Peter Inchley. The Expansion Chamber. 1963.


5. The Expansion Chamber


Peter Inchley, whilst working as the Competition Development Manager for Villiers, worked on his machines at the Cotton premises, with their Competition Team Manager and development rider, Fluff Brown.

One of the major skills and expertise that Peter brought to the workshop was the black art, of the expansion chamber, Fluff brought amongst other things, Cotton’s expertise, as the master’s of frame design.

Web 032

Better Breathing / Peter Inchley

There’s hidden power in the exhaust pipe, and Peter Inchley, competitions development manager for AJS, shows us how this power can be harnessed, to increase output, this was one of the key elements in his success with the Villiers Special.

The flow of intake and exhaust gases is one of the main factors, which control the running characteristics of an engine. This pulsating flow produces certain phenomena, which can be used to help both cylinder filling and scavenging. There are known facts, both empirical and theoretical, about the behavior of gas pulses moving through different shapes of pipe and orifice, and theories have been developed from these facts to explain the effects known as “ram” and “resonance”.

Four-stroke and two-stroke engines are both affected by these phenomena, but because of the limitations of its design, the two-stroke is more susceptible to intake and exhaust modifications.

This article will deal with the effects on two-strokes, mainly competition models, as the effects are used to a much higher degree on high-output engines. First, a gas pulse moving through different shapes of pipe produces different effects – in a plain pipe there is no reflection of the pressure waves (a reflected wave travels in the opposite direction to the pulse and may have a high pressure – a positive wave – or low pressure – a negative or suction wave).

When the pulse reaches a diffuser or megaphone (an increase in section) there will be a negative reflection produced. Similarly a nozzle, or decrease in pipe area gives a positive reflection – a ram wave.

Web 033

This is quite straightforward and the effects can be calculated, but unfortunately there are, secondary reflections set-up and these can completely change the calculated pressure at any given point.

The only means of predicting how the complex of reflected wave will behave is to use a digital computer – not the sort of thing you find lying around the workshop. However, the basic theory of wave reflections is generally good enough to form a starting point, and from there it is a matter for experiment.

The way in which these pressure waves can increase the engine efficiency is first in the scavenging, where a negative reflection arriving at the exhaust port around bdc will help suck out the remaining gas.

This wave should not be brought back too early as it may induce the crankcase blow down to pass straight from the transfer port to the exhaust, instead of swirling up into the cylinder. In the way the scavenging can be improved to such an extent that some of the incoming charge will be lost into the exhaust towards the end of the exhaust period.

This brings us to the second way in which the pressure waves help the engine – the “lost” gas can be retrieved by arranging for a positive reflection to reach the port just before it closes.

The fresh mixture is pushed back into the cylinder and the port is effectively closed earlier than it’s geometric timing, resulting in a higher cylinder pressure on the compression stroke and a higher mep.

Finally, the expansion box itself has to be scavenged – the backpressure needed to set-up the reflection waves may also “clog up” the box and it is essential that the system is clear of gas before the next pulse arrives or the engine will simply choke itself.

The nozzle and tailpipe have therefore to be designed very carefully. The general effect of the exhaust is to increase the engine’s scavenging and delivery ratio, resulting in a higher output.

The exhaust system can be varied to give maximum torque at virtually any engine speed, although this will be lower than the absolute maximum available from the engine.

Web 034



In a similar way the intake can be made to fill the crankcase more efficiently, but the effects are not so marked as with the exhaust. This is because the gas is at a lower energy and the short intake stubs, especially those used with rotary disc valves, do not show inertia and pulsation effects to the extent of the exhaust system.

The only certain means of improving intakes is to reduce flow resistance and losses to a minimum. As the engine speed increases the exhaust port has to cope with a greater flow and so it either needs to be open longer or to have a larger area.

This consideration is known as the time-area of the port.

  Web 035

Villiers Starmaker 250cc / 1963

There is a maximum width of port (70 degrees) which can be used without causing ring failure and so any further increase in the time-area involves making the port higher – this also increases the timing period and reduces the effective stroke of the piston.

However, it is possible to regain any power lost here and to keep the volumetric efficiency to an acceptable value by using the exhaust system phenomena already described. It must be realized that the exhaust pipe cannot be used to the full unless it is designed in relation to the complete gas flow system, intake, scavenging and timing periods.

The theory of the resonance effects is explained by the fact that each exhaust system has its own natural frequency and the period of this frequency should be made about equal to the period of the scavenging process at the speed where the effect is required.

In other words, the exhaust is resonant with the engine at that particular speed. The velocity of the gas also explains the “ram” effects, in that its inertia can be used to set-up reflected pressure waves when it meets a change of section in the pipe.

Generally, two-stroke expansion boxes are constructed with a plain pipe leading into a diffuser followed by a plain section and a nozzle (converging tube) and finishing with a relatively narrow tailpipe.

Web 036

Rex Avery and Peter Inchley / EMC 125cc / Brands Hatch / 1963



As the exhaust pulse reaches the diffuser it slows down and its kinetic energy is changed into pressure energy. This also results in a negative reflected wave being sent back along the pipe.

The gas in the expansion box then reaches the nozzle and the resulting throttling action sends a wave of high pressure along the pipe.

The first wave helps during the scavenging period and the second pushes back any lost mixture into the cylinder and effectively seals the port, before it is close, by the piston.

The taper of the diffuser part of the expansion box is fairly critical as the flow of the expansion gas is unstable and if the taper is too steep, it may give an undesirable amount of turbulence.

The throttling action caused by the narrowing section at the end of the box also tends to obstruct the flow and although the degree of taper is not so critical as the diffuser, it must not be so sharp that the gas cannot get out through the tailpipe.

These are the effects produced by the exhaust system; now, co-ordinate them, to give the desired effect on the engine.

Web 037

Peter Inchley IOM TT 1966

As the engine speed changes, so does the time-area of the port and the frequency of the engine, but the dimensions of the exhaust system remains the same.

This means that the wave action will only be “right” for a very limited speed range and the beneficial effects will be confined to a narrow rpm band.

This poses the problem of whether to aim for the maximum power possible or to settle for less power but speed over a wider range.

This can be done by reducing the effects of the exhaust and by arranging the peak intake speed to be slightly different to the speed where the exhaust effect is greatest.

On some machines it is acceptable to have useful power band of only 1000 or 2000 rpm – it is the racing driver’s job to keep the engine within that range by using the gears, even if he needs six or more.

On the other hand a moto-cross machine needs enough power to pick up without “fluffing” from almost tick-over – absolute power is less important than keeping the engine running without “going off the megga”.

Web 038

Ernst Degnar / Suzuki 50cc / Brands Hatch / June 1962

The effects of exhaust pipes can be pretty hairy, anyone who’s been to a Grand Prix will tell you, and the explanations of these effects have been given over the last few pages.

The big question is how do you make the exhaust produce these effects at the right time.

It is accepted that the exhaust system virtually dictates the shape of the torque curve of a high-output two-stroke. The intake affects it, but to a lesser degree.

The way in which the size of the front pipe, and expansion chamber, are related to the exhaust and crankcase blow down periods is the controlling factor over the engine’s output.

Also wherever there is a boost in power there will be a corresponding loss somewhere else and it may be necessary to forsake the absolute maximum in order to have a wider spread of power.

 Web 039

Derek Minter / King of Brands

Maximum Torque


On the road racing machines it is usually possible to aim for maximum torque at the highest speed at which the engine will safely run.

For torque at high speed, the exhaust would probably need a short front pipe, a small expansion chamber and large time-areas on all ports.

On the other hand a scrambler needs tractive power down to very low engine speeds and to get torque low down, the first pipe would need to be relatively long, the expansion chamber large and the inlet port time-area would need reducing.

Apart from actually giving an increase in power the exhaust system governs the rpm range where peak torque occurs.

The power spread also tends to be wider as the peak torque is taken further along the rev scale from the point where absolute maximum power would be.

This change in torque can be seen in the graph. The engine is a 250cc single cylinder scrambler giving vastly different torque curves after altering the front pipe length by 2in. The graph marked (1) is the performance with the shorter pipe.

Web 040

Expansion Chamber


The other diagram shows the arrangement of an expansion chamber and gives a good idea of just how many dimensions there are to vary, and how many combinations of different shapes and sizes one could devise. The mind boggles.

However there are one or two empirical formulae, which relate some of these dimensions, reducing the number of combinations to a less mind-boggling figure and providing a basic starting point from which to carry on experiments.

In the diagram the lengths I0, I1,I2, etc., refer to the centre line distance from the piston skirt, a1, a2, d1 and d2 are the area cross-section and diameter respectively, of the portions indicated.

The length of parts of the exhaust can-be roughly related to the engine speed by the following formulas;

I1 – 900/n (E – ½ T)

I3 – 900/n (E + ½ T)

I4 – 1800/n E

Where n is the required engine speed for maximum torque and E and T are the exhaust and transfer periods respectively (in degrees).

The volume and length of the expansion chamber also play an important role and as a general rule the angle of the cones of the diffuser should be no more than 7 degrees.

If this angle is kept within the range of 4 to 6 degrees then the length I3 – I2 of the parallel portion of the chamber should tend to zero.

Along with the length and angle of the cone, the area cross-section of the chamber in relation to that of the front pipe is fairly critical. In the diagram the ratio a2/a1 should be between 4 and 8.

These dimensions have now been interrelated in such a way that the shape of the exhaust is already determined – this is the basic size which then has to be experimented with to get the best performance from the engine.

The only part of the exhaust system, which hasn’t been covered, is the tailpipe. In fact there is not a great deal known about the function of this component, particularly about its length.

In 1910 a Scott was the first two-stroke motorcycle ever to complete a full TT course under race conditions and in 1911 a Scott ridden by Frank Phillip gained the TT lap record with an average speed of just over 50 mph.

Web 041

Walter Kaaden



The general effect of a small diameter pipe is to produce a larger “ram” wave with loss in scavenging, and a large diameter pipe has the opposite effect. 

As a basic guide, the diameter of the tailpipe should be about half that of the front pipe for racing machines, and slightly larger for scramblers.

For these machines the ratio d1/d2 should be about 1:6.

These figures are generally accepted mainly from experience and experiment with competition machines and are quoted by such authorities as Dr. G.P. Blair, of Queens University, Belfast, who has a strong connection with the new racing two-stroke being developed there for B.S.A.

“This is the basic theory of the exhausts as practiced – the complete story is simply not known; even using a series of pressure transducers in the exhaust and a computer to analyse their readings doesn’t eliminate the need for final experimentation, on the track to get the best results.”

by Peter Inchley,

Web 042

Kaaden and the MZ GP Team

Return to Contents.