MoTeC Global Forum

I would like to be able to log cylinder pressure through I2 however as of now I have an SDL, M800, and E888 the maximum logging rate of the bunch is 500hz. My engine can turn 10000rpms which makes this a fairly useless rate to log cylinder pressure. I see that the ACL can log at 5000hz which would start to get into an area that could potentially be useful but still a little slower than I would think ideal (not to mentions I would have to buy an ACL for one sensor). Is there an easy work around I am not thinking of? Surely someone has done this before?

Ben-S wrote:I would like to be able to log cylinder pressure through I2 however as of now I have an SDL, M800, and E888 the maximum logging rate of the bunch is 500hz. My engine can turn 10000rpms which makes this a fairly useless rate to log cylinder pressure. I see that the ACL can log at 5000hz which would start to get into an area that could potentially be useful but still a little slower than I would think ideal (not to mentions I would have to buy an ACL for one sensor). Is there an easy work around I am not thinking of? Surely someone has done this before?

I only have ideas as I have not done this - but it is interesting.

12000RPM equates to 200Hz.
Of course you would want to measure it at the right time in the cycle - which becomes problematic.

I see two options:
1) An integrator circuit to hold the peak long enough to measure it.
2) A "sample and hold" circuit.
If you need the pressure before the firing or after, then perhaps you could have the sample and hold trigger off of a spark or f/i trigger???

Would take some work, but seems easy/doable as a Gedankin experiment.

I understand your request, but in reality, there isnt going to be an easy solution for you.

Yep, the fastest that we can log is 5000hz, that is with an ACL and a VIM, as the VIM has the inputs capable of reading at that speed.

At 10,000 RPM, you have 166 Rev Per Sec, so at 5000hz, you would get 30 samples per crankshaft revolution, so only 1 sample per 12 deg of crank rotation, so it would be only just ok I suspect.

of course, at 5000 rpm, that comes back up to 6 degrees of crank rotation, so if you are happy to measure at a lower RPM, it would be reasonable.

In reality, to measure every degree at 10,000rpm, you are going to need to log at 60,000 hz, and that is currently only in the realm of the very expensive industrial data logging equipment, and not something that we do.

Regards

-Jamie

Ben,

What sort of cylinder pressure measurement device do you have?
Do you need to log the data in a datalogger, or just record pressure while on a dyno?

The reason I ask is that there are some sound cards that record at 196kHz/24bit.
You can get them with 2,4,6,(and maybe 8?) input channels. I have a 6-input channel one.

I haven't bought one yet (want to make sure I can record it first) but the ones I have been looking at output a 05v signal. It looks like some sound cards will do 10V? I have to admit I'm not very familiar with any type of stereo equipment.

I work in OEM engine developement, where in-cylinder pressure measurements are routinely done during the engine mapping/calibration process (ie: setting ignition timing, cam timing, etc.). The sensors to do the job PROPERLY are made by AVL or KISTLER (type 6023 or 6125A are the best), and must be fitted in a very specific location and manner in the chamber to avoid resonance and aliased readings. The sensors start at $3050 each, and are fragile enough to be very easily damged by excessive detonation (6125A better). Also, they can be bought or adapted for use as spark plug mounted sensors, where a grrove or channel is cut down the side of the sparkplug, and a thin stainless tube brazed in the groove, terminating above the sparkplug hex with a boss to screw in the Kistler sensor. As you can imagine, on a modern DOHC head with deep spark plug wells, this is a difficult proposition...and mounted in this way, there is a fair loss of resolution and accuracy due to the addtional volume of the tube, and the distance from the chamber.

These type of sesnors are all charge-based sesnors....output is in coulombs, and therefore they need a charge amp for each sensor, this then outputs a voltage into a VERY high speed DSP (dig sig processor)...the encoder is fitted onto the front of the crank, ususally with 0.1 or 1deg resolution, to provide index for mechanical TDC. The DSP samples the pressure waveform at VERY high speed, does FFT analysis, and provides statistical data for cycle to cycle variation......a lot of things which make the in-cyl pressure measurement useful.

I have used high speed data logging to log the output of a charge amp & kistler, but it is of little values...the cycle-to cycle variation is way bigger than you could imagine in most cases.....thus the DSP looks at a larger number of samples, and then outputs an average of between 20 and 720 cycles (revs essentially), making the data meaningful.

If you stop to think about it, 1 degree resolution (the minimum acceptable to be useful for the pressure data) means generally at least 4 point per degree, preferably 10.....at such hihg speed (100MhZ sample rate), even running the engine for a minute generates a HUGE number of pressure peaks, and all of them slightly different, and therfore difficult to make sense out of.

If you sampled at the rates suggested, you'd miss all the action....the peaks are very steep, with fast rise time.

In cylinder pressure measurement is VERY DIFFICULT.....this is why the equipment is so expensive. Every manufacturer and specialised engine developement group (FEV, AVL, Ricardo, ProDrive, Roush) are always tring to improve the affordability of the kit...typically starts at 85000 for 4cyl kit.....not many high end race teams even have access to this stuff...i remeber showing one of the premier V8 supercar chiefs through the benefits of the system, and how easy it makes mapping & finding gains, and they were amazed.....but they still don't have the kit.!

The only afforadable way of measuring cyl pressure has been via the ionisation energy of the spark plug...but it needs a fair bit of characterisation for the engine you are using. Harley davidson use this method for knock detection on air-cooled engines (too noisy for knock sensor), but has to be modelled for spark plug temp, etc.
Saab also uses similar methodology....
I'll see if i can find the info on Harley's & saab's system...i have it at work somewhere.

What do you want to measure cylinder pressure for?

I think i know why you want to do it, but i think you are resigned like the rest of us that it is not in the everyday sphere of engine developemnt / modification.

Sorry to be the bearer of bad news....

Isnt the BMW V8 found in the M3 also using the ion method for knock measurement? I didnt know Saab and Harley used it...Im very surprised to hear " Harley"

Ooops, maybe the Harley system isn't in production yet!!! It was part of a vendor tech presentation that i went to, and i have some other tech papers written by Harley engineers (unfortunately i can't share them), talking about the difficulties of air-cooled engines being to "noisy" for reliable acceleromter based knock detection. Don't forget that motorcycles (even HOGS!) are being subjected to increasingly stringent emission requirements, all driving sophisticated engine control solutions...FINALLY tech is coming to the bike world....high tech engines being fitted with carbs! In 2011 timeframe, it looks like Haeley will have to abandon air-cooling in favour of water cooling....due to emissions/radiated noise requirements! 102-odd years of low power, large displacement HOGS, and finally being dragged into at least the 20th century!

Up !

I'm more and more interrested by this system ...

What about the TFX kit ?

http://www.tfxengine.com/

Optrand also make some spark plug pressure sensor , cost is nice but i'm not realy conviced ...

http://www.optrand.com/