ESE's works engine tuner

[wobbly]

Here are the initial results of mods to Robs GP-NSR110.
The cylinder porting I left alone,as this was capable of around 36Hp at the crank,the way Rob sat it on the case.
In the new sim the cylinder duct is changed,as is the pipe,the ignition, and the PV curve to suit the pipe.
This result is shown against the digitized dyno curve with +12.5 % to show crank power , and the sim as Rob delivered to me.

The new sim is run at .88 Combustion Efficiency , not the .92 as was needed to get the original sim to match.
This I believe is more realistic as it shows the crank power the transfers are capable of theoretically,running the new pipe at .92 shows closer to 40 Hp - I dont think so.
Later I will go thru the sim , page by page to show the general thinking behind the result.

The Mach in the header and the stinger are on the money,the TuBMax is close enough with no deto,so the combination of bmep , compression , ignition lead , and stinger size
are all working together.
Maybe by messing with the PV curve the small dip could be eradicated, but cant be arsed at this time as it doesnt show up in the power curve as an issue.
The pipe is very basic and I havnt even messed with 3 cone rear, as the current shape isnt too shabby.
At the last shot I added 30mm to the inlet,thinking that as it will be EFI the wave action wont affect the the jetting badly, this only added about 1 Hp @9000 but i left it in.

[husaberg]

Here are the initial results of mods to Robs GP-NSR110.
The cylinder porting I left alone,as this was capable of around 36Hp at the crank,the way Rob sat it on the case.
In the new sim the cylinder duct is changed,as is the pipe,the ignition, and the PV curve to suit the pipe.
This result is shown against the digitized dyno curve with +12.5 % to show crank power , and the sim as Rob delivered to me.

The new sim is run at .88 Combustion Efficiency , not the .92 as was needed to get the original sim to match.
This I believe is more realistic as it shows the crank power the transfers are capable of theoretically,running the new pipe at .92 shows closer to 40 Hp - I dont think so.
Later I will go thru the sim , page by page to show the general thinking behind the result.

The Mach in the header and the stinger are on the money,the TuBMax is close enough with no deto,so the combination of bmep , compression , ignition lead , and stinger size
are all working together.
Maybe by messing with the PV curve the small dip could be eradicated, but cant be arsed at this time as it doesnt show up in the power curve as an issue.
The pipe is very basic and I havnt even messed with 3 cone rear, as the current shape isnt too shabby.
At the last shot I added 30mm to the inlet,thinking that as it will be EFI the wave action wont affect the the jetting badly, this only added about 1 Hp @9000 but i left it in.

So what does the pipe look like?
Don't keep us in suspense....ya big tease

[F5 Dave]

It's thin at one end. Then a bit of a bulge. Then even thinner at the other end than the first bit.

[Vannik]

SwePatrick, as I said the hump in the roof is part of reducing the sudden area expansion within the duct where the Aux ports intersect.
And as far as flow is concerned the humps angle change is so gradual that the returning gas flow stays attached so it does not create any drastic turbulence or reduction in stuffing efficiency.
In any case the very quickly reducing area above the piston is so small , if you draw a tangential line from the hump to the port face , that at that stage , any inward flow is practically nil.
Proof of the above assumptions on my part is that if you grind out the roof hump in a TM cylinder power is lost , if you add a hump to say a PVP superkart ( read Aprilia copy ) cylinder , power goes up.

There are a number of issues that come into play with the exhaust port window and passage shape and its interaction with the pipe:

1. To maximize outflow during blowdown we want the flow to stay attached to the roof and currently the best solution seems to be the Frits radius and Jan's 25° roof angle.
2. Then we want to minimize the mixing of the fresh gas that escapes with the burnt gas to have the purest fresh gas pushed back into the cylinder during the plugging phase. To do this we want to minimize the interface area between the two gasses which requires a small cross sectional area which leads to the small effective duct diameter (currently about 75% of the total window area for a tripple port) and we want to minimize the turbulence caused by separated flow, we already have the roof down at 25° so the other area is to add a hump in the floor to allow the flow to attach.
3. We do not just want the plugging pulse slug of gas to be as pure as possible, we also want to have the maximum mass pugged back in and to help that we need to maximize its density. To do this we want to cool the duct walls as much as possible but only for the required distance to keep the plug of gas cold and not more as that removes unnecessary energy from the pipe and add it to the cooling system.
4. Next we want to have the best flow into the cylinder from the plugging pulse so we are looking at attached flow to the roof and floor which our previous down angle roof and ski-jump floor gives us. The issue is the sides on a tripple port layout. The way is not clear yet although Jan has stated that shortening the divider improves the power and Wobbly is busy experimenting in this area.
5. A final issue on the gasdynamic side - the boundary between two gasses of different composition or of different temperature is known as a "contact discontinuity" and when a wave travels through it it changes shape. This is quite a dramatic effect that plays a major role in how a high performance pipe works. For those of you with Blair's two stroke or four stroke book there are experimental results shown in chapter 2 that is quite enlightening. If you simulate a pipe and ignore this issue the results are very different. Any empirical formulas ignore this effect which is why they get you close but never exact. There is no magic formula. To get back to the topic, the exaust port duct that is developed according the preceding points also increases the size and definition of the major contact discontinuity at the duct and through this greatly influences the working of the pipe. Fortunately it improves it.

[SwePatrick]

Proof of the above assumptions on my part is that if you grind out the roof hump in a TM cylinder power is lost , if you add a hump to say a PVP superkart ( read Aprilia copy ) cylinder , power goes up.

When grinding out the roof, do you add the same area/volume at the floor or walls to maintain the speed?

[lodgernz]

It's thin at one end. Then a bit of a bulge. Then even thinner at the other end than the first bit.

...and this is my theory. And what it is too. Ahem ahem ahem...

[Frits Overmars]

molded air bubble or too much iron particle in the alloy that was attacked by the acid for nickasil plating ?

That cylinder never got near an acid bath. Trying to nikasil it would have been throwing money away. Instead it is now on display as modern art
(you can present anything as modern art).

It's not my picture, it's Frits

Yeah, it was my picture, but not my cylinder. It was the first attempt of a couple of mates to cast a scaled-down RSA cylinder.
They learned a lot (like listening to me ). Flettner nailed all their mistakes (no surprise there).
On the other hand, I always say: 'Everybody's got to make their own mistakes'.

[Frits Overmars]

To maximize outflow during blowdown we want the flow to stay attached to the roof and currently the best solution seems to be the Frits radius and Jan's 25° roof angle.

Honesty compels me to admit that the 1968 Bultaco engine that I used as a basis for my 500-class race bike, originally had a generous exhaust top edge radius, which I enthusiastically removed, together with the inner curves of its transfer ducts (using a hammer and chisel, my special tools in those days). It's a miracle the result still produced about 55 hp from 372 cc. The conversion to rotary inlet, a home-built 40 mm carburetter instead of the 32 mm original, and a fat pipe may have helped a bit.
Boy, did I have a lot to learn...

[jbiplane]

We all learn...Attachment 336345

Hi Fritz. We finally make (paramotor) engines with opposite exhaust, would intensively test.



The first impression works very well, though we have not correct assumptions how make best possible geometry.

[tdc211]

Even with water close to plug and close to sq.
And msv up at 40
I am still at. .88 to .89 combustion efficency in sim. For 39sim power. And even most small engines I do, end up around this combustion efficency

What's the trick for better combustion efficiency?
I have tried many times to improve on .89.
I can't.

I am guessing Neels post has alot to do wirh that

[wobbly]

I wasnt planning on showing the pipe dimensions till after explaining all the sim details,and the reality is that I chose the length,messed with the header
to get the % correct,and then went bigger on the stinger as the Mach was too high.
End of story.
I have been down the path of over 200 sim runs to get a pipe " right " but I was being paid to do that exercise.
Anyway here is the first guess result for those of you that cant wait.
The cylinder duct is extended 20mm,and this could ( should ) be cooled,then there is the 30mm slip on oval to round transition, as the first pipe element.
The 32mm exit size is a little bigger than the 75% rule, but the 36mm header is equal to the port effective area.
I would change the stinger tube to 1 1/8" by 1.2 wall so the ID would be 26.2 and make the nozzle 10mm long with only a 5mm expansion angle out to the stinger ID.

[wobbly]

Regarding the TM duct roof hump - sadly the rules allow removal of material, not adding any so I have not been allowed to make changes
like lifting the floor ski jump.
But my opinion would be that quite possibly a 25* straight roof would work better together with the floor lifted above BDC a few mm as Jan did
on his cylinder.
This would then make the ski jump higher on the floor and reduce the volume.
But remember that lifting the floor only works if full attention has been paid to matching the blowdown to the transfer STA numbers.
In a full out design, this then involves cutting the Aux around to bore center,and this leads onto adding small end plugs,nothing is simple at this level.

[jamathi]

There are a number of issues that come into play with the exhaust port window and passage shape and its interaction with the pipe:

1. To maximize outflow during blowdown we want the flow to stay attached to the roof and currently the best solution seems to be the Frits radius and Jan's 25° roof angle.
2. Then we want to minimize the mixing of the fresh gas that escapes with the burnt gas to have the purest fresh gas pushed back into the cylinder during the plugging phase. To do this we want to minimize the interface area between the two gasses which requires a small cross sectional area which leads to the small effective duct diameter (currently about 75% of the total window area for a tripple port) and we want to minimize the turbulence caused by separated flow, we already have the roof down at 25° so the other area is to add a hump in the floor to allow the flow to attach.
3. We do not just want the plugging pulse slug of gas to be as pure as possible, we also want to have the maximum mass pugged back in and to help that we need to maximize its density. To do this we want to cool the duct walls as much as possible but only for the required distance to keep the plug of gas cold and not more as that removes unnecessary energy from the pipe and add it to the cooling system.
4. Next we want to have the best flow into the cylinder from the plugging pulse so we are looking at attached flow to the roof and floor which our previous down angle roof and ski-jump floor gives us. The issue is the sides on a tripple port layout. The way is not clear yet although Jan has stated that shortening the divider improves the power and Wobbly is busy experimenting in this area.
5. A final issue on the gasdynamic side - the boundary between two gasses of different composition or of different temperature is known as a "contact discontinuity" and when a wave travels through it it changes shape. This is quite a dramatic effect that plays a major role in how a high performance pipe works. For those of you with Blair's two stroke or four stroke book there are experimental results shown in chapter 2 that is quite enlightening. If you simulate a pipe and ignore this issue the results are very different. Any empirical formulas ignore this effect which is why they get you close but never exact. There is no magic formula. To get back to the topic, the exaust port duct that is developed according the preceding points also increases the size and definition of the major contact discontinuity at the duct and through this greatly influences the working of the pipe. Fortunately it improves it.

The roof angle of 25° (?)(I never measured it!) was completely STD ROTAX, I never changed it.
What I did first was enlarging the auxiliary ports, with good results.
But they were too high already.
When I could cast my own cylinders I made them lower and wider
And of course the 'ski jump' was introduced immediately after good results with other engines.
I saw that when a cylinder seized the divider between main and auxiliary frequently cracked.
So I tried to make an 'indestructible' cylinder, by making this divider bigger and longer.
With a nice round corner around it.
I lengthened the divider until almost the exhaust flange.
As I had seen on a Krauser 80cc cylinder.
The result was a very good central exhaust duct flow.
And also a really good separately measured auxiliary duct flow.
But the total exhaust flow diminished, and the engine did not rev.
By gradually shortening the divider wall I got over-rev and power back.
Until the best result was reached by coming back to original ROTAX dimensions.
I was very displeased, away went my dream of an 'indestructible' cylinder......
A pity, but I had to live with it.
It took years of work to reach the best auxiliary duct and ports dimensions.
The main port was made a little bit 'squarer' and narrower around 2001, and the auxiliaries widened to the inside.
The exhaust main port was unradiused until I tried a new exhaust pipe which gave very good low and midrange power,
but wouldn't rev.
Making a radius made it rev, but we lost most of the gained mid- and low range.
We decided to leave it like that....as it was good for the piston ring.
Later we started to raise the exhaust port floor, it looked quite good, but then I retired.
The goal was to reach a point at which the A-ports could have been made still wider.
I don't know if this was ever tried later.
I heard about a 'Honda' type exhaust duct being tried, it lost 1,5HP.....

[Muhr]

Now the first prototype is being manufactured! It has not been easy to solve the increase in cooling of inside transfer. The prototype produced now has a 60% chance of success(survive the casting). Just keep my fingers crossed now …

images from the caster!

[Tim Ey]

images from the caster!

Fingers crossed!
If I were you I would also chose to 3D printing in alloy - since 40% of failure would not satisfy me...