When embarking upon any aspect of engine development some of the fundamental, overarching questions must be, what are we asking and expect, can it even be fully accomplished. In this case it is the exhaust port and duct that are the subject for analysis. In turn, that further begs the question, what is actually happening within the duct, do we have smooth efficient turbulent free flow or chaos, do conditions change significantly along the length of the duct, and can its geometry be modified to advantage? Some might see this as a simple question to answer, just a big hole and passage to let gas out, but that conclusion would be to underestimate the true nature of exhaust dynamics as investigated then applied subsequently to two stroke, therefor Bantam race engines, from the latter part of the 20th century up to today.
Clearly the overriding consideration is the need to get rid of spent combustion gas as rapidly and completely as possible during cylinder blowdown. Exhaust gas is very hot and contains no oxygen so as such is useless and needs to depart the premises rapidly before the next cycle begins. There are two major influencing events within the cylinder on the behavior of that gas. One is the fresh, cool mixture column entering at the rear of the cylinder flowing from the transfer ports, progressing upwards then curling over at the cylinder head helping to push the old gas out. Secondly, wave functions from exhaust pipe also act in drawing the spent mixture out. The balance here is to prevent the fresh, transferred mixture from combining with, and thus contaminated by the old muck, sadly however, some mixing is inevitable. Next, is to prevent fresh combustible mixture from being over scavenged by the pipe`s diffuser and drawn too far into the exhaust header pipe to be lost before the exhaust port closes.
The exhaust duct outlet, where it meets the exhaust header pipe only requires an exit area commensurate with the maximum quantity of spent gas the engine produces during each firing cycle at high rpm, and, within the flow potential of the entire duct. That quantity is in turn proportional to the amount of power produced at peak torque rpm. It might also be suggested that the outlet needs only to reflect the degree of flow during blowdown, anything extra could be excessive, and at lower rpm the blowdown STA will be too much anyway…….. Should have persevered with that proposed exhaust valve installation, something I increasingly came to regret! It is crucial therefore that the exhaust duct volume is not made any larger than that which gets the job done efficiently, bigger is certainly not better! For a Bantam that volume needs to be no more than around 75% of the cylinder volume, for a 125 that’s 125x.75= 94cc, the larger engine, 186x.75= 140cc, that 46cc difference is quite significant in respective volumes. This is particularly relevant when one considers that the average larger capacity Bantam engine has about the same power as a good 125 so the proportionally larger duct will not offer any increased benefit, quite the reverse in all probability?
Irrespective of which type of exhaust port configuration is adopted there are fundamental requirements common to them all. We need to absolutely minimize the mixing of fresh gas, that escapes during scavenging, with the burnt gas to enable the purest, coolest fresh mixture to be pushed back into the cylinder during the pipe`s plugging phase. Not only do we want pure mixture but also the greatest possible mass of that mixture, so to help with this we need to maximize its density, big volumes do not encourage dense mixture. Density is mass divided by volume. The cylinder has a fixed volume so increasing charge density will also increase the mass of mixture in the cylinder, thereby offering the potential for making more power. The easiest way to negate mixture density is to allow it to heat up and expand, by allowing this to happen both pressure wave and gas particle velocity are reduced. Imagine dumping that mixture into a dust bin size pipe; you wouldn`t! Working backwards then to relate that to the exhaust duct and header volume, thus geometry, gives some idea as to the negative influence of excessive volumes.
It is situations such as these which emphasize the crucial interrelationship of all of the engine`s functions. Starting from the atmosphere at the carb bell mouth profile through to the tailpipe/silencer exit to atmosphere, all events must be synergistic, where the combined effects are greater than the sum of individual parts. It is always much easier to lose power than to make it.
At this point in proceedings I had machined the exhaust port profile into the liner and the corresponding cylinder wall, so all was set to try and create the short profiled connecting duct from the 78mm diameter of the barrel stub fins to the inside of the water jacket.
The area of the profiled port including the bridge and radii, as near as I could measure and calculate was 988 sq.mm, equivalent to a diameter of 35.5mm. The blowdown area was 590 sq.mm giving a diameter of 27.5mm, providing for a useful 95sq.mm (19%) improvement over the single port example. I can`t say with any certainty that these numbers are precise to the very last square mm; the blending of 4 different size radii on each side of the bridge was tricky to say the least. But the potential blowdown advantage over a single port is quite clear to see, and I was content with that. I was also mindful of the fact that the bigger the exhaust port the greater the potential for serious charge loss at lower rpm thus reducing power band width and power numbers. I say it time and again, the handicap of three gears has to be of considerable influence when designing ports and pipes.
It might be useful to point out and provide some context, that the Aprilia GP engine does not even begin to open its exhaust power valve until 10,000 rpm so the initial timing of the main port is more akin to a roadster engine; food for thought?
I was only too well aware of the obstruction to flow of the port bridge and its corner radii and didn’t want the additional interference of an immediate, steeply angled duct roof to further disrupt that first, high velocity exit of gas right at the very beginning of blowdown. I felt that by cutting the liner and barrel square to the cylinder axis for their combined thickness of just 12mm, then blending with the downward sloping duct roof, could on balance assist gas flow and give rise to a greater exhaust pulse amplitude by the rapidly opening port. It is this pulse which travels through the mass of exhaust gas at the local speed of sound and which forms the basis for pipe length calculations. The sides of the liner port and barrel were cut parallel to the bridge/cylinder centre line offering little initial resistance to mixture flow. There was no practical way I could quantify the hoped for benefits accruing from this line of thinking but as the engine achieved very good results from its very first race, I have to place some credence to it. Just about the only potential drawback I can visualize is the returning plugging mixture might be aided by an upward sloping roof to guide mixture up, over the rapidly moving piston crown closing the port, then into the cylinder. The reverse is the case at exhaust port first opening with the combustion gas above the piston crown and the plugging mixture below the crown. Swings and roundabouts I guess, however with hindsight, I might now opt to incline the port roof few degrees?
Taking an over view of the exhaust operation from port to pipe, most of the short comings that can be seen in the original BSA design have been addressed and those areas which could be improved have had modifications implemented.
One of the last modifications to be made was to re-locate the water inlet low down and adjacent to the exhaust duct. Here the coolest and fastest pumped water could flow under and around the duct and up over the front of the barrel. The hope was that the plugging mixture waiting in the duct and header can be kept cool and dense and so augment the following combustion event. But before that, the duct has to be made!
I started off with an overlong piece of mild steel tube, which I machined from solid bar, of initial 34mm bore tapering down a couple of mill and 3mm wall thickness. The tube was squeezed in a vice, with more emphasis to larger port window end, to roughly resemble the desired shape but was not wide enough to accommodate the port width. The sides adjacent to the port edges were built up with weld and extended around 20mm or so. The squashing process forced the underside to bell outwards, deforming metal has to go somewhere. The belling was ground out and the hole filled with weld and dressed back, inside and out. From making the drawings I had an angle that the sides tapered inwards to then create the oval shape at the water jacket inside wall. I wanted to mill these angles in but this irregular shape could not be held securely so I opted to weld it to a base plate and clamp this in a machine vice and canted over at the appropriate angle. This process of welding, milling, using various diameter cutters, re-welding, machining, grinding process carried on until I had an acceptable duct matched to the port profile at the 78mm diameter and which more or less maintained the 3mm wall thickness. The two duct ends have their respective radius roughed in with an angle grinder then finished with high speed air tools. The downwards angle to the duct was taken care of when roughing in the end radius to match the 78mm barrel diameter and the 117mm diameter of the jacket inner diameter. When I was satisfied with the blending of the duct to the barrel port profile it was brazed into place, at the same time the bridge was brazed into position with a mill or so projecting into the bore to be then machined back at a final stage. After cooling the over length duct was turned down to the same diameter as the barrel lower flange so the water jacket could still be slipped on and off. It was then a simple matter to open out the water jacket wall to the inner profile of the duct, plus a weld prep for final brazing later in the project. The final piece of the duct length, namely the external 50mm long exhaust stub/pipe joint, was a tube taper bored from the height of the oval duct to just under the exit diameter then hand ground to form the oval section at the water jacket to smoothly blend towards the final exit diameter. Overall the duct centre line came out at around 90mm from piston skirt to pipe inlet.
The barrel reconstruction at this stage was only lacking some sort of inlet arrangement, part 3 will describe my approach to that part of design.
Enjoy the Bank Holiday, no such thing as too many holidays!
Cheers Trevor
Home 
Sun Apr 30, 2023 8:55 am