A little light Christmas reading

Conceptual Design Analysis of a Bantam Race Engine                
Part 1, with emphasis on preliminary transfer duct design and construction.
                                                                                                                                                  Towards the latter part of last century, well 30 years ago at least in 1993, after an enforced period away from any form of association with Bantam racing, I started to get the initial thoughts sketched out as to the design and construction of a new, and for me, a more advanced version of a Bantam race engine, and it was in all probability going to be the last one I would ever tackle, life`s priorities move on. It had to be water-cooled of course and have reed valve induction, these were prerequisites, and to be of square bore and stroke dimensions. I spent a lot of time considering and trying to eliminate as many frictional and mechanical losses that I could. The gearbox deserved all needle roller bearings, including the sleeve gear. A new indestructible all metal clutch with hardened steel plates was on the list, its body made from en24t. Westland helicopters donated the billet of alloy steel, but I doubt they knew much about it! I skimmed the chain-wheel true and parallel on both sides then milled lightening holes, after assembly the complete unit was dynamically balanced, another of those make, fit, forget upgrades. Chunky 25mm journals for the new crank shaft and bigger main bearings were essential, flexing cranks cost power and the ratio of crank diameter to bearing diameter is an important factor. The full shopping list initially was far more extensive but time constraints prevented any additional work, but I now wish I had persevered beyond the drawing board and card templates stage with the exhaust power valve, what a difference that could have made! My biggest disappointment at the time was that we had no cash left to buy a modern alternative retarding ignition system, the old straight line Motoplat ignition had to be used which was a severe performance handicap. A decent rev counter was another unaffordable piece of equipment and one that would have proved very useful.
It took me quite a long time, working in the main single-handedly, to develop the engine design from those initial thoughts and pencil lines on paper to subsequently progress to machining work and welding and eventually to the race track. Drawings are very forgiving, mistakes can be easily erased before getting to the metal cutting stage and a better design can often emerge, that happened a lot for me. However, the protracted exercise ultimately proved to be a rewarding one and the end result produced immediate and dramatic success at its inaugural outing at lydden in June 97.
The theme driving the push towards high bmep demands that all sequences of engine cycles are as efficient as possible, that at least seems obvious. Most of the limiting restrictions of the original Bantam engine design are dispensed with and replaced with modern efficient alternatives, particularly the cylinder. What must be avoided at all costs is the sort of scenario that produces the exhaust timing of a GP racer, the blowdown time/area of a lawn mower, the transfers of a sport roadster and inlet of a commuter engine, or any other permutation you care to make. You must have all the ducks lined up, everything must be synergistic, where the combined effect is greater than the sum of their separate effects, you can`t make up for later what was missing at the beginning! All has to be right from the get go. Ensuring that all the tough homework is done correctly will be rewarded with an “A star” performance!
Inlet and exhaust ducts are a relatively straight forward proposition, ostensibly straight in and out, a somewhat simplistic notion I know but I`m sure you get the general idea. However, the multiple transfer ports and their corresponding feed ducts were a different challenge altogether, a triple compound problem, I would have to construct them from ground zero. As always for me, drawings were also a pre-requisite, but before that came the initial design concept, the back of a fag packet stuff. I have always felt there is also a place for the interrelated concepts of form and function, for instance, Italian women, architecture, cars and motorcycles, always have a style, flair and grace about them and their bikes perform extremely well, Spitfire, Concorde and E-type Jaguar are each classic British examples, they even looks very fast. If after due consideration your judgement suggests that it looks right, then it probably is.
The cannibalised D7 iron barrel eventually consisted of only top and bottom flanges with a length of cylinder wall tube, with holes in, separating them, no ducts were going to survive the machining onslaught and the remnant fins were mere vestigial stubs left to maintain structural integrity whilst leaving a tad more surface area for cooling. Should anyone be sufficiently interested, then, some time ago I posted a selection of dimensioned drawings and images of duct components here on the Bantam Forum site, just dig around and you should be able to find them! In fact those images could well explain things more eloquently that I can describe them here. Alternately, send me your email address or postal if preferred and I`ll send them off to you, plus a few items not originally seen on the forum, where appropriate and as it`s Christmas, I`ll even cover the postal charge!
The first question must be, what are we trying to achieve, do we have an agenda, an ultimate goal, a target or just a dream? Then, perhaps just as crucially, can this vision realistically be produced within the context of the cast iron Bantam barrel, and of course along with the making and installation of a cast iron liner? The obvious and fundamental requirement is to try and make more power than the opposition, and be able to consistently circuit the track in a shorter elapsed time than the competition.  I had established a fair assessment as to the potential of the main opposition and felt we could create something to be at least the equal of that.
Clean sheet design, for whatever application it may be, is a huge challenge and you don`t get too far into a Bantam project like this before compromises are forced upon you. Engineering is a discipline which is predicated by compromise, and in so many instances the compromise is forced somewhere down the line necessitating some degree of re-design. In many respects then, it is easier to alter an established design by simply adding to the parameters of those existing features. But let`s face it, no one would contemplate the construction of a race engine by starting with a Bantam barrel! The limits of potential achievement fall within the two extremes of; what you would like to create and what you are actually able to achieve. As is so often the case, that gulf can be just too much for the average person to reasonably overcome. However well-intentioned you may be, if you don`t have access to welding gear you can`t make an exhaust pipe!
Within the constraints of what was realistically achievable and a finite time frame but continually hampered with a severely limited financial budget for construction of the new engine, a series of initial thought concepts as to duct designs began to formulate. The transfer passages were clearly going to be the most challenging, and at times certainly proved to be so. The exhaust duct for instance, whilst complex and at times tricky to make, was just a single item, similarly the inlet reed housing is not much more than just a simple welded steel box. The transfer ducts comprised of many individual parts, inner and outer with side plates and dividers, and there were two of them with left and right hand orientation. At times the entire exercise took on the mantle of a nightmare scenario, attended by demon spirits that stalk the mind at night! By complete contrast producing the crank assembly seemed a far more relaxed exercise, coming after eventually finalising the transfer ducts and seeing them fixed into place.
The primary function of the t-ducts is to deliver as much fuel/air mixture into the cylinder within the time restricted duration of transfer port open period. The mass of mixture contained within that column in the duct equates to; specific mass x duct length x duct cross section area. To move that mass into the cylinder in the port open period requires sufficient energy to be put into it to initiate movement and achieve a certain flow velocity. However the velocity is inversely proportional to the duct`s cross section area and length, and, if this wasn`t tricky enough the pressure difference required is proportional to that velocity! Flow velocity depends on the pressure delta between the entry and exit of the duct. Flow acceleration from zero eventually reaches its maximum velocity when that pressure difference equalises. The flow velocity in a duct, be it transfer or exhaust, is governed only by the pressure difference over that duct, but that then only applies if both the flow velocity and pressure difference are constant, that is never the case in a running engine! The pressure difference varies during mass transposition of gas from one end of the duct to the other. Compounding this are the other influences of pipe action and volume changes to the crankcase and cylinder as they each come into and out of play.
The rate of mixture movement within the t-duct as we have seen is dependent on the pressure difference between the crankcase and the cylinder, and the gas inertia per mm^2 of cross flow area, accelerating from zero to its eventual max velocity. Complicating this is the fact that piston motion from bdc begins to lower the case pressure which in turn will try to slow mixture movement. At the point of t-port opening the cylinder pressure is always higher than case pressure, one of the reasons why staggered port openings are effective where the first to open, the A port in most cases, is the last to flow. Two ways to help with this, increase case pressure and/or reduce cylinder pressure, cylinder blowdown time/area and pipe effects heavily influence these cylinder conditions. That’s part of the reason why at 2/3 of max torque rpm a too early returning positive exhaust pulse is so damaging. The lower inertia within the transfer mixture columns offer much reduced resistance to being shoved back down their ducts and into the crank case, and recovery takes time. That is just one example where an extra couple of gears would be most helpful.
Historically, and up to this point in the progression of Bantam race engine development and tuning, it was all about trying to get the cylinder`s elementary port/duct design to work a bit better. I did some pretty elementary investigation, doing a few sums, into the relative transfer duct flow potential taking a race 175/186 type engine barrel and this new one which only existed as images and numbers on paper then compared the two. I had to make a few assumptions but they were the same for both and the redesign was better by about 18%, very elementary unscientific but the advantage of one over the other was quite clear.
The original Bantam`s t-ducts are basically of the “elevator” type, straight up with a tight turn into the cylinder and high flowing mixture simply refuses to make that tight turn, they will get the job done if originally you only needed commuter level of power. Of their type they are not too bad and a quadrupling of BSA D1 power is a relatively achievable exercise but that is not terribly impressive for a race engine but bulk flow at higher rpm is not their forte, and as such form a limiting constriction. In all types of similar ducts the flow coefficient is very poor and their rudimentary geometry creates massive pressure absorbing turbulence, compounded exponentially as flow velocity tries to increase but mass mixture flow can`t do the same. Cramped transfer ducts are a lose-lose situation, Bantam tuners however by their very nature are an ambitious lot and more considered and adventurous porting was sure to happen, so fast forward from the pioneering days of the 60s to 1993.
All ducts irrespective of whether they are inlet, transfer or exhaust should provide an easy path for the transit of fresh charge, and hot, spent exhaust gas, with as little resistance to flow and pressure loss as possible. As we have seen, the geometry of the t-ducts, namely their length and cross-section, all play a role but the duct radii are especially crucial. The inner radius being critically important, it should be as large as possible to avoid flow detachment. Keeping flow clinging to the inner duct surface means there will be less pressure absorbing turbulence, and that flow will transport a greater, denser mass, experience less resistance and deliver it where you want it to go. The outer duct profile is not as critical as the inner providing there are no hazards where flow can be impeded and disturbed. Flowing mixture within the duct is forced by pressure action to follow the contours of the outer wall, smooth gently guided flow is far more efficient. Pressure action influencing the outer duct flow also assists in keeping flow attached to the inner duct, which is why transfer duct profiles are so very important.
The radius of inner and the outer duct profiles tighten as the combined narrowing effects turn inwards at the proximity to the cylinder port window. As flow velocity increases due to the duct`s narrowing, the inner wall radius effectively gets smaller and trickier to negotiate so flow begins to detach and containment of flow direction decays. In a perfect world that inner radius needs to increase in size as it nears the port window, instead we have to content ourselves with a relatively constant radius inner duct, which, it must be remembered, is also the floor of the approach to the port window. The transfer duct/port floor geometry at bdc is a super critical area for any high performance application and getting this absolutely correct should be standard practice. The relatively shorter turn radius and the flow exit angle combine to maintain a semblance of coherence within the exiting mixture streams.
It is at and around bdc with maximal port area that the majority of mixture transfer occurs, with the combined effects of high case pressure and a correctly dimensioned and timed exhaust pipe working at an appropriate rpm. However, it may be that the mass of transfer mixture is still accelerating from zero, so flow velocity may not be at its peak even when the piston is on its way upward towards transfer port closure. Even after 20*or so of crank rotation abdc the transfer flow could still be accelerating. Coincidentally, it is around this point in crank rotation that the reeds might just be thinking about lifting off their seats to initiate inlet flow movement, responding to the combined influences of atmospheric pressure in the inlet tract and increasing crankcase volume with its pressure going sub-atmospheric. The part cycle of case fluctuations has the positive pressure accelerating the transfer flow, to a negative pressure trying to slow it back down to zero, hopefully the mixture column has enough inertia to continue to flow, even though it is slowing down approaching port closure. Relative con-rod length can have an influence on crank dwell and reed action and acceleration for similar crank angular rotation around bdc.
I can`t over emphasise the critical importance of correct alignment of the t-port floor in the liner, the original remnant barrel wall and the edge of the inner duct radius. I arranged for the lower outlet angle of 9*to form a tangent with the piston crown squish band area at bdc, the need is that all of the edges co-align and have no steps to disturb transition flow and maintain optimum attachment of that flow over the inner duct wall, out and over the piston crown. Bdc is potentially the point of maximum unrestricted mixture flow into the cylinder so there is no place for shoddy alignment, flow coefficient is at the point of no return! Keeping the piston crown cool is an important and very welcome by-product of maintaining the transfer stream`s use of Coanda effects to exert some degree of control over cooling flow—when this doesn`t happen everything turns to a very nasty form of ordure!
Having gleaned as much relevant information from such material as was available at the time, ostensibly from sources other than specific Bantam, coupled with a few years of Bantam race engine construction experience, plus my own engineering background, the real work had to begin, prevaricating any longer seemed more like a protracted example of indolence.
It seemed obvious that to enable both t-ducts to be identical, they would need to be formed at the same time, port directional symmetry is vital in the case of mixture transfer into the cylinder. Getting this feature wrong can cause mixture to swirl around the cylinder and be lost exiting the exhaust port. The solution I came up with seemed to satisfy all criteria that could realistically be applied, at least it was doable. A steel ring was made with inner and outer diameters corresponding to the barrel stub fin diameter and a location lip skimmed into the inside of the lower barrel fin. A 6*angle was machined outwards which blended into a 22mm radius which stopped at the stub fin diameter. That 22mm rad size could eventually have been larger but possible revised size is not annotated on the original drawing, I just can`t recall precisely which? The wall thickness of the remnant barrel and liner then completed the top edge transition upwards at an appropriate angle and rearwards into the cylinder. The outer radius was free handed on the lathe leaving a 3 “ish” mm wall thickness, when all was finished I just parted the ring off the remaining lump. The rough free-handing left a course, grooved surface of greater area to hopefully shed the odd degree more heat into the coolant water. It was then a simple matter to cut two segments from the ring for left and right hand. Profiled 3mm steel side plates were then welded to each duct piece and, hey presto, two identical, elegantly profiled, free flowing ducts. The typical 3mm thickness steel used for fabricating the various ducts is a repeating theme throughout the design and construction of the engine. I wanted rigidity of the whole fabrication and no vibrating flexure, for that just induces fatigue cracks then leakage and reliability must paramount. 3mm wall thickness also provides for a useful weld prep to be formed as required on any edges.  Make it once but make it right, was the mantra I applied.
The inner duct radius profile was again made in the form of a ring but this time in aluminium, with the inside diameter corresponding to the liner o/d. That way segments could be cut out of the ring and fixed to the liner after its installation along with profiled plates forming the septum between the two pairs of side transfer ports, namely A and B, with the rear 5th C port being the domain of the reed valve assembly. Angling of the septum plates then provided suitable area ratios for the two ports of differing sizes. When eventually I milled the ports into the liner I cut a small shallow groove into the liner o/d between the two transfer ports, these then locate the septum plates within the ducts and it was then a simple matter to precisely locate and position them against a common template for accurate, repeatable area/volume control for both ducts.
I always include about half length of the case radius in any consideration of mass flow up through the ducts, for they do make up part of the duct, no other part of the case contributes. Under no circumstances should the lower part of the liner transfer cut-away be knife edged, I know, we have all done it in our time, a generous radius of the liner thickness is way more efficient. The only time a knife edge will work without creating massive turbulence is if you can guarantee the flow direction is dead parallel with that edge sides and not at an angle to it as is the situation in the crankcase. Even then small radii improve flow over and around sharp edges.
 When reviewing the length of the fabricated outer duct profile we have part of the large case radius flowing into the short slightly outward 6* taper blending tangentially to the 22 mm turn radius, the whole duct wall takes a gentle, ostensibly curving path flattening progressively to the angle leading to the port roof in the liner. The various angled side and septum plates dictate the direction that the mixture streams take into the cylinder, and the degree of taper from the entry in the crankcase to the port window in the bore. It is important that the inner and outer duct radii have asymmetric centre point positions this will encourage the ducts to naturally converge as they approach the port window. A common centre point just produces concentric circles and that doesn’t work efficiently, with the inner duct having a considerably smaller blending radius than the outer, thus requiring a different centre position.
 All of this description is a loose interpretation of my thinking back at that time but the results we achieved are I feel, proof of concept, scrutinising the on-line forum drawings or hard copy will clarify things a lot. Incremental changes were progressively made along the way and the final result was a much improved, user friendly power delivery, more in harmony with the restraints imposed by only having 3 gears available. Throughout all of the design processes I always tried to place emphasis on breadth and intensity of the mid-range power delivery even to the sacrifice of sheer top end power, although I don`t think we lacked that either. The eventual replacement of the Motoplat for a second hand RS Honda ignition set-up was transformational and made a massive performance difference and graphically confirmed my contention that the engine`s power delivery was emasculated by using a single point ignition system. Mark was delighted and declared the difference was like having an extra gear! There are always more slow corners than fast straights on race circuits and it is there that good drive delivered from a wide power band can save precious slivers of time, and taken over many corners and many laps the accumulation of savings can be considerable.
The next instalment could well be the exhaust duct, but the last couple of days have been difficult for I have just tested positive for Corona virus so not a lot will get done before the New Year. As I am on the vulnerable list of patients with compromised immune systems, I have had six covid jabs but even so I have succumbed to the damned bug. Although our family Christmas get together has now been postponed I hope you all have a great time and enjoy a rewarding New Year.
Trevor.