ESE's works engine tuner

[Jannem]

In the present case, 72cc the Derby, which is not mine, was already tested on the dyno with 31.9 hp at 13.900 rpm
Photo's of the graph are visible on Racehelden forum https://www.deraceheldenvanweleer.nl...p?topic=2007.0
Beside it was already written here by Yannik, but most people apparently do not want to see this.

Please. It's not a matter of not wanting to see it. The link you provide requires a login and gives zero information unless you have the login credentials. This has been already mentioned here a few posts back, and it doesn't help your cause to post the same link again with that kind of comment. Just pisses off everyone who's left still actually wanting see information from you.

[katinas]

Have you made any changes to get similar in exhaust duration after you have finished the radius?

No, this last important adjustments not made nor SX65, nor SX85, so exhaust/transfers duration with radiused piston slightly higher than with std piston during all tests.
After looking at all graphs, I really keen to see how engine react to this adjustment and maybe later, he will try to do this with SX65.

[Muhr]

No, this last important adjustments not made nor SX65, nor SX85, so exhaust/transfers duration with radiused piston slightly higher than with std piston during all tests.
After looking at all graphs, I really keen to see how engine react to this adjustment and maybe later, he will try to do this with SX65.

It would be very interesting, if we take 1mm it could be something like 4 ° - 5 ° exaust which would correspond to the difference you have in rpm peak hp.
Geat work

[wobbly]

I can suggest a way better option to doing MatLab calculations, especially when those results are only logged at a tiny number of discrete points.
There is a app from NT- project called Virtual Dyno.

http://www.ntproject.com/software_vd1eng.htm

This uses all the inputs from the several types of dashboards available that use GPS data to generate a track map.
The data ( from up to 18 satellites concurrently ) is then integrated with the speed/time/mass/inertia info, and produces a Hp curve of the engine thru a whole lap of the track.
My inertia dyno of the best KZ2 engine showed a peak of 48.8Hp at 13800.
The Virtual dyno showed 48.5 Hp at 13600 with an identical curve shape.
My take on the difference is the pipe cooling from 132Km/Hr as every other variable is identical such as egt,water temp,and acceleration rate.
If it wasnt for the hassle/cost of doing track testing, I would ditch the dyno altogether.
As the Virtual Dyno uses a huge number of data points that are averaged to produce a real world representation of the engines performance.

[lucf]

I can suggest a way better option to doing MatLab calculations, especially when those results are only logged at a tiny number of discrete points.
There is a app from NT- project called Virtual Dyno.

http://www.ntproject.com/software_vd1eng.htm

This uses all the inputs from the several types of dashboards available that use GPS data to generate a track map.
The data ( from up to 18 satellites concurrently ) is then integrated with the speed/time/mass/inertia info, and produces a Hp curve of the engine thru a whole lap of the track.
My inertia dyno of the best KZ2 engine showed a peak of 48.8Hp at 13800.
The Virtual dyno showed 48.5 Hp at 13600 with an identical curve shape.
My take on the difference is the pipe cooling from 132Km/Hr as every other variable is identical such as egt,water temp,and acceleration rate.
If it wasnt for the hassle/cost of doing track testing, I would ditch the dyno altogether.
As the Virtual Dyno uses a huge number of data points that are averaged to produce a real world representation of the engines performance.

Thanks for the links Wobb, This will be the new way to measure the power of any vehicle in the future.
And it is done on the road which is much more usefull that any other way.
Your KZ engine has less power on the track than on the dyno, and what do you want to know on the dyno or on the track?

New GPS action camera's from GoPro or Garmin will do the same. With a drone above the circuit you could analyse the power of a F1 car.
The weight is know, the track is known, the weather circumstances are known. So you only have the analyse the G-forces and you can calculate the horsepower.
And if the downforce is known still more accurate.

[jonny quest]

Not trying to make light of your accomplishments LucF.
It's easier for a smaller engine to make more HP per cubic inch than a bigger one. So your percentage is probably off in relation to the RSA.
Smaller bore, better flame travel, less internal friction...

[lucf]

Please. It's not a matter of not wanting to see it. The link you provide requires a login and gives zero information unless you have the login credentials. This has been already mentioned here a few posts back, and it doesn't help your cause to post the same link again with that kind of comment. Just pisses off everyone who's left still actually wanting see information from you.

I never force anyone to have a look.
And there are enough people here who has a login.
And if you don't want to have one, piss of yourself with your useless reply.

[Carel H]

.
Difference = Maximum - Minimum;
if (Difference > 0)
{MAPout = Difference;} // MAPout becomes an indication of air flow through the motor.


}while(IgnPulse != true); // keep looping until there is an ignition pulse.


analogWrite(17, MAPout); // pressure difference as a MAP value sent to the EFI CPU.
int Maximum = 0; // Reset the Maximum Variable so as to find the next crankcase maximum pressure.
int Minimum = 1023; // Reset the Minimum Variable so as to find the next crankcase minimum pressure.
(IgnPulse = false); // reset IgnPulse for the next cycle.
}

With analog sensors there will very often be a numerical difference, so you will be constantly setting Mapout. So there is no edge detection. Edge detection meaning when you want to go from scenario 1 to scenario 2.

Solution 1: make difference larger because of the noise.

Solution 2: use a moving average filter (MAF) to suppress the noise, giving a lower difference (which will be still larger than 0).

And then there is the threshold, which may also be a variable: above this point everything is OK, below this point not. With this, you can have an Arduino based on/off detector, not mapped.

[lucf]

Not trying to make light of your accomplishments LucF.
It's easier for a smaller engine to make more HP per cubic inch than a bigger one. So your percentage is probably off in relation to the RSA.
Smaller bore, better flame travel, less internal friction...

This is corrected in the graphic I showed many years ago for the first time.

https://www.deraceheldenvanweleer.nl...ch=59295;image
https://www.facebook.com/photo.php?f...type=3&theater

[wobbly]

What I am going to do is datalog the pipe surface temp ( in the header, the mid and in front of the stinger entry ) at the track, as I have done
on the dyno for Neels.
This was done as we wanted to change the EngMod code to enable these real world temps to be input.
Once I know what the real temps are at the track, then I will go back to the dyno with sufficient air blowing over the pipe to simulate the track result.
I can easily do 50 dyno runs in a day, with a change to some variable for every test.
Every result is started and ended with the same water temp with auto cooling control,and egt temps, by changing the jetting.
This is much much easier to do on an inertia engine dyno than trying to get all the other variables lined up, such as tyre temp etc on track.
Its a nightmare in time and money to get sufficient data on track, then have to ignore that run just to change a jet.
But if the time and money constraint wasnt an issue, then I would be at the track every day using just the Virtual dyno app.

[TZ350]

With analog sensors there will very often be a numerical difference, so you will be constantly setting Mapout. So there is no edge detection. Edge detection meaning when you want to go from scenario 1 to scenario 2.

Solution 1: make difference larger because of the noise.

Solution 2: use a moving average filter (MAF) to suppress the noise, giving a lower difference (which will be still larger than 0).

And then there is the threshold, which may also be a variable: above this point everything is OK, below this point not. With this, you can have an Arduino based on/off detector, not mapped.

Thanks for your input, I am enjoying learning a bit about signal processing.

On each cycle (engine revolution) Max cant decrees and Min cant increase, they just go to their respective extremes within each cycle. Then after a trigger pulse is seen they are reset to their respective starting points.

I like your idea of a minimum difference or a moving average, otherwise the difference is the difference between Max & Min with noise.

Mapout is written to the EFI CPU only once a cycle as the cycle is completed. Maybe I could apply a mean moving average to MAPout outside the loop over a number of cycles. Or a moving average to Max and Min within the loop. I will run it in its simple form first with a scope, to see what needs to be done. .... thanks for the ideas.

[Jannem]

Nothing pro, but useful entry level tool to make back to back power comparisons on wheels to try things out:

http://atom007.heimat.eu/tmt/gsf_dyno.html

Does have temp and baro corrections though. You can just record the exhaust sound with a cell phone during the run. Beats the guesswork every time. Coupling that with deto and ex.temp sensor is already a lot for a beginner.

[lucf]

This is what Jan wrote when I published the graphic for the first time.

Jan Thiel, tuner of the fastest twostroke ever exist in GP racing till 2007, the Aprilia RSA 125cc, wrote next.

Quote from : Jan Thiel on October 18, 2012 , 06:17:55
Beautiful graphics Luc ! That 'red ' abilities are 100 % definitely achievable. It will , however, most of the time, and especially money, are missing.
To meet the RSA power was used 5 days per week by 6 people and with it the aid of the drawing room, and electronic engineers. And that over about 12 years is not that simple ! We started in 1995 with 46.5 hp, and ended in 2007 with 54hp, so 0.6 to 0.7 HP per year on average. Actually seems very little , but it was a lot of work ! We made about 300-400 cylinders per year. So to get that far, there are at least 3600 cylinders made. But maybe 4800 !, I did not count them .... Of course, a lot of cylinders were equal to each other ! But we surely tested 30 different types of flushing channels, and also exhaust conduit 20 different types. Estimated that another 100 different cylinder heads and of course lots of exhaustpipes!
Overall it was pretty special, and will not arise so quickly again!

Quote from Jan Thiel November 19, 2015
Calculations with a specific program are certainly very useful!
I wish I had something at Aprilia, then I had definitely become further!

https://www.facebook.com/luc.foekema...5454134&type=3
When pictures could be load up directly, I would do so.

[koenich]

thought Jan has actually no idea?

Pity that you keep coming back attempting to collect claps for nothing...

[Michael Moore]

It can be more convenient to blow an engine up on the dyno in the shop where you've got all the tools/parts available to fix it vs going to the time/effort of dragging everything out to the track to blow it up.



cheers,
Michael