Heads up you guys!

Design Analysis of a Bantam Race Engine
Part 4.
Cylinder Head
The cylinder head described here is the final example of several heads used over the years, the chamber profile and volume changed more frequently than the physical head did. I see no point in describing the ones which have been superseded as the final iteration was superior in every way. That included the design of the physical structure, material spec, water flow and combustion chamber geometry. There is one small caveat in that the liner flange diameter, providing head location, was dictated by the diameter of the cast iron tube we were able to source, after preliminary machining to arrive at a cleaned up size. I would have wanted a slightly larger diameter but as we had got hold of the foundry off-cuts for the price of `a drink` we were in no position to complain. As you can see, still holding to the old, traditional Bantam Racing Club ethos of promoting cheap racing, by making our own barrel, liners, heads, exhausts and so on!
The aluminium alloy for the head was an off-cut from a very large sheet/ block of high spec, high strength, high tensile alloy from which aircraft components are produced. The sort of material that has to be visibly marked with grain flow direction and have specification and heat treatment documents accompanying it at all times for true identification. Needless to say, a local helicopter company very generously `donated` the piece of alloy and thereby waived any charge that might have accrued to me?
 I roughed the square billet down to get the best diameter I could of 107mm, to a width of 35mm. and stuck a pilot drill of 11.5mm through for the plug thread. A spigot of 35mm diameter was machined in some 13mm long. Next a flange was produced by machining down from the o/d to leave the flange 12mm wide with a flat surface for clamping nuts and washers. This left a mass of material from 65mm diameter to hand rip down in a sort of flattish radius which then formed the outer skull of the combustion chamber, blending from the 65mm to 35mm diameter. The next operation was to machine a location spigot to protrude through the top of the outer water jacket, this was 25mm in diameter and 9mm long an `o` ring 3.5mm in width and 25mm inner diameter sealed everything up nicely. Sitting under the `o` ring is a 2,25mm spacer of 38mm diameter, more on this when we get to the combustion chamber design. The final machining operation was to cut a 3.5mm wide groove with a round nose tool along the external roof of the combustion chamber to bring water in very close proximity to the lower, hence hotter, plug threads, all helping to keep excess thermal loading in check.
I know this whole machining sequence may be difficult to imagine and I probably haven’t described it too well, so for those of you that might possibly be confused or maybe interested enough, I have a number of photos available of the head alone and also sitting on the barrel showing the water access openings. These pictures will illustrate everything very clearly, a good example of where an image can convey a thousand words. As always message me, or post on the site, your email address and I can squirt them to you, royal mail is also an option if I have your postal address.
It is not until one delves a bit deeper into the whole, and still evolving topic of combustion that the realisation forms that the subject is unimaginably complicated and you can`t even see it or feel it, like trying to grasp smoke!
The amount of energy available to the combustion process is finite, how that energy is distributed depends upon the parameters and priorities the tuner decides upon. Do you transfer heat energy into the engine metal, piston, head, barrel, cooling medium etc. or allow the pipe to use it? However, you don`t have to delve too far into the subject to realise that turbulence in the combustion chamber is an absolute and critical necessity. Vigorous gas motion exposes increasing surface areas of mixture to be enflamed. Kinetic energy is taken from the moving mixture and converted to turbulent spinning eddies which then breakdown to ever smaller eddies, and at this scale combustion proper can begin. It may appear somewhat perverse to actively encourage turbulence in one part of the engine and go to huge lengths to eliminate it everywhere else?
Because the combustion sequence cannot be instantaneously completed and the volume and shape of the combustion chamber is constantly changing then burn rate and chamber geometry significantly influence how much useful work can be created during the whole combustion process. By not igniting the mixture at an appropriate time the heat release process would not be efficient enough to capitalise on the small chamber volume at top dead centre, resulting in a loss of maximum brake torque (MBT). Back in the day all that was demanded was to buy one of George`s 12:1 cylinder heads, set the squish band, screw in a 9 grade plug and go racing, it was as simple and undemanding as that. The combustion chamber section may illustrate the development of the whole combustion subject which guided this design process when developing the engine from initially back in 1993 up to the present day. It is to our collective, great good fortune that science and engineering does not rest on its laurels but constantly moves ahead in the relentless quest for knowledge and improvement, guided at all times by peer reviews, always asking and answering questions. Enterprising design engineers within our Bantam racing world can grasp, interpret, exploit and capitalise on such developments as they appear.

Combustion Chamber
From the moment the exhaust port closes to the point it re-opens, the so called `closed cycle`, whatever mixture is contained within that constantly evolving volume is all you are going to get. It is therefore imperative that the maximum amount of energy, in the briefest time frame, that can be extracted from burning the fuel contained there is achieved, the exhaust pipe will also thank you for the energy it eventually has to work with. The actual amount of power which can be extracted is almost entirely dependent of the amount of oxygen contained in the cylinder prior to ignition. Leaving aside the geometry which will be addressed later, there are two variables here, the fuel type and the stoichiometric ratio of air (oxygen) to petrol. The quantity of air is pre-determined by the breathing efficiency of the engine systems but the amount of fuel is controlled by the carburettor function and the persons responsible for jet selection and as soon as the fuel is atomised and has left the carburettor and is on its way, that’s it.
For complete combustion the ideal combination of fuel to air is regarded as 14.7:1, a lambda value of 1.0, meaning that 1kg of fuel requires 14.7kg of air to fully combust. It is widely considered that to ensure that every molecule of oxygen is successfully mated with a molecule of fuel that a slightly richer mixture, lambda .86, is recommended, which means for every kg of fuel we need 12.64kg of air. Lambda .86 might ensure that all of the oxygen is consumed but does not guarantee that all of the fuel is consumed also. It is important for power production to create a chemically correct mixture, the A/F ratio has a very significant influence on flame speed (turbulent burning velocity) with the lowest flame speed obtained with leanest mixtures. The combustion process should be viewed as both a mutual chemical and a physical process, one won`t occur without the other.
I guess that the majority of Bantam racers will by now be using Avgas 100LL, some will maybe use unleaded fuel, or a blend of unleaded and Avgas and it is quite likely that a sprinkling of you will use some form of illicit race fuel  C12/C14? Avgas however has some disadvantages, in addition to its lead content having a detrimental impact on our environment and health. It is slow burning, far more so than unleaded fuel, and has low energy density, that is until the CR is jacked way up to around 15-16:1, 100LL then begins behaves more like pukka leaded race fuel. In fact all fuels are slow burning without active, turbulent action in the combustion chamber. Comp ratios like that will in all likelihood push the thermal envelope too far for air cooled engines, at least ones that are making any power, to safely operate at and don`t even think about it for cast iron race engines!
In general terms unleaded fuel hates compression, but loves timing and makes best power with rich mixtures, Avgas is the opposite when optimised, loves compression and lean mixtures and late ignition timing. Just in terms of reaction, these two fuels are completely different.
Octane numbers should not be confused with energy content, it is a common misconception that an octane increase will mean an automatic power increase, an octane number only allows for an increase in compression ratio without the fear of fuel detonation. Octane is a measure of how much pressure the fuel can be placed under before it detonates. The only way to liberate extra performance from such fuel is to raise the compression ratio, in doing so the cycle efficiency is improved giving more energy. There is also a direct correlation between bmep and compression ratios, however I cannot say that it is any way strictly linear so a pro-rata approach could be dramatically un-suitable. Power comes from bmep x rpm, with rpm potential pretty much determined by pipe action and port angle/areas, but neither are influenced by comp ratio, thus power must come from bmep. As always the trick is balancing rising heat against, in the case of air-cooled cylinders, a constant therefore limited and restrictive cooling potential, thermal runaway must be avoided by every means possible!
During the process of searching my archive for information I came across an invoice from George Todd, hidden from sight by being interleaved between some other Todd memorabilia, one that I had completely forgotten about, it was a very long time ago. The date was July 1970, 50 plus years ago and referenced a newly revised cylinder head combustion chamber shape with a request to try it and give our (Brother David and myself) impressions. As the new head was free of charge, we felt obliged to give it a fair test. With a squish clearance of .8mm the comp ratio was 14:1, very high for the period we are talking about, but the 40* conical chamber was very deep so the flame travel length was much longer. The cone dia was 34mm leaving a squish band width of 9mm on the 52mm bore! Typical 50% squish area ratio (SAR) for a 52 bore is 7.6mm, remember we were still in the long- stroke days in 1970. Thinking back, and with hand on heart, I can`t say there was any significant performance increase over the traditional hemi head profile at 13:1. It does show that George was thinking that comp ratios could be pushed a bit higher. It also shows, perhaps with hindsight, that without a realistic chamber profile, high compression numbers alone won`t always give the hoped for results.   Another interesting footnote then in Bantam Racing history and one which I was very pleased to be part of, the learning curve never seems to flatten.
There are numerous possible consequences to be considered regarding compression ratios. The higher the ratio the more the hot combustion gas in the cylinder expands prior to the exhaust port opening. As has been mentioned before, when hot gas expands it cools so loosing density and energy, (density is mass/volume), eventually to dissipate to nothing, for maximum power we need maximum charge density. Exiting exhaust gas will transfer less heat to the cylinder walls, suggesting maybe that a higher compression could yield a cooler cylinder? Turbulent gas will transfer more heat to surrounding metal than quiescent gas, increasing the rate and intensity of heat transfer.  The downside is that air cooled engines will inevitably run hotter than water-cooled ones will.
Fast combustion has a twofold effect, one is less risk of detonation and less time for heat transfer to all of the surrounding metal of barrel and head. The most fundamental requirement for effective combustion irrespective of the actual chamber shape and volume is a very rapid burn rate, the faster the better. One of the most effective ways of promoting a rapid burn is through turbulent squish action, and burn rate increases with an increase in turbulent activity. Burn rates also increase as pressure and temperature rise within the combustion chamber, a shorter burn duration will again also help limit the amount of heat transferred to surrounding metal, much more productive to let the exhaust system exploit it.
Make the squish band parallel to the piston crown and the squish gap as small as possible. By calculating 14.64% of a cylinder bore diameter you can conveniently arrive at the universally accepted 50% squish band width, for example, 54mm x.1464 = 7.9mm and, 64mm x.1464 = 9.4mm. If you have a good, rigid crank and mains, and a non-stretchy con rod set up (stretching forces rise with the square of rpm) then about 1% of the stroke is the number to aim for and around 35m/s squish velocity. Just shy of the piston touching the head at peak revs is what to aim for, a heavy piston set up and chunky con rod don`t help here. Where the squish band meets the chamber there needs to be a sharp edge, just a rub with wet and dry after machining is enough, definitely no radius, this edge will create a lot of turbulence which is crucial for a thorough and rapid burn. Squish action begins to be felt around 20* BTDC and has petered out by TDC, with having only that brief period of crank rotation it becomes imperative that squish action is as effective as can be achieved. When I first built this engine I went for .5 mm clearance at the bore edge and .76mm at the squish band edge to the chamber with a 50% squish area. The thinking at the time was that by reducing the end-gasses out at the cylinder wall meant that more mixture was available for combustion and the risk of detonation was reduced. The engine was fine with that but in subsequent heads I went for a straight taper that impinged at 4 mm from the piston edge and a gap of about half a mill, it seemed to work well and the engine never had any problems with detonation. Different fuels will permit different compression ratios, but two stroke engines, unlike four stroke ones, are not dependent on high comp ratios for power, so it is unwise to blindly ramp up the comp just because the fuel might be safe, we have the exhaust pipe to do the heavy lifting for us.
Strong turbulence in the combustion chamber is absolutely critical to promoting a rapid and complete burn and the radial mixture inflow of a symmetrical squish band is an essential component in achieving this. Another significant benefit of rapid combustion is that ignition can be started later, reducing pressure rise before TDC creating less negative work thus robbing energy from upward piston motion and peak temperature is present during a shorter time frame. Peak cylinder pressure, at peak torque, should ideally occur just as the piston begins its rapid descent from tdc. Something else worth thinking about with regard to heat input into an engine is not so much the rate of turbulence that you can influence but managing the gas temperature at the moment of maximum turbulence: when the exhaust port opens! However, no matter which way the piston may be moving as long as there is combustion the pressure in the cylinder will rise, but slow combustion can mean that the piston is way down the cylinder and it is not until combustion has finished that expansion of gas can begin, hopefully before the exhaust port opens.
In a typical full spec 125 race engine the plug sparks at around 14* btdc at 12,000 rpm. Combustion is initiated between the plug electrodes at a laminar rate, another 10* later and turbulent combustion can be detected, 40*after that it is all over. The rub here is that at lower rpm there is far less cylinder filling and combustion is a lot slower. As a fundamental, spark plug quality should have the base line of an NGK 10 iridium, no exceptions, this engine runs a 10.5.Big fat sparks will rip off the fine wire electrodes much more readily than any other plug configuration and will require less energy to do so. Back to back tests show that they out perform any comparable plug. As a rule, a two stroke race engine responds well to an ignition curve reflecting the inverse of the bmep curve. As pipe efficiency increases the dynamic compression increases dramatically, so when all of the engine variables synchronise at peak torque rpm the ignition timing tends to end up at around the 14*/15*advance mark.
Just to illustrate that two strokes are fickle things and are tricky at nailing down to generalisations, an 11 :1 compression ration with early ignition timing may well behave just like a 15:1 compression ratio with late ignition, but I know which one the pipe would prefer!
A domed piston is more effective in taking advantage of the Coanda effect for both transfer flow and inside the chamber for piston cooling, it is here that flat top pistons are at a disadvantage. Mixture ejected from the squish band flowing up over the piston dome will impinge into the heart of the combustion flame, creating turbulence so that every molecule of fuel is efficiently mated to an oxygen molecule and the whole conflagration creates massive pressure and heat to act upon the downward motion of the piston.
It might be timely to mention a few words about when it might be beneficial to reduce compression. As bmep rises so by definition do the values of Delivery ratio, Trapping and Scavenging efficiencies. These variables in concert with pipe effects ultimately increase the amount of clean mixture in the cylinder at exhaust port closure creating more power, and more power means more heat. As these combined effects become more efficient in cramming ever more mixture into the cylinder so the Dynamic Compression increases, thus the law of diminishing returns begins to adversely affect the optimum static compression that the engine can cope with. Air cooled engines are thermally limited by the fin`s ability to get rid of heat efficiently, something has to give and reducing compression ratio reduces heat input. Water cooling offers so many greater alternatives to manage more heat so the comp ratio can remain higher along with a greater thermal safety margin. Remember, compression ratios are just a geometric number which make no reference to the quantity of mixture actually trapped within the cylinder at exhaust port closure. The simple ratio is the maximum volume divided by a minimum volume.

Flat tops
A domed piston will by design angle squish streams as near the plug electrode as is practicable, the plug face should then protrude by a mill or so into the chamber. Flat top pistons can`t take advantage of this, however there is a reasonable solution, a toroidal chamber shape which will bring the plug much lower into the chamber reducing the flame path length. The bigger capacity Bantam engines with 64mm bores will have larger and deeper combustion chambers for similar ratios as their smaller siblings, a toroid could also help here even though they use a domed piston. Doing the comparative sums for the two capacities will soon tell their own story.
The inward, radial action of squish flow is symmetrical, but the flowing scavenge column of fresh transfer mixture is not symmetrical, being by design very much off centre at the rear of the cylinder. That mixture column moves upwards then loops over at contacting the cylinder head expelling old spent gas in the process. An intrusion suspended from the combustion chamber roof as in a toroid`s central boss could possibly impede the effectiveness of the scavenge process. It may well be then that too much is detrimental to efficiency, there is evidence to suggest that this is the case, but how much is too much, only experimenting will reveal that? I have never felt the need to use a flat top piston in a Bantam engine, I never saw it as a common sense alternative, but have been involved with RS Hondas that have used flat tops and/or toroid’s and a lot more unbelievably weird stuff, suffice to say reverting back to conventional profiles seemed eventually to follow.  

  Final combustion chamber geometry
Set against the background of decades of Bantam Racing experience, successes, awful failures and poor decisions, this final head design found its way to the drawing board. It might be more accurate and honest to suggest it was in part, plagiarised! However there are enough changes to the Aprilia original (Jan Thiel design) to enable me to say I have at a minimum, paid homage to a superior concept. When I first saw the images of the chamber geometry and read the explanatory text the thinking behind the profile was clear, to promote maximum turbulence which will enable all of the oxygen present to be consumed. As a consequence I felt sufficiently impressed, and persuaded enough, to try and replicate it in a Bantam context.
It is impossible to ignore the fact that the Aprilia 125 engines in both guises, when measured back in 2007, achieved an astonishing 16+ bar bmep at 13,000rpm with reliability and consistency. The combustion chamber design therefore must be a significant, contributory element in that achievement. Prior to the banning of leaded fuel in a GP context, CRs were regularly up at 18:1 running Elf 130 octane leaded rocket fuel, the sort of stuff you need the protection of a respirator to safely get anywhere near. The design geometry details, now freely available on the net, has the CR down to 15.6: 1. The remainder of the design is fairly conventional but the chamber roof itself is flat with a radius blending at 90* to the 50% SAR having a clearance of .7mm from the 190mm radius piston crown which in turn clears the plug face by 7.6mm.
It would have been convenient to be able to simply reproduce the Aprilia profile in a new head, but the cooling potential of my engine is vastly inferior to that of the Aprilia by a number of litres of water per minute so some concessions to thermal survival had to be made, whilst also not losing sight of the limiting factor that the cylinder barrel is still cast iron! The now accepted standard for adequate water flow is, one litre, per crank horse power, per minute, so for the Aprilia engine that means 60 Ltrs/min!
The crown radius on our CR piston is 147mm giving a height of 2.5mm and the Aprilia`s is 190mm giving a crown height of 2mm, quite clearly some alterations were going to have to be made to make the new design viable. A plan was conceived to start with a modest, 12.5:1 CR and incrementally raise the comp if the engine remained reliable, problem is the whole chamber would normally need adjustment as volume went down. The arrangement I came up with was to lower the chamber volume by introducing a short parallel section where the chamber met the squish band, consequently reductions in volume become one simple calculation. So if I wanted to raise the comp by the equivalent of .5mm to 13:1 just a skim on the gasket face, another to the squish band and we are good to go, the chamber itself remains untouched. There is a combustion chamber drawing available for clarification, (there are always drawings) for anyone interested in getting hold of the drawing contact me in the customary way and a copy can be yours.
One configuration which I recently thought about was to try the toroid head in conjunction with a domed piston machined flat, on a 54mm piston, to a diameter of 38mm on the on the crown thus leaving an 8mm wide squish band, could be the best of both worlds, bringing the plug much lower into the path of the energetic turbulent squish flow. It may well then be the case that a domed piston in conjunction with a flat roof chamber has a similar effect as a flat top piston and a toroid chamber, so many avenues to explore so little time!
Although it is not one of mine, I do also have a toroid drawing for a 54mm bore with a 2mm high, 190mm piston crown rad, using a 10 grade plug the volume is 9.75cc so the comp ratio is 123.7+9.75/9.75= 13.7 Shouldn`t be too much of a problem to adjust the chamber volume to reduce the compression ratio or to accommodate a different piston crown radius. Either way, it makes for interesting scrutiny, if anyone reading this would like to obtain the drawing then you know by now what to do.
It was my original intention to describe both the cylinder head and liner in one article but this instalment has grown so large as a consequence of which, the liner and porting stuff will now be delayed until the next episode.
Cheers for now, Trevor