Making the power-2

Some Bantam Racing enthusiasts might be puzzling over the choice of pipe which produced the results shown by the dyno readout posted in the topic` Making the Power`. How could it be that a design calculated for an earlier spec could prove to be so wrong back then and yet the same pipe for a different spec be so right for what is ostensibly the same engine?
The answers to some of these questions might be found in the numbers below, and may serve to illustrate that simply using one of the many generic pipe calculation formulas doesn`t always reflect what might be thought to needed for the particular engine in question? As someone once pointed out to me, you can put a race pipe on a trials bike and you still have a trials bike but put a trials pipe on a racer and you no longer have the race performance.
Notwithstanding all of that, the numbers highlighted perhaps give an idea as to how these seemingly contradictory results can come about. Indeed, changes in two-stroke engines can at times have unintended consequences not even in mind, until the results are examined, and a modification may give rise to a positive effect which in turn may lead to a false conclusion!
Power production is not a linear process and peak power alone will never define the optimum engine characteristics needed to fulfil specific criteria.

These examples use the formula for the tuned length, and the other component lengths are calculated as percentages of that length. However, I always caution that Bantams have just 3 gears and so breadth of power range is a major consideration in any pipe/ torque/power design. A pipe needs to be designed around the intended purpose and not just for peak horse power at constant rpm! Not being slow anywhere around the circuit is more important than just terminal velocity, and should influence the choices available!

The following examples are calculated using the formula: s/s x ex/t x constant, divided by: rpm
Where s/s is speed of sound in mtrs/sec. Ex/t is exhaust port timing duration. Constant is 88. rpm for power at maximum revs.
550 mtrs/sec is a typical, ball park figure, all the other factors are self-explanatory.
190* is a typical number for exhaust port timing. For this look at things only the s/s and rpm will change:

550 x 190 x 88/11,000 = 836mm. 550 x 190 x 88/10,500 = 876 So just by reducing max power rpm by 500rpm we have a 40mm longer tuned length and all other length dimensions will follow suit.
What happens if we reduce the speed of sound which by definition is a reduction in pipe temperature, let`s try 525mtrs/sec but retain the other numbers:

525 x 190 x 88/11,000 = 798mm. 525 x 190 x 88/10500 = 836mm.

So next let us increase the speed of sound to 575mtrs/sec.

575 x 190 x 88/11,000 = 874mm.

So let us next increase the exhaust duration to 192*

575 x 192 x 88/11,000 = 883mm

It is pretty plain to see that juggling the input figures around in the formula results in some not too dissimilar tuned length values. Equally it might explain why, by producing more power, thus heat and at similar rpm, the initially rejected pipe when subsequently fitted to the engine with an improved breathing capacity, can produce the effects represented in the dyno graph! It might equally suggest that the original design assumptions as to the compatibility of pipe and engine were badly flawed?
On the other hand, none of the foregoing may be entirely relevant: Wave speed (s/s) varies with engine rpm, engine power, combustion, cooling, ignition timing, mixture strength, and compression ratio. Then it varies with every crank rotation and along the length of the pipe profile and finally the out-going pressure pulse travels at sonic speed and returning pulses are subsonic! So the s/s value to input into the formula is at best an approximation, but you still have to start somewhere in the design process.
There is also another problem that can confuse pipe design and might at first, even appear contradictory?
Residual pressure waves continue to resonate through the pipe after being reflected from the piston skirt and rear cone. At around 2/3 of the optimum engine revs the pulse arrives back at the cylinder when the transfer ports are still open. In a worst case scenario the transfer flow maybe reversed and at best simply delayed a bit, neither are good for power. The further away from ideal exhaust resonance of 180* exhaust port duration strays, the more likely this is to occur. In this situation diffuser action during transfer port opening is in apparent conflict with the rear cone sending spurious returning pressure pulses. This situation can be improved by opening the exhaust port later, the exhaust pulse starts later so returns later, or install a power valve and adjust opening to counteract this tendency. A bit tricky on a Bantam though, but an ATAC chamber of appropriate dimensions and timing might possibly help?
Additionally, if this same resonance phenomenon can be used when exhaust port is open at around max torque rpm, a superposition situation can occur where the out-going wave is augmented and amplified, then we have potential for a power increase if the pipe is proportioned correctly!

So, what was done to the engine in question to make the improvement, very simply and in the main, by concentrating not so much with the struggle to make power, as to avoiding wasting power, a slightly alternative approach? By looking then at smooth mixture and gas flow in all ducts; avoiding pressure and power sapping turbulent flow; by creating a coherent in-cylinder transfer flow regime, both radial and axial, and by making sure the combustion process is most certainly turbulent and rapid! Therefore, and by definition, power delivery improves across the whole spectrum of the engine`s engine operating band.

Trevor

so Trevor, is there a factor that can be applied to the mathematical calculation of exhaust system design for the bantam engine ? i would think that you may have already have been able to deduce this from the exhausts used on the dyno for the engine in question.

The Bantam race engine is unique in the two-stroke race engine world in that it has just 3 gears, it can`t even be compared, say, with direct drive karts that have no gears, for their engine spec is very tightly controlled right down to individual dimensions that all manufactures have to adhere to before they can compete. The enforced operating rev range of these engine is absolutely massive but no generic pipe formula could reflect that specific requirement, as a single ignition point is mandatory no assistance can come from that direction! The same is certainly not true of the very lose rules of Bantam racing.
The open class engines like the Hondas, TZs, Aprilia and so on can call upon a multitude of optional internal ratios for their six-speed gearboxes and can as a consequence have a suitable gear ratio for any corner and straight on just about any circuit. Consequently, their engines can be maintained in the optimum power ranges for most of a race lap. That luxury is not available to the Bantam, so consequently is severely disadvantaged, another strategy has to apply and here the exhaust pipe has a significant part to play. Rider bulk and weight also have an impact on the race potential of an individual machine`s power to weight ratio, trying to accelerate a 90kg rider out of a hairpin is a real struggle compared to a 63kg rider. Here again a different strategy needs to be applied if overall lap times are logically thought about.
By designing into the original concept of the engine spec the need for a broad span of usable power the inherent handicap can to some extent be compensated for. Reference points taken from higher bmep engines must be regarded in the context of what they are, and of what a Bantam is capable of!
In the previous article it was only the tuned length that was examined and a few examples calculated, but even these numbers are conditional on the influence of other functioning parts of the pipe.
All two-stroke race engines produce a finite amount of gas energy and the Bantam engine has a comparatively modest amount to exploit. That quantity is at a maximum immediately after combustion has completed, from then on it is dissipating at an alarming rate! So it follows that if your engine produces less power, it produces less gas and heat energy, less energy means lower gas speeds even though some of the remaining data to input into your preferred formula may be the same as a more powerful engine.
What is crucially important is to ensure that the maximum quantity of energy arrives at the diffuser at the prescribed point of crank rotation, and at the rpm preconfigured into the pipe calculations! The exhaust port window, the profile and length of the duct, and the header pipe all need be constructed with flow coefficient in mind.
What is important to consider is the way exhaust gas is converted into suction by the diffuser, whilst retaining enough of that finite quantity of energy for the returning positive plugging pulse. Pipe performance is all about finding a balance. It is of no practical use to have the diffuser pull a ton of mixture through the engine into the duct/header and for the plugging pulse to be so weak that it can only return a fraction of that mixture to the cylinder. Equally it is of no use either if the positive return pulse not only shoves the pulled through mixture back but along with it, a large amount of hot spent gas as well! So it is very clear that the various formulas for establishing an efficient tuned length is dependent on more criteria than is actually asked for, and so are at best, an approximation.

Nigel,
There is a consistent theme that runs through multi pipe dyno testing sessions that I have seen and participated in and is that the longer tuned lengths offer better performance up to a certain point and then begin to fade at higher rpm. This is very clearly shown on the dyno print out, knowing that this happens, my approach for my Bantam engine pipes is to calculate the tuned length for 500rpm lower than the engine`s natural rev peak for maximum power. This is a sort of fudge factor that provides initially for better performance to help with the lack of gears, then the ignition retard can work in concert with carburation to boost over-rev as part compensation for the power fade below the peak. Lowering the compression ratio helps to give the pipe more energy at higher revs. However, there are also other engine features that are detrimental, such as the reducing blowdown time-area as rpm escalates. Trying to use a shorter pipe length to get more revs will eventually just lose you performance. One slightly more contentious development, but one I am personally convinced works well is to raise the floor of the exhaust duct/port window above bdc. The gains as I see them are; the obvious one of reducing the opportunity of mixture short-circuiting, from the nearby transfer port, out of the exhaust port. The reduced cross section area of the duct will offer an improved outflow coefficient because of less turbulence and that will aid blowdown flow and maintain particle velocity, there will be less mixing of the spent gas and fresh mixture in the smaller duct, better flow coefficient for returning plugging pulse. All of these incremental improvements are to the benefit of overall pipe efficiency and as such should be welcome, particularly as the duct is an integral part of the pipe system and features in the tuned length calculations. This is another instance where: what was not there in the first place can`t be made up for later!

Trevor