Second Epistle to the Bantamites

Mixture streams flowing into the cylinder do so along two vectors, axial and radial, both of which are variously controlled by the design geometry and position of the transfer port windows, duct orientation and proportions, and to a lesser extent, the piston crown profile.
Primarily what we hope for is that the various transfer streams angling rearward away from the exhaust port, conjoin on the rearmost fore and aft centre line of the cylinder, slow each other down somewhat and form a single, high density mixture column of reduced velocity. The column then flows up the rear cylinder wall towards the head whilst expanding into that ever- reducing cylinder volume. Above all else what we do not want in those scavenging streams is turbulence, nothing destroys flow coherence more completely than turbulent shock waves. Turbulence here will encourage the premature mixing of the fresh, in going mixture with the old exiting exhaust gas and we just don`t want that. Save all that for the combustion chamber where turbulence is essential for a rapid, efficient burn!
The actual sequencing analysis of the transfer/scavenging phase of mixture flow within the cylinder is one which to a large degree is interpreted by imagination. It is ephemeral, impalpable, and impossible for us `garden shed` tuners to get a handle on, be able to see or touch. We don`t have access to trick technologies, and investigative laboratory facilities, such as Lazer Doppler Anemometry, are not available to us. So, the best we can do is to read, listen, observe, experiment, even copy, then try to emulate in our own engines what we hope will work well. What we must always be guided by is the acknowledgement that our 3 speed, home tuned specials are just that, being based on a cheap, basic 1930s German design commuter bike for the masses. As one very well respected, international two stoke expert once advised, “Stop trying to work outside the box until you understand what`s going on inside the box”, I can`t help thinking that is a prescient piece of advice to take heed of.
In its entirety, transfer duration lasts for around a 1/3rd of one crank revolution, the briefest duration of any of the cylinder ports. Even then, the port windows are for the most part only partially open, which means that the actual flow coefficient of the transfer ducts becomes crucially important. Way back in the 90s it was just this which helped encourage me to make both outer and inner duct profiles for multiple ports when drawing up the design for this barrel. Given that the whole transfer mixture flow process is so brief and compromised it’s a wonder then that it works as well as it does! It becomes therefore, incumbent on the designer/tuner to provide for the mass movement of a mixture of air and fuel from the crankcase, into the cylinder to be as efficient as possible!  
It might be timely, or merely a refresher, that:  Air has mass, mixture even more so, flowing air doesn`t like to turn tight corners, ports and ducts need to be aerodynamically `clean`, no sharp edges, air in motion seems to have an opinion, balking at where it is intended to go, straying just where it wants to, bit like a stroppy, wilful toddler. On top of all that, we only use the small, 20 or so percent of oxygen contained in that volume of trapped air at exhaust port closure!
So, what is happening here?
Flow cycles begins with an initial velocity of zero, mixture sitting adjacent to the port window moves first with the entry mixture in the case moving last, there is therefore always a time delay to mass movement of mixture during which remember, the piston is still in motion.  Flow velocity will increase when pressure in the duct entry from the crankcase is higher than the exit into the cylinder. When the entry and exit pressures equalise, flow velocity has achieved its maximum value. From this point, entry pressure in the case will fall and cylinder pressure will still rise, slowing the in-flow down. In a perfect world, flow will stop just as the transfer ports are being closed, but what goes on inside of a two-stroke engine is anything but perfect!
It could quite reasonably be imagined that flow velocity will be at a maximum midway between transfer opening and closing; at BDC. Problem is, the internal flow regimes of a two- stroke engine all overlap each other, perfect chaos. For instance, different conditions exist at either end of a transfer duct with both acting simultaneously, the entry being open 24/7 and the exit barely ever fully open. Normal mixture transfer reaches its maximum flow velocity some 10* to 20*after BDC, at 30* ABDC it is really slowing but perhaps still flowing. So, at BDC, flow can be well short of its maximum value!
One point of reference here: Pipe lengths are fixed, rpm isn’t. As crank speeds rise and fall so does wave intensity, the position of maximum depression therefore will move from before to after BDC, bulk mixture flow will fluctuate around BDC in response to those pressure demands. The exhaust pipe can only develop a suitable depression over a fixed duration in time, but as engine revs alter when the engine accelerates through the power band the whole dynamics are continually in flux. Again, the lack of gear numbers can force certain pipe design parameters into a compromise direction to compensate for that fact. To be able to arrange for the widest depression in the pipe, the proportions that the header pipe and diffuser lengths become relatively fixed percentages of the tuned length. Superposition resonance effects, (very desirable in a Bantam scenario), factor in the tuned length and exhaust port timing. A particular effect that may be required can therefore be “tuned in” by manipulation of various percentages of lengths. The exhaust pipe for this engine errs on the side of power band width rather than sheer top end speed and seems to work rather well!
Irrespective of what dimensions are adjusted during pipe experiments one factor will invariably change, that is the total energy available, and it is finite. Making a gain somewhere will always come at the expense of somewhere else and energy is constantly being depleted. Where Bantams are concerned the relatively modest bmep produced at export opening is all you’re going to get, so use it wisely!  
Exhaust open duration down at 190* will offer greater superposition pulse enhancement, going higher will reduce the effects, going over 200* almost nullifies any advantage. Put simply, superposition is the adding of a new pulse to an existing pressure pulse, consequently, has a greater amplitude, in turn, this creates a wider depression around BDC, thereby significantly enhancing the whole scavenge cycle.
Flow from the crankcase up through the ducts and into the cylinder begins with the piston on the down stroke, causing pressure to rise in the crankcase, reed valve inlets will therefore, close sooner than race timed piston port inlets. The rate of descent, the duration of piston dwell at bdc and its rate of acceleration away from bdc are all functions of the crank stroke and con rod length. The same is true with piston motion up to, through and away from TDC.    Certain design parameters for this engine build were determined by consideration of those influences, and their most advantageous use within the confining context of Bantam Racing. It should also be constantly born in mind that we are not trying to build a grand prix engine here so component selection and state of tune should reflect that fact.
Lots of talk of velocity there, but is high velocity what we really need, particularly as the scavenging streams will be slowed down in the cylinder? In fact, what Is essential is to achieve the maximum mass movement of cool, fresh, dense mixture transferred into the cylinder. If all the mixture waiting in the transfer ducts could be transferred during a complete transfer cycle (and retained) then a huge amount of power could be realized? However, mass transfer of mixture is defined as, flow velocity x mixture density x flow cross section area. There are enough caveats contained in that short definition to confound the most optimistic engine tuner, for instance; velocity is inversely proportional to the cross- section area of the duct!
There is also the school of thought which suggests that a smaller duct cross area will initially move its inert, stored volume contents earlier and accelerate more rapidly into the cylinder than a large volume duct and overall, flow more mixture in each time span, that could be a pointer for low bmep engines like our Bantams. Try measuring the swept volume of the combined transfer ducts in your engine, compare that to the mass of mixture transferred and retained to produce, a generous, 7/8 bar bmep of the average Bantam engine. One can only conclude that what mixture is contained within the ducts is of a greater volume than is retained in the cylinder and used for combustion, even assuming an optimistic, 100% volumetric efficiency!
Ensure that your calculations include the large radius in the mouth of the crankcase which directs mixture to the barrel ducts, for they are an integral part of the transfer flow regime. When the piston moves upwards from bdc the space it vacates must be re-filled with mixture, does then, some of the duct contents reverse flow into that space with cylinder pressure higher than the case, an interesting thought as to the possible consequences of this happening?
Anywhere there is mixture movement, kinetic energy is present (kinetic energy is the energy of motion) and, in a Bantam context, kinetic energy can convert to very useful static energy. A moving fluid has three types of energy associated with it: kinetic (velocity), potential (pressure) and thermal (heat). The sum of all these energies is a constant: an increase in one can only come at the expense of the others. As local velocity decreases local pressure increases, that is a pressure recovery. If a decrease in fluid velocity is accompanied by an increased turbulent flow, there will be an increase in heat loss at the expense of pressure recovery! Don`t lose too much sleep over all of that that, but I find it is always useful to learn a bit about this stuff, and don`t forget as always, knowledge is power, so never stop learning, the more you learn the more questions you can ask. And with questions come answers!
Taking a somewhat wider view, what does all of this say for large case volumes, regarding reed valve-controlled inlet systems compared to piston-controlled ones? Both of which can be seen in the Bantam sphere of competition.
The duct entry to exit ratio has always given rise to some debate with general opinions over the years agreeing that 1.5 works well and so has become a sort of de facto working average. In their GP engines however, Aprilia had the larger B transfer port ratio of less than unity. As that engine produced close to 60 crank hp, it does challenge preconceived assumptions that bigger ducts are better, and consequently, to prioritizing flow efficiency instead.
Whilst we are on to the subject of duct entrance, never knife edge the cutaway entrance, doing so will cost power, we`ve all done it in the past, but it is counterproductive, a full round-over will smooth out otherwise turbulent flow and make power! There should be no sharp edges anywhere inside an engine, they will only impede flow and create power sapping turbulence. Of course, there is always turbulent flow in engines, any motion will have turbulent consequences, but it needs to be micro and not macroscopic, so that then becomes the responsibility and a challenge for the engine designer!

Cheers for now, Trevor