[asoftaaja]

Pocket bike madness gets worse.

I have made a Pocket Bike Dynamometer.
It's a 15 inch car tire filled with concrete, weighing about 70kg or 150lbs. It's mounted on a peugeot car wheel hub which is bolted and welded to a frame.
There is a speed sensor in the wheel hub which can be used to input a signal to my laptop and then I can use a dyno program to get a power curve.

I'm going to use it to test some common cag engine mods found on this forum and provide some cold hard numbers instead of making claims and debating.
I want to try to disprove some theories that I think are questionable and promote my own tuning advice instead...we will see what I can get out of it ha ha.

A word about using a concrete filled car tire as a dynamometer roller. It can be extremely dangerous. Since tensile strength of concrete is poor, it will disintegrate if accelerated too much. I have made some rough calculations that the critical speed for this tire would be in range of 120-150km/h. I have actually made it 2 years ago when I wanted to dyno my other moped (70cc am6 engine). But it's too powerful for this. If I want to get sufficient load for the engine, the speed will be over 100km/h and it frightened me too much so I stopped using it...but I figured I could well use it for this pocket bike since this cannot do over 50km/h (except if I change gearing). So I just want to say that if you want to make something like this, be careful.

It's basically ready for pulls, except I'm missing a cable plug that is used to connect the roller sensor to my laptop. I need to get that done and then can start experimenting...

 

[asoftaaja]

Today I had some fun time trying it. I just did a base measurement to see what it's got and then a bit of carb and ignition tuning.
Didn't take any picture but testing scene looks just like above except a laptop is there.

Blue curve is the first, a whopping 3.2 HP. I think the dyno might be on conservative side but I guess it is better to show too little than too much.
On the green curve you can see a later run but with cold engine. The difference is massive when cold, it can rev much more (but still sputtering). Just 2 pulls later it dies a little over 10000 and sputters.

I tried to drop the main jet size from 82 to as small as 76 and to my surprise it still seems not lean and there is a hefty gain on midrange, but no help for top end. Except when the engine is cold, there the main jet made a big gain. Usually I have seen that top end getting worse when hot is due to running too rich but seems it's not the case.
I'm quite surprised how it can run with such a small main jet 76 and the carb is a 19mm dellorto. Usually it requires around 90 main jet on gear moped engines. I think that this engine has a high crankcase compression ratio, so it may create a strong vacuum which pulls a lot of fuel from the carb. Goes to show how large the main jet can be and the bike still seems to run "ok", but would waste a lot of fuel. When I first put the 19mm carb there it had something like 88 main jet and that worked too.

Last I tried to advance the timing at high revs. I had a curve where under 7000rpm it was 25deg, 7-10k 20deg and at 10k drops to 10deg. I thought the 10deg drop might be the issue and put it back to 20 but no, there was no change. Actually from that I got the best run 3.45HP but that may also be due to engine temperature difference since it doesn't even really hit the switch point.

Next time I think I'm going to compare the stock reed valve if my custom one is worth anything.
Any suggestions what is causing it to lose topend power when hot? I have a few suggestions but I want to see what others think...

 

[MrKitty]

Awesome news and I'm really happy about this I hope that you can post some YouTube videos as well about how you built your dyno and what boards and sensors you use I also love to hear pocket bike engine screaming up to RPMs

 

[asoftaaja]

(deleted)

 

[asoftaaja]

I'm going to post a video where you can hear it screaming in the dyno soon. I think I'm not going to do a dyno build video because I kinda feel like I don't want to show people detailed instructions how to make something that could potentially be very dangerous if used improperly, lol. But let's just say that it has a drum brake car wheel hub that is mounted from the wishbone bushing holes onto a welded support that I made from box section pipe. The frame has to be able to carry the 150lbs tire + weight of the moped and some more.

The dyno electronics are very very simple. There is an original ABS sensor of the car in the wheel hub which has just 2 wires. The wires are connected to a voltage divider (2 resistors) which drop the voltage a bit, and then output is wired to a stereo plug which goes to the microphone jack of the computer.

There is no separate input for engine RPM. The software can calculate it from the gear ratio. Actually because a pocket bike has a centrifugal clutch, it is slipping at first so the power readings at low rpm are not correct(higher than should be), but at about 6000rpm the slipping starts to stop and the power reading becomes correct versus the rpm. And I have swapped the stock clutch springs back to make it engage at low rpm for dyno tests.

Today I spent whole day trying to debug it. First I tried to reduce the main jet from 76 to 72 as that was the next smallest. Didn't make much difference. Then I tried the next smallest that I had, 62. With that, the first pull was good and it sounded really strong, but then started to bog down so that was too small. So the main jet seems not solving the problem.
Then I started to think what if the issue is that the fuel starts foaming inside the carb because of the crazy vibration and it makes it to go lean/too rich. There was a slight evidence of this because when it gets hot and starts misfiring, there is a little bit of fuel dripping out of the overflow tubes(but not much).

Then I wasted couple of hours making a new intake manifold, which would allow to mount the carb with a rubber pipe which could hopefully reduce the vibration at the carb.

I didn't find a better diameter pipe so I was forced to make one that reduced the intake diameter from 19 to 15mm. I was not very happy with it, also the carb hits the frame which doesn't help with the vibrations. Then I tried it and a bummer, it didn't fix the problem and actually lost over 0.5hp everywhere! That was a total waste of time... With this manifold, the jet sizes were affected and 76 became way too small, it started to bog. I swapped a 82 jet and the power was way down, it was probably still lean but I didn't try to test it further. I don't believe the diameter reduction caused it because 15mm should still be plenty for this size engine, but it was probably the jetting.

Then I finally found some larger pipe and remade the manifold and now it's 19mm inside diameter.

Because the rubber piece is so short, it is still pretty stiff and probably doesn't dampen the vibration by much..so I cannot even fully verify this theory. At least there is no change. This manifold now works well and performance is same as before.

The chinese had done a terrible job of finishing the reed valve...

I went and cut excess rubber off and did a little bit of smoothing to it, but couldn't do much because on the other side the rubber was actually half detached and I was afraid that it could detach completely if I tried to port it further. Very bad quality part.

Then I had one more try on the ignition timing. This time I gave it 5 degrees more advance for high rpm so it is 25 degrees at low revs and also at powerband. Hoping to solve the sputtering problem. Well turns out that it doesn't help either. Here is some curves that I got.

Hot engine. Green is with the new intake manifold, filed reed valve and 5deg more advance. Blue is previous best with 20 degrees. Seems a little less peak power, either it is temperature difference or 25 degrees is too much at that rpm. But on low end there is a good gain.

Cold engine. Red is today's modifications, 25 degrees advance. Blue is previous best with 20 degrees.

And here is again the difference between hot and cold engine, both curves are with today's mods.
Now there is no loss on lowend torque any more but at high rpm it is still the same.
When cold it sounds really strong except right at 10000rpm there is slight stuttering but it still pulls over 11k. I'm sure it could do very close to 4hp if it didn't have this symptom.

Now I'm starting to believe the problem is the steel reeds. I think at 10000rpm the reeds start to flutter and that causes reverse flow which pulls extra fuel from the carb and then later that gets sucked in the engine, so in effect it has a rich mixture surge at that rpm. It kinda would make sense because generally too rich mixture causes worse performance when hot. But on other hand it doesn't because the behavior is the same with all main jets that I have tried and hot engine performance stays totally same. Also I cannot see any cloud of fuel coming out of the carb. I have seen that before on other engines when steel reeds cannot keep up. Strange.

 

[asoftaaja]

Here is the video about first dyno tests, english subtitles available:

Yesterday I spent again the evening tuning it. Not much point but it's a hobby ha ha.

I bought a sheet of carbon reed petal material to cut new reed petals from. But before I did that, I wanted to test a theory that by reducing the reed petal stopper distance, it could help the reeds to not flutter. The reason a reed petal flutters is because it has too much mass and moves too slowly to keep up and cannot close in time. By reducing the stopper distance the reed cannot move so far, so it has more time to return to closed position.

Here was the original stopper distance, about 7mm

I reduced it by about 2mm

This was the result. Red curve is previous best with 7mm opening and blue curves with 5mm opening. It seemed to run now rich and now the midrange got worse when it heated up.
But as you can see, from 6000rpm there is practically no difference, actually the smaller opening seems to do a slightly better job at higher rpm.
But..it does not fix the problem. Still dies at 10000 when hot.

Then I cut new petals from the carbon sheet.

I was pretty sure that the carbon petals would fix the problem but I was again wrong...


Blue and purple is with the carbon petals. On low end there seems to be a big loss compared to steels, but from 6000rpm it is better. Now it roars out 3.6hp.
Blue curve is with cold engine. There is a large dip at 10000rpm, at that rpm there is something strange going on in the engine.

So the issue is not the reed fluttering. Then the only possibility I can think of, is that the spark has not enough power to ignite at peak torque rpm. The pipe is working hard at 10000 and pushing mixture back into the cylinder, and that causes the cylinder pressure to be so high at ignition time that the spark cannot fire. A cure to that which sometimes works, is to make the spark plug gap smaller. So I changed it to pretty minimal 0.3mm. And that too didn't help, but it did give some power improvement. The best hot engine run is now 3.7hp!!

Then I swapped the stock reed valve in to do the comparison, whether my custom reed valve made with great effort is worth it.
Except it is not fully stock, there is a malossi carbon reed which I have put on there many years ago.

This is how the intake manifold looks with the stock valve. Actually it is not that bad, the carburetor is already at a good angle for the flow to go through the reeds.

And here is comparing to the best 3.7hp curve.
It ran really lean with the stock reed valve, I had to keep the choke on and it was still lean. So this is not really a fair comparison.
But the stock reed valve is pretty close. It does just as well or even better at low rpm. At high there is not nearly as much power, but I suppose if I had changed a larger main jet, it would have got really close to the minarelli reed valve.

Stock reed valve pull was on cold engine so it again revs more. When it got hot, it started to bog down so badly that I couldn't make a pull. Two strokes are pretty strange machines. After startup and bit of idling, the first pull is partly very rich due to residual mixture in crankcase and at high rpm it goes leaner. When the crankcase clears from extra fuel, it will go too lean if the main jet is too small, and that's what was happening with the stock valve. Then when it heats up it will go richer again.

In conclusion, the stock reed valve is not that bad of a restriction as I would have thought. Would be interesting to adjust the carburetor properly for the stock valve, and then try porting the valve, but it is a lot of work to fiddle with it and swap parts back and forth so I didn't want to do it now.

There seems to be this myth of cag engine sensitivity to crankcase volume, that you should not enlarge it. I think this pretty well busts it. My custom reed valve block clearly adds volume, but it still works just as well and makes more power at least in this test. In my view, you can do porting on the crankcase totally freely and don't have to worry about the volume at all. If using a tuned pipe, a large crankcase will only help it, it will provide more mixture that the pipe resonance can suck into the cylinder.

Here is comparison to when I made the first dyno pull. It was 3.26hp and now is 3.68. Actually the record is almost 3.8, that was one pull with cold engine but forgot to picture that.

Pretty good improvement so far. It felt already powerful, must feel now even more insane. Basically it is smaller main jet, carbon reeds and more ignition timing. There is probably more to be gained from the ignition timing, so far I have just made some rough adjustments just trying to improve the hot engine misfire problem.

I'm a bit frustrated because of the problem. If it cannot do more than 10000rpm it's not going to have any kind of totally insane power that I hope for. I still have one thing left to try which is to improve the ignition box. I can improve the circuitry to produce more spark voltage. Although the small spark gap did not cure it, it improved slightly and I have a feeling that it might still require increasing the voltage from the ignition box. If that doesn't fix it, then I have to admit that I do not know how to tune toolbased engines as CAM2 has suggested, lol.

 

[MrKitty]

Wow this is a lot of information... Amazing work... I'm definitely going to have to sit down on a PC and study all this information that you have brought forth on us...

 

[FutureWorld]

Amazing work! And thank you for the english subtitles very much appreciated.

Have you thought about testing stock ignition for comparison?

I have had the same thought of crankcase volumes. And just as you pointed with a tuned pipe it's even a benefit for making power with larger volume.

Keep up the good work and keep us updated 🙂

 

[MrKitty]

And I just watched a YouTube video about building a voltage doubler for ignition systems the circuit was really simple and just consisted of some capacitors and diodes unfortunately I'm unable to find it in my history.

I have also seen members convert these engines to capacitor discharge ignition using Hall sensors and other typical motorcycle parts with the benefit of being able to change the timing by moving the hall sensor Rcexl Universal type Ignition Hall sensor KIT for Engine Ignition | eBay

There is a guy on YouTube his channel is called 2 stroke stuffing he builds crazy high-reving high horse power 50cc engines in the 14K rpm range you might want to talk to him and ask him about his ignition systems

One more option would be to try to adapt a coil off of a BZM or Polini engine Chrisi ( a member here ) would know more about the possibilities on doing that I believe those engines hit 10 to 12 K stock maybe more.

These are just brainstorming ideas that I'm just throwing out there.

 

[asoftaaja]

Thanks for your comments. Seems my thread is now pinned, that's cool.

Basically my ignition box should be capable enough. Just by looking at the spark I can tell it is stronger than most stock systems on any moped I have seen. But maybe the high compression ratio I'm running combined with the pipe requires a really hard spark. Although I have now come up with some other possibility what could be wrong. Will keep posting.

 

[asoftaaja]

I didn't modify the ignition box yet, as I started to suspect that the issue might also be related to so called "squish velocity" inside the engines combustion chamber and how the worn out crank bearings might cause too high squish velocity.
Basically it means that when the piston comes up and squeezes the mixture on the very small squish gap, the mixture will be pushed out with x velocity. The smaller the squish gap is and the larger the squish area is, the higher the velocity. Then I speculated that if the velocity is too high, it might have undesirable effects on combustion. Maybe it would somehow "blow out" the spark. As my crank bearings were worn out, it would cause the squish gap to get smaller at high rpm, because the piston would pull on the bearings and reach higher. There is evidence of this, marks on the piston crown hitting the stock head back when it still had it. With the new head there was no new marks.

So I decided to change the bearings. I again went with low budget and bought some chinese bearings from hardware store and just ripped the seals out of them. Might be a mistake to use general chinese bearings here ha ha..

At least the new bearings felt good. There is a lot of play on the old ones and new ones can feel absolutely no play.

When I assembled the engine and tried to start it, I immediately noticed that the engine now turns a lot smoother and with less noise. I always wondered why the engine is so noisy and seems to have this high pitched rattle when running. I thought that it is coming from the pull start, but it was actually the bearings...

It also seemed to start and run better. I had high hopes but unfortunately the dyno said different.

The red curve is with new bearings. If anything, there is about 0.2hp loss on peak power and no change on the problem.
On green curve I got the most power with cold engine so far, 4hp. That was after I tried to clean the spark plug.

Speaking of the spark plug, that was the next idea to try. Maybe it would be that simple? The spark plug was years old and not very clean. First I tried to just clean the old plug by heating it with a gas torch burning old carbon away. There was a noticeable improvement at first(the 4hp curve), but then it seemed to reduce back to old level. Then I bought a new spark plug to confirm it. That didn't make a difference either.

Maybe too high compression ratio, or still too high squish velocity.
I took the head off and slightly enlarged the combustion chamber, making squish area smaller.

No change.

Then I started to try various adjustments. It feels like it's still a bit rich with the 76 main jet, but I'm missing a 74 jet, 72 was too small. I leaned out the needle and mixture screw if that could make it slightly leaner. Well, it only lost power as a result. Then I adjusted ignition timing again. It was 25 degrees constant curve, first I tried to increase it to 30. That didn't make much difference, there was maybe slightly more low end torque but more misfiring at high. Then I tried 22 and 20 degrees and with those, it started to act even stranger. Cold engine power slightly reduced and with hot engine, a huge power loss (about 0.5hp) at all rpms... it just didn't make much sense. I also tried a larger 78 main jet with the 22 degree timing, because it felt like the timing changed it so that it went lean when hot, but that was also a wrong assumption, it got only worse.

Then I sighed oh well, and thought that maybe I just continue with the original tuning plan improving the parts and see how it changes. Sometimes these things are so random that it's impossible to predict what change makes it better or worse so I might as well change something that shouldn't relate to the problem. So the next thing to improve was the squish clearance. It was around 1mm, so I took the cylinder off and mounted it in my lathe and machined 0.5mm off the top.

Turns out that I had measured the 1mm squish clearance in a hurry. It was actually less and now the piston was hitting the head lol.
To fix it, I slightly machined the head, but that was not bad since the squish band angle was already wrong, I could fix that at same time.
After that I measured that the squish clearance is now about 0.32-0.35mm. That might be too little. A rule of thumb for minimum squish clearance is 1.5% of stroke, and since the stroke is 32mm, this is now around 1%. But it was getting late so I left it. Might be a bad mistake, the piston may hit the head at high revs, I have to check for that.

Then, testing the new squish clearance.
Obviously the compression is now higher. Bike sounds much more snappy at around 2000-5000rpm. But again a bummer, after I dynoed it, it became clear that it is again a backwards step..

sorry about the bad image, had to take it with a gopro

The highest curve is the best overall hot engine curve from last week(3.7hp).
Pink curve is with cold engine with the 0.32mm squish gap. Barely 3hp. If I remember right, it still had 3.6hp just before the change.
Then when hot it just gets worse. This thing is annoying. It seems that I could easily spend all my spare time tuning it if I want to get great results.
Although these curves are still with the 22 degree ignition timing. I'm pretty sure that the worst drop in performance came from changing it from 25 to 22. That's when hot engine power dropped under 3hp.
There is a slight improvement in the high rpm region though. Now it continued to rev over 11000 even after couple of pulls although still with some misfiring. But it might just be that it didn't get so hot, because it was making less power overall.

I have to put the timing back to 25, and also test settings between 25 and 30. Although I'm a bit afraid that there might be something else going on. The cylinder is also pretty worn, I think that may also contribute to power loss when hot and makes it harder to tune.

Here's a couple of pics of the cylinder. The exhaust port is raised(cannot remember how much in mm), I have ground a boost port and slightly filled the transfer corners with JB weld, but that was a poor job and it really needs much more filler. Planning to redo it at some point. I also plan to do some proper porting on the transfers, but I would like to get the engine running well and fix the misfiring problem first, to really see the effects of porting.

I have to say I had much higher expectations at first about a cag engine power capability with tuning. Seems not easy at all. I'm still keen to improve it and solve the issues, but it seems to steal so much time that I may need to take a break..

 

[FutureWorld]

Interesting results.

Have you considered using e85 and some oil that mix with alcohol?
Will help with pre detonation too.
Might cool the head?

Do you still have the stock intake, reeds and carburetor to try if the problem is the reed and reed cage?

You will find the problem/problems.
🙂

 

[asoftaaja]

E85 test would definitely be interesting. It would surely cool the engine a lot and help with the issue. But it will bring more problems with tuning. I read that it is especially hard to get the low throttle settings in tune with ethanol. I think I want to get it working with normal fuel first.

I have the stock intake parts. I have already tried it (posted about it). It was badly out of tune, but it seemed like the behavior was totally same regarding the misfiring problem. But I may give it another try if I cannot figure it out otherwise. Didn't try the stock carburetor, because it leaks. I think that is the worst carburetor I have ever seen. It was always leaking and very hard to start

 

[FutureWorld]

Okay.
I've ethanol converted my moped 12-15 years ago.
I had no problem with low end tuning.
It needed a very big main yet though 😃
Smelled wonderful

 

[MrKitty]

IDK I think for experimental the pumper carburetor is the way to go. No rejeting just turn the screw

 

[FutureWorld]

Hi! I have thought about your engine and to make the engine rev more..
I think that your squish area limits the maximum rpm.
If you read about squish area, it says that low rpm engines use a big squish area to compensate/increase squish velocity.
High rev engine use less squish area to keep the squish velocity at the recommend max 15-30m/s.

But a narrower 15-45% area squish band makes less torque but more rpm.
A wider squish band 45-60% area makes more torque less rpm.

i could of course have wrong but I think it is plausible to be, if not the issue a part of the issue.

sincerely

 

[asoftaaja]

So which one you think, my squish area is too small or too big?
I just calculated it, should be 47%. I don't think it is the cause, because I already did many changes that should affect the squish velocity but there was no significant change.
I have an other moped which has a 70cc AM6 engine. It has a head with around 50% squish area and has no problem doing over 14000rpm.
Basically here are my notes regarding the cylinder head/combustion chamber that I can remember:

1. with stock head, the squish area looks very large, probably around 60-70%. Also the squish gap must have been less than 1mm, because I saw the marks on the piston as it had hit the head. I used a very thin 0.25mm base gasket, so that made the squish gap smaller than stock motor should have had (and the worn bearings also effect like I described earlier). So from that, we can conclude that squish velocity must have been really high on the stock head. It didn't rev more than 11000 and got worse when heated, but I don't remember by how much as I didn't do so much testing on the stock head.
2. with my new machined head, the squish area was first maybe 55-60% and gap around 0.9mm, I guess slightly larger than the stock head because I didn't see new marks on the piston any more. Max rpm the same.
3. enlarged the combustion chamber to bring the squish to 47%, couldn't tell any difference
4. lowered the cylinder top, squish gap to 0.3mm(increases velocity) and remachined the squish angle for more angle to match the piston better(decreases velocity). With that it sounded more powerful at low rpm, but had much less peak power for some reason. But still did over 11000 when cold. I have got it to 12000 by playing with the throttle, seems that part throttle helps it.

Also, it doesn't sound like it runs out of power, but it hits a misfire which prevents it from revving more. Basically at first it seems like it is definitely a too rich mixture issue, but I have tried even a too small main jet but it doesn't help, and also shutting the fuel valve(leans it out) doesn't make a difference either.

Currently I have 2 options which needs investigating.
First is the ignition box, maybe it doesn't produce enough voltage or there is some other problem at high rpm. I have built an improved version.
Second is the carburetor. I read that a normal venturi carburetor has a tendency for the mixture to get progressively richer as rpm goes up. I read from a finnish forum that someone had drilled an air jet hole in the carburetor larger to fix an issue like this. Maybe the cag engine sucks so hard from the carburetor, that the mixture gets way too rich at high rpm, even though the main jet is small or fuel valve is shut off. At least I'm going to try and see if it is possible to drill the hole larger

 

[FutureWorld]

I thought too big.
But you did change it to a smaller area and that didn't help. So it must be something else i think.

Try with the improved ignition and drill the holes.
And see how it goes from there.

What fuel are you using?
Octane rating?
Can it be too low octane rating and it gets too hot and missfire's occurs

I have also thought if the transfer limits the rpm, but you did have quite high transfer duration for a cag engine so it shouldn't be that either.

Edit.
If your right about the crank compression and it sucks too much fuel you would need a bigger crankcase volume?
Then people must done it wrong when they adding a full circle crank and reduce the crankcase volume?
Or at least in your case it would be that

 

[MrKitty]

When I can't get the RPMs I'm looking for I try playing around with the timing as well.

 

[asoftaaja]

I use normal 98 octane fuel with 2% motul kart ester oil (the most expensive oil I could find, over the years I have noticed that it really makes a difference to use best oil lol).
I never worry about octane rating. Because it's hard to even tune an engine to such a high level that the octane level would matter on avoiding detonation. The cag engine is far far away from such high cylinder pressures that it would detonate.

The pumper carb is a good suggestion... I might be ordering one. Is there one with about 18-20mm bore size from aliexpress or ebay?

Today I tried the carb air hole mod.
This is the hole

It feeds air to a main jet mixing tube. Theoretically, the harder the engine sucks fuel from the main jet channel, the harder it will also suck air from this hole, so I think by enlarging this hole, it should make the mixture to go leaner at higher rpm. And help with the problem, IF the issue is too rich mixture at high rpm.
At first the hole was about 0.8mm, I could fit a 0.8mm drill bit through. 1mm didn't go through and didn't have 0.9mm drill. So I drilled it first with a 1mm drill.
Tested. It seemed a little bit better, but not a huge difference. With cold engine, peak power interestingly moved higher and I got 3.6hp at almost 11000rpm, but unfortunately still when it heated up it was pretty much the same as before.
I thought maybe make it still larger and I took a 1.5mm drill bit and drilled that through...
I was almost sure that the carburetor would be ruined now, but no, it still runs and again seems like not much difference, certainly not any better.
So that theory was wrong too...

Then I again swapped different mainjets and tried different ignition timings with not great results. At least I got some power back, I could now get about 3.5hp but still down from the previous best 3.7hp. I opened the cylinder head and saw marks on the piston as it had hit the head, like I expected. Put the head on lathe and machined a little bit away from squish band to prevent it.

Then I thought that just forget the heatup issue and just try to make it rev more, as my original target is 15000rpm max revs.
How to do that? Basically should be easy, just shorten the expansion chamber, if the engine is otherwise capable...
So then I sawed the expansion chamber in half, and removed 20mm from it.

The pipe is really horrible looking, I made it when I was 15, and after that the pipe was modified several times to fit various vehicles. It has leaks and is smeared with unburnt oil, especially this pocket bike does it as it never really gets the pipe seriously hot. I welded it back together and tested...

Red curve is the previous best 3.7hp.. blue curve with chopped pipe under 2hp
Well..I guess if anything, the power is easy to get down.
And absolutely no more rpm. It means that the engine just cannot do that sort of rpm, the cases and cylinder cannot flow enough. I was pretty much expecting this, but didn't expect the power to go so badly down. Basically the pipe length was originally about 700mm, and with that, it was able to provide some charging of the cylinder at 9-10k rpm. If the rpm was not restricted to 10500, it should have had the power peak much higher and have more power. But now the pipe length is 680mm and power peak moved higher, so there is less power at lower rpm, and the engine just cannot rev enough for the pipe to start to work.
A bit sad I have to say. It just seems to get worse no matter what I do.

I think the cylinder is too worn and it has too much compression loss. It still starts well and pulling the rope there seems to be compression, but maybe it still leaks too much. Looking at the cylinder bore, the plating wear spots looked worse. I think the power losses are due to this, and from experience there was one engine which ran okay when cold, but when heated up it was really sluggish. After a new cylinder kit, it worked perfectly. I'm not sure if the misfiring can be due to that, but I'm going to replace the cylinder now. I actually ordered one about a week ago. Fortunately they are so cheap..

From these observations, I think the cag engine runs pretty much out of steam after 10-11k rpm, even with raised cylinder ports and better intake valve. When designing a tuned pipe for a cag engine, I think it should be designed for around 7500-8500 peak power rpm. A 10000rpm+ pipe will be doomed for failure as the cag engine does not have the flow capability for that sort of speed. I think a piston port engine should be much better capable of doing crazy rpm. A low rpm pipe should work much better in this reedvalved cag engine. But, I of course don't want to accept that so I will see what I can do to improve it..