A serialized version of this text commenced in the MODEL ENGINEER,
volume 134, number 3344, of 3 May 1968.


by Edgar T Westbury




In a series of articles published in M.E. over a year ago, I reviewed the various aspects of the miniture four-stroke engine and its progress over half a century of development. As a sequel to this series, it will, I trust, be appropriate to deal in a similar way with the two-stroke, which, in the size and form used by model engineers, is at least as popular as its more sophisticated rival. One of my friends, referring to the above articles, said "Very nice for those who like to play with valves and cams and things, but why have you neglected the two-stroke for so many years?"

True enough, I have not produced a new design for a two-stroke engine for a long time. This is not due to lack of interest, or even ideas for development of design; but the fact that miniature two-strokes have been produced commercially in great numbers and variety, and are readily available to those who want them, makes it more difficult to persuade model engineers to undertake their construction. Let me make it clear that I have nothing against the ready-made engines or those who use them for propulsion of various kinds of models; most of them work very well, and provide a source of power far more efficient than anything previously available. They also enable impatient enthusiasts to buy time, to say nothing of the elimination of hard work and skill involved in construction.

I have on many occasions given technical assistance to the manufacturers of engines, and many of the features of design which are now common practice were first applied by amateur designers, among whom I venture to include myself. But I have also tried to encourage individual constructors (when you come to think of it, this is really what model engineering is all about), and this is sometimes interpreted as a private vendetta against commercially made engines. Such an idea is clearly absurd, as the number of engines likely to be made in home workshops, even at the most optimistic estimate, could never be a serious menace to manufacturers. In same respects, any individual enterprise may even be helpful to them, by demonstrating the application and possibilities of miniature engines to the uninitiated more convincingly than the most expensive publicity campaign.

In this review, therefore, I propose to concentrate on the home-produced two-stroke, and trust that I shall not offend any persons or parties by doing so. Of necessity, I propose to indulge freely in reminiscence (or as some may consider, ancient history) about two-strokes in general and miniatures in particular, but I shall also do my best to consider modem problems and to answer some of the innumerable questions which have been put to me by readers.

Mainly historical

I have always found it helpful, in trying to explain the development of an engine or other mechanical device, to give at least a brief review of how it originated, and its early development. There are many who consider this unnecessary and, in support of their opinions, are fond of quoting the famous aphorism "History is bunk!" ascribed to the great Henry Ford. But nobody can deny that everything in the present and future has its roots in the past, sometimes a long way back. One of my objections to many modem text-books is that they often assume that the reader knows all about the foundations of the subject beforehand—or ought to —before starting to read them.

One of the questions I am very frequently taked is "Who invented the two-stroke?" This is a very simple question, but, like many such, not really so simple to answer. The early "atmospheric" internal combustion engines, such as the Hugon and Lenoir, were two-scokes, in the sense that combustion took place once par complete revolution of the shaft, i.e., two strokes of the piston. But these engines were completely superseded by the Otto engine, in which the mixture charge was compressed in the cylinder prior to ignition and nearly all later i.c. engines follow this principle, though not necessarily the identical working cycle. This includes two-strokes, in which it became necessary to incorporate some form of air pump, compressor or blower to force the charge into the cylinder as distinct from sucking it in during one complete piston stroke, as in the Otto cycle engine.

Most authorities credit Dugald Clerk with the invention of the precompression type of two-stroke, and it is probable that his engine was the first of such to be employed to my great extent. But Clerk's patent of 1881 was predated by that of James Robson in 1879. Besides individual features of design, the major difference between the two engines was that Clerk employed a separate pump cylinder for charging, while Robson utilised the reverse side of the piston, in a cylinder like that of a double acting steam engine, for the purpose.

In Fielding's patent of 1881, engine construction was simplified by using an enclosed crankcase as a charging pump. The engine patented by Day in 1891 carried this idea still further, besides eliminating mechanically operated valves and their operating gear. In Fig. 1, based on Day's parent specification, it will be seen that the charge of air/fuel mix (carburetion and ignition details are not shown) is taken in through a large disc valve in the end plate and, after precompression in the crankcase, is transferred to the cylinder through an automatic valve in the piston crown, while the exhaust is released through a port uncovered by the piston. Both these events take place dining the later part of the down stroke and the beginning of the up stroke.

In later development by Day and some other engineers, the automatic valves were eliminated, and all there events controlled by the piston through ports in the cylinder, as shown in Fig. 2. A deflector is fitted to the piston head with the object (only partially achieved) of directing the fresh mixture upwards and preventing it from escaping or mixing with the exhaust gases. In the "three-port" engine, as it is called, the mechanical system is reduced to the minimum of three working parts, and this has been the pattern for the great majority of two-stroke engines ever since. The diagrams are explanatory only, and not complete or exact in detail, but they are approximately correct in preportions and design of the actual engines.

Before the end of the nineteenth century, many individual inventors filed patents for "improvements in or relating to" two-stroke engines, but whatever their merits might been, manufacturers have not been enthusiastic about any departure which involves complication of design or construction. Much the same applies to the innumerable attempts to improve thes engines in more recent years and, apart from engines of quite large size, the simple three-port engine is still predominant, though it has bean considerably improved in detail, and its performance has been enhanced by long and patient development by many individual experimenters.

As a rival to the four-stroke engine, however, the two-stroke has many limitations and disadvantages. Although it is sometimes thought that an engine which fires twice as frequently as a four-stroke (other things being equal) ought to produce twice as much power, this is rarely, if ever, achieved in practice, and it is even difficult to attain parity of performance. Two-stroke engines have to answer for many sins; they usually more noisy, less economical, and more difficult to control than font-strokes; in some cams they are liable to overheating, tricky to start add generally temperamental.

Sophisticated engineers, for the most part, have tended to regard two-strokes as unworthy of serious consideration, and have neglected them in the general scheme of engine research. I have even encountered a rooted opposition to them, and have been told "We wouldn't have anything to do with two-strokes at any price!" This may partly account for the fact that in the development of full-size aircraft engines, few attempts have been made to use a two-stroke engine, though some promising designs have been evolved at various times. The late Professor A. M. Low once facetiously remarked "These wretched little (two-stroke) engines have no right to existence; they ought not to work at all, but, surprisingly, they do!" At the time when I became interested in motor cycling, two-stroke machines were considered definitely inferior to four-strokes, and, in fact, their riders were regarded as declasse. Things are very different now, and over 90 per cent of two and three wheeled vehicles have two-stroke engines.

I have always considered that the two-stroke should be given due consideration in technical literature, if only because most mechanically minded youths obtain their first introduction to i.c. engines by way of two-strokes of some kind or other. But many of those who use them have no more than a vague idea of how they work, to judge by some of the very naive questions they ask about them.

Fifty years of experimental work

I have been personally interested in the design of both four-stroke and two-stroke engines in small sizes almost as long as I can remember, but the first that I actually designed were of the latter types. They included an air-cooled motor-cycle engine, a three-cylinder engine for light aircraft, and a watercooled stationary (or marine) engine, all of which worked fairly well, but failed to achieve commercial success for reasons unconnected with technical merit. Most of the engines designed since that date have been "models" in the sense that they were intended to drive or propel models of some kind, though it is often contended that such engines are not models in the true sense of the term, as they do not emulate full-size engines in scale proportion. But this applies to nearly all miniature i.c. engines up to the present day; though it is not absolutely impossible to build them more or less accurately to resemble full-size engines of certain types, it is not generally expedient to do so, if they are required to work as efficiently aspossible.

In common with four-stroke engines, miniature two-strokes in the early days were not only few and far between, but also of dubious design, and no information on their construction was available. The first of my engines to be described in M.E. was Atom 1, built in 1925 with the intention of installing it in a half-scale model of the Comper CLA3 light aeroplane constructed by Cranwell RAF Apprentices. The engine ran quite well, except for sparking plug trouble (the miniature plugs then available were not at all satisfactory), but the 'plane never became airborne. As a "guinea pig" for further experiments, however, Atom 1 served a very useful purposes though it was too large to suit the popular sizes of model aircraft. I tried to stimulate interest in the development of suitable engines, but found it extremely difficult to persuade aeromodellers that anything better than rubber motors could be produced for this kind of work. It was not until the early '30's that I succeeded in obtaining the co-operation of (then) Capt. C. E. Bowden, and was able to prove that engine-powered model aircraft were practicable.

During the 1920's there was a great upsurge of interest in the development of model speed boats, and although flash steam was then regarded as the most efficient means of their propulsion, petrol engines were beginning to challenge their supremacy. From over the channel came that wizard of two-strokes, M. Gems Suzor, to astonish the natives with the performance of Canard and, later, Nickie 11. The M.P.B.A [rc: Model Power Boating Association] and the Modele Yacht Club de Paris got together to set up an International class of petrol-driven speed boats with engines limited to 30 c.c. capacity, and this was undoubtedly a milestone in the history of model power boats.

This application of small petrol engines offered, at the time, a more promising field for their development, and by taking advantages of lessons learnt by experiments with Atom I, I designed an engine of 30 c c. capacity, having features which I considered suitable for efficient speed boat work. It was named Atom II, and gave excellent results on the test bench, but, unfortunately, ended its career in a spectacular breakdown under load, due to a faulty crankcase casting. Its successor Atom III, of more robust construction, though similar in its general design, withstood the most stringent tests, and produced speed and power beyond my expctations; I still regard it as one of the most successful of my two-stroke designs.

Most model exponents of petrol-driven speed boats, however, preferred four-stroke engines, and regarded two-strokes with suspicion. I can remember only a very few of them who had any faith in the possibility of achieving high boat performance with a two-stroke engine. Notable among them was that great individualist from the North, Mr A. D. Rankine, whose Oigh Alisa dynasty of boats put up spectacular performances; another was the late Bill Rowe, of the Victoria M.S.C., whose boats were mostly noted for high speed, long distance endurance runs. My own attempts to produce results in this sphere were handicapped by lack of knowledge of hull design, and usually ended in crashes or capsizes.

Following up the success of Capt. Bowden's engine-powered aircraft, in demolishing records in their class which had been held for many years, I designed a two-stroke engine of 15 c.c. capacity in 1932, which I called the Atom Minor. In power and reliability, demonstrated in Bowden's Blue Dragon and other aircraft, proved beyond all doubt that all existing records in endurance, speed and altitude of model aircraft would become things of the past.

I shall refer later to the particular features of these early engine designs, which, although now becoming antique, are still capable of producing good results and for this reason cannot be dismissed as completely obsolete. While I have never claimed brilliance of design or super-efficiency in any of my engines, I think I can justly claim that many of their features, in structural methods and functional working parts, were innovations which have influenced the progress of design even up to the present date, including that of commercially built engines. I have always sought to exploit as fully as possible the potential advantages of two-strokes, including low frictional or other mechanical losses, and low weight, while reducing their inherent disadvantages such as imperfect charging and scavenging.

Up to and during the last war, I made many experiments in design of two-stroke engines from 50 c.c. down to 2-1/2 c.c., most of which, but by no means all, have been described in M.E., and nine of these designs are still available from the M.E. Plans Dept.; but others, and mutations (or mutilations) of them, have never been recorded, beyond references to them in a series of articles in Improving the Two-stroke published during the war years.

When the decision was made to produce an aircraft engine smaller than the Atom Minor 15 cc, the opinions of interested aeromodaillers were consulted, to ensure that the engine would be well suited to their requirements. Not only the capacity of the new engine, but also relevant features of in design and construction, and the mode of installation in the airfrme, were discussed in detail. The outcome of all this was a specification for an engine of 6 cc, which was described as the New Atom Minor, though I preferred to class it a "Mark II," became the term "New" has a short-term significance, and often comes all too soon outdated or unfashionable.

Though this engine fulfilled its puprpose, and gave good results in model aircraft of managable size, it was soon brought home to me that there was vary little interest in the construction of engines for this particular purpose. Ready-made engines were becoming available, and have since become almost universal, not only for the propulsion of aircraft but also for other kinds of models.

During the war, I was invited by a well-known firm of motor-cycle engine manufacturers to submit a simple engine design, suitable for production in moderate quantities, for a specific purpose not entirely unconnected with the war effort. The engine I produced was an improved version of the 6 cc Atom Minor. It gave results which satisfied requirements, and a number of "tool room" prototypes were made, but I was never fully informed what they were used for! (Official secrets and all that, you know.) The Atom Minor Mark III has never been described in M.E., but a handbook on its comstruction was published after the war by Percival Manhall & Co, and was fairly popular, though 6 cc has since proved to be an unfashionable size. An external view of this engine is shown in Fig. 3

But in the chronological order of things, the Mark III is out of its proper place, because in 1937 I produced the 5 cc Kestrel engine, which was not only successful from the functional aspect, but also appealed to many constructors who had not previously attempted to build a petrol engine. It featured in several early attempts by model engineers whose names later became well known in connection with model cars and boats.

The first model racing car built by Jim Crankshank was, I believe, powered by a Kestrel, and the design formed a basis for many modified and improved engines by other individual constructors. (Fig. 4.)

In post-war years, a 5 cc engine specially intended for model racing cars, known as the Cadet, was designed, together with another of 10 cc known as the Ensign, for either cars or boats. The early International class hydroplane engines of 30 cc Atom II and Atom III were followed up by Atom V, which had several unorthodox features of design. A great interest arose in the use of small engines for attachment to pedal cycles, and I designed the Busy Bee, 50 cc engine which was described in M.E. in the early 1950s and has been built and used by many readers. Some of these Bees, I understand, are still buzzing! A modified version, the Bumble Bee, was designed for light stationary work. Other post-war designs include the Zephyr 2-1/2 cc engine, the Ladybird 2-1/2 cc twin (C.I.) [rc: Compression Ignition, aka "diesel"] and the Cherub 10 cc twin.

It is not my intention to bore readers with a catalogue of engine specifications, but the above engines are, I think, worthy of mention as examples of experimental design. To illustrate their salient features of design, drawings of some of them have been reproduced. In a few details, these differ from the drawings originally published, to incorporate minor modifications introduced in the course of development. The word "experiment" so frequently used in these descriptions is often loosely applied, and may mean anything or nothing in particular instances, but generally it indicates some attempt to improve design in respect of functional efficiency, power-weight ratio, or materials and methods of construction. There is no advantage in altering design simply for the sake of being "different," but it is equally futile to adhere slavishly to conventions which, for all one knows, may be based on ancient fallacies.

The dominant component in an enclosed engine such as a two-stroke is (in most cases) the crankcase, which may incorporate other items in order to reduce the number of parts, reduce all-up weight, or simplify machining operations. In the early days of motor-cycles, a form of crankcase was developed which was well suited to the enclosure of internal flywheels; it was split on the vertical transverse centre-line, with main bearings in the two halves, and machined flat on the top to form a cylinder mounting platform.

This design has persisted, even when its original purpose no longer holds good, but I have rarely used it in two-stroke engines, as I consider that it involves unnecessary problems in machining the concentric register of the two halves, and the cylinder seating and register after bolting the halves together.

The blind-ended or "cup-form" crankcase is suitable for overhung-crank engines which do not need a power take-off on the end remote from the flywheel. But, the "barrel-form" crankcase, open on both ends, may be better suited or even necessary for assembly in certain cases. The former type was employed in the Atom I engine, in which the crankcase was extended upwards above the level of the exhaust and transfer ports; this was done mainly to reduce engine weight as much as possible. A cylinder casting (in iron) with ports and passages in it would inevitably have been heavier and might also involve molding difficulties. By forming the ports, with flanges, and also the transfer passage, in the light alloy crankcase castings, weight was reduced, and machining simplified. The cylinder was machined all over from fine-grained cast-iron, with fins on the upper part, and an external fine thread below the port level to screw, into the crankcase.

This method of construction, though mechanically sound, involves the problem of keeping close clearance in the port region to avoid leakage from one port to the other, and also in the orientation of the ports when the cylinder is screwed fully home. It is not generally convenient to fit a ring nut to lock the cylinder in the correct position, though this been done in some cases. The use of four (or more) long screws or bolts to secure both the cylinder and its head, as in the Kestrel and Atom Minor Mark II and III, is more common, but is open to the objection that differential expansion may cause the bolts to become loose (this has not happened in my engines).

If the crankcase or body casting is carried still higher, and provided either with fins, or an annular space to form a water jacket, it must necessarily be open at the main bearing end, or it would not the possible to assemble the connecting rod and piston. In the Atom Minor, engines, both ends of the crankcase were open, and apart from convenience in machining and assembly, this made it possible to reverse the engine from left to right, if this should be necessary, such as for paired engines running in opposite directions. In other engines, such as Atom II and Atom III, separate cylinders were employed, and the crankcase incorporated the main bearing housing. The Kestrel engine also had an integral crankcase and bearing housing, but the casting was carried upwards to a point below the ports, which were formed by a narrow belt pressed on to the cylinder skirt; a unique feature which undoubtedly simplified construction and contributed to the success of the engines.


In the smaller sizes of engines (less than 10 cc for instance) saparate cylinder castings, with flanges for ports and seatings are liable to come out heavy and clumsy, and it is therefore an advantage to design the cylinders so that they can be machined all over to a fairly light section, either as a plain sleeve or with fins on part of their length. The former type are sometimes objected to on the grounds that heat conduction may be imperfect, especially in view of differential expansion of the metals employed. But as the heat generated under working conditions always comes from the inside, it is clear that the liner will always be hotter than the outer part, and expansion will look after itself, provided that excessive local variations do not occur. In my early engines I took great pains to machine the parts to shrink fit limits, but later, I found this unnecessary, and easier fits were found quite satisfactory.

I have tried out various materials for cylinders (or their working surfaces), but I have yet to find anything better than good cast-iron for this purpose. Its low tensile strength is compensated for by high wear resistance and anti-scuffing properties; it will retain oil films and work with lower friction than most other metals. Steel, or at least the readily machinable grades of it, is much less satisfactory in my experience. I have, however, employed high carbon and other steels in some engines.

The Atom Minor Mark III had a low carbon steel liner which, after partial machining, was carburised on the inside surface to produce a "case" of high carbon steel. In the Ensign engine, the daring experiment of casting in a thin high tensile steel tube worked out successfully; bronze, and light alloy working surfaces were also tried with some success, but in the latter case at least, scuffing and general roughing were liable to occur, and one could not expect a long working life from such cylinders. Nowadays it is possible to electro-deposit a layer of hard chrome on light alloy cylinder walls, but I have not been able to get this done satisfactorily in my own engines.

Whatever kind of metals or form of construction is employed, it is of the highest importance that cylinder bores should be highly accurate and smoothly finished. It is not impossible to obtain the desired results directly from the cutting tool, but it is generally mere satisfactory to lap or hone the bores by methods which have often been described in ME. Provided that these methods are duly observed, the main ingredient in success is patience rather than skill. No pains should ever be spared, however, to ensure that cylinder bores are as truly circular and parallel as it is possible to make them. Final lapping should be carried out after ports are cut and liners inserted.

Port design

The size and location of exhaust, transfer and inlet ports is one of the most critical factors in the success and efficiency of a two-stroke engine. All engine design involves compromise between the ideal and the practically possible, and nowhere is this more true than in the two-stroke. The earliest engines with piston-controlled ports were restricted in both timing and area of port opening, as may be seem from the diagrams, Figs. 1 and 2. These engines were not required to run at very high speeds (500 to 600 r.p,m. was considered very daring at the time), but when higher speed and performance became desirable, considerably larger ports became necessary to allow for the passage of the gases in the very brief time permitted.

If there could ever be such a thing as "ideal" timing for a two-stroke engine, the exhaust port would begin to open at the instant the energy of the falling pressure in the cylinder became too low to be mechanically profitable; and the transfer port would begin to open when the cylinder pressure fell sufficiently to be balanced by that in the crankcase. It is clear that these conditions could only be obtained at one load and speed, and would be upset by changes in either respect. Moreover, the "ideal" closing times on the reverse stroke of the piston, to ensure the maximum charging efficiency and avoid undue wastage through the exhaust port, may call for different port locations.

Very little is known about the conditions of pressure and gas flow in small engines, and therefore a good deal of experimental work is called for to get the best possible results from a high speed engine. Unless the transient pressures in both the cylinder and crankcase can be recorded, there is no scientific basis that I know of that could be applied to the calculation of port sizes or timing. Apart from the actual timing of the ports, which is mainly determined by their height in relation to the piston stroke, their circumferential width is a factor in port area, and it is often possible to increase this with advantage; but the width of a single port is limited by practical considerations, and it may be necessary to increase the number of individual small ports to attain the same results. When piston rings are employed, the ports must be kept fairly narrow to avoid trapping the ring; and the bars between exhaust ports must be substantial enough to avoid overheating, with risk of distortion.

The actual period of port opening, in terms of crank angle, is also influenced by the length of the connecting rod in relation to the piston stroke. A short rod has a large angle of swing, and the exhaust and transfer ports are open for longer periods than their height in relation to piston stroke alone indicates. It is also possible to modify port timing to some extent by offsetting the cylinder in relation to the crank axis, and the same feature can be exploited to reduce side thrust on the cylinder walls during the firing stroke. The desaxé principle, as it is called (Fig. 5), has been used in several of my engines, including the Atom Minors; it appears to be most advantageous over the lower range of speeds, and for engines which run in only one direction.

Admission of mixture to the cylinder in the "three-port" type of engine, by means of piston-controlled ports, at the upper end of the stroke, also involves problems. It is clearly desirable for the ports to open long enough and wide enough to induce a full charge to enter the crankcase in the short time available, and also to reduce the unprofitable work of producing a partial vacuum therein before the ports open. But if the ports open early, they must inevitably close late, resulting in some of the charge being blown back on the return stroke, and this not only lowers volumetric efficiecy, but also impairs carburation, by reversal and shock in the air flow.

In engines of moderately high speed, it is possible to time the inlet ports of a "three-port" engine to give quite good results, and air flow can be kept more or less unidirectional by good design of the carburation and induction system. But above a certain speed—between 6,000 and 8,000 r.p.m., say—timing becomes critical, and inevitably aspiration or "breathing" becomes increasingly difficult as the port opening time becomes shorter. Some means of advancing both the opening and the closure of the inlet port is then desirable, and various devices have been introduced to do this; I shall deal with them later.

As I have many times been asked to give a suitable timing diagram for a small three-port engine, I have showed one [here], which may be used as the basis for experimental work. It differs from some of the diagrams I have given in the past, as it is based on port dimensions in relation to piston stroke, instead of in degrees of crank movement, which, as stated above, vary to some extend with connecting rod length. The inlet period is deliberately kept on the short side to avoid blowback at low and moderate speeds and give good acceleration under load—features which are valuable in practice, though often neglected in engines intended only to run at maximum speed.

The early two-port two-stroke engines, in which the air/fuel mixture was admitted to the crankcase by some form of automatic valve, behaved quite well within the limits of speed for which they were designed, but the valve usually restricted their efficiency at higher speeds. A valve which relies on difference of pressure for its operation must not only offer some restriction of flow, but also be subject to inertia, which causes delay in both opening and closure. Light spring loading may be employed to assist quick closing, but it increases the pressure necessary to lift the valve, and thus further restricts air flow. The valve may tend to be noisy in action, though this can be remedied to some extent by fitting a muffler in the air intake.

It is possible to incorporate the admission valve with a simple form of carburettor, by attaching or connecting to it a needle valve to control the fuel orifice; this was done on both two-storke and few four-stroke engines, and formed the basis of several automatic carburetters of early date, The "mixing valve" type of carburettor was used on both the Atom I 50 cc engine and the Atom Minor Mark I 15 cc engine in their original form, though as the latter was a three-port engine, it was not used for the direct control of admission to the crankcase. I shall deal with the subject of carburation in greater detail later on.

In the effort to obtain more rapid and efficient action of automatic valves, they were often made extremely light, in the form of discs or reeds similar to those used in high-speed air compressors. These had somewhat limited success, and in the case of reeds which were required to withstand elastic deflection, failure by fatigue was not uncommon. In recent years, interest in reed valves has ben revived, and they have been used successfully in some racing engines, including those by Mr S. H. Clifford in his Poly Ester, and other hydroplanes. The thin flexible reed generally favoured is the "petal" type—virtually a multiple-ported valve—made of a fatigue-resistant metal such as beryllium-copper alloy (Fig. 7.) I was informed by Arthur Weaver, of the Viewing M.S.C. some time ago that he was experimenting with a fully-floating rigid disc "rigid" disk valve, but I do not know what results have been obtained with it.

In passing, it may be mentioned that automatic valves are still used on some large marine and industrial two-stroke engines; these are mostly of multiple form, composed of reeds or free flaps. But crankcase compression, in these engines, has largely been superseded by displacement or exhaust turbo-blowers, and in any case, the rotational speeds involved bear no relation to those of the small engines used in model work, which are required to run over a hundred times faster in many cases. I must confess that my own experiments with reeds or other forms of automatic valves, in this particular application, have not been very successful, and I have found other methods of admission more reliable and consistent.

Mechanically timed admission

Many inventors have at various times made determined attempts to improve the "breathing" of the high-speed two-stroke engine by fitting some form of mechanically operated admission valve to the crankcase. The types of valves include sleeve or piston valves operated by eccentrics, and poppet valves operated by cams, but the simplest, and generally most satisfactory, are rotary valves of one form or another.

It may be said that rotary valves, in their general application to functional operations in steam and i.c. engines, hold a very poor reputation; they have all too many times been tried and found wanting in practical merit. The main objections to them are, that they are difficult to fit and maintain in pressure-tight condition, or to lubricate properly; they are also liable to pick up grit or carbon particles, to cause scoring of the working surfaces. But these do not apply in the case of an admission valve for a two-stroke engine, where neither the pressure nor the temperature are very high, and the surfaces can be kept continuously lubricated without complication or special provision for oil feed.

The first application of a rotary admission valve that I have been able to trace is that in a patent by L. Cordonnier (nationality not stated) in 1901. This refers to an overhung crankshaft engine, the journal of which was hellow, and provided with a port in its outer circumference, which registered with a similar port in the bearing, over an angular period timed in relation to the crank to admit mixture to best advantage.

A somewhat similar crankshaft valve was used a few years later in a small motor-cycle (or possibly an engine attachment for a bicycle), known as the Ixion (The name, incidentally, featured in Greek Mythology, and signified "the slave of the wheel", which seems quite appropriate). Several inventors produced their own versions of thr rotary admission valve, including Alfred Scott, a versatile pioneer whose motor-cycles were years ahead of their time, and set up a cult which still has its devotees at the present day. A French motor-cycle known as the Gillet employed a rotary valve, which if I remember rightly, was not actually formed in the crank journal itself, but was located in the other side of the crankcase and driven by a follower crank.

During the early days of my experimental work, I kept up discussion (by correspondence) with enthusiasts in many parts of the world. One of these was a young experimenter in California named Bill Atwood, who expressed great interest in rotary valves, and later became the manufacturer of a well-known miniature engine which was the first that I knew of to employ this form of admission. It was known as the Baby Cyclone, and had a cylindrical valve in the crankshaft as described above. Since then, many of the engines intended for use in model boats, cars and aircraft have adopted similar principles, with equal success. (Fig. 8.)

In the course of experimental modifications to Atom I I found it convenient to fit a rotary valve of the disc type to the crankcase, and this enabled the output of the engine to be substantially increased. A similar valve fitted to a Velocette U.S.S. 250 cc motor-cycle, produced equally impressive increase of speed and general performance. Most of my subsequent engines in which speed was of high importance have been fitted with disc rotary valves, with one or two exceptions where the cylindrical type was more convenient in the scheme of design.

The major advantage of any mechanically-timed admission valve is that it enables both the opening and closing of the port to be arranged where required in relation to crank angle, rather than being rigidly tied to piston stroke. A longer opening period is possible, as the port can begin to open early in the upward stroke, as soon as the transfer closes, or even earlier if this is found desirable. As this reduces the partial vacuum formed in the crankcase prior to the opening of the port, pumping losses are also reduced, and mechanical efficiency thereby improved. Other incidental advantages include less sharply interrupted flow of air, or abrupt reversal of flow, in the induction system, which tends to simplify carburation problems.

Both cylindrical and disc rotary valves give good results in high stated engines, but I have definite reasons for preferring the latter type. It is very readily adapted to most small overhung crank engines by a slight alteration to the blank end of the crankcase, and does not entail any enlargement of the crankshaft or modification of the bearing housing. The valve needs to be little more than a flat plate, posting on a central pivot, and driven in any convenient way from the crankpin. There is room for a port of adequate size in both the valve and its seating face, and tuning can be checked visually. As the valve is continually washed with lubricant (in engines which ran an oil/fuel mixture), friction is low, and wear very slight; it is also self-seating, but if and when the working surfaces become worn, they can both be re-lapped to restore them to good condition. Last but not least, port timing can easily be modified, or a different valve disc fitted for experimental purposes.

I have always had some misgivings about using a cylindrical valve which is combined with a heavily loaded crankshaft and its main bearing. Even though perfectly fitted in the first place, the sealing of the valve is liable to be impaired by wear, and cannot easily be removed. To alter port timing, a major component needs to be replaced, and exact checking of the timing may be difficult. The size of port is limited by the diameter of the crankshaft and the cavity in the shaft increases the crankcase clearance volume. It is sometimes claimed that the port in the shaft helps to balance the crankpin, but this is doubtful, because it is usually unsymerectical to the angle of the pin, and may therefore do more harm than good. In one make of engine, crankshaft fractures, due to the weakening in the area of the port, were not uncommon. Nevertheless, it must be admitted that thousands of small petrol, glow plug and diesel engines with this feature have given very good results.

It should also be noted that, with few exceptions, most two-stroke engines of commercial design, employed for motor-cycles or other vehicles, and also industrial purposes, have always been, and still are, of the three-port type, and the advantages of rotary or other valves have not been considered worth the slight extra complication in manufacture. But many prominent firms have made experiments with these and other improved features at some time or other, to my certain knowledge. By the way, just in case some well-informed reader tells on, that Scott motor-cycles did not have rotary valves, this is true so far as production models were concerned; but they were fitted to several experimental and racing machines.

Applications of rotary valves

The first of my engines having a rotary admission valve as a corporate feature of its design was Atom II (30 cc), which was described in M.E. in 1929, though, as already mentioned, it was the outcome of several years' development in this and other features. In both this and Atom III, the valve was in the form of a cast-iron disc with a projection to engage in the end of the crankpin, and mounted on a shaft with end-play adjuatment. These details apply also to the later 30 cc engine, Atom V. The Atom Minor Mark I was designed as a 3-port engine, but when used in a boat, a new rear endplate incorporating a rotary valve was fitted; and the smaller engines, Mark II and III, conceded the superiority of the rotary valve, as proved by many tests, by eliminating the piston controlled port. In the Kestrel 5 cc engine, the valve was located between the crank web and the main endplate in order to allow the fuel reservoir to be incorporated in the rear endplate—another innovation at the time, though it has since become common.

All these features can be seen in the drawings of the vaious engines including the Ensign (Fig. 9.) and it should not be necessary to point them out in detail. The disc valves in the smaller engines were were made from carbon steel gauge plate, which, being ground on the surface, only needed to be lapped on a glass plate to produce a true, hard wearing seating on the light alloy endplate. They worked best when pivoted on a stationary bolt, rather than attached rigidly to a rotating shape. It was, however, found that the centre hole in the disc tended to wear oval, indicating that it was out of balance, and the addition of extra weight in appropriate location enabled this to be remedied. The driving pin was flush riveted to the disc, and, provided that any projection was carefully eliminated, this has proved quite satisfactory.

I have dealt with the subject of admission valves in some detail because they are in my opinion one of the most important functional features of high speed engines, and have a far reaching effect on efficiency. My early efforts to encourage the use of rotary valves have aroused a good deal of controversy among those interested in two-stroke design.

Before dismissing the subject of rotary admission valves, a brief mention may be made of unusual applications of them. These include gear or chain driven valves, of the hollow cylindrical type, which are suited to engines of more than one cylinder where there may be difficulty in locating the simpler forms of valves. Conical valves have been used in a few cases, and I have reason to believe that the late Mr Heath, of the Victoria M.S.C., employed this type. The combination of a disc crank with a rotary valve has also been tried, but there are objections to making the valve integral with or rigidly attached to a major working part which may be subject to deflection under load. It is better that the valve should be free to float, but capable of end play adjustment so as to work with the minimum clearance. Spring loading, to keep the valve up to its seating is undesirable on the grounds of iincreased friction.

It has been mentioned that mechanical control of exhaust or transfer port timing is liable to be difficult and complicated; but so far as the transfer port is concerned, this problem did not defeat the wit of the redoubtable Gems Suzor. In his 10 cc engine, as used in his first post-war hydroplane Mdlle. Sylla, the rotary valve controlled both admission and transfer ports in the crankcase. The latter led to a forked pipe which injected mixture into two of the open exhaust ports at just the right moment before they closed. This ingenious device was undoubtedly successful, but unfortunately could not be used with a closed exhaust pipe and silencer.

Crankcase compression

The limitation in volumetric effeciency of two-stroke engines is primarily due to the incomplete charging of the cylinder by using the crankcase as a pump. Apart from the loss of effective volume through imperfect port timing, the necessary clearance space to accommodate working parts imposes a limit on the pressure which can be produced in the crankcase. Many designers have taken staps, sometimes drastic, to reduce crankcase clearance to the bare minimum in order to improve both suction and pressure effects.

While these measures are often beneficial, they may in some cases defeat their own purpose by increasing the load involved in charging. The function of the crankcase is not so much that of a pressure pump as a displacer, and work done in producing pressure is a dead loss from the aspect of mechanical efficiency, which is one of the virtues of the simple two-stroke. Unfortunately, a certain amount of pressure is necessary to charge the cylinder in the very brief period allowed for transfer port timing, in a high speed engine; but, again, too high a transfer pressure may cause turbulence in the port and impair scavenging.

Attempts have often been made to improve the volumetric efficiency of two-stoke engines by using an oversize charging pump of some kind. This may take the form of a separate charging cylinder, as in the original engine by Dugold Clerk, or a rotary blower. Alternatively, crankcase displacement can be increased by adding a crank or or eccentric driven displacer piston, or using a stepped main piston. This was done in the Dunelt motor-cycle, which had a certain vogue in the 1920's, but did not prove the advantages of the method. Such engines have sometimes beeen described as "supercharged" but this term is more correctly applied to engines in which the cylinder is charged at more than atmospheric pressure.

This is clearly impossible in normal two-strokes in which exhaust and ports are open simultaneously. They can, however, be "super-scavenged" by an oversize charging pump, often with some advantage, but at the expense of economy, because wastage of fuel through the exhaust port is inevitably increased. This may be tolerated in racing engines, and is no disadvantage in diesel engines of the injection type, which are charged with air only. In my experiments, however, I have failed to obtain any substantial increase of performance in two-strokes by increasing the charging pump volume.

Piston design

It may be said that the piston of a two-stroke engine has a fourfold function, as it must act as a cylinder seal, a pressure transmitter, a piston valve and a crosshead. Unlike a four-stoke engine piston, which may be fitted to the cylinder with a large clarance if spring rings are used to provide the seal, a close contact of the piston to the cylinder wall is essential to avoid leakage at the ports. The effect of differential expansion of the parts is therefore pronounced in two-stroke and where light alloy pistons are used, it is advisable to use a low-expansion alloy, and also to design them in such a way that heat is conducted away front the crown as rapidly as possible. In some respects the air flow on the underside of the piston helps to keep it cool.

In most of my early two-stroke engines, cast-iron pistons were used, became it was possible to fit them to fine clearance, without risk of seizure when hot. It was generally possible to obtain ready-made piston rings in sizes of 1 in. dia. and over; the bore size of some of my engines was determined by the standard sizes of rings available. For the smaller engines, no ready-made rings could be obtained during the war, but they could be dispensed with by observing the highest possible accuracy, and finest clearance, in finishing the cylinder and piston. I still recommend ringless pistons on engines of say 3/4 in. bore or less, as I have one or two engines in which these pistons have given good results for many years, and still hold good conpression. I do not propose to discuss the making of piston rings here, as it has been dealt with by several other contributors.

It is often necessary to make small pistons from solid material, owing to the difficulty of obtaining accurate castings. The main problem involved in doing so is to machine the inside to reduce weight as much as possible, while retaining adequate strength. Having made many pistons in this way, I can recommend the methods shown in Fig. 10. The material (preferably cost-iron stick) should first be machined parallel on the outside, so that it can be chucked truly, and, if necessary, rechucked for subsequent operations. After facing and centering, it is drilled and bored to dimension D1, and finished to required depth by a flat-ended drill or and mill. The dimensions D2 and D3 can be bored out by a round-nored boring tool, which may be made from round silver steel bar and held in a split holder. Depth measurements can be taken from the end of the holder, and undercut diameter D2 can be measured by means of the cross feed index. If no rings are to be fitted, D2 can be opened out to the same size as D3. Outside diameter can be finished at the same setting if desired, but it is generally better to carry it out in a later operation.

Before parting off the piston (at the height of the deflector tip) the chucking piece can be retained for convenience in holding it while cross drilling for the gudgeon pin, and milling to remove surplus metal from the centre belt. For shaping the crown of the piston to form the deflector, either by hand or machining, I recommend making a split holder, bored to fit the piston, which will hold it firmly in the vice without marking or distortion.

Composite pistons

It is sometimes convenient to make the piston in two or more parts, and perhaps [of] two different materials. The methods which have been employed in full-size practice are not well suited to very small pistons, and the usual method is to make the shell in the form of a light cup, machined all over, and fit a yoke inside it to carry the gudgeon pin, as shown in Fig. 11. Sometimes the yoke is secured by a single screw in the centre of the crown, but two screws are to be preferred, especially in view of the need to locate the line of the deflector in relation to the gudgeon pin centre. It is not desirable to allow the screw heads to project above the crown, and if the deflector is thicker in the centre, they may be sunk well below the surface and filled flush with plastic metal. To prevent any risk of the screws becoming loose, a sealant such as Loctite or Casce ML may be applied to the threads on final assembly, which of course must be carried out after the gudgeon pin and connecting rod are fitted.

Pistons of this type can be made very light, especially if there is no need to thicken the upper part to take piston rings. They also solve the problem of restraining end movement of the gudgeon pin, to prevent risk of scoring the cylinder walls. Their heat conducting properties, however, are inferior to those of integrally cast or machined pistons, and this may be a disadvantage in engines which are required to develop their maximum power for any appreciable length of time.

In small piston casings, it is often difficult to core the inside contour, and the gudgeon pin bosses, accurately. Generally, a corebox is required, and it is advisable to provide a large core print at the base, with a key to ensure alignment with the deflector on the outside of the crown. It is possible to dispense with the corebox by carrying the bosses right up to the inside of the crown and allowing liberal taper for draught, but moulders do not like this method of moulding as a rule.

I have succeeded in making quite small pistons accurately by a simple method of die casting, and saveral other model engineers I know have employed metal moulds of some form or other, in a similar way. The core of the mould shown in Fig. 12 is machined in steel or cast-iron integral with or attached to a base plate, which should preferably be made so that it can be clamped to the bench, or in the vice. Recesses are milled in both sides to form the gudgeon bosses, and these, together with other surfaces, should have a draught angle of 5 to 6 deg. inclusive. The outside of the mould is a thick-walled steel tube, with an inside diameter giving allowance for machining the outside of the casting. It is registered concentrically by a spigot at the base of the core and should be quite an easy fit.

One of the secrets of successful die casting is to avoid trapping air or gas anywhere in the mould—in other words, proper venting. This is assured in the mould shown by cutting fine grooves in the register spigot with a point tool or engraver, and making notches in the end face of the tube, to allow air to escape freely. An ejector screw, of large enough aim to avoid excessive local pressure on the piston crown should be provided. To prevent adhesion of metal to the mould, the outside of the core and the inside of the tube should be coated with grate polish, or a paste composed of clay and powdered plumbago (or graphite) in water. This should he baked on, and the entire mould heated to such a temperature, that if you spit on it, it spits back!

It is equally important that the metal used for die casting should flow freely whern melted; any old bit of scrap metal just will not do. The aluminium alloys intended for die casting have free flowing qualities, but do not machine so freely as most other alloys. An iron ladle, on a gas ring, may be used for melting the small amount of metal required, but the inside should be coated with the mixture as above to prevent contamination of the alloy. When poured into the top of the tube, it should fill the mould properly by gravity, but it is possible to apply pressure by the method well known for "lost wax" casting, namely by applying a cap containing a damped asbestos pad to the top of the tube, or by pushing the molten metal down with a loose ramming plug—but beware of splashes! It is possible to improve on this method by using a closed mould to form the deflector top, but this is more complicated, and I have found it easier to machine this, as already described.


The high vertical deflectors used on the early two-stroke engines (see Fig. 2) performed their primary function of directing the fresh mixture upwards, and deterring its escape from the exhaust port, fairly well; but they were not good heat conductors, and often became overheated to such an extent as to cause distortion of the piston, if the engine was hard pressed. In later engines designed to run faster and produce more power, it was necessary to thicken the deflector to avoid local heating, but excess piston weight could be avoided by coring the inside of the piston to follow the external contour. The shape of the deflector is widely varied in different engines, and I have often been asked "What is the ideal shape for the deflector?", but this is another of those simple questions which are not so simple to answer. I have tried out almost every conceivable shape of piston crown in my experiments but have not arrived at any definite conclusion.

In full-size practice, deflector top pistons have largely been superceeded by the flat top type, the major practical advantage of which is that it reduces the heated area of the piston to the minimum, and its symmetry reduces the tendency for it to become distorted. Engines can therefore be run at full power without distress, and the design of the combustion chamber is also simplified. In small engines the flat top piston is equally practicable, but it makes the design of the cylinder ports and passages much more critical, and thus it may be more difficult to ensure that the engine is effectively scavenged, and wastage of mixture from the exhaust port avoided.

Engines of especially high compression ratio, such as those working on glow plug or compression ignition, may have insufficient clearance space in the head for a deflector piston, and a flat top—or something like it—may be a neccessity. But apart from this, some of the engines which have put up the highest performances, in sizes of 10 cc, or less, have pistons with deflectors, though often only very rudimentary ones. This tends to support my personal belief that any deflector is better than none, and that scavenging is generally made easier by some internal means of directing the transferred mixture upwards. The exhaust side of the deflector appears to be of less importance, but may be influenced by compression ratio and combustion space. Imperfect scavenging is evidenced not only in general performance, but also in carburation which tends to be critical in adjustment, and narrow in speed control.

The "Zephyr" 2-1/2 cc engine

This engine is illustrated not only to exemplify some of the features discussed, but also methods of construction in an engine made without castings. Many constructors have built engines by fabrication from stock materials, or machining from solid, or a combination of both methods. While the functional design of such engines may be more or less satisfactory, it is not at all easy to obtain a balanced distribution and economy of material which is essential to really good engineering design. It can be done—but few constructors, in my experience, are prepared to carry out the intricate machining operations involved in putting just the right amount of metal in the right place, and no more. No doubt it can be argued that it does not matter what an engine looks like so long as it works; but a good craftsman will always seek to make the outward appearance of his product a sign of its hidden inward merits.

The major parts of the Zephyr (Fig. 13) were machined from solid duralumin, which by reason of its high tensile strength, enabled the section of metal to be kept to the minimum without risk of weakness. Operations on the upper part of the body were mostly simple internal and external turning, but the lower part, besides boring at right-angles, required shaping on the outside by part-circular milling, planing, and a certain amount of hand filing. The main bearing housing was milled externally to a near-cruciform section, to form four strengthening ribs; they could have been dispensed with by making the circular part of heavier section, if a little more weight was tolerated. This applied also to the rear housing, in which a cavity was milled around the centre boss and carburettor stub.

The parts were fixed together by screwing the joints 24 t.p.i., and no screws or bolts were used in the entire structure. A slotted locking collar was fitted to the rear housing to enable it to be located correctly, and this, incidentally, allowed of adjusting the timing of the rotary valve if required. The main bearing housing, however, was screwed firmly home to its shoulder, and a frictionally mounted contact breaker, similar to that of the Atom Minor, Mark III, was fitted, and operated by a cam formed on the back of the propeller boss.

A somewhat unusual feature was the transfer passages which had upper and lower ports, the latter communicating with the crankcase through ports in the piston skirt. The piston of a two-stroke engine usually needs to be a little longer than its stroke, in order to avoid opening of the exhaust port at the top dead centre. Them are exceptions to this rule, when the intention is to utilize the velocity of the exhaust flow to produce an extractor action, or conversely, the back pressure to boost the crankcase pressure. The benefits in either case depend on several factors which are difficult to calculate, and from my own experience, I am of the opinion that exhaust leakage into the crankcase is best avoided, if only because it contaminates the fresh charge and makes carburation more critical as a result. Experiments under closely observed test conditions will show whether any advantage can be obtained by this or similar measures.

Within reason, the longer the piston is, the better, as it acts as an efficient crosshead, lowers the side thrust pressure, and helps to retain the lubricating oil film. But a long piston calls for a long cylinder, and in most small engines, compactness is a desirable virtue. The tendency towards the short stroke or "over-square" engine has to some extent been influcenced by the fact that it enables the maximum cylinder capacity, to be obtained within limited overall bulk and weight. Two-stroke pistons must maintain close contact with the cylinder walls throughoort their length; they cannot be, "waisted " to relieve the centre, or grooved to assist oil sealing, beyond mere scratches, or interrupted grooves which pass only partly round the circumference. Continuous grooves would form a communicating path between exhaust and transfer ports, and besides affecting the piston seal, would cause oil trapped in the grooves to be blown out and wasted. All these details are simple and obvious, but it is surprising how often designers or constructors defeat their own attempts to improve engines by perpetrating obvious fallacies.

Piston head transfer valve

In the original Day two-stroke illustrated in Fig. 1, the transfer of mixture from the crankcase to the cylinder was carried out by means of an automatic valve in the piston head. In some respects this might be considered a very good arrangement, as the valve, being operated by difference of pressure, world open as soon as the cylinder pressure became lower than that in the crankcase, and remain open until the conditions were reversed. But this ignores the effect of inertia, which causes delay of both opening and closing, besides possible erratic action by the location of the valve in the moving piston, especially at high engine speed.

The piston head transfer valve, in various forms, has been used in many engines in the past, but appears to be almost completely obsolete now. An interesting variation of the principle has, however, been employed in recent years, and is capable of working efficiently at very high speed. As seen in Fig. 14, the piston comprises two parts, one being a symmetrical outer shell, and the other, fitted to slide freely inside it, having bosses to carry the gudgeon pin. The innner part, or yoke, has a conical head, which forms a valve fitting a seating in the underside of the circular aperture of the shell.

While there is positive pressure on the hard of the shell, during the compression and firing periods, the transfer valve remains closed, but when the exhaust ports (which are located in the cylinder wall, as usual) are opened, and the cylinder pressure drops below that in the crankcase, effort on the shell ceases and it slows down or stops, so that the transfer valve is opened by the continued (positive) movement of the yoke. In order to guard against the risk that the two parts may become completely disconnected, if the shell becomes stuck near the top of its stroke, a circlip or other retaining device is fitted inside the shell, to limit the play allowed in opening the transfer valve. The speed at which this device will work properly is limited by the inertia of the shell, which must be made as light as possible consistent with strength.

Where pistons have one or more ports in the skirt wall to provide communication with lower transfer ports in the cylinder, care should be taken to avoid undue weakening or reduction of the thurst area. If the ports are at the sides (i.e., in line with the gudgeon pin axis), they are not affected by the thrust load and it is possible to use a slipper type piston as in the Ensign 10 cc engine. But cutting away the lower rim of the piston may cause it to spring out of round, and this tendency should be carefully guarded against.

I have on more than one previous occasion described the means I recommend for finishing the outside diamatcr of the piston, after all other essential operations have been caried out, It consists, as shown in Fig. 15, of a disc of metal which may be held in the lathe chuck or made with an internal thread to fit the mandrel nose, having a truly concentric spigot to fit the machined inside of the piston skirt closely, but without any tendency to expand or spring it. The centre disc is drilled and tapped to take an eye bolt, or alternatively this may be made in the form of a long draw bolt to pass right through the mandrel, and take a draw nut on the outside. When the piston is fitted to the spigot, a temporary, shortened gudgeon pin is passed through the bosses and the eye of the bolt. The piston can then be drawn firmly home against the shoulder of the spigot far final machining, with no risk of distortion through chuck pressure or other causes. Measurement of clearance is facilitated, or the cylinder may be offered up to the piston far testing the fit.

Crankshaft design

Overhung cranks are fitted to the majority of small two-stroles, as it is generally convenient to transmit the drive from the flywheel end, and power take-off at the other end is not required, except for light auxiliary drive in a few cases. There is rarely any advantage in fitting a "full" crankshaft with journals running in bearings at both ends, and it may in fact complicate construction or assembly, by necessitating a split big-end bearing, which, however well fitted, is often a potential source of trouble in a high speed engine. All the single cylinder two-stroke engines from 50 cc downwards which I have built have had overhung cranks, with the exception of the 15 cc Pheonix, the Busy Bee, and the Bumble Bee. Engines with more than one cylinder obviously call for more complicated crankshaft construction; these include the 10 co Cadet and Craftsman twins and the Ladybird 2-1/2 cc (C.I.) twin. In all but the last mentioned engine, however, the crankshafts were built up, and capable of being disassembled, so that split big-ends were not necessary.

I have explained in several constructional articles that there are many ways of making crankshafts, and all of them are good if properly executed. The traditional way of making a sound crankshaft in one piece is by machining from a forging in high tensile steel, but this is not obtainable at present so far as I know. I have experimented with castings in special steels with some success, but these are also difficalt to obtain, and more often than not, "solid" crankshafts must be machined from bar stock, with considerable wastage of material and expenditure of time in the process. I generally recommend this method of construction because it eliminates uncertainties in mechanical soundness, and is conducive to accuracy if correct machining principles are observed.

For "permanent" fabrication of crankshafts, I consider brazing (which includes silver soldering) to be most satisfactory, though it limits the choice of material to some extent, and the effect of heating may be harmful to some grades of steel. It is posssible to braze up an assembly with the crankpin and journal pre-machined to size, so that no after-machining is necessary; but if this is not so, the necessary setting up for finishing these parts, in addition to other operations, may absorb as much time as machining from solid material.

Construction of crankshafts by simply pressing the journal and the crankpin into bored holes in the web is quite practicable, but it demands special accuracy in ensuring correct limits of interference, and also a suitable press to assemble the parts. It is generally necessary to make the web thicker than normal to ensure sufficient area of seating surface. These conditions apply also to crankshafts in which the parts are clamped or taper fitted; constructions which involve the need for riveting (i.e., burring over) or expanding the shafts are not recommended as they involve the risk of forcing them out of truth, or "mushrooming" the journal ends. The amount of interference for a secure press fit should not exceed 1-1/2 thou (0.0015 in.) per inch of diameter.

For mounting flywheels on shafts, I have never found anything better than a well-fitted taper fit, with a draw nut, but without a key or other positive location, unless it is necessary for timing or other essential purpose. It is better that a flywheel should slip in emergency, such as a sudden stop or excessive torque overload, than that the shaft or bore should be torn or damaged, but this does not mean that one can afford to be careless about the proper matching of the tapered surfaces. Engines of the larger sizes may have the taper turned directly on the ends of the shafts, but in many of my engines, I have found it better to employ a tapered split collet, fitted either to a plain shaft, or (preferably) to a slightly reduced part of it which gives positive end location. The collet should on no account be loose on the shaft, as this may cause some distortion when it is compressed; it should fit tight enough to allow it to be finish-machined after mounting on the shaft by friction alone, if necessary.

The collet provides additional bearing area in the flywheel boss, and avoids weakening the end of the shaft or reducing the size of the securing nut, which is often too small for safety, especially when it is combined with a screwed shaft coupling. Only one sawcut through the collet is (or should be) necessary, but two or more slots partially through it may be helpful in some cases. Finer limits of accuracy must be observed in the machining of the components, but they are by no means beyond the skill of "average" model engineers or the capabilities of their lathes.

It will be seen from the sectional drawings of typical engines that most of the crankshafts are in one piece, but the exceptions to this rule may be of some interest. In the Busy Bee engine, for instance, the web and crankpin were machined from solid, though a brazed version was tried successfully. This was bored to fit a taper seating on the main shaft, keyed, and drawn up on it by a sunk head set screw. One of the rasons for this unusual form of construction was that in the experimental stages, the shaft design was subject to variation, to suit differences in the method of drive; it was therefore regarded as the "expendable" part, which could be changed without difficulty.

In the Bumble Bee engine, which was identical to the former in most of its internal design, the requirement of a double-ended shaft was met by making it in five pieces—two journals, two webs, and one crankpin. The parts were made a light press fit to each other, and further secured by partially splitting the webs on the centre line, and fitting a high tensile dowel bolt, the hole for which cut into both the journal and the web, thus forming an effective key to prevent movement of the parts when assembled. This construction allowed of using a hardened crankpin, while leaving other parts soft, with less risk of distortion. This applies also to the crankshaft of the Phoenix, but in this case each journal and web are in one piece, and the crankpin is a light press fit in one web, and a lighter fit in the other, so that it can be detached without great force. It may be made of mild steel, brass or aluminium alloy, and is easily renewed if necessary; it may also serve as a spacing or oil sealing sleeve, also to clamp a ball race if required. The taper I have found suitable is about 10 degrees inclusive (5 deg. on the swivelling slide) for both direct and collet fitting. There used to be reamers available for angles of this nature, including the standard magneto taper, but if available they should only be used for light finishing of pre-machined bones.

The choice of material for crankshafts is of great importance, as they are subject to heavy working stresses, besides being subject to running wear. Special steels have been developed for crankshafts, but it is not easy to obtain them, and for my engines, I have found it necessary to use mild or low alloy steel, with or without locally hardened surfaces. High tensile steels where available, are undoubtedly to be recommended, but it is most important to know what the steel really is, in physical properties, and the effect of heat treatment. Odd pieces of scrap steel are not to be trusted, unless they can be tested or analysed. It is worth while to normalize any samples of forged, hot rolled, or cast steel before final machining, by heating to redness and allowing to cool slowly, unless the sample calls for special treatment. This relieves internal stresses, avoiding the risk of the component springing out of truth after machining and reducing the risk of distortion, if hardening or other subsequent heat treatment should be required.

High carbon steels, in general, are not to be recommended for crankshafts, unless they are of a grade intended to cope with heavy mechanical stress, such as for crane hooks or locomotive braw bars, as distinct from tool steels. Silver steel [rc: drill rod], though very useful for many workshop purposes, is not the best steel for crankpins or journals, especially if built up by brazing. In its normal state it is better than mild steel in hardness and durability, but inferior to it in toughness, and if hardened, is liable to become brittle. Local hardening of bearing surfaces, though not necessary for engines of moderately high duty, is generally worthwhile, and if the heating involved in case hardening is undesirable, electrolytic deposition of a thin layer of hard chrome will produce an equally satisfactory result. This process, which is widely used for the reclaimation of worn components of car engines, is undertaken by many firms specialising in this class of work.


There is a choice between plain (bush, sleeve or split) bearings and rolling (ball or roller) hearings for engine crankshafts, and it is not easy to state definitely which is the best in all cares. The bearing loads in two-stroke engines are usually lower than thase in four-stokes, and it might be expected with good lubrication, as provided by "petroil,", almost any old kind of bearings would give good results. If properly made and finished, plain bushes of gunmetal or medium bronze will work well and long in engines of high speed and duty; phosphor bronze, so often regarded as the ideal bearing metal, should only be employed if the journals are hardened. The properties of aluminium alloys, as bearing metals, are much under-rated, and the only disadvantage of running shafts dircutly in the crank-case housing is that it makes renewal of worn bearings more difficult, if it becomes necessary.

In the engine with all overhung crank, most of the journal load comes on the ends of the bearing. especially the inner end. It is therefore prudent to make the bearings fairly long, or to insert separate bushes from both ends, to give support against the rocking effect produced by the crank. If the total distance between the ends of the bearing is too short, wear is liable to be more rapid than it should be. I have generally specified adequate effective bearing length in my engines, as can be seen from the drawings of examples reproduced; little is gained either in the bulk or weight of engines in skimping the length of bearings or their housings.

By using ball or roller main bearings, they can be placed closer together and the shaft is more rigidly supported against stress or impact load. But it is only possible to use light duty type races if excess bulk and weight is to be avoided, and the bearings do not work under the best possible conditions. I was once advised by manufacturers to employ only perfectly sealed ball races in small engines, but this is difficult to carry out in practice, and liable to increase friction. The necessity to avoid compression leakage from the crankcase introduces a further problem in the use of ordinary ball or roller races. In some of my engines I have used a ball race at the inner end of the main bearing and a long sleeve (which serves as a fairly good low-pressure seal) at the other; this feature has often been criticised, but it gives very good results, and avoids the extra friction which is almost inevitable when using standard forms of seals, such as those of the spring-loaded annular or face types.

Connecting rod bearings

The crankhead or big-end bearing is undoubtedly the heaviest loaded bearing in any engine, and all too often it is inadequately lubricated. With the "petroil" mixtures usually employed for small engines, there should be no lack in the quantity of oil applied to the bearings, but there is often much to be desired in the mode of application. By mixing oil with the fuel, the intention is to bring it into contact with all working parts, but it is obviously in a diluted from, and, moreover, a good proportion of it fails to deposit in the crankcase, and passes through the engine, to be wasted ultimately through the exhaust ports. It is therefore important to give the oil the best feasible chance of getting where it is really needed, that is the actual bearing surfaces of crankpins and journals.

A fundamental principle m bearing lubrication is that the oil should enter the bearing at or near the point of lowest load pressure. It is often assumed that (in the absence of positive pressure feed) oil collected on the connecting rod will flow downwards and can be admitted to the bearing through a hole in the top of the bearing. But, in fact, gravity plays but a small part in the oil feed in a rapidly moving bearing. Except in slow running engines, the best possible position for the hole in the big-end bearing is in the underside or low load side; the little end, of course, should be lubricated from the top. In both cases, the oil holes should be of adequate size, and may with advantage be countersunk or elongated into slots.

The question of fitting rolling bearings to the big-end is one which is frequently discussed by designers, and in many engines single or double row ball races have been fitted with some degree of success. Roller bearings of the "square" type, as used in several motor cycle engines, have often been fitted, but the "needle" type, in which the rollers are long in relation to small diameter, has advantages in compactness and can be obtained in sizes suitable for small engines. I do not deny that these bearings can be used to advantage in certain circumstances, but they nearly always involve problems in design, including balancing, and tend to increase the volume of the crankcase. Theoretically, rolling bearings do not require lubrication, but in practice they run very harshly without it, and it is not easy to feed oil properly to their working surfaces.

I have tried all these bearings with fair success, but have never found that they improve engine performance or durability to an extent which justifies recommending them to the constructor, who demands that designs should be kept as simple as possible; especially as bearings of suitable type are not always readily obtainable. I remember that many years ago, a design for an engine was published in which the big-end bearing had rollers made from silver steel, working in a steel connecting rod eye, and all parts, including the crankpin, were hardened. The crank journal ran in a plain bush, which was not long enough to give really adequate support. For several rasons, this bearing arrangement was doomed to failure from the start.

One of my engines which was taken up commercially on a limited scale was "improved" by fitting two ball races to the journal in place of the original long plain bushes. I have had the opportunity of making comparative tests, but the only effect that I could see in the alteration was that it added a good deal to the weight of the engine without adding anything to the power output or increasing reliability. In the latter respect, there may be more reason to question their use, because ball races have been known to disintegrate or become pitted or corroded; a good plain bearing, on the other hand, will always "get you home." Though it is fairly safe to use properly fitted ball or roller races on a main journal, in a rigid housing, as I have done in several engines, the pros and cons of the particular application should always be carefully considered. Incidentally, the poor little-end bearing never seems to be given V.I.P. treatment, though it carries just a much load as the big-end; it survives only because the movement at this point is oscillatory, and oil films are more easily retained.

Connecting rods

Rigidity without excessive weight is the most essential feature of the connecting rod, which theoretically works mainly in compression, though, in fact, inertia and other forces also have their effect in engines which run at high speed. If the rod flexes sideways in action, it upsets alignment of the bearings, increasing friction and causing rapid wear. In full-size practice, connecting rods are usually made from steel forgings of I-girder section, to enable weight to be tailored without weakness, and bushed at the eyes, except where a rolling bearing is fitted at the big end. It is not usually possible to obtain forgings for the rods of model engines, and if steel or other stock is used, they have to be machined from solid. For the larger size of engine, it is worth while to mill the flutes in the rods to something approaching the typical crosssection, but the amount of weight which can be saved in the smaller sizes hardly justifies this, and the shanks of the rods are generally left in rectangular, or occasionally, circular shape.

In my early designs, I specified steel connecting rods for all high performance engines, as I considered that the use of any other material would be risky. But as a result of many experiments, I found that good quality castings in bronze, Meehanite, or even aluminium alloy could be quite safely used. The high tensile rolled aluminium alloys, including duralumin, must be machined from solid, but they are more reliable for engines of high performance. These metals do not necessarily call for bushing of the connecting rod eyes, and if the bearings do wear out in the course of time, it is often just as easy to renew the rod as to make and fit new bushes. It is, however, most important that the eyes of the rod should be bored in exactly parallel alignment, in both horizontal and vertical planes. I have shown methods of doing this in several previous articles. Lateral alignment of bearings is generally simple, if other engine components are accurately machined; it is usual to allow side play on one of the bearings, preferably the little-end to avoid the need for internal lining-up.

The gudgeon pin does not need to be press fitted to the piston, but if allowed to "float" (which does not mean a sloppy fit), the ends should be prevented from scoring the cylinder walls by fitting soft pads to them; it is not usually practicable to fit circlips, or other locating devices used in full-sise practice. A mild steel pin, preferably hollow to reduce weight, and case-hardened or chromed on the outilde, outside is recommended. Silver steel, though it may serve its purpose, either in the normal or hardened condition, is too brittle for maximum duty, and I have known gudgeon pins of this material to fracture, even at what seemed to be only moderate load and speed.


I have often been asked to give a "formula" for the most suitable flywheel to suit an engine of a given size, but while I am sometimes inclined to retort "There ain't no such animal!" it is perhaps more polite to say that it is difficult to find a basis for calculation. Apart from the fundamental problems of finding the radial centre of mass, and the factor of momentum in relation to peripheral velocity, it would be necessary to know the friction and back pressure involvad in carrying the engine over its idle strokes, at the slowest speed at which it is required to run. Speaking for myself, I should find this a longer and more tedious job than applying the old-fashioned and much despised "rule of thumb" in other words, trial and error. The flywheels in the various examples of engines illustrated with these articles have all been found stuitable and adequate for their purpose.

In model aircraft engines, flywheels, as such, are not necessary because the momentum of the airscrew, at or near its tip, is exerted at a much greater radius than in a normal flywheel.

It is possible to run engines at their full speed with little or no flywheel beyond a propellor boss or a flange coupling, but this makes slow running, or starting by normal means, almost impossible. Logically, a two-stroke should require only half the flywheel momentum of a four-stroke, but in either case, the major weight of the flywheel should be concentrated in the rim, at the maximum permissible radius. A comfortable sized flywheel will help to reduce cyclic variation, that is, slowing down between power strokes, and may therefore be is partial cure for rough running, or tendency to stall when load is suddenly applied. On the other hand, the inertia of a heavy flywheel tends to retard acceleration, though I have not found this to be a disadvantage, even in racing engines, as plenty of time is allowed for them to reach their full speed.

The Atom V engine

This design, illustrated in Fig. 16, was produced as a result of experiments made daring and after the war, and was my last attempt at a 30 cc racing engine. The popularity of engines in this class has declined considerably in recent years, as the power they are capable of developing in a model hydroplane is almost too much for the competitor to manage single-handed, and boats with smaller engines have produced equal or better results in racing events. In some respects, the cylinder and piston design of this engine is very unconventional, and as originally designed, the port arrangement did not conform with popular ideas. Several constructors of this engine made more or less drastic detail modifications to it (as they do to nearly all published designs). They have a perfect right to do this, and nearly all constructors have ideas of their own, in these matters. So far as I am able, I design engines to be adaptable, and capable of minor mutations to suit individual preference. But constauctors who have altered the design can hardly blame the original designer if the results obtained do not come up to their expectations.

The Atom V engine is the only one in which I incorporated a self-winding "recoil" starter and a metering oil pump as an integral part of the design; these are optional features and, of course, are applicable to other engines. The starter has a drum (7) with an internal torsion spring and a ratchet wheel attached to its face. Three pawls (9), only one of which is shown, are pivotally fitted inside the flywheel rim, so that they engage the ratchet wheel when stationary, but are centrifugally retracted when the engine is running. The cord attached to the drum (not shown) is normally kept in the wound position by the torsion spring, ready for rotating the engine when pulled, A bridge clamp (13) is fitted to the main bearing housing to brace it against the pull of the cord and improve stability of the mounting, The reason for the tapered flywheel is to enable the engine to be fitted as low as possible, with the shaft inclined up to about 10 degrees, in a cambered or V-bottomed hull.

I am a believer in supplying clean and undiluted oil under pressure to the big-end bearing, but the methods used in four-stroke engines are not readily applicable, as the oil cannot be recirculated, and the amount supplied must be accurately metered to suit the requirements of the engine. The pump, which has been offectively used on the Atom III other engines, is not shown in detail in this drawing, but it has been fully described in Model Petrol Engines It has a long vertical plunger, which is rotated by worm reduction gearIng (4) from the rotary valve shaft, and at the same time given a reciprocating motion by a face cam. The bottom end the end of the plunger is partially cut away D-shaped so that it controls the suction and delivery ports of the pump (6); no valves are employed. To control the output, while maintaining constant oil pressure, the face cam can be partly rotated by the lever (11) to alter its phase and thereby control the oil pump port timing. The oil is delivered through the feed pipe (8) to the long sleeve main bearing, and thence by drilled passages to the big-end. This method of oil feed has proved to prolong the life of bearings, and only a small proportion of oil (about 1 in 30) needs to be mixed with the fuel to lubricate the rotary valve and other minor parts. The disc valve works in contact with an inserted bronze or cast-iron face, and the inlet from the carburettor is taken by way of the flange (10) through an annular surge chamber (3).

Phoenix 15 cc engine

This is a straightforward "three-port" engine, the design of which was evolved from a self-contained generator unit of 1 in. bore x 1 in. stroke, by increasing the stroke to 1-1/8 in., and making other parts to suit. It was never intended to be a high efficiency engine, but it worked well and reliably; several constructors have used it in cruising boats and have been well satisfied with its performance. The main object of the design was to demonstrate that a successful engine could be built by any compotent model engineer, despite lack of experience in this class of work. It had a three-piece crankshaft, running in two long plain bearings, with a pressed-in crankpin, and a simple machining jig was employed to ensure that the assembly could be accurately lined up. Other features of the design included a carburettor (made without castings) which could be adjusted to give a fair range of throttle control, and an enclosed contact breaker, havinf a Bosh type rocker arm. The crankcase was made in two halves, jointed on the vertical transverse centre-line, and an optional feature was the fitting of a drain valve, the knurled head of which can be seen in the photograph. This was mainly for the benefit of the user who gets the crankcase choked with excess petrol, so that the engine cannot be started.

The Busy Bee 50 cc engine

In this, the largest engine to be illustrated in the present series, the object was to assist model engineers to construct an engine of practical utility, which could be applied to the propulsion of a small cycle. This idea, of course, was as old as the hills, and it may be said that the first motor-cycles, or most of them, were really motor-assisted bicycles. In the days immediately after the last war, when mechanical transport vehicles of any kind were difficult to obtain, a new demand arose for small engines which could be attached to cycles. Although a great deal of ingenuity was devoted to their design and the methods of driving either on the front or rear wheel, the project was obviously doomed to be short-lived, because they were bound to be superseded when more efficient and comforatable means of transport became readily available. Nevertheless, these unpretentious little engines served a very useful purposes for several years.

My contributions to their developments included several designs which were adapted to different drives, or attachment to cycles, including some ideas which were submitted for commercial production. The Busy Bee, however, was designed purely for individual construction in a modestly equipped workshop. It was, in its original form, a simple three-port engine with no special features except that the main bearing housing was in the form of a bridge casting intended to straddle the bicycle wheel, canrying the engine at one end and the flywheel magneto at the other. In the centre was mounted a friction roller to make contact with the tyre of the wheel. The entire unit was spring-loaded to provide the friction pressure, but could be lifted by a lever to leave the wheel free when required. The only controls fitted were the throttle (the carburetter at first used was a miniature Amal, though a special one was designed later) and a decompressor valve in the cylinder head, both operated through Bowden cables from a single two-way handlebar lever.

The power developed by the engine proved to be ample for its intended purpose. In experiments with an early racing "kart," a tuned version with a rotary valve was produced with success. To satisfy the demand for a small stationary engine, the same internal design wes adapted in the Bumble Bee. Many M.E. readers built the engine in one or other of its forms, and I have heard from several of them in more recent years, telling me that they have given long and reliable service, One reader said that he had renamed his engine "Charley's Aunt, because it is still running!" Another, now living in Dublin, wrote to me recently, and told me that "my Busy Bee, which ran first in 1951, is still going strong.... I gave it a complete overhaul last winter-no replacements necessary, not even piston rings... the wear all round is negligible."

Carburation and combustion

In all i.c. engine of small capacity, carburation is one of the most important functions, and its correct adjustment is more essential in two-stokes than in four-strokes. This is because in the former, some residue of combustion gases remain in the cylinder to mix with the fresh charge, and thereby narrows the range of mixture strength that can by ignited. On the other hand, the charge is very effectively churned up, and to some extent warmed, in the crankcase before entering the cylinder, so that it is more homogenous, and less liable to condense or precipitate after leaving the carburettor than the charge supplied directly to the cylinder in a four-stroke engine.

I have, in the past written several articles on the theory and practice of carburation, but at the present time very few readers appear to be interested in the serious pursuit of this subject, and prefer to make carburettors as simple as possible, both in design and function. For the present, therefore, I do not propose to say much a this subject, except to answer one or two of the queries which are constantly cropping up. The first relates to the means of fuel supply to the carburettor and the position of the fuel tank in relation to the jet, in cases where no float feed or other means of controlling the level or "head" at this point is provided.

In the quest for reliable and consistent running, one of the first aims is to reduce or eliminate variable influences as much as possible; there are all too many of them, including capillary attraction, viscosity and gravity or other pressure acting directly on the fuel. The best results are usually obtained when the static pressure at the actual jet orifice is neutral (i.e. atmospheric, or 0 on the normal gauge). Under, this condition, fuel will not flow from the jet, until it is induced to do so by a reduction of pressure, in other words, suction, applied to the jet, as a result of drawing air through the air passage. It will therfore be clear that if the level of fuel is substantially below that of the jet orifice, the fuel will not flow until the suction is powerful enough to overcome the negative gravity head; the jet output is therefore dependent on the degree of suction beyond this point, and may be said to be uncompensated. By raising the static fuel level above that of the jet, the discharge of fuel is influenced by positive gravity plus suction, and is thus is compensated to a certain extent. This principle is utilized in many types of automatic carburettors, including the Zenith and Solex, but if applied without fuel level control (in direct feed "floatless" carburettors) the compensation is variable, and the jet will overflow, to flood the system, while starting or after stopping the engine. For this reason, it is generally found most satisfactory to locate the fuel reservoir slightly below the level of the jet, to a sufficient extent to prevent overflow, but not more than otherwise deterimined by convenience in filling.

As the level in the reservoir varies between full and empty, it is desirable to keep it fairly shallow, but not to an extreme which might involve risk of interrupting fuel flow by surging or agitation. In boats or other mobile installations, the effects are often prevalent, and in circular-course hydroplanes, both inertia during acceleration and centrifugal force have a very powerful influence on fuel flow. It is usually necessary to locate the reservoir forward of the engine, and toward the inner side of the running circle, so that these forces tend to increase fuel pressure at the jet rather than otherwise. A good deal of experiment has bean devoted to finding the best shape and location of tanks, and the layout of the pipe system.

Other questions which are often asked relate to the function of the "choke tube," and whether it is necessary in carburettors for high efficiency engines. Same local reduction in the cross-sectional area of the inlet passage is generally desirable, in order to ensure high air velocity, and also suction, in the region of the jet; and this forms the equivalent of a choke tube as a separate component. In carburettors which have a jet tube across the air passage, this serves much the same purpose as a choke, and no reduction in the diameter of the passage is necessary; but where there is little or no obstruction of the passage by the jet or other projections, it is necessary to make the "throat " of the passage smaller, or at least on no account larger, than the rest of the entire induction system.

It does not necessarily follow that reducing the diameter of the throat restricts the rate of air flow through it. A double-tapered or flared air passage, generally termed a Venturi tube, has a much greater flow efficiency than a parallel tube of the same minimum diameter, and this feature can be used to advantage in carburetter design. But this applies only when the air passage is open and unobstructed, which is not always practicable, especially when a throttle of any kind is fitted. Restriction of flow is the normal function of the throttle, in regulating the speed of the engine, but at full bore, it should present a "clearway" with as little obstruction as possible to the air passage. In several carburettors which I have designed in the past, some attempt at automatic compensation and speed control has been made, but the objects of the designs can easily be defeated by (apparently) small modifications of air and fuel passages. The fact is that all the functions in small carburettors are critical, and influenced by any variations in design or adjustment to a much greater extent than is generally realised.

Combustion chamber design

It is well known that a great deal of research has been devoted to the design of combustion chambers of full-size engines, with the main objects of promoting turbulence, improving the efficiency of combustion, and reducing thermal losses. These factors become increasingly important as the compression ratio is raised, because the combustion space becomes more flattened or attenuated, with a greater tendency to combustion knocking or "pinking", and local overheating. I do not propose to discuss the complicated theory of combustion head design here, as a great deal has been written about it by designers of automobile engines, but would point out that many of the problems involved in these engines apply equally to models, though not necessarily for exactly the same rasons. For instance, pinking becomes less liable to occur as the size of the combustion head is reduced, and is almost non-existent in engines of 15 cc or less per cylinder, even with very inferior or low octane fuels.

There is rarely any lack of turbulence in small cylinder heads, but it is not always employed to the best advantage, and the internal shape of the head is not always conducive to efficient combustion.

The deflector on the piston, in conjunction with the internal contour on the head, can be designed to make good use of turbulence without impairing scavenging effect, and also to confine combustion to a compact portion of the head, which is all to the good. It is also important to locate the sparking plug so that it is in the path of relatively "clean" mixture. For most engines, a symmetrical internally-machined head, with the plug set vertically in the centre, gives very good results, both with spark and glow plug ignition, but there is always room for experiment in both the location and reach of the plug; usually the end of the body may be set flush with the inside surface of the head, but sometimes it may be shielded or even pocketed with advantage.

Sometimes the plug is located on the transfer passage side of the head, with the object of keeping it relatively cool by the flow of fresh mixture, but this also makes it more liable to oiling up, and it may be better to place it in a hotter zone. If the cylinder head is symmetrical, there is a choice of positions for an inclined or eccentrically located plug, for experimental purposes.

Another point about combustion chambers is that their internal surfaces should be free from either local projections or internal corners, which might be overheated, or trap hot gases, tending to cause pre-ignition. In nearly all my engines, I have found it a sound principle to keep the combustion space entirely above the cylinder, without the internal locating spigot which is common in certain types of small engines; this certainly helps to conduct heat away from the inside surface of the head.

Twin two-stroke engines

Not many amateurs have attempted to build two-strokes with more than one cylinder, and from the purely practical aspect, it may be considered that there is no great incentive to do so. As a means of propelling a boat or other model, the single-cylinder engine does all that is generally desired, and increasing the frequency of power impulses, in engines where they are already very high, does not appear to compensate for the extra complication, and possibly more tempermental performance, as the size of individual cylinders is reduced. There is however, a certain fascination about a twin two-stroke (as many readers who have experience with motor-cycles having such engines will testify) and I have often been urged to pay more attention tp this type. Three such engines, varying in size and details of design, have been described in M.E. by me, and have been successfully built by many readers.

The Ladybird compression-ignition engine (right) is one of the few exploits I have made in the design of engines working on this principle. Though they may appear to offer a prolific field for development, their success depends more upon the accuracy of construction than in the finer points of design—and that is something one cannot put on the blueprint. A large number of inexperienced readers have started out to build engines from quite sound published designs, but have partially or completely failed to achieve success with them. In the circumstances, I have considered that discretion is the better part of valour, and have preferred to keep quiet about most of my own attempts at designing them.

In the Ladybird, however, I believe that I have been able to make some contribution to design, and though it calls for just as much accuracy and care in construction as other engines of its type, it is relatively straightforward in machining and fitting. Basically, it follows the three-port principle, with the inlet ports on the inner sides of the cylinders, and the induction passage between them in the centre of the main casting. The transfer passages are on the outer sides of the cylinder, with twin exhaust ports at right angles to them. A fairly long stroke in relation to bore is employed, and the cylinder, which is machined from steel and internally carburized, has a fixing flange above the level of the ports, and a finned "bonnet" screwed on to its upper part.

The crankshaft is machined from steel, preferably a tough high tensile grade, and in order to avoid the need for split big end bearings, U-shaped straps are riveted after assembly to the feet of the dural conencting rods, the shanks of which are of turned circular section. Contra-pistons are used for compression adjustment, and the lever arms of the screws are oriented so that their angular positions conicide when the compression is the same in both cylinders. Fully detailed drawings of the engine are obtainable from the M.E. Plans Dept., catalogue number P.E. 22.

The Cherub 10 cc twin (right) employs two of the body castings for the Cadet 5 cc engine, with a split centre section in between them, incorporating a cylinderical rotary valve housing and induction passage. The crankshaft is in three piecs, the centre one having crankpins on each end at 180 degrees to each other, and engaging with bored seating in the outer crank webs. To form the passages in the centre piece of the crankshaft, it is drilled from each end with an acute-pointed drill, leaving a blank portion in the middle. Rectangular ports are milled in the journal on opposite sides, each breaking into the respective passages, and faired into it by a riffler or rotary file. The ports are timed to line up alternatly with a single port in the centre bearing, to admit mixture to each crankcase in turn.

Spark-ignition in twin or multi-cylinder engines presents certain problems. For a twin in which the firing intervals are evenly spaced at 180 deg., one of three systems can be used: (a) single coil, single breaker, and h.t. distributor, (b) two separate coils with individual breakers, or (c) special double-pole coil and single breaker. Provision is made in the design of the Cherub engine for fitting any of them. The original Cherub engine had an h.t. distributor fitted at the flywheel end, with the insulated rotor inside the flywheel rim; but a more accessible position for it is at the other end, mounted on the contact breaker flange. No details of the ignition arrangements are shown in the drawing reproduced, but full details of alternative components are given in the separate M.E. Plans drawing P.E. 28.

The horizontally-opposed Craftsman Twin 10 cc engine (Fig. 19) is designed to give improved balance, by setting the reciprocating forces against each other. Though the offset of the cylinders produces a slight "couple", the balance is much better than that of a single cylinder engine at all speeds. The cylinders fire simultaneously, so that the impulse frequency is the same as that of a single-cylinder engine, but is no great disadvantage at the speeds normally obtained. A special double-pole coil as specified above (c) is necessary for producing sparks simultaneously in both cylinders; magnetos capable of doing so have also been produced. Detailed drawings for this engine are available only from Craftsmanship Models Ltd.


The series concluded with the design details and plans for
ETW's "new" and final two-stroke, the
an over-square 10 cc air-cooled, spark-ignition, rear-rotary valve engine.
These plans are not reproduced here.