 23, is taken from a later and improved Otto, as made by Messrs. Crossley Bros, for the Society of Arts motor trials. Fig. 22 shows the operation of the cycle in the original Otto engine. It will be readily followed by imagining that the two circles are folded upon each other at the point A F, like a figure 8 cam groove cut upon the peripheral surface of a cylinder, in which the line at C continues on the second circle from C to H, and ( 48 ) having described that, returns to the first circle again at A, and so on regularly round each circle, describing the complete cycle of the engine. In the indicator card, Fig. 23, and the diagram, Fig. 22, A is the com- mencement of the admission stroke, from B to C is the return inward com- pressing stroke, immediately upon or at the dead point C, or at a certain distance from it, varying in different engines, and indeed in the same engine slightly from stroke to stroke, the ignition takes place. In the indicator card, Fig. 23, it will be seen the pressure rises suddenly to 220 lb. per square inch, the working stroke occurs during the expansion of the ignited charge from D to E. But at E in both Fig. 22 and Fig. 23 the exhaust valve is opened, the pressure falling until the dead point F is reached, Fig. 22, when the cycle re-commences. The conditions laid down by de Rochas as necessary to economical working of the cycle indicate a remarkable prescience on his part, since they are com- pletely endorsed at the present day. With regard to his first suggestion, it is known that the larger the engine, with certain limits, the more economical it should be. His second suggestion would be endorsed with the proviso that the velocity of the working fluid in the ports and through the valves is not such that the cylinder is not fully charged, or such as to produce an appreciable rise above atmospheric pressure in the cylinder during the exhaust stroke. The third condition is too obvious for discussion, provided the expansion be limited in practice to that point at which the pressure in the cylinder is only just sufficient to overcome the friction of the engine. Every engineer will endorse de Rochas’ fourth proposition : in gas engines there has been a progressive rise in the degree of compression for some years past, and the limit is not reached yet. The limit of compression is that degree of temperature produced by compression that will ignite the charge. This limit can be controlled somewhat by a free use of cold w r ater in the jacket. If, however, the fuel, gas, or oil be delivered to the compressed air at or near the highest compression, there is hardly any limit to the compression. See Fig. id, Type 4. In oil engines, however, we are already confronted in using a high compression, with the difficulty of spontaneous or premature ignition, owing to the heat of the compression, plus the heat the charge receives from the cylinder walls. There is also a limit of piston speed at which advantage would be turned to disadvantage. For while a high speed in an oil or gas engine conduces to greater ( 49 ) economy owing to (i) a more rapid combustion and expansion, (2) less cooling of the charge at the moment of ignition by the passage of heat to the cylinder walls, and (3) the entering charge not having so much time to receive heat from the contained products and hot walls, and consequently to expand and reduce the quantity drawn in. Yet on the other hand we have as disadvantages resulting from high speed (1) a tendency to take less than a full charge during each charging stroke, owing to the resistance of the ports and ’ valves, so that the pressure in the cylinder at the end of the suction stroke will be below atmo- spheric, thus reducing the power per explosion, and therefore the power of the engine, and also by adding resistance to the movement of the piston by outside atmospheric pressure : (2) the walls of the cylinder and the contained products of previous combustion will be at a higher temperature, a less fraction of time having elapsed since the previous ignition, the entering charge is more heated and therefore more expanded, which also tends to a reduction of the quantity of charge per stroke. And this is in spite of the fact that the entering charge has less time in which to expand. Nevertheless, the balance of advantages is in favour of high speed in internal combustion engines of this cycle, always provided that the ports and valves are so designed that no throttling occurs in the admission of the charge. The velocity of the explosion through the charge appears in both gas and oil engines to increase with the compression, to increase with the temperature of the charge at the moment of ignition, and also to increase with the purity of the mixture, i.e.) its freedom from products of combustion.* The de Rochas or Otto cycle, or some modified form of it, is the only one that has had continuous success, and it at present holds the field, for while there are a few at present made upon two cycles giving an impulse every revolution, their minority is so small in comparison to the overwhelming majority upon the de Rochas cycle that it may be said that the de Rochas cycle practically stands alone. Several years ago I made a considerable number of gas engines upon a modified form of the de Rochas cycle. * Berthelot and Vielle’s experiments give the result that there is no difference in the velocity of the explosion wave in tubes having an internal diameter between 5 mm. and 15 mm. Now, 5 mm. is rather less than the internal diameter of the ^in. gas pipe in Great Britain, yet makers of internal combustion engines frequently find that the |in. gas ignition tube fires more quickly than a £in. tube. Berthelot also says that the velocity is independent of the pressure. I cannot help thinking that if the makers of gas and oil engines given to observation and experiment were canvassed upon this point, the majority would hold the opinion that the velocity of ignition in an internal combustion engine increased with the compression. I am not in any way impugning the results arrived at by M. Berthelot. I have no doubt there are other reasons which convey the contrary impression mentioned, but I am using this as a peg upon which to hang the com- plaint that scientific men draw up the results of their experiments in such a way that they are frequently of little use to the manufacturer — that is, that they are drawn up more from the point of view of pure science rather from that of applied science. They are consequently of little avail to the observant constructor— in a word, they are too academic and theoretical rather than applicable and practical, ( 5° ) Roots’ Second Explosion Engine. The patent specification No. 9310 and date June 26th, 1895, describes Roots’ second explosion engine. Fig. 24 is a vertical section of a vertical gas engine made under this patent. A is the cylinder, B the piston, C the second explosion chamber, whose port D is uncovered by the piston at about one-fifth of the up-stroke. E is the admission valve, F is the gas cock, G the governor spindle, H the exhaust valve, I the exhaust valve lever operated by the excentric rod J, worked by two to one gear on the main shaft on the other side of the frame standard. K is the ignition tube, M is the Bunsen burner, N a projecting piece of tube found by experiment to be necessary for ignition, P is the lubricating oil feeder, Q is the crank. During the commencement of the first out-stroke of the cycle, that which is usually entirely admission stroke, the exhaust valve H remains open until the piston reaches the edge of the port leading to the second explosion chamber. Immediately the exhaust valve closes the piston uncovers the port D, and the move- ment of the piston, tending to produce a vacuum, causes the admission valve E ( 5i ) to be opened by atmospheric pressure, and the charge of gas and air flows through the chamber C into the cylinder, sweeping before it the products of combustion. The charge is compressed during the second or in-stroke into the second explosion chamber and into the clearance space equally until the piston covers the port D, the pressure in C then remains stationary, except for the heat the charge may receive from the chamber walls. The piston continues the compression of the charge in the clearance space until the dead point is reached, immediately after which this portion of the charge is ignited ; the piston commences its working out-stroke, the pressure as shown in the indicator diagram — Fig. 25 — rises to 240 lb. Immediately the piston uncovers the port D, a portion of the pressure in the clearance space is used to further compress the charge contained in the second explosion chamber, the flame simultaneously igniting it. The fall of pressure after the opening of the port D, followed by the rise after the ignition of the second charge, is clearly shown in the indicator diagram. Near the end of Roots Second Explosion Engine the working stroke the exhaust valve is opened, and held open during the next in-stroke and for a portion of the following admission out-stroke. The cycle then re-commences. It will be observed that the charge is divided into two portions — a small portion in the clearance space, and the larger part in the second explosion chamber; and that while the compression in the chamber is only about 40 lb. above atmosphere, the compression in the clearance space suddenly, upon the ( 52 ) closing of the port D, rises to 88 lb. At the beginning of the compression stroke the contents of the second explosion chamber are relatively pure charge, while the gases a little below port D, and in the clearance space, must be very largely products only. The violent agitation of the contents of the cylinder, owing to the speed of the piston, ensuies a certain quantity of new charge mixing with the products in the clearance space. Nevertheless, this portion of the charge is so extremely diluted that it would not fire without a fairly high compression. By keeping the exhaust valve open to the point shown on the diagram, the cooling of the charge in the exhaust pipe tended to create a vacuum in the clearance space, enabling a larger quantity of air and gas to be drawn in as soon as the port D was opened. Fig. 26 shows the cycle. For the indicator diagram, Fig. 27, I am indebted to Mr. Clerk’s paper on “ Recent Developments in Gas Engines,” read before the Institute of Civil Engineers. This card shows the great advantage of high compression, and in a less degree that of clearing out the products of combustion from the cylinder and replacing it by air. Efforts to clear out the products of combustion in this cycle have been numerous, as is shown by the names marked on the table, page 10. The methods of effecting this are numerous also. This indicator diagram is taken from a Crossley gas engine of n *97 brake horse power, at 200 revolutions per minute, the indicated horse-power is 14, and the gas consumption is stated to be 17 cubic feet per brake horse-power per hour. The diameter of the cylinder is 7m., and the stroke 13m., and to obtain n '97 brake horse-power out of an engine of this size certainly marks an epoch in the history of the gas engine. This Crossley variation in the de Rochas cycle is as follows : — During the work- ing stroke the exhaust valve is opened sooner than usual in the ordinary de Rochas engine, and is held open later ; that is, instead of being closed at or about the dead point at the end of the exhaust stroke, it is held open for a portion of the following suction stroke. The admission valve — air only being allowed to enter — is opened earlier than the usual time, before the dead point and during the exhaust stroke. In the Crossley engine the admission valve is opened by a cam and lever ; but this is not absolutely necessary, as, with a light spring, the admission valve will open by atmospheric pressure. The effect of this arrangement is that, by the sudden cooling of the products in the exhaust pipe, a vacuum is produced in the pipe, causing the pressure within the cylinder to be reduced below the atmospheric, and air flows info the ( 53 ) cylinder, sweeping the larger portion of remaining products in the clearance space into the exhaust pipe. Both Mr. Atkinson, in his paper on this subject read before the Manchester Association, and Mr. Clerk, in the paper before mentioned, attribute this vacuum to the “ energy of discharge of the exhaust setting the column of fluid in the exhaust pipe in motion,” but I have no doubt this is solely due to the cooling of the' steam and products in the pipe, the con- traction causing a scavenger flow of air to enter the cylinder, by creating a partial vacuum. This I proved by means of a length of pipe about 15ft. long, and with several bends placed in the water tank of the engine. I have obtained a vacuum in a cylinder exceeding 2 lb. with engines of this type by this means. With a continuous spray of water into the top of the exhaust box placed about 1 oft. from the engine, and the pipe continued for another 10ft., having four, five, 01 six bends in it to delay the inflow of air from the open end, a still greater vacuum may be produced. An apparatus similar to a Kdrting’s injector, fitted to engines of this type, is also very effective in producing a vacuum for clearing out the products. I agree with Mr. Clerk that only a small percentage — I think 8 to 10 per cent. — of the economy and high mean pressure shown by Messrs. Crossley’s engine is due to the scavenging, and it is no doubt chiefly due to the compression. The natural corollary of my experiments in scavenging was the endeavour to make the cooling of the products in the exhaust pipe perform the whole of the suction for the engine, i.e., produce a sufficient vacuum in the cylinder to draw in the working charge. Hence the one-revolution cycle without pump of the type already described, from the specification of patent No. 3972 of 1889, page 42. Undoubtedly, in a gas engine it is of benefit to remove all the products, and to replace them by air. The advantage, however, in an oil engine of so doing, E ( 54 ) it appears to me from my own observation, although considerable, is less than in a gas engine. The reason of this difference is one of those mysteries to be yet solved. The solution of it may be associated with the facts (i) that it requires a much lower temperature to ignite an oil charge than it does a gas charge, and (2) the flame passes through the oil charge with a greater rapidity than through a gas charge of similar strength, provided the oil forming the mixture is completely vaporised outside the cylinder, and the charge is thoroughly mixed before ignition. Fig. 28 illustrates the Roots horizontal oil engine, which works upon the de Rochas or two-revolution cycle. The system is that of outside vaporisation, in which the oil is vaporised and the vapour mixed with air simultaneously before they reach the admission valve of the engine. A is the cylinder, B the side shaft, C the charge admission valve, Ci the admission valve lever, D the exhaust valve, Di the exhaust valve lever, E the governor, F the exhaust cam provided with two steps, one for starting only ; G is the oil pump, H the air pump for the pressure lamp, I the main oil tank cast in the bed, J the lamp oil tank, K is the burner, L the air heater and vaporiser, M is the oil feeder and measurer, N O are the two skew gear wheels by which the side shaft B is rotated at half the speed of the engine. The valve C is opened by its lever Ci, and air is drawn through the holes in the top of the air heater L between the double casing surrounding the ignition tube, and through the oil feeder M, the spindle of which has just delivered a groove full of oil. The heated air sweeps the oil from the groove in the spindle, and the groove is returned to the oil space to be re-filled. The air and oil together pass through the vaporiser, and thence through the open admission valve C to the cylinder. The return in-stroke compresses the charge ; at the dead point it is ignited by the ignition tube within the casing L heated by the burner K. The explosion impels the piston on its working-out stroke, the third stroke, just before the end of which the lever Di opens the exhaust valve D, the excess pressure escapes into the exhaust pipe, and the next in-stroke of the piston expels the products of combustion. The engine works exceedingly well, and with great economy and steadiness. The indicator diagram Fig. 29 is taken from this engine. ( 55 ) CHAPTER VII. The Two-Revolution or De Rochas Cycle. Numerous inventors propose the use of steam or water vapour to increase the power or efficiency in a gas or oil engine, and this idea was supported by Professor Unwin during the discussion at the Institute of Civil Engineers, following the reading of the paper “ Recent Developments in Gas Engines.’ Others, again, have wished to increase the efficiency in a gas engine of this class by lining the combustion chamber with some non-conductor, or by the use of a regenerator, or in some other way endeavouring to raise the temperature of the incoming charge at the expense of the outgoing products of combustion. This latter is no doubt up to a certain limit of temperature necessary in an oil engine to vaporise the oil used — a temperature depending largely upon the flash point and the specific gravity of the oil, and possibly is a useful proceeding in some engines of Class 3 and Type 10 having continuous combustion. The following- experiments made a few years ago will show, I believe, the futility of such before mentioned ideas in this cycle. A vertical engine upon the de Rochas cycle, having a cylinder diameter of 6in., with ioin. stroke, compression 52 lb. per square inch above atmosphere, indicating 4*5-horse power, and giving 3*6 brake horse-power, was adjusted to run normally with its governor at 210 revolutions per minute. The governor was disconnected and taken off the engine, the brake was slowly loaded to the maximum power of the engine, and the speed was maintained fairly regularly at 215 revolutions per minute for about a quarter of an hour. The consumption of gas was 307 cubic feet per brake horse-power per hour, and it remained practically constant throughout the experiment. The air or suction pipe was placed at the bottom of a pail containing ice, the pail being perforated at the bottom, not only to allow a freer access of air so that there should be no throttling, but also to let any water drain off. The number of revolutions of the engine rose steadily to 228, and varied between 228 and 232. The temperature of the ingoing air near the admission valve was 36 deg. Fah.* The ice pail was then taken away, and the air suction pipe, a piece of ijin. gas pipe about 5ft. long, was then heated about 2ft. from its end to a very dull red with a Fletcher gas brazing burner. The engine slowed down to 182 revolu- tions and laboured heavily. The thermometer, placed through a cork plug in a * The thermometer used was marked to 650 deg. Fah., was, I believe, of German make, and was fairly accurate for boiling water, and for the freezing point. E 2 ( 56 ) hole drilled in the ijin. gas pipe next the admission valve box, showed a tem- perature of 255 deg. Fah.; the gas pipe was still further heated until the thermometer registered a temperature of 370 deg. Fah., when the engine, after becoming slower and slower, finally stopped. Although these results are what might be expected by anyone having much experience of gas engines, I have given them as a warning to those sanguine inventors who think that improvement can be made in gas engines of this type by means of a prior heating of the charge or the use of some form of regenerator. I do not believe that advance in this type is possible on the lines of regeneration; the combination of the hot-air engine, a low-pressure engine, in which heat and power are developed slowly, with the internal combustion engine, an engine of high and increasing pressures, in which power is developed by explosion, is the harnessing together of the hare and the tortoise. Another experiment made upon the same engine, but at another time, seems to indicate that the drier the air entering a gas engine the greater the advantage. The engine had its governor taken off as before mentioned, and was loaded with the same weights upon the same brake and spring balance. The engine on this occasion ran with the load at 216 revolutions per minute. The end of the air pipe of the engine, lengthened as in the last described experiment by means of a piece of 1 Jin. gas pipe, was fitted in a hole in an improvised kind of box tray, 36m. by 36m. by 2jin. deep, having the end opposite to the closed end, through which the pipe was placed, open. Unslaked lime was spread over the bottom of the inside of the box, covering the whole of the surface. All the air entering the ijin. gas pipe was therefore drawn across the surface of the lime, and must have been practically devoid of water vapour by the time it reached the pipe. The speed rose to 219 quickly, and remained at 219 to 220 for twenty minutes, but during the course of the next half-hour the speed diminished steadily to 216 again. Upon the same engine upon another occasion a further experiment was made to determine the effects of an excess of moisture mingled with the ingoing charge. The same brake, spring balance, weight, and gas adjustment were used as before. The top of the tank which contained the cooling water of an oil engine that had been running some hours, had a board fitted within it 2in. less in diameter than the inside of the tank, thus leaving iin. all round it for the passage of air. The board was suspended by iron strips from the top of the tank 2 Jin. above the water ; it had a hole in the centre through which a 1 Jin. gas pipe bend was fitted. This was connected by i|in. gas piping to the admission valve of the engine. The temperature of the water was 195 deg. Fah. The air drawn into the cylinder must necessarily pass over the surface of the water and carry with it the vapour which was being freely given off from it. The engine had been running for the previous 25 minutes, making 215 revolutions per minute. The same length of pipe being used as before, the elbow at the end was inserted through the board within the top of the tank. The engine slowed down at once to 202, and misfired about 1 in 5 charges. In about 10 minutes, however, the misfires ceased, and the number of revolutions increased to 210. The rapid current of air passing over the surface of water had cooled it down to such an extent that but little vapour was given off. As this was an unsatisfactory experiment, I prepared for another on the following day with steam instead of vapour. A water tank of the size used lor J-brake horse- power oil engines was raised off the ground upon a tripod, and five gallons of water were placed in it. A Fletcher gas brazing burner with full flame was fixed beneath it. The engine had been running half an hour, making 215 revolutions per minute, loaded as before. When the water in the tank boiled, the suction pipe was placed through a hole tightly fitting it in a circular board. The pipe was flush with the surface of the board, and the board was placed 2in. above the surface of the boiling water, but before the pipe with its attached board could be securely fixed the engine pulled up from repeated misfires. The engine was started again, and the board and suction pipe end placed 6in. above the boiling water, but the engine again pulled up before the pipe could be fixed. The board was then taken away, the engine started again, and the brake replaced as before. The pipe was placed at ioin. above the boiling water, the engine carried its load for six minutes, missing fire a great deal, and running at about 150 revolutions per minute, and then pulled up. The suction pipe was then raised i8in. above the surface of the water. The engine now ran at 180 revolutions per minute, missing one in nine on an average, but varying a good deal. The following card — Fig. 29A — shows the behaviour of the firing in the cylinder of another engine under similar conditions to those described, viz., steam being mixed with the working charge. The results of these experiments are what any man experienced in internal combustion engines would probably expect. Not only will the maximum and mean pressures be reduced by the use of steam or water vapour, but the ignitions will be rendered very uncertain and difficult. Many makers in practice have had a cover joint leak, or a spongy place in the cover or liner, or some other defect, by which the water will get into the cylinder, and they will know well the result of such a leakage, however small it may be. The following experiment will probably not coincide with what would be expected : — A hole was drilled through the cylinder jacket and liner walls of a horizontal engine, 4 Jin. by pin. stroke ; the hole was tapped and fitted with a piece of fin. gas pipe, screwed on the outside. This pipe was provided with a screw gland close to the outside of the cylinder. A thermometer, registering up to 650 deg. Fab., was fitted through it, and the bulb of it projected into the combustion space, at a point Jin. from the piston at the inner end of its stroke. The space between the thermometer and the tube was packed with asbestos, and the gland was packed with cork ; that is to say, the pressure exerted by the gland upon the glass was by means of cork. The thermometer bulb was projecting into the combustion space at a point where it was expected it would be least affected, or cooled by the entering charge. The engine had a compression of 52 lb. above atmosphere. It was started, and the thermometer closely watched, while I kept my hand upon the gas cock in readiness to turn it off if the mercury should rise too rapidly and show any signs of breaking the glass. But, to my surprise, the mercury on reaching 540 deg. simply jumped to and fro between 540 and 550 with each explosion, and con- tinued to do so for the quarter-hour the engine ran. The cork gland remained perfectly tight. The thermometer was then placed in the exhaust valve channel. It was fitted in the same tube with its cork-packed gland, the tube being screwed through the unjacketed wall of the exhaust valve channel fin. above the exhaust valve, and projecting beyond the inner wall. The engine was started ; I caught a glimpse of the mercury at the third explosion, and the engine ran for a quarter of an hour without my seeing it again. The engine was stopped and the tube unscrewed, when I found that the projecting part of the bulb and all the mercury had disappeared. The glass had fused. In the first experiment no doubt the jacket protected the thermometer to a considerable extent, even although it projected ; nevertheless, I am convinced, from other experiments since that time, that in the second experiment, even if the exhaust valve had been jacketed, the same result, viz., the fusing of the glass, would have followed, for in this position it has a stream of flame directed upon it during combustion, and passing it afterwards on its way to the exhaust valve. Before commencing any experiments having for their object the ascertaining of the temperature in an internal combustion engine, one is confronted by the want of a reliable pyrometer. A great many methods of measuring high temperatures have been devised, but 1 do not think any pyrometer has yet been produced that can be called trustworthy. I carried out some experiments to ascertain approximately the temperature of the combustion in a gas engine of this cycle, and I will give these briefly. They do not coincide with what one from calculation and theory would expect, because any test by melting metal can only be approximate as the metal is subjected to the high temperature for only a little more than one-fourth of the whole time ; that is to say, it is heated for one stroke during the combustion, and to a less extent during a large part of the exhaust stroke ; it is cooling for the remainder of the exhaust stroke, and also during the suction and compression strokes, and the metal therefore is cooling for about two and a-half out of the four strokes. It appears to me that this inherent objection would apply equally to the use of any pyrometer. The degree of cooling during the exhaust stroke is small, especially in the exhaust port. The amount of cooling in the admission stroke is considerable in the admission port, and in the ignition or clearance space, but ( 59 ) very slight in the exhaust port, provided the same channel does not serve for both valves. In the compression stroke the cooling is greater in parts of the exhaust port, as the new charge is being brought into contact with a portion of its surface. This is speaking relatively to the temperature of explosion, as the actual cooling will be in proportion to the difference of temperature between the enclosed gases and the enclosing walls, and this latter temperature is largely dependent upon the jacket water. I have carried out numerous experiments in gas and oil engines, with the view of arriving at the temperature of combustion in different parts of the com- bustion space and ports. In all cases these have been by means of fine wire of different gauges and different metals. To enter into the details of these would make this chapter much too long, but the following is a condensation of the results. The engines in which the experiments were carried out were always of small size, never exceeding 7m. diameter. Very fine wires of silver, brass, copper, iron, and German silver have been placed at all points in the combustion space and ports of gas and oil engines. The compression pressure in no experiment exceeded 70 lb. per square inch above atmosphere ; it was generally between 50 lb. and 60 lb. In all cases the German silver wire fused rapidly, and in many cases disappeared altogether. Frequently the wires of the other metals were broken, and in many cases showed evidence of rapid oxidisation, and once or twice appeared to show symptoms of fusing; but on examination under the microscope these were seen to be due to the action of oxidisation or some similar rapid corrosive action, and not to fusing. In not a single experiment have I seen any clear evidence of fusing or melting of the metal wires other than the German silver. The fusing temperatures of the wires used are as follows : — Brass, 1650 deg. Fah. ; copper, 2050 deg. Fah. ; silver, 1830 deg. Fah.; iron, 2912 deg. Fah.* The composition of the German silver alloy varies so much that it is impossible to give the fusing temperature, unless the sample used had been specially tested. It is, however, always considerably below that of brass. A piece of solder screwed upon the centre of the explosion face of the admission valve would sometimes not melt during a 15 minutes’ run in a gas engine, starting cold, but had melted at the end of 20 minutes. In a high-speed oil engine solder screwed on to the admission valve had usually melted within three minutes, and sometimes within one minute, after starting. Solder would melt practically immediately after starting the engine in the clearance space or in the exhaust port if not placed too near the cylinder walls in either a gas or oil engine. The fusing temperature of this solder, according to Hutton, is 441 deg. Fah. Tt must not be supposed, because the time of the combustion stroke is only one-fourth of the whole, that combustion takes place in the cylinder for only that proportion of time. It is generally supposed that when the exhaust valve opens in an engine upon the two-revolution or de Rochas cycle, combustion ceases Hutton “ Works Manager’s Handbook.” ( 6o ) with the sudden fall of pressure ; but this is very seldom so. Combustion prac- tically always continues in the cylinder for one-fourth of the return or exhaust stroke — unless the engine has a slow piston speed — even in a thoroughly well- jacketed cylinder after the water is in full circulation. If it is a high-speed gas engine, it will continue for one-third of the stroke ; and if a high-speed oil engine, I am fully convinced by careful watching and timing that if the air has its full charge of oil it continues to fully one-half of the return exhausting stroke, and occasionally a little more than one-half. To show this, an unjacketed exhaust valve box should be fitted to the engine, and the engine allowed to run for fifteen minutes at 320 revolutions per minute. The exhaust valve box becomes a dull red, and with practically every ignition a tongue of flame shows from the valve box. I determined the time by attaching a piece of wire to the reciprocating valve shaft, and so adjusting the end of the wire that at exactly half stroke it should be just level with the edge of the exhaust port. The flame generally continued until the wire reached this level. If the engine is run at a higher speed than 320 it becomes too fast to watch it, and it is advisable to adjust the engine at 250 first, and gradually accustom your sight to the rapid movement, as it is easier for the eye to determine whether the wire is ever in contact with the flame, than it is whether they are simultaneous in movement. It might perhaps be possible to obtain a clear photograph of the flame, the wire effecting the exposure. The exhaust port near the valve, I believe as a result of my experiments, is the spot in which to place metals for the highest fusing temperature, while the space within the exhaust valve box, but immediately beneath or outside the valve, has generally a much higher temperature in gas engines than the inside surface or top of the admission valve. In some further and more recent experiments carried out upon a high-speed oil engine, 5^in. by 6in., I tried silver, German silver, brass, iron, and copper wires. For obvious reasons the finer the wire, short of the point of being broken by the rapid out-passage of the exhaust products, the more likely is the wire to become fused in that small fraction of a second during which the explosion takes place, if the temperature be high enough. The engine ran at 420 revolutions per minute, and the wire was placed, in the first experiment, in the exhaust port outside the exhaust valve, and afterward inside, within the exhaust port. The gauge of the silver wire was 40 B.W.G., of the German silver wire 34 B.W.G., of the brass wire 28 B.W.G., iron 32 B.W.G., and of the copper 30 B.W.G. Again, only the German silver wire was fused. The results of these and similar experiments upon the temperature of the gases in oil and gas engines lead me to the conclusion that dissociation does not occur in gas and oil engines as is commonly supposed. It may possibly occur in a very limited way in engines of very large size, or in an imperfectly jacketed engine. The products of combustion, steam, and C 0 2 , immediately they begin to be formed render the combustion more or less comparatively imperfect, even at the very commencement of ignition, and the act of ignition itself, the first shooting of the flame from the igniting point into the mass of the working charge, brings with ( 61 ) it in its train more steam that further and further hinders or delays complete combustion as the flame spreads throughout the charge. In addition to this the jacket at the moment of ignition rapidly cools the explosion flame. These two things, viz., (i) the effect of the steam and products produced by the explosion itself, plus the steam and products of the previous combustion, and (2) the cooling by the jacket are, I think, sufficient in themselves to account for the temperature in a gas engine falling so far below the calculated temperature without seeking for an explanation in dissociation. Nevertheless, for some unexplained reason, the steam in the products in an oil engine does not affect the combustion to so great an extent as in a gas engine, nor is it quite so beneficial to the brake horse-power to clear out all the products. Water vapour in the air at ordinary temperatures, however, will necessarily reduce the power of the engine, as it will not only affect the combustion, but also the vaporisation. The benzoline spirit motor will be more affected in both respects than the oil motor, as the heat necessary for vaporisation in the latter also raises the temperature of the water vapour, so that at the moment of ignition less heat is wasted in converting it into steam.* The tendency of further improvements, with a view to economy, will, I think, be upon the following lines in their respective order : — 1. Further expansion, not necessarily in the same cylindrus. 2. Higher and still higher compression. 3. Removal of all products.! 4. A more reliable means of ignition always exactly at the dead point. 5. Greater piston speed.! CHAPTER VIII. Modified Two-Revolution Cycles Having Increased Expansion. Type 7 is a modified form of the previously described de Rochas cycle, which is characterised by a further expansion of the explosion products in the same “ cylindrus ” in which the explosion took place. In the previous de Rochas or Otto cycle, Type 6, the full cylindrus of working charge, or as much of it as the relative proportions of the admission valve and port to the piston speed will * Users of vehicle motors running with benzoline spirit have been discovering this fact lately. + Mr. Frederick Grover, in some published experiments made upon a gas engine, discovered benefit from the retention of the products ; but clearly there was some oversight which vitiated his results. + Possibly the present vehicle oil and spirit motors are near the limit of beneficial increase of speed. The Roots car oil motor runs at 530 revolutions per minute, while the Daimler spirit motor runs at 700 to 750 revolutions per minute. The removal of all products would, however, no doubt enable these to run with advantage at 1000 revolutions per minute. Some of the smallest sizes of benzoline and gasoline spirit motors run at a higher speed than this. ( 62 ) permit, is drawn into the cylinder on the out-stroke, compressed upon the in-stroke, and being ignited is expanded to the same volume — or a little less, according to the time of the opening of the exhaust valve — as it occupied before compression. There are three ways of effecting this “ further expansion.” In this column, Type 7 of the table or chart (q.v.) are, therefore, those modifications of the de Rochas cycle in which (i) a portion of the charge, or of the contents of the cylinder, is displaced during the in-stroke before the compression, i.e., a portion of what in the de Rochas cycle is the compression stroke is used to expel from the cylinder a corresponding portion of the charge or cylinder contents, after which the admission valve closes and the compression commences. (2) At a certain point in the suction stroke the admission valve is closed, the piston continuing the suction stroke, expands the contained charge, tends to produce a vacuum in the cylinder, and lowers the pressure below the atmo- spheric. On the return or in-stroke of the piston, the piston first reaches that point, or about that point, at which the charge was cut off in the preceding stroke, and the atmospheric pressure is reached before compression commences. (3) At the desired point in the suction stroke the admission valve is closed, and the working charge is cut off — up to this point, as in (2) — the exhaust valve is then opened, and for the rest of the stroke products are drawn into the cylinder from the exhaust pipe, the exhaust valve is also kept open on the return stroke to the same point of the in-stroke or compression stroke at which it was opened on the out-stroke or suction stroke. When it is closed compression commences. The object of these three variations of cycle is to effect a greater expansion of the working charge, i.e., to ignite and expand to a greater volume than the charge occupied before compression when at atmospheric pressure. Say, for instance, that two-thirds of a cylindrus of new charge plus the con- tents of the clearance space are ignited, they are expanded relatively to the new charge, 33 per cent, more than in the ordinary de Rochas cycle. Properly carried out, increased efficiency and economy should result ; there should also follow a lower terminal pressure and a quieter exhaust. Of the three methods of carrying out this modified form of the de Rochas cycle for effecting increased expansion, for many reasons — some of which are obvious — the first is undoubtedly the best : the other two are placed in order of merit from the purely practical point of view. It would appear in the third method that some of the new charge must inevitably find its way out of the exhaust valve, unless some mechanical division be provided in the combustion space to prevent it. The Charon gas engine representing this cycle, first variation, is reputed to be one of the most, if not the most, economical gas engine in France. There are some published tests of this engine by M. Witz, which show it to be a very economical and efficient engine. Louis Charon applied for a patent in this country on August 28th, 1888, No. 12,399, entitled “Improvements in gas motors with variable expansion.” Some years ago I had the pleasure of seeing one of these engines at work at M. Charon’s works at Solre le Chateau. The quantity of charge fired is varied in proportion to the work to be done. This system of varying expansion cer- tainly produces an exceedingly steady running engine ; but to put it in practice effectively it is necessary to employ electric ignition. Fig. 30 is taken from the patent specification. It is an end view of the cylinder and cover showing the valves. It is only necessary to show the part of the engine required to carry out this modified cycle. A is the chamber consisting of a coil of pipe into which a portion of the charge is displaced during what is in the ordinary de Rochas cycle the com- pression stroke. B is the side shaft rotating once for every two revolutions of ( 6 4 ) the crank shaft. C is the admission valve in the admission valve box Ci- D the exhaust valve in the exhaust valve box E a rocking shaft operated by cams on the side shaft, having arms projecting from it which open the valves C and D at the desired times. F is the cam operating the commutator for electric ignition, the wires passing through the top of the valve box Q. G is the gas valve supplying gas to the admission valve C by the pipe G x . The gas is fed to the passing air through perforations in the seating of the valve C. H is the governor which slides on a feather a pair of stepped cams on one sleeve — not shown— along the side shaft B. One stepped cam controls the gas valve G by means of the lever G 2 , and the other the admission valve C. On the piston commencing its suction out-stroke, air is drawn through the pipe A, together with any charge which may have been displaced into it during the previous compression stroke, and gas by the pipe G 1 , through the inlet valve C, into the cylinder during the whole stroke. During the in-stroke or compression stroke, for that portion of the stroke determined by the governor and the stepped cams, according to the work the engine is doing, the piston displaces into the tube or chamber A a certain quantity of the contents of the cylinder. This quantity is always one-fifth, and may be more — see diagram, Fig. 31 — up to two-fifths of the stroke. When the valve C is permitted to close, the contents of the cylinder are compressed and ignited by the electric spark, and the piston is impelled on its working stroke. Whatever the volume of the charge may have been at the commencement of compression, the charge expands to nearly the whole length of the stroke. The exhaust valve is opened just before the end of the stroke, and the products are displaced through the valve D. At the com- mencement of the next suction stroke the charge displaced into A is first drawn in, and the cycle is repeated. The volume of new charge compressed may therefore be only two-fifths of the cylindrus, which expands to nearly the whole stroke. Obviously for an engine ( 65 ) having a cycle of this type, with a varying expansion, electric ignition is the most suitable, as the pressure of compression is always varying from stroke to stroke. I am much surprised that the variations i and 2 of this cycle have not found more favour in this country, especially variation No. r, for a properly-designed gas or oil engine having this series of operations must necessarily be more economical than the ordinary de Rochas cycle, especially in large engines. The only objection to the cycle is that engines constructed upon it are rather larger for the same power than those upon the ordinary de Rochas cycle. It would be advisable to run such an engine at a considerable speed, and by having a complete cut out of fuel by a hit-or-miss governor, and by closing the admission valve at one fixed point for every compression stroke, say at half stroke, a greater rate of expansion would be obtained than in the Atkinson cycle engine, and ordinary tube ignition could be used. As the combustion space would be so much smaller, the products remaining in the cylinder and mixing with the new charge would be relatively to the expansion much reduced. An engine constructed as here indicated would have an expansion to double the volume occupied by the charge before ignition, and if adjusted to run with a high piston speed, and a high compression be employed, should be more economical than the Crossley scavenging system; for any method of scavenging by the cooling of the products in the exhaust pipe will of necessity be more or less uncertain and untrustworthy. CHAPTER IX. Compression Cycles of Three Revolutions, Type 8. Griffin’s patent, No. 4080, August 23rd, 1083, has been selected as the first representative of this cycle of three revolutions, Type 8 in the chart, because it is the most widely known three-revolution engine. Fig. 32 is a horizontal section produced by combining two of the drawings in Griffin’s patent, including those parts necessary to explain the cycle. It is single-acting ; many of the Griffin engines subsequently manufactured were double-acting. A is the cylinder open at the front end. B is the connecting-rod which communicates motion from the piston C to the crank D. On the crank shaft is fitted the toothed wheel E which gears with the toothed wheel F. The larger wheel F being three times the diameter, and having three times the number of teeth of the smaller wheel E, it follows that F will rotate once for every three revolutions of the crank shaft. In the wheel F is fitted a crank pin G which gives motion by means of the rod H to the slide valve I. There are cams driven by the wheel F for operating the exhaust valve and gas valve. W is the port leading to the exhaust valve — an ( 66 ) ordinary mushroom valve opened by a lever and cam. J is the slide valve cover, and K its tightening screw ; D x the igniting flame chimney, O the gas valve, and N the rod operating the gas valve. With the piston C in the position shown at the back end of the cylinder, and the inner end of its stroke, the exhaust valve has just closed, the slide I is just opened to admit air to the cylinder. During the first out-stroke of the piston air only is drawn into the cylinder through the slide I. This is shown in the diagram Fig. 33 from A to B. At the end of the stroke the exhaust valve opens, and during the in-stroke from B to C the charge of air is displaced from the cylinder sweeping with it the products of combustion remaining in the cylinder from the previous combustion. On the next out-stroke from C to D on the diagram — Fig. 33 — the working charge is drawn into the cylinder for the whole stroke. During the next in-stroke from D to E the charge is compressed, and at the dead point F, commencing the third revolution and circle, or a little after, is fired, propelling the piston outward on its working stroke from F to G ; a little before G the exhaust valve is opened, retained open until the end of the working stroke, and also during the next in-stroke and last stroke of the cycle, while the products of combustion are being displaced through the exhaust port from G to H, when the cycle commences again at A. Fig. 32a is a part section plan of the double-acting engine, with an explosion on both sides of the piston, as subsequently made by Messrs. Dick, Kerr, and Co. I have given in this description the cycle — Fig. 33 — as employed in the engine as subsequently manufactured ; but the inventor includes in this patent the drawing in the charge for a portion of the first out-stroke of the cycle, from A to B. Such a modification of the cycle would certainly be a disadvantage, as there would be considerable difficulty with such an arrangement in preventing a portion of the new charge from passing through the exhaust valve during the next in-stroke, from B to C. ( 6 7 ) Fig 33. First Revolution Second Revolution Third Revolution ( 68 ) Fig. 34 is an indicator card, or rather cards, taken during the Society of Arts trials from the Griffin engine. It was the double-acting engine, Fig. 32a; the com- plete cycle was carried out, and explosions took place on both sides of the working piston. So far as the cycle is concerned, it is only necessary to consider the left- hand card taken from the back end of the working cylinder, as the operations on both sides of the piston were precisely alike. The difference, however, between ( 69 ) the areas of the two diagrams is instructive, as showing the effect of a small additional cooling surface, exposed to the working charge during combustion. In this case the reduction of pressures is stated to be chiefly due to the piston- rod passing through the front combustion space, and no doubt is partly so, although I cannot help thinking that there must have been some other occult cause to produce so great a difference in the mean pressures shown, viz., 6 i' 5 lb. per square inch in the back of the cylinder, and 47*5 in the front. Possibly the back was in some way getting more gas than the front. This cycle was probably originally devised to evade the Otto patent, which had been upheld by the Courts, as we now know, more or less wrongly, the de Rochas patent having anticipated the Otto patent, so far as the cycle was concerned, a good many years previously. When the Otto engine became so pronounced a success, strenuous endeavours were made by inventors to discover some new cycle, or to devise some addition to the existing Otto cycle, which would enable a gas engine to be put on the market to compete with it. The inventor of this three-revolution cycle was C. Linford, whose name stands first on the list of this Type 8, Class 2 (see table) and some of the Linford patents display marvellous ingenuity and fertile inventiveness. The patent 330 of 1880 contains no less than twenty-seven figures, all of which are well thought out and thoroughly studied designs. I do not mean by this that they would be necessarily successful in practice. There are, however, two engines which I know of still at work in London, and one is giving full satisfaction to its owner, constructed according to Fig. 3 of the before-mentioned patent 330 of 1880; one of them has had tube ignition fitted. The student of internal combustion engines would do well to study the Linford patents. The next representative of Type 8 is the engine described in Roots’ patent specification, 16,220, November 9th, 1888, “ Improvements in Gas Engines.” The diagram, Fig. 35, shows the operations of this cycle. There are three revolu- tions or six strokes to complete the cycle, in which there are two working strokes and one full or complete stroke each of suction and compression. The cycle consists of first stroke (out-stroke), a complete suction stroke ; second stroke (in- stroke), complete compression stroke ; third stroke (out-stroke), ignition of half the charge previously compressed and working stroke with expansion to double the volume the charge occupied before compression ; fourth stroke (in-stroke), exhaust for a part of the stroke and then admission of the other half charge to F ( 7 ° ) the cylinder from the chamber, Fig. 36, and compressing it again ; fifth stroke (out- stroke), ignition of the second half of the charge, working stroke, and expansion to double volume ; sixth stroke (in-stroke), a full exhaust stroke. Referring to Fig. 36, which is largely a diagrammatic section of a portion of the engine to show the cycle, A is the cylinder, B the piston, C the chamber for receiving half the charge drawn into the cylinder ; A x the cylinder inlet port, A 2 the cover, D the valve closing communication between the chamber and the cylinder, D x the gland through which the spindle of the valve D passes, Dg the tappet of the lever operated by a cam upon the side or valve shaft. The side shaft — not shown — makes one revolution for three of the crank shaft. E is the inlet valve to the chamber C, which valve also covers, when closed, the gas ports E x in the valve seating, supplied by the gas pipe E 2 .* F the port to the exhaust valve for * It is remarkable how many inventors have had the idea of this method of supplying the gas in a gas engine ; how beautiful and simple it appears as an idea, yet how generally unreli- able and troublesome it is in practice, without an additional gas valve to prevent the explosion pressure from occasionally entering the gas supply pipe when the admission valve does not close in time, or is prevented from completely closing by a piece of grit under the seating. ( 7 1 ) the last or complete exhausting stroke, opened by a lever and cam on the rotating side shaft. G is the exhaust valve for the first working stroke of the cycle ; it is operated by a lever from the rotating side shaft. The tappet end G x of the lever is shown. This second exhaust valve was not absolutely necessary, and might have been avoided. I is the ignition tube kept at a red heat by an ordinary atmospheric burner. The suction out-stroke of the piston B draws in a charge of gas and air through the chamber C by the two valves D and E, the valve D being opened by its lever and cam on the side shaft, the valve E by atmospheric pressure. At the commencement of the compression in-stroke the valve E closes ; the valve D being kept open by the cam during a part of the in-stroke, a portion of the charge is returned to and compressed into the chamber C, as well as in the cylinder and clearance space, the valve D is then permitted to close by the movement of the lever D 2 . The half charge in the cylinder and clearance space is further compressed, and immediately after the dead point and at the commencement of the third stroke of the cycle is exploded, doing a full working stroke, and having a double expansion, that is, expanded to double the volume the charge ignited occupied before compression. The cubic space of the chamber C must bear a correct proportion to that of the whole cylinder. Close to the working stroke end the exhaust valve G is opened and a part of the exhaust gases escape. A little after the dead point on the return stroke the valve D opens by its cam and lever, or by the pressure within the chamber C, and the half charge then pours into the cylinder, displacing more of the exhaust during the in-stroke, until the piston has covered the exhaust port of the valve G, which is then closed, and during the remainder of the in-stroke the second half charge is compressed together with the remaining products. At the commence- ment of the fifth stroke and the third revolution the second half charge is fired, Fig. 35, performing the second working stroke with double expansion, near the end of which the exhaust valve in the port F is opened by its cam, and the cylinder — except for the very small clearance space allowed — is completely exhausted on the in-stroke and last stroke of the cycle, when the series of operations commences anew. No engine upon this cycle was made, the patent has now lapsed, and therefore the cycle is open for anyone to use. Its advantages — (i) two working strokes in three revolutions, (2) in each working stroke the charge is expanded to double its volume when at atmospheric pressure, as compared with one working stroke in two revolutions in the de Rochas cycle, expanding to the same volume the charge occupied at atmospheric pressure — are so obvious that I need not further dwell on them. The power used in the process of compression is only slightly greater in this two-in-three revolution cycle than it is in the one-in-two, or de Rochas cycle. I believe that this cycle will be heard of in the future, but it will be an expensive one to work out and perfect. F 2 ( 72 ) CHAPTER X. Class 2, Type 9, Explosion Compound Engines. The leading idea of the next column, Type 9, is that of a second or further expansion of the working charge after it leaves the working cylinder in another cylinder, and by means of a second piston or second cylindrus, for further expan- sion may take place in the same cylinder on the other side of the same piston. There are some patent specifications prior to the first one mentioned in the table, q.v ., apparently describing compound gas engines, viz.: — Boulton, No. 766, of 1877 ; 2525, of 1878, and some others later, which appear to be constructed largely upon scientific principles known only to the inventor, and for that reason they were omitted. There is also a specification, Halliwell, 1450, of 1879, describing a non-compression engine, the charge of which is further expanded in another cylinder, but in the table, q.v., only compression compound engines have been included, as the compound engine is essentially the engine of the future. There are also some — for instance, Maynes, 1882 — whose chief characteristic is continuous combustion, and these are therefore included in the next column, Type 10, Class 3. There is a patent of Mr. Dugald Clerk’s, No. 4948, of 1882, which should possibly have been included in the table, but the inventor clearly intended to work it with very little or no compression. There are numerous engines described in which two cylinders are used — the one or even both serving for pumping purposes to draw in the charge — the pistons in both of which cylinders receive the impulse of the explosion simultaneously or nearly so. There are also two or three specifications in which it is the intention of the inventor to effect another working stroke by so cooling the working charge as to create a vacuum in a second cylinder. For obvious reasons these have not been included in this column, Type 9. Only those specifications are given in the table the engines described in which are true compound engines. The definition of compound, as applied to internal combustion engines, has been taken to be, and this column only includes, those compression engines in which the charge is ignited in a cylinder, and after doing the whole or a part of a working stroke therein, it, or a portion of it, passes to another cylinder or cylindrus to expand and do a further working stroke. In expanding to effect a second working stroke, the charge or a portion of it in which combustion is generally still going on, may pass to a second cylinder, to exert pressure upon a second piston, or it may pass to a second cylindrus, or expanding and contracting chamber, upon the other side of the same piston in the cylinder in which the ignition took place. One of the engines described in the specification of Holt and Crossley, 1884, ( 73 ) has been selected as the first representative of the compound engine, Class 2, Type 9, of the table. Fig. 36A is a section through both cylinders of the first engine described in this patent. The cylinder A is the high-pressure cylinder and B the low-pressure. The high-pressure crank is set from 70 deg. to 90 deg. in advance of the low-pressure. C is a valve which closes communication between the two cylinders ; the passage to the cylinder A is through a grid opening D. E is the exhaust port of the Fig. 36a Holt and Crossley's Compound Engine. Section through both cylinders of the first e?igine described in Holt and Crossley’s Specification of 1884. cylinder A. F indicates the position of the port for the admission of the gas and air. F should not really show in this sectional drawing, as it is on the top of the horizontal cylinder A. The high-pressure piston has a port H which registers with the port F in the cylinder A. The projecting piece K on the high-pressure piston is intended to prevent the passage of new charge through the exhaust port E. The low-pressure piston B is at the beginning of its out-stroke, a little later, the piston of A near its working stroke end uncovers the port D, and the valve ( 74 ) C has been opened and is held open by a cam, the products in A pass through to B and exert pressure on the low-pressure piston. The high-pressure piston continues its out-stroke and brings the port H in the piston opposite the port F, “ and owing to the reduction of internal pressure caused by the continued move- ment of the piston of B, a charge of gas and air is drawn ” into the cylinder A. The piston of A on its return or in-stroke closes the ports F and D, and then compresses the charge and products remaining in A, and at highest compression the ignition port I in the piston registers with the port in the cylinder leading to the igniter, and the charge is ignited and the high-pressure piston does another working stroke. A port L communicates between the exhaust port E and the low-pressure cylinder to allow of the exhaust products being expelled on the in-stroke of the low-pressure piston. W is the water-jacket. This engine was sufficiently ambitious ; a great many things were attempted in its construction: — (i) compounding; (2) the controlling the admission, ignition, and exhausting of the high-pressure piston, and the exhausting of the low-pressure piston by means of ports uncovered by the former, and the doing away with mechanism for valve opening other than that of the valve C ; (3) effecting the indrawing of the new charge without a pumping stroke. One of the chief objects of this construction was to do away with the slide — a source of great trouble in all gas engines at the time. The almost obvious questions which arise in considering this construction are : What becomes of the benefit to be derived by compounding and the “ propelling the piston of B ” by the products, if the continued out-stroke of B is to draw in the new charge ? Would the new charge enter in sufficient quantity to fire while F registers with H ? If it did enter in sufficient quantity, would it not flow out of the exhaust port in spite of the projection K on the piston ? Or, on the other hand, would not the exhaust products be in excess of the new charge during compression ? The idea of compounding is familiar as applied to steam, but no attempt has yet been successful to overcome the difficulties to be encountered in making compound internal combustion engines. Nevertheless, it will be seen by refer- ence to the column, Type 9, of the table that the attempts to effect compounding are more numerous than one would expect. Only one engine having any similarity to a compound internal combustion engine, as far as I am aware, has been placed on the market, and that only for a shoit time. This was, I believe, called the “ Acme,” produced by Messrs. Burt and Magee, of Glasgow, and was on view some years ago for a short time in a window in Broad-street, City of London. I have not been able to obtain further particulars of this engine, but I think it was made in accordance with the patent specification 14,578 of 1886, which is not a true compound engine, and has not, therefore, been placed in the column Type 9. It is really of the de Rochas cycle, modified by the fact that the explosion pressure is exerted upon two pistons simultaneously — the crank pin of the one piston makes one revolution while the crank pin of the other makes two by means of two-to-one gear wheels. The second piston is merely a large piston valve, so for as the de Rochas cycle is concerned. ( 75 ) To attempt compounding in internal combustion engines with two cylinders and . pistons is, in my opinion, to court failure, except in large engines. To convey the hot products of combustion from the hot cylin- der to another and cooler one is necessarily to rapidly cool down the products and give to the walls of the second cylinder a large part of the heat which it is intended to use as power. This loss would be much reduced in a high-speed engine. A judicious heating of the second cylinder, or parts of it, by an exhaust jacket might remove a part of the objection, but even then there are the increased friction, lubricating oil used, and first cost entailed in the use of the second piston and cylinder. A further difficulty that has to be encountered in internal combustion engines before compounding can be attained with advantage, is that of the cooling and expansion of the products by their rapid passage through a more or less contracted port to a cylindrus having a much lower pressure. I venture to think that the engine shown by Fig. 37 — specification No. 17,308 of 1894 — will overcome these difficulties. Fig. 37 shows a horizontal section through the cylinder of this engine. The de Rochas cycle operates with one variation at the back end of the cylinder, the same cycle, but without ignition, and therefore with only that amount of combustion arising from the flame which passes through the intermediate ports, takes place in the front end of the cylinder on the other side of the piston. In Fig. 37, A is the jacket cylinder, B the liner, C the cover, D the exhaust valve for the explosion end of the cylinder, E the admission valve for the mixture ( 76 ) of the air and fuel, F the second expansion exhaust valve, G the admission valve for air to the second expansion end of the cylinder J, D x the lever for opening the exhaust valve D, Ej the lever for opening the admission valve E. The two valves for admission and exhaust, F and G respectively, are opened in a like manner by similar levers and cams on the side shaft rotating at half the speed of the engine. H is the piston, H x the piston-rod, I the rotating side shaft, J the second expansion cylinder end, J x the piston-rod stuffing-box, K K are the two ports by which the hot gases while still in a condition of combustion pass from the end C to the second expansion end J of the same cylinder. The explosive charge is drawn through the admission valve E during the suction stroke, at the end of which, as the pressure on that side of the piston is as usual slightly below the atmospheric, a small quantity of the air and gases at Roots Fig. 38. Compound Cycle atmospheric pressure on the other or second expansion side J, which are then being exhausted, flow in through the ports K, the charge is compressed by the return of the piston H to a high compression, and is ignited at the dead point by a hot tube placed over the valve E, or other method, propelling the piston on its working stroke. Immediately the piston uncovers the ports K, near the end of the explosion working stroke, the gases and flame pass through the ports K K into the cold air which has been compressed on the other side of the piston J. Directly the piston covers the ports K K on the return or next in-stroke, the exhaust valve D is opened, and the products are exhausted in the usual way. On the second expansion side J of the cylinder, the hot exhaust products and flame, which flow through the ports K K, not only raise the temperature of the cold air, but also by increasing the total quantity of the gases, increase the pressure on this side J. This pressure impels the piston on another working ( 77 ) stroke, near the end of which the exhaust valve F is opened, and remains open in the usual way during the next exhausting stroke of the cylinder end J. During the next stroke air is drawn in through the valve G to the cylinder end J, which is compressed on the return stroke, and when the ports K K are uncovered, the flame and products pass through, heating the air as before described, and the double cycle continues. With the differences described, both sides of the piston carry out the operations of the de Rochas cycle, fuel being mixed with the air on the one side only, the combustion on this side serving to heat the charge of air on the other or compound side. Now if the compression piessure before ignition in the ignition and combustion end C of the cylinder be ioo lb. per square inch above atmosphere, then the pressure when the ports K K are opened by the piston near the end of the working stroke would be about 55 lb. per square inch above atmosphere. The air on the second expansion side is compressed to 36 lb. above atmosphere; the higher temperature and pressure from the other side of the piston will therefore raise the total pressure on this side very considerably to enable it to perform its working stroke efficiently. Moreover, during compression of the working charge, a certain amount of leakage will take place by the best fitting piston rings ; this is not lost as in the ordinary engine, but is retained on the second expansion side of the piston, and must be burned by the high-pressure flame entering from the other side. Fig. 38 shows the operations of the Roots compound cycle. Fig. 37 is to scale, and is really a small scale copy from the working drawings ; the piston rings have been omitted. Thus, the larger part of all that temperature and pressure usually passing out of the exhaust valve in the ordinary de Rochas engine, which is completely lost and wasted, is by means of this cycle utilised to obtain another working stroke. This, it must be remembered, is effected with the same consumption of fuel, and without any corresponding disadvantage beyond the additional pair of valves. This engine has not yet been completed, but it will, I believe, show a greater economy than any other type of engine. Since making the experiments described in Chapter VII., I have repeatedly exposed a thermometer marked to 630 deg. Fah. to the products of combus- tion in the exhaust pipe of a 4-horse power high speed oil engine. The thermometer was placed 2^in. from and outside the exhaust valve, and the bulb projected into the passing products \ of an inch. Within three seconds the mercury ran up to 630 deg. Fah., and if not taken away quickly would have burst the glass. Four seconds would be sufficient exposure to burst it. Numerous experiments which I have carried out at different times, and various things which I have observed in running gas and oil engines — which would make this chapter too long to enter into here — convince me that the temperatures hitherto allowed for the exhaust in most of the published tests of various engines on the de Rochas cycle are much short of the actual. Professor Robinson says that the exhaust temperature “ may reach as high as 500 deg. ( ) Cent, sometimes.” Dr. SlaLy gave the exhaust temperature in an Otto gas engine as 432 deg. Cent. In the Society of Arts Motor Trials — and no published tests of gas or oil engines since made in this country have been so trustworthy and thorough — the temperature is given as 1866 deg. Fah. absolute when the exhaust valve opens ; this, allowing a fall of temperature of 406 deg. Fah. for the passage through the exhaust port and valve, makes the exhaust temperature 1000 deg. Fah. If we allow it to be no more than this, then if we fully realise that a temperature of 1000 deg. Fah. and a pressure of about 40 lb. is being always wasted at the end of every working stroke in a de Rochas or Otto gas or oil engine, we must see conclusively that this cycle, though an extremely convenient and reliable one, is a very bad and wasteful cycle from the economical point of view. Some other promising and recent attempts at compounding have been made with three cylinders on one crank shaft. The two outside cylinders are high- pressure, and expand into the larger central one. The exhaust from each high- pressure cylinder exerts pressure alternately upon the opposite faces of the piston in the low-pressure cylinder. CHAPTER XI. Class 3, Continuous Combustion Engines. All the types and cycles hitherto considered were explosive engines, i.e., ignition of the whole working charge in the combustion chamber or clearance space of the engine is effected at once, and in such a fraction of a second as to be relatively almost instantaneous, although extreme dilution of the fuel by either products or air will render ignition slower. In this class, however, the working charge is ignited as it begins to enter the combustion chamber or clear- ance space, and the combustion flame goes on during the whole time of delivery of, and from the port of admission of, the fuel. The Brayton petroleum engine, one of the earliest and best known engines of this class and type, is selected as the first representative of the class. The patent was taken out in this country February 10th, 1872, No. 432, Gas Engines, by Brayton, of Boston, Massachusetts, U.S. Fig. 39 is a vertical section of the Brayton engine from the patent specification. A is the working cylinder provided with a water jacket — not shown — B is the work- ing piston connected by the rod C to the pump piston D ; E is the pump cylinder ; F is the pump admission valve ; and G the delivery valve. H is the inlet pipe for air, and I the supply pipe for gas or petroleum vapour from a carburetter, both of which are fitted to the mixing chamber J ; K is the receiver or pressure ( 79 ) chamber into which the fuel and air are pumped and retained under pressure ; L is the screw valve which controls the admission of the mixture to the cylinder ; M is the burner placed at the bottom of the working cylinder, it consists of a number of layers of wire gauze ; N is a second valve controlling the admission of the mixture, operated by the cam O, and returned by a spring. The cam O, by means of the lever O l5 opens the valve N during the working stroke of the engine. P is the combustion chamber ; Q is a mass of non-conductive material* — the specification says soapstone — to protect the piston from the high tempera- ture of the combustion chamber; R is a “ V-shaped channel ” cut into the valve seating of the valve N to form a bye-pass for the mixture, to maintain an ignition flame at the burner M during the non-working or exhausting strokes ; S is the crank shaft, the crank and connecting-rod are not shown. The exhaust valve does not show on this section, but it is opened by a cam on the crank shaft in the usual manner. Either gas or the vapour of petroleum spirit may be used as fuel, and the inlet pipes are fitted with screw valves to adjust the respective quantities of fuel and air that may be required. If petroleum be used the petroleum spirit, naphtha, is dropped upon a sponge through which air is drawn by the pump to vaporise it. Of ordinary gas it is stated that one part to twelve of air forms the best proportion for complete combustion. The gauze M serves the double purpose of a burner and to prevent the flame passing back to the receiver K. To start the engine, the fly-wheel is turned by Such a provision appears in numerous specifications, but is probably of very little use. ( 8 ° ) hand and the pump piston D draws in its mixture of gas or petroleum vapour and air through the chamber J and the valve F into the cylinder E. The return or upward stroke compresses it through the delivery valve G into the receiver K, to a pressure — a gauge is fitted to the receiver— of at least 60 lb. to the square inch. The screw valve L is opened by hand, the mixture flows through the bye-pass channel R to the gauze diaphragm burner M, and a light is applied to it through the exhaust port. The working piston is placed in its lowest position, the valve N is opened by the cam O and lever Oi, and the mixture for the working stroke flows through to the gauze burner, where it is ignited by the flame, and expanding in burning drives the piston upward. The mixture continues to flow through and burn at the burner through the greater part of the stroke. On the return or down-stroke commencing, the valve N is already closed, the exhaust valve is opened, and the products of combustion are expelled by the piston from the cylinder. At the next upward stroke the cycle re-commences. Although Brayton does not specify a cut-off period in the original patent, in most other specifications of engines upon this cycle it was usual to cut off the charge at about half the stroke, and Brayton subsequently did so in the engine exhibited in this country, which engine was also otherwise modified. The pistons of the pump and working cylinder were placed above and attached to opposite ends of a rocking beam, but this mechanical re-arrangement did not affect the cycle. Fig. 40 is a diagram showing the later Brayton cycle of operations. Fig. 41 is the working cylinder and pump indicator diagrams taken from a Brayton engine made in this country by Messrs. Simon, of Nottingham. This firm ultimately made this Brayton engine to run with gas instead of petroleum spirit, and attached a small boiler to the cover of the working cylinder. The exhaust products passed through a coil in the boiler and raised steam, which was used in the working cylinder. Any benefit obtained by the steam pressure would be more than counterbalanced by the cooling of the flame by the steam, and the conse- quent incomplete combustion. ( 8i ) In many specifications describing this class and type a regenerator is proposed to be used. In Siemens’ patents for continuous-combustion engines a regenerator is usually described. With regard to the value of the regenerator, theoretically there ought to be a considerable benefit from its use, and one would certainly expect a well-arranged regenerator, that without throttling saved heat, would effect some economy of fuel ; but it is significant that the only one of these engines — the Brayton — which had any commercial success was constructed without one, and it is very doubtful whether any benefit will be found in practice from the use of one. Great things have been expected of continuous-combustion engines, owing to their theoretical perfection of cycle, but I do not think it is possible to make an engine of this class that will excel in general convenience and reliability some of the types of engines of Class 2 — a good compound, for instance — or even in economy if the only actual and commercial test be applied, that of con- tinuous daily and all-day work. It is upon this principle that the Diesel engine is constructed, but instead of having a cycle of one revolution, as in the Brayton engine, it has two revolutions, and the cycle is in other respects similar to the de Rochas, except that con- tinuous combustion, the distinguishing feature of this class, is substituted for explosion. The Diesel engine, patent a.d. 1892 in this country, has been described by the inventor as “ novel in principle.” On reference to the table, Class 3, Type 10, it will be found that there are a considerable number working upon the same principle, namely, continuous combustion, and the inventor possibly made this claim not knowing what had been done previously. There does not appear to be anything new in the engine beyond the very high compression employed, and this is decidedly original. As far as I am aware, no English engineer had contemplated the use of so high a compression. But such a high compression will necessarily have certain disadvantages to possibly com- pensate for the advantages. Durability must be taken into consideration as well as economy. Fig. 42 shows a sectional elevation of the Diesel engine at right angles Fig 41. Motor Diagram Brayton ( 82 ) Fig. 42 ( 8 3 ) Fig. 43 ( 84 ) with the crank shaft, and Fig. 43 a sectional elevation through the crank shaft. C is the cylinder, P the piston, b the connecting-rod, c the crank pin, d the crank shaft. Q is the air pump (water-jacketed), driven by the levers X and links Z. The air pump is used to pump air under pressure into the receiver or reservoir L, Fig. 42. The air at a high pressure is then employed to spray oil into the air in the working cylinder through the nozzle or small orifice D. The valve shaft W is rotated at half the speed of the crank shaft by means of bevel wheels. Cams are fitted on the valve shaft, and its exhaust and air valves are opened in a similar manner and in like relative times to those in an ordinary Otto or de Rochas engine ; the chief point of difference being that in the Diesel engine the fuel is delivered to the air at the commencement of and during the working stroke, whereas in the majority of Otto engines the fuel is mixed with the air just prior to entering the cylinder. In Fig. 43 V 1 is the air admission valve, V 2 the exhaust valve, n the oil feeding valve, each opened by cams upon the valve shaft W. U is the lubricator supplying the annular tank T, into which the lips R on the piston dip. Y is the valve which is opened by a cam on the shaft W to admit the air pressure to the piston to start the engine. The cam for this valve is thrown out of action when the engine is started. In working, after the engine has been set in motion by the air pressure “ kept at forty atmospheres ” in the receiver L, the first out-stroke from the position shown in Fig. 43 draws in air through the valve V 1 ; the piston returns and compresses it to a high compression sufficient to ignite the oil. At the point of highest compression, air from the reservoir L sweeps the oil delivered from the petroleum pipe by the valve n through the nozzle D into the air in the cylinder, the temperature 6f which is raised sufficiently by compression to ignite it. After the piston has moved through one-fourth of its stroke, the oil is cut off and the charge expands ; but; no doubt combustion continues throughout the working stroke, and after the working stroke the exhaust stroke follows in a similar manner to that in an Otto engine. This engine lias created considerable interest, and it is advisable to correct some little misapprehension which possibly has existed owing to greater claims being made for it than were warranted. As has been said, there is not anything novel in it except the exceedingly high compression. There is no new theory — no new system. It has also been stated that “ one of the most important characteristics is its small dimensions as compared with explosion motors.” The diameter and stroke of the engine illustrated are 9‘8in. and 15 7 in. respectively, and the brake horse-power during two separate trials were 19*87 and 17*82 ; the mean of this is 18*84. I think most of the makers of gas or oil explosion engines in this country would be able to obtain the same power from the same diameter and stroke, and probably some of them a little more with a compression of, say, 75 lb. per square inch— a compression very considerably less than that used in this engine of Class 3. I think, moreover, that the economy of a properly designed explosion engine ( «5 ) of Class 2 of these dimensions could be made to compare very favourably with any continuous combustion engine of Class 3. It must be remembered, however, that at the present a great deal more time, thought, and labour have been devoted to the explosion engine of Class 2 than to the continuous combustion engine of Class 3, and that therefore there is scope for a more rapid improvement in engines of the latter class. The Hornsby oil engine has some points of resemblance to the Diesel in that it has two revolutions, the same four strokes, suction, compression, explo- Fig. 44 Williams and Baron j , sion, and exhaust, also an injection of oil into the cylinder ; but in the Hornsby the oil is injected earlier in the cycle, and it is an explosion engine of Class 2. The engine described in the patent specification of Williams and Baron No. 2 of 1879, has been selected as the other representative of this class. Fig. 44 is a section through the cylinder and fuel pump of this engine. A is the reservoir or chamber into which air is compressed and stored under pressure, D is the water- jacketed cylinder, B is the piston having the trunk extension E of less diameter, within which is attached the connecting-rod. Surrounding the trunk E is the annular space F, into which, by the movement of the piston B in the cylinder, G ( 86 ) air is drawn by the up-stroke and delivered by the down-stroke to the pressure chamber or reservoir A. G G are the valves by which the air is drawn into the space F, and H H are the air-delivery valves to the chamber A I is the pipe which conveys the air under pressure to the other end of the cylinder through the valve J 1} and the perforations S S in the grating J. K is a slide valve con- trolling the port U leading into the combustion space, it also controls the ignition. R is a gas pump, Q its piston, which delivers gas through its valve Ri, to the perforations S S under a pressure greater than that of the air. Z Z are the exhaust ports uncovered by the piston near the end of the stroke, and also controlled by a slide — not shown — which latter would appear to be a wholly unnecessary complication. The upward movement of the piston B in the cylinder D draws air into the annular space F through the valves G G, the down-stroke of the piston delivers this air through the valves H H to the annular chamber A, where it is stored under pressure. The air from the chamber A passes by the pipe I through the valve J 1? and the perforations S S in the grating J, controlled by the slide valve K, to the upper or combustion side of the cylinder. This delivery of air takes place towards the latter part of the upward stroke of the piston, by which it is further compressed. The higher pressure in D then closes the valve J 1} and at or near the dead point of the piston B and highest compression, the gas is delivered by the piston Q in the pump R, through the valve R 1} and through the perforations S S in the grating J. On issuing from the perforations, the gas is ignited by a flame in the slide valve K. The piston commences its down or out-stroke, mean- while the gas is continuously delivered through the perforations S S. At three- fourths of the stroke the gas is cut off, and the combustion products expand until the piston near the end of its working stroke uncovers the exhaust ports Z Z. More of the products are swept through the ports Z Z until they are again covered, on the return of the piston, by means of a flushing or scavenging quantity of air, which is permitted to pass through the valve K from the pressure chamber A. The piston then continues its up-stroke, and the cycle re-commences. “The engine may produce its own supply of compressed air, or the compressed air may be supplied by an independent pump.” To start the engine, “ the air may be supplied from a receiver, where the air is stored under pressure.” The points of resemblance in the Diesel engine to this engine are manifest. Neither the engines nor specifications describing engines having continuous combustion are sufficiently numerous at present for the same division as has been made in Class 2, but almost all the engines of the different types and divisions of Class 2 might also have applied to them the continuous combustion of Class 3 instead of explosion, and the majority might have this change of system effected with but very little change in mechanical construction. Generally an additional pump would be required for the delivery of the fuel under pressure. For instance, Williams and Baron’s engine, except for the substitution of con- tinuous combustion for explosion, is similar in cycle to Type 4, Class 2, while in the same way Brayton’s engine as at first made, and also as subsequently made, is similar in cycle, with the same exception, to Type 3, Class 2. I have been sometimes somewhat doubtful if it would not have been better to have numbered the first type of Class 2 in the Table as Type 1 of that class, thus making Type 9 Type 7, and to have omitted No. 10, leaving Class 3 for future division in the same way and into the same types as Class 2 is divided, when the engines become more numerous, as no doubt they will do. But — thought, idea, and even expression relating to the cycles of the internal combustion engine, especially in relation to one another, were chaotic, and the need for classification was crying.