Exhaust System

On the trials and roadster, original header systems were nothing more than a straight pipe, with a proprietary silencer stuck on the end. This set up did little for the power output of the engine, but in those days it did not matter as other runners were also dogged by the same lack of expertise. In the mid sixties trends altered, and saw the introduction of the long tapering systems which improved power outputs, but nothing like the bulbous systems around today.

The Scrambles or Motocross scene saw the biggest change, for originally systems were little more than a short length of pipe, cut off at a specific length to give the right power characteristics. At one time a works machine was seen to be using nothing more than a 4 inch deflector plate sticking out from the finned exhaust nut, more in an effort to stop paint being burnt off the frame. With no noise regulations in operation in the sixties these systems were painful to the ear, with noise levels reaching 115 to 120 Db . As time progressed the pipes became longer and flatter, resulting in the under crankcase squashed megaphones, which were prone to suck in paddock floor debris on the negative pressure waves. As these pipes generated a couple of extra horsepower they were tolerated by riders, but not by landowners and local inhabitants. To counter this problem a switch to the long tapering expansion chamber was made, but as yet unsilenced. As time progressed and the world became more environmentally friendly (to the non-sporting humans), the lowering of noise levels and the introduction of mandatory silencers was made. This had little or no effect on power outputs as two-stroke pipe technology had advanced beyond the wildest dreams of the average rider. Silencing today is a legal requirement (110 Db in 1990 and 105 Db in 1993, a 50% reduction in a very short period of time), and often achieved by the use of an absorbtion muffler, made from a thick wall alloy tube of sufficient volume, loosely packed with glass wool, around a central perforated tube. See chapter on silencing for a more detailed discussion.

If there is no cosmetic limitation on the type of system to be fitted, then recourse to one of the many tuning manuals will give some formulae for an adequate system. Alternatively, one could be made by one of the many fabricators that specialise in computer designed, exhaust pipe manufacture. The formulae presented in this publication have been verified by the authors with qualified success. We do not pretend that it is easy, or simple, just beware of any offer that promises easily achieved success whatever your engine.

The evolution of the exhaust pipe has been charted on the previous page, and for simplicity, each shape has been loosely associated with the decade in which it was popular. The 1960 shape is a straight pipe header, no belly section, and a stinger which is too short and too fat. In 1970 the straight header pipe is retained, but we have a belly section giving better compression wave timing, and a longer thinner stinger giving much higher back pressure. There is no attempt at silencing.

The 1980 pipe has learned of the more efficient tapered header pipe, generally at 1o or 1.5o and is often joined to a two stage diffuser. Some exhausts were very very loud in the early years, and when some silencing was required the forecast drop in power did not materialise as many had feared. The 1990 pipe has a tapered header with a multi-stage expansion cone, sometimes as many as four or five sections, and occasionally supported with a variable exhaust port timing. The Yamaha "power valve" system was successfully introduced in the second half of the 1980 decade. Silencing is very much with the competitor today, road racing is restricted to 105 dB.

With a modern high performance system attention must be paid to its mounting. No solid joints as of 30 years ago. The port fixing should be of the double skinned, spring loaded sliding joint type, with the rear bolted to a rubber anti vibration mounting, or fracturing will occur. Because modern pipes are made to resonate at a set frequency, which can cause heavy vibrations, a frequent examination of the system must be made to ensure serviceability, as any crack will bleed away pressure or even suck in air when the system is at a negative pressure, resulting in a drastic loss of power.

Early expansion chambers of the double cone type were usually a compromise between cost of manufacture and efficiency, resulting in pipes that were down on power and very peaky. If the accompanying sketches are studied the trends of the following decades will be noticed. The modern multi diameter type is more complex in both design and manufacture, but they are more efficient in power production as well as making the motor more flexible, in that the usable rev band is increased. One side effect of the modern system is that the effective compression ratio is increased, which calls for a reduction in the amount of ignition advance, down to as little as 16o BTDC on a well tuned racing 200 or 250. A really efficient pipe can actually require a compression ration below that of a standard roadster.

A very simple look at the early exhaust theory might provide the following scenario. The piston opens the exhaust and out pops a sonic wave together with a lot of flame, smoke and noise. The early part of the exhaust is essentially a tube along which the sonic wave roars. It reaches the start of the divergent cone at the start of the expansion chamber proper, and a divergent cone reflects the sonic wave with a negative sign (ie: as a depression) so the steeper the angle of the divergent cone the stronger the negative reflection, so the sonic wave runs the length of the divergent cone giving off negative reflections along the way which travel back up the pipe to the exhaust port. The first negative waves reach the exhaust port to assist the extraction of gas from the chamber and to draw mixture up through the transfer ports. This assistance is not insignificant in any way as the pressure in the exhaust port and early sections of the exhaust system can easily drop well below atmospheric, and reminds us that air leaks in the exhaust are not just inconvenient but detune the exhaust.

The sonic waves take a fixed amount of time for their activity for a given shape and exhaust size, but the benefits need to be synchronised with a similar time set by the engine rpm, exhaust port height and the blow down period until the transfer ports open. Negative sonic waves from the end of the divergent cone arrive at the engine at or about piston BDC where the exhaust suction is about 0.5 atmospheres and the crankcase pumping is similar at 0.5 atmospheres, giving a full 15 psi or so to blast the new mixture up into the cylinder. By now the sonic wave has travelled along the belly section of the exhaust and now meets the convergent cone, and is reflected with a positive sign, sending positive sonic reflections back up the chamber. The convergent cone is a much steeper angle that the divergent cone and therefore the positive waves are stronger and shorter lived. The blast of high pressure sonic waves hit the unburnt fresh charge which has been sucked up the transfer ports and out into the exhaust, and rams it back into the engine after the piston has closed the transfer ports and just before the piston closes the exhaust port. This ramming pressure, just like a super-charger, materially effects the combustion pressure and hence our prior comment about combustion pressure being more important than the geometric compression ratio. With a really excellent exhaust, the road going CR might be high enough to give you detonation at maximum power. The final part of the exhaust is the outlet pipe, known as a stinger, and is there to regulate the internal pressure and bleed the gas out of the chamber while not disturbing the sonic waves. The exhaust pulse diagram shows this graphically.

A slightly more considered view of the same exhaust theory follows. The piston opens the exhaust and a high pressure sonic wave exists into the early part of the exhaust which is essentially a tube, either parallel or expanding slightly at 0.5o or 1.0o , along which the sonic wave moves. A slightly tapered pipe is preferred as the friction between the gas and the pipe wall is less, and that means there will be more energy in the tuned pulses. It reaches the start of the divergent cone, which reflects the sonic wave with a negative sign (IE: as a depression), and the steeper the angle of the divergent cone the stronger the negative reflection. The limit is about 8o included cone angle, at steeper angles the gas flow starts to separate away from the cone wall leaving eddy currents, which rob the sonic waves of their power. The sonic wave runs the length of the divergent cone giving off negative reflections along the way, which travel back up the pipe to the exhaust port. Note that the sonic wave travels relative to the speed of sound and not at the speed of the exhaust gases, and the sonic speed slows as the temperature of the gas drops. A minor discrepancy which we don't acknowledge is that the expansion (negative) waves travel slightly slower than the speed of sound, and the compression (positive) waves travel slightly faster than the speed of sound.

The first negative waves reach the exhaust port to assist the extraction of gas from the chamber and to draw mixture up through the transfer ports. This assistance is not insignificant in any way as the pressure in the exhaust port and early sections of the exhaust system drop well below atmospheric. The sonic waves must be tuned to arrive at an appropriate time in order to assist the extraction of the exhaust gases. Exactly when the first negative pulse should arrive is open to debate, but a good starting point is just before the transfer ports open, when most of the cylinder pressure has gone, when the exhaust port is 25% open, as measured in crankshaft degrees.

The last of the negative sonic waves from the end of the divergent cone arrive at the exhaust port sometime after piston BDC. We describe the exhaust as a single pulse, but it is in fact a double pulse composed of a primary pulse from the exhaust port opening followed by the residual exhaust gases blown out by the new charge released by the transfer ports opening. The secondary pulse from the transfer opening has about half the energy of the primary pulse, and is strong enough to create its own compression and expansion waves in the exhaust pipe. The primary pulse does unfortunately contain at least 10% of new gas which has short circuited directly from the transfer port to the exhaust port without doing any scavenging. This short circuit can be much more than 10%, and is depressingly high for those seeking maximum power.

The first positive waves have been created by the start of the convergent cone which reflects the sonic wave with a positive sign, sending positive sonic reflections back up the chamber, travelling slightly faster than the speed of sound. The convergent cone is a much steeper angle that the divergent cone and therefore the positive waves are stronger and shorter lived. The blast of high pressure sonic waves hit the unburnt fresh charge which has been sucked up the transfer ports and out into the exhaust, and rams it back into the engine after the piston has closed the transfer ports and just before the piston closes the exhaust port. The first positive waves need to be effective as the transfer ports are closing, and any waves which arrive after the exhaust has shut are obviously not going to help very much. Dr Gordon Blair of QUB has shown that the maximum cylinder filling is achieved when the exhaust compression wave hits the exhaust port some 40o before the exhaust closes. That is why we are so concerned to match the engine ports with the pipe and the expected power range, and whether the gearbox ratios we have chosen will give us a useable machine.

Blair's research also shows further subtleties that are too complex for the time being, and that the exhaust waves start out with a rising pressure wave but as it travels up the pipe, the crest of the wave catches up the base to create a steep fronted wave (see Fig 42 ), rather like a wave at the beach coming to a peak then a crest then over toppling to a "roller". In sonic terms the roller does not occur, but the vertical fronted wave is a sonic shock, and is unsuitable for our needs. It can persuade particles of gas to achieve sonic speed locally, and produces a great deal of heat. The effect of the sonic pulses is to increase the pressure in the cylinder by anything up to 25%. Combustion pressures and therefore power are proven to be proportional to trapped cylinder closing pressures, and therefore proportional to the ramming pressure of the exhaust.

This chapter is continued in the book with the following major sections.

Fully worked calculations
Cut your own steel
How to trace and measure ports
Basic time area diagram and time-area calculations