Extra, Extra , Reed all about it, headline news!

Conceptual Design Analysis of a Bantam Race Engine

Part 3, with emphasis on the Reed valve Inlet System.

The reed valve inlet control concept could almost have been developed exclusively with Bantam race engines in mind, the benefits it brings in terms of easier starting and general, clean, unfussy running and power band enhancing action, all of which compliments the Bantam`s limited gear options ideally, are quite remarkable. The dynamics of this type of inlet control provide many tangible and useable advantages over traditional, symmetric, fixed timing of piston port inlets. Back in the early 1990s, reed induction was at the top of the list when I first jotted down my fundamental priorities and requirements for proceeding with this engine design and development. A project which I have to admit that initially I was quite reluctant to embark upon with my enthusiasm for `things Bantam`, at a minimum, and this would certainly be quite a lengthy endeavour. However, once started, by ferreting through my archive of Bantam stuff for inspiration and impetus and some initial basic research my enthusiasm, at least for this new engine, was rekindled.
It is all the more surprising then that such a simple device as a reed block assembly can operate so effectively, when it is ostensibly little more than a conical, one way check valve controlled by pressure differentials, the inertia of mixture flow and reed material stiffness, with one pressure element simply being atmospheric. The great appeal of reed intake control is that its opening and closing points are not pre-determined, but are actioned by the engine`s actual air-flow demands which of course fluctuate with throttle openings and engine speed. The opportunity also arises to incorporate a useful fifth (or third) “C” transfer port, with the cylinder communicating with the area volume around the reed block. Additionally, with crank rotation through bdc, pipe effects could persuade the reeds to lift and the cylinder to communicate with the outside world via the carb. Taking all of these positive functional and design features into consideration, reed valve selection over the piston port inlet alternative seemed to be a total no brainer, but what to choose as hardware?
After a bit of a selection process we ended up choosing the Honda CR125 reed assembly above other possible candidates, probably because we got it all at a knockdown price, second hand maybe as hard cash for the team was always tight? One nice addition was that of a streamlined plastic insert which fitted neatly into the back of the reed block and helped to promote smooth mixture flow and maintain air velocity, pretty basic though when compared to today`s sophisticated `aerodynamic` offerings. Standard RS 125 reeds would fit ok so we initially went with that. A quick `off the cuff` calculation of skirt and tip area at maximum reed lift, times six reeds, suggested that the CR block should flow as much mixture as any Bantam engine would ever need. Assuming that is that the reeds can in reality be encouraged to achieve full lift for any meaningful length of time?
Out came the by now much used drawing board again, and I drew out a partial view of the drive side crankcase looking inside, with the barrel basics drawn out in situ on the case mouth. By including such details as crank/cylinder centre line, rod length, piston crown and skirt position, etc. at bdc, important reference points are established. A stiff card template of the reed block profile taken along the flow centre line together with the curved reed stops in place was made which when laid on the drawing enabled the final compromise position to be selected which accommodated the bulky carburettor at a sensibly angled location. Eventually though we were fortunate enough to get hold of a carb manifold rubber 30mm shorter and straighter than the original CR and were able to squeeze the whole lot in together, one advantage of having drawings to adapt and not having to re-cut metal!
The reed block housing was mocked up using suitably shaped, 3mm balsa wood sheet side plates and stiff card for the curved top and bottom plates, which cleared the reed stops by a couple of mm or so, the flange was thicker 6mm sheet balsa. All of these bits were tack glued together using cyano, when I was happy with them and were rigid enough to offer the reed block into for appraisal `in the hand`. Eventually when all was approved, it was a simple matter to cut the 3mm steel side profiles on a band saw, radius the top and bottom plates, machine up, drill and tap the 6mm thick flange and weld the whole thing together. Once again, I called on the expertise of the guys in the fabrication department of the engineering firm where I worked to `persuade` curves into the top and bottom plates, they opted to use a small hydraulic press. One guy, rather modestly, characterised the sequence as, a nudge here, an over push there, flip it over, regain the profile, another tweak and so on, ten minutes for both plates in the hands of those experts and it was all done.
The inlet side of reed valve induction doesn`t in reality even need a port in the commonly accepted sense, the reeds themselves control that. What is crucially needed however is piston support, this is no more evident than in the 125 RS Honda engine, although it must be noted that Bantam engines are by and large cylinder inlet and not case entry. The port I ended up with was far bigger than anything I had used before in piston ported cylinders and was open for 360* of crank rotation. I have to admit at the time to being a little nervous of piston survival. Distortion and subsequent failure with the piston skirt repeatedly flexing in and out of the big port induced by lack of adequate physical support on the major thrust line I felt was a very real danger. To eliminate that eventuality I incorporated a narrow central bridge in the liner; and I am very pleased that I did. There can be no doubt that the bridge would disrupt flow to some extent but that small deficit would be more than compensated for by an inlet cycle which could be open for a period of up to 265*of crank rotation, depending on the rpm and pressure ratios produced during each engine cycle. This is way more than any conventional piston port could sensibly offer in a Bantam race application. Piston ports always suffer backflow due to their long open period after top dead centre whereas the inlet closes when reed stiffness asserts itself, which in turn self-adjusts for changing rpm pressure levels.
 As in the situation of the exhaust port, a short inlet spacer was needed, the reed block housing was to be brazed to the outside of the water jacket so the profiled spacing piece was again from the inside of the jacket to the barrel stub fins. I won`t go into a blow by blow account of how that was done as it was a quite similar sort of procedure to the exhaust duct, take a chunk of metal carve it about, weld up the mistakes, carve more. What I couldn’t do was produce compound curves at multiple angles, that is the domain of casting and modern CNC machining, at the time I had access to neither, but I did have access to Devcon and Plastic Padding! There can`t be many Bantam race engines out there which don`t have some filler inside somewhere!
Reed lift opening, thus initiating crankcase inflow, is dependent only on the pressure differential across them and not with piston motion, it is perfectly possible that given enough pipe suction at really low rpm and a descending piston for the reeds to be activated. Reaction by the reeds to varying pressure levels will also be influenced by reed “stiffness” plus mixture inertia. The inlet tract column length of mixture has mass so won`t move unless forced to do so and as such creates a time lag in flow within the inlet tract. The initial flow velocity is zero and only then reacts, through acceleration, to the increasing pressure differential and the gas inertia per sq. mm of cross section area. The pressure difference varies because of mass transportation of mixture from one end of the duct to the other. These pressure changes only propagate at the local speed of sound with temperature being the major influence, and the inlet is cooler than just about anywhere else in an engine. Add in other influences like pipe action, volume variations in the reed block, and energy loss by rapid expansion at the increasing area just after the reed tip exit into the crankcase and so on. Mass flow rates reduce at any expansion in cross section area, and this applies throughout the engine and the expansion chamber, hence the name. The exhaust pipe diffuser, by means of sending a negative (suction) pressure wave signal via the transfer ports to the crankcase, also motivate the reeds into action. In most race engines the reeds will start to lift somewhere after bdc and transfer port closing (tpc) and more likely closer to tpc. What opens the reeds initially is the case going sub-atmospheric, and the intake wave going positive, these combined influences plus the rising piston progressively dropping case pressure, slowing down transfer flow in the process, then prompts and maintains the whole cycle action. It can only be then the pressure differential across the brief period that the transfer ports in the bore are open that is transmitted to the crankcase. This also then dismisses the notion that there is in one cycle, flow through the carb, the intake tract, the reed block, into and around the case, up through the transfers and into the cylinder. It doesn`t happen, there is just not enough available time, that’s for lawn mowers. For instance, in this engine there is as much mixture volume in the combined transfer ducts as the cylinder swept volume, it is this reservoir of fresh mixture that is drawn on by pipe action, even accounting for a possible 100%, or better, volumetric efficiency. The crankcase mixture reservoir replenishes the transfer ducts in readiness for the next repeating revolution which takes a miniscule fraction of one second to complete. It is always worth keeping in mind that the transfer ports are open for the shortest of all port`s period of crank rotation but the ducts are accessible 24/7. Imagine trying to persuade mixture to move from the bottom of the crankcase, under and between the lower crank wheels, upwards through the transfer ducts and into the cylinder with transfer ports that are already closing rapidly, an unrealistic ask?
Reed valve equipped engines do not respond at all well to `large` crankcase volumes, typically well below 1.4. Bigger volumes given the right conditions can have a positive effect, but this is significantly offset by the change in Helmholtz frequency and that affects the natural resonance at reed opening. To get any sort of response requires thinner and thinner reeds to get resonance working. But as soon as rpm begins to climb toward the 1st harmonic which might be around 15% less than peak torque rpm, the thin reeds flex and flap uncontrollably causing havoc with power delivery. Stiffer reeds can reduce uncontrolled reed action but to the detriment of low speed running. Behaviour like this can be clearly seen in simulation and during testing on the dyno. Any race engine in the lower power capability range (read Bantams) will respond far better with case volumes at 1.4 or even a tad over. The smaller case will speed up flow so filling the case, and emptying it, more effectively and giving a much more positive signal to the carb which will aid low rpm pull from slow corners. A `large` case volume is just not a critical need in our low bmep engines, where potential losses can outweigh any supposed benefit. Fundamentally, a generous case volume transmits its contents quite languidly and thereby degrades throttle response where reeds won`t lift as far, whereas a tight case is prompt and stronger in action. The crucial problem is in getting all of that mixture into the cylinder, via less than efficient transfer ducts, and of course keeping it there. It was just this reasoning which was the prime mover for me in making my own transfer ducts with their greatly improved flow potential, and synergistic directional control of mixture stream flows into the cylinder.
There are a number of “don`t do’s “associated with Bantam reed installation, the most glaring of which is to, not even consider using the reed assembly from a 250 MX single engine in the hope that bigger is better, it won`t be! As an example, this was initially tried by Yamaha for the TZ 125 single that was to be a competitor to the Honda RS 125. It didn`t take long for tuners in Europe to junk the 250 block and substitute the RS items, the sides of the housing were lined with suitable alloy packing plates to take up the excess volume created, and guess what, power went up across the range!
Another installation to avoid is the angled, off to one side, manifold, a lot of MX engines feature this and some Bantams have the same manifolds sourced from MX cylinders. Anywhere you have asymmetric flow through the reed entry you lose power, there is higher velocity flow out sideways than directly on the centre line axis. The flow killing turbulent spill-over flow through one side is very apparent, but it is to the loss of flow to the other side reeds, and to centre flow. Some compensation could be made by adjusting reed stiffness across the horizontal plane to attempt to redirect the biased mixture flow back to being more symmetric across the 3 petal curtain areas. But the complete cure is a straight manifold where all centre line elements of the inlet tract from carb bell mouth to port are aligned.  
From the very first proving run of the new engine and going forward, it was always intended that there was going to be a program of reed optimising through pragmatic testing. By reed testing I mean the actual petals themselves, at the time we must have tried every available type and combination of material, plastic, fibre glass, carbon fibre, single petal, duals, waisted petals, the list goes on. Even homemade arrangements, with thin carbon petals, a rev plate made from an old reed, short tapered carbon back-ups and a much shortened stop plate all clamped together. It worked just fine, but not noticeably better than anything else did, I can`t say with any certainty that any one particular set up was clearly superior to any other, this might be a symptom of low bmep engines. Even having different stiffness petals on top and bottom of the block to encourage flow upwards seemed neutral. This all seems at first to be out of kilter in that that the Young`s Modulus of Elasticity and petal density vary between the different petal materials, carbon fibre seemingly having some small advantage, suggesting that carbon fibre is the way to go. The physical properties of various reeds were of course determined at Queens University it was found that glass fire reeds had modulus of 21.5 GN/m^2, carbon fibre 20.8GN/m^2 and densities of 1850kg/m^3 and 1380kg/m^3 respectively.  
As soon as the reed length/lift ratio approaches 1/3, petals may bounce around in some odd form of rhythmic pattern. Any improvement by two or three percent of the typical Bantam crank power is unlikely to be felt much at the back wheel.  The accumulation of test data and hands on experience tends to suggest with this engine, and perhaps many other competitive Bantam race engines, that any limit, or bottle neck, to improved power production lies not with the reed inlet system but somewhere else, yet to be discovered. For example, Queens University carried out an enormous amount of development work for Yamaha including comprehensive reed valve testing. What was discovered, by using high speed cameras peering through a Perspex window on one side of an LC 250 reed block with the engine on a dyno, buzzing away at 9,500rpm, and connected up to strobes and every measuring gizmo known to man, was that the greatest influence on reed action came from systematic alterations to the exhaust system and crankcase volume; so get your hacksaws and calculators out! Subsequent investigations led to a test rig being developed which allowed rpm up to 13,000 rpm to be used.  Reeds are certainly not as uncomplicated as may first be assumed.
A little bit of investigation may illustrate this more eloquently than several paragraphs of dull text. The natural frequency of a reed for best overall efficiency is about 80% of engine speed. So at 10,500 rpm that will equate to 140 cycles per second, .8x10500/60. However, frequency is proportional to reed thickness, but inversely proportional to length-squared. So double the thickness doubles the frequency, but double its free length cuts its frequency by a factor of four! It may be that when playing with reeds this could be the reason some combinations produce so little difference in engine performance?    
What did make a dramatic difference to the power production with this engine was when we were eventually able to acquire and install a second -hand ignition system from a Honda RS. The retard facility available transformed the engine behaviour, the old Motoplat straight line timing system proved to be a serious bottle neck restriction. The extra ignition advance at low rpm really did have a positive influence and consequently affected reed function, for the better. What we didn`t bargain for was the extra rpm gain at peak power rpm as the ignition retarded past the old static advance setting allowing the engine to rev on as the extra combustion heat `dumped` into the pipe and not to the cooling system, effectively shortened the wave length by speeding it up. Such a dramatic change to the engine behaviour also affected the reeds and carburation so a certain degree of further optimising was called for. It is quite clear that altering the reeds changes the air flow through the inlet tract and it is air flow which then influences the fuel flow. The pressure drop in the carb, which persuades fuel to rise and exit the spray tube, is dependent on the speed of air through the carb. Much more crucially is that it is proportional to the square of the speed. A small change in air velocity gives a much bigger change in speed-squared. For example an increase in speed of say 2.5% will create a pressure difference of over 5%, and that change is sufficient to influence the operation of the reeds. As is so often the case, change one thing and a whole raft of other things get altered and need re-optimising.
I don`t carve extra inlet holes in the skirt of the pistons used in this engine, after saying that however, there was one exception, I did experiment by milling a hole high up below the piston ring to hopefully allow trapped hot mixture to exit from beneath the piston crown under pipe influence. But that was a bit different, the intention was to let hot, stagnant mixture out rather than to encourage fresh mixture in. Again, no discernible difference was detected so didn`t bother again, The Honda CR125 piston we use is unmodified except for a gentle de-burr with a fine needle file, more or less out of the packaging box into the engine.
One effect, which is not often referred to which none the less comes into play is that by including, or omitting, piston skirt holes will affect the cavity resonance of the reed box and in turn will influence reed resonance behaviour. The question must be, is your engine actually, or even likely to make enough power to realistically need extra inlet flow area? There are three main variables, the port effective area, the fully open petal curtain area and the 1st mode natural harmonic of the petals in bending. One more factor to think on is the Bernoulli Effect, when air moves some of its pressure energy is lost, the small narrow openings between the reeds are like small venturis and the flow through them generates a partial vacuum this tends to draw the reed petals closed, it`s always interesting to ponder and think about these things, but don`t lose too much sleep over it.
Quite clearly there is far, far more at play here than just a few reeds flapping up and down, it all gets way too complicated, way too quickly, the possible permutations seem to run to infinity. As in reality, “suck it and see” is the only practical way to evaluate a working rule of thumb for your individual, home tuned set up, but it may take some time, with a certain element of “mix `n` match” thrown in for good measure. Perhaps even attempt to align the forcing frequency of the inlet tract to the reed frequency when experimenting with differing reed thickness and material? Or, stick in a cheap, generic set of reeds forget all else and just go racing, and that is what it is all about!

Instalment 4………cylinder head and liner.

Cheers, Trevor.