I have continued troubleshooting the bike.
After changing crank bearings, drilling the carburetor air jet hole and shortening the exhaust were all failures, next suspect was leaky crankcase seals.
I read from a twostroke guru forum that a case leak actually makes a motor run rich, not lean. It is pretty hard to explain and I cannot explain it fully but one thing to consider is that if there is a case leak, it will also blow out air(and fuel) the same way as it will suck air in, so it cannot be directly said that a case leak makes it to go lean. But if it really goes rich, it could explain the sputtering and spark plug always being dark/black.
So I went out and changed new crank seals...
After that, it again seemed to start and run noticeably better. I had to disassemble the dynamometer temporarily so I had to test by driving. Seemed to have a lot of power, no signs of the <2hp that the dyno showed after I shortened the exhaust. Strange.
Then after just a short drive it started to bog down. It was clearly lean now. At first I thought that the carb must be running out of fuel, so then I went to drill out the carburetor needle valve...
Basically I drilled two side holes on the needle valve channel, so the fuel has a shorter path around the needle. This improved fuel flow from 170cc/min to 200cc/min. But didn't fix the problem.
Then I realized that the main jet must be too small now. Either it is because the crank seal leaked and really made the motor to run rich and the 76 jet was ok, or then the carb air jet hole that I drilled on previous page made it go lean and I didn't realize it when running on the dyno. This would also explain the power losses, so the shorter exhaust didn't really cause a power loss. It just didn't do the bogging on dyno, so I didn't see that it was lean.
So..then changing a larger main jet. Now I put there a 80, too lean, 85 too lean, then 92, much better and then 90 was the best.
Then I also made a new exhaust mounting flange that uses springs to hold the exhaust, so the exhaust got about 20mm length back. At this point, the bike now had the previous power level back, and I could get 3.8hp on cold and 3.6hp on hot engine, but still the problem that it loses about 1000rpm of revs when hot.
Then there was the ignition box. I made a new CDI with about 20% more spark power and thoroughly tested it in test bench that it should be able to provide full power to around 15000rpm. Didn't help. Still one thing to check with the ignition is the actual timing. I guess it's possible that the trigger coil goes haywire at 10k, and throws the timing off.
Ok. I again scratched my head and tried to come up with new theories. I again came up with a new idea which seemed that it could be likely. Maybe the problem is that the exhaust is way too short, and actually does not even get to powerband, but it only gives a power dip at 10000rpm. The "power dip" is a symptom that occurs in all two stroke tuned pipes, that at 2/3 of peak power rpm, there is a torque dip, due to exhaust pressure wave returning to cylinder at a wrong time and pushing mixture back to crankcase. So, if the pipe length would be suited to around 14000rpm, then there should be a power dip starting at around 9500rpm. If you look at some of my power curves, with the cold engine curve there is always a dip at 10500rpm and then it gains power back and continues to rev over 11000. Maybe when the engine is hot, this dip is too much for it and it cannot rev past it.
Now there was just a question of is the pipe really that short for this to match? I have usually seen that a pipe length of around 750mm will produce a power peak at around 12000rpm. The pipe on this pocket bike was 725mm in length, that would mean 12500-13000rpm. So it doesn't quite match... but I figured I still give it a try.
So...then I took some old car exhaust pipe, chopped the exhaust in half and welded a generous 100mm elongation to it.
Now the pipe extends way out..looks pretty funny. I thought that if it works, I will remake the pipe with bends to fit it better inside the bike.
It was again very exciting. If the calculations match, this pipe having now a tuned length of around 820mm, should produce a power peak at 9000rpm. So the engine should get a crazy boost.
Again it turned out that this engine just doesn't behave the way I expect. The red curve is with the short 720mm pipe and blue curve elongated 820mm pipe (graph description says 800mm).
Almost identical peak power and practically no change in peak power RPM! I was completely amazed at this..I was sure that there should be a significant difference at least on peak rpm...
Then the green curve is cold engine power with the short pipe. Cold engine power with long pipe is less and it cannot rev over 11000. At least now the heatup problem is "fixed", because now it doesn't rev as much when cold. So clearly the exhaust plays some part in it, I just don't get it right. It seems like the exhaust pipe is not doing much at all. I might as well make a test with no pipe at all and see how the curve looks like.
Not going well. I was hoping that it could be doing well over 4hp at this point. But at least I have learned something.
I still believe there may be something wrong in it which causes problems at higher revs, because it still sounds it is misfiring. Next I'm going to check the ignition timing with a strobe light.
Then, the next tuning modification. I have read a lot about porting the cases on this forum. I gave that a go...
Basically I rounded out the edge on the cases, ground out a "channel" around the bearing and then cut the cylinder skirts, smoothing the path to transfer ducts.
To my eye, there is not much that is improved by grinding. There is not really so much flow obstructions. Maybe the exhaust side cylinder skirt corners were a bit on the way.
Then I cleaned it up and put back together. Then a test..this was again very exciting. According to a member on this forum, the case flow porting will gain hugely. Maybe I could get over 4hp now.
The result, hot engine (blue: stock cases, 820mm exh.; purple: ported cases, 820mm exh.; red: stock cases, old shorter pipe) :
Ha! No difference! This was again quite interesting, as the peak power is identical which is remarkable because usually even the slightest change may show up on the dyno...
There seems to be some gain on the lower end, and I think it might have sounded a bit better on takeoff, but in conclusion it is really not worth much to grind the cases...
Then another point, the crankcase volume.. some guys seem to worry about the effect of diminishing crankcase volume. Well there is some reference for you.
Next plans to get to over 4hp and over 11000rpm power band...
Unless there is some problem with the ignition timing, it seems like the engine is not going to give any serious power beyond 10000rpm.
I believe the back reed valve is a bad restriction, with the turning crank webs acting like a rotary valve blocking the inlet flow.
I have designed a new reed block:
By grinding out the crank halve locating pin boss, there is just enough space (I think) to mount a reed valve block on with a circular flange matched to the crankcase shape. It will require a lot of grinding and sanding to seal it, but I think it can be done. I 3D printed a first model, but it is too large, have to shrink it further.
Then some failures.. On a test ride, the freewheeling clutch broke and it didn't engage any more. I fixed it by welding it solid on the wheel hub.
Then pretty soon after that the clutch bell holding nut snapped off (just as I did the test for the case mod) and the engine locked up. I thought that it seized and was pretty annoyed, but fortunately it didn't. I tried to fix it by welding the bell back on, but I didn't screw it back on fully before welding so it is now binding on the clutch and have to grind the weld out and redo it...
Pocket bike tuning continues...