VEMS triggerlog on 7A showing intermittent loss of ~3 cycles

[ejfluhr]

Can anyone help us troubleshoot a bonkers problem shown by this VEMS triggerlog? Our NA 7A race engine runs fairly well with recently installed VEMS (prior to that we were using the factory Hitachi ECU which ran completely fine), but often experiences a single misfire in the RPM range of 4.5K - 5.5K. Other than that it idles OK and pulls well to the rev limiter ~7200RPM. We are running the factory flywheel sensors combined with the hall sensor in the distributor. The flywheel is an 034Motorsport aluminum flywheel with steel ring gear and crank pin.

After some debugging, we've got this triggerlog example of VEMS losing all the triggers for 105-107 teeth several times in a row when revving to ~2K RPMs. Sometimes we rev and see no problems, sometimes it's massively worse than the log I am including here. I don't understand how the engine continues to run after losing cycles worth of triggers, which makes me wonder if the data logging is bad rather than the actual sensors.

First batch of errors starts at sample 6615, second batch starts at 13187. There are a couple of one-time errors at 9430 and 11538 missing only 80 teeth.

Any recommendations/ideas??

Thanks,
Eric

[ejfluhr]

Just an update on this problem. We still have that random misfire at high RPMs. It turns out that VEMS throws a "Less Primary Trigger" error whenever the misfire happens. We also often get a "Bad Tooth Length" error when <= 3.5K RPMs, but no signs of misfire or other problem when that happens.
I've concluded that the trigger logging doesn't work after RPMs climb to something above 2000...not sure where it gets lost but everything is perfect below that point and crazy above it, while the engine runs just fine.

The primary trigger is the flywheel tooth counter. As I understand it, we are using "Audi divide-by-9" which turns the 135-tooth flywheel x2 rotations per combustion cycle = 270 teeth into a 30-tooth virtual trigger wheel. That is combined with the flywheel pin sensor & hall sensor to identify TDC. From everything I can tell, the secondary trigger (that flywheel pin sensor) works great (i.e. no errors and always looks perfect in the trigger log). That partly agrees with the fact that our 034Motorsports aluminum flywheel has a much larger pin than the factory flywheel.

We are using a custom ECU harness, pretty high build quality and matching (as best we could) to the VEMS documentation. We have swapped ECUs, swapped sensors, shimmed the crank sensor in and out from stock, wired a separate "positive" line from the sensor to the ECU, added grounding at (1) the sensor connector and (2) the ECU common ground, Ohm'ed the sense/ground/shield wires between sensor connector and ECU, and retimed the distribution/cam/crank very-very carefully. Nothing has made a clear difference.

As an example of how intermittent the trigger error & misfire is, our typical track lap time is ~1m35s with 8 instances of revving in 3rd and 4th gear from ~4000rpms to our shift point >= 6700rpms. We may experience the misfire between 0 and 3 times per lap, very commonly 0 and 1, less commonly 2, and rarely 3. The engine easily revs to the redline at 7200 (?) rpms, even right as the misfire happens.

So we are very very stumped. I know there is a lot of older info stating issues with VEMS and the Audi factory sensors, but my dad is running VEMS on a 3B engine at 400hp using the factory sensor config, though with the Motronic harness on a Fidanza aluminum flywheel. I thought the trigger issues were sorted out by now, but here we are.

Pending further ideas, we are going to take out the wiring harness and triple-check the wiring. I may also look for a portable oscilloscope to see if I can monitor the sensor signal and look for anything funny. VEMS is throwing a primary trigger error so it clearly detects something is wrong, we just cannot figure out how to get it to tell us specifically -what- went wrong. Maybe there is some error logging we haven't figured out?!?!

Thanks for reading...[almost] any input is helpful!
Regards, Eric

[varia]

Just an update on this problem. We still have that random misfire at high RPMs. It turns out that VEMS throws a "Less Primary Trigger" error whenever the misfire happens. We also often get a "Bad Tooth Length" error when <= 3.5K RPMs, but no signs of misfire or other problem when that happens.
I've concluded that the trigger logging doesn't work after RPMs climb to something above 2000...not sure where it gets lost but everything is perfect below that point and crazy above it, while the engine runs just fine.

The primary trigger is the flywheel tooth counter. As I understand it, we are using "Audi divide-by-9" which turns the 135-tooth flywheel x2 rotations per combustion cycle = 270 teeth into a 30-tooth virtual trigger wheel. That is combined with the flywheel pin sensor & hall sensor to identify TDC. From everything I can tell, the secondary trigger (that flywheel pin sensor) works great (i.e. no errors and always looks perfect in the trigger log). That partly agrees with the fact that our 034Motorsports aluminum flywheel has a much larger pin than the factory flywheel.

We are using a custom ECU harness, pretty high build quality and matching (as best we could) to the VEMS documentation. We have swapped ECUs, swapped sensors, shimmed the crank sensor in and out from stock, wired a separate "positive" line from the sensor to the ECU, added grounding at (1) the sensor connector and (2) the ECU common ground, Ohm'ed the sense/ground/shield wires between sensor connector and ECU, and retimed the distribution/cam/crank very-very carefully. Nothing has made a clear difference.

As an example of how intermittent the trigger error & misfire is, our typical track lap time is ~1m35s with 8 instances of revving in 3rd and 4th gear from ~4000rpms to our shift point >= 6700rpms. We may experience the misfire between 0 and 3 times per lap, very commonly 0 and 1, less commonly 2, and rarely 3. The engine easily revs to the redline at 7200 (?) rpms, even right as the misfire happens.

So we are very very stumped. I know there is a lot of older info stating issues with VEMS and the Audi factor sensors, but my dad is running VEMS on a 3B engine at 400hp using the factory sensor config, though with the factory Motronic harness on a Fidanza aluminum flywheel. I thought the trigger issues were sorted out by now, but here we are.

Pending further ideas, we are going to take out the wiring harness and triple-check the wiring. I may also look for a portable oscilloscope to see if I can monitor the sensor signal and look for anything funny. VEMS is throwing a primary trigger error so it clearly detects something is wrong, we just cannot figure out how to get it to tell us specifically -what- went wrong. Maybe there is some error logging we haven't figured out?!?!

Thanks for reading...[almost] any input is helpful!
Regards, Eric

https://www.s2forum.com/forum/techni...55#post1282155

its a crap shoot. it works or it doesnt, not much you can do.
I wasted many hours chasing this issue once, and every since I always install a 60-2 seal/sensor combo to avoid this (best way to avoid it, not to use vems)

[ejfluhr]

Thanks for the feedback, varia. I've seen that and many other pages when searching S2Forum and the internet for similar problems. Most are old or are focused on difficulty with the secondary trigger, i.e. the flywheel TDC pin. Our issue only calls out primary trigger errors, according to the ECU messages. Problems with that sensor seem to be much rarer.

[ejfluhr]

Just posting an update...we have not yet solved this problem. I've triple-checked all the wiring and added grounding points on the shields. I've played with the spacing adjustment of the crank sensors and put an oscilloscope on the sensors to demonstrate it creates a good quality signal. We still get the intermittent single misfire at > 4000rpms, coinciding with "Less Primary Trigger" errors, as well as intermittent "Bad Tooth Length" error when between 2000 and 4000rpm. We have a mild belief that it correlates to temperature, as it often does not happen in the first 10+ laps, then gradually shows up at increasing frequency as the car gets hotter and the day gets hotter.

I have swapped to a second ECU, hoping to test that later this year at the track. VEMS works great on my father's 20v 4ktq pushing 400hp, and this is a measly NA 7A making ~160hp.

Part of the problem may be my lack of skill in data logging/debugging. I've done ignition trigger logs and some basic track lap logs of the ECU, but don't know what other logging options may be helpful. Any recommendations on that front would be greatly appreciated...

Also Marc S. gave us a base tune, and then we have been relying on an experienced EFI tuner (who doesn't know VEMS) to provide ignition and fuel injector updates. This is where the ignition map is at now...I'm curious if anyone knows whether this is on the aggressive or conservative side? Our compression ratio should be about 9:1, according to FRP who provided the pistons.


Vems3.2_ignition_table.JPG

Thanks, Eric
Team U/A 1991 90q20v

[ejfluhr]

I've looked at a couple of other ignition tables for MAP vs. RPM and our values seem to be 4-7 degrees conservative. E.g. at 5600rpm and ~100kPa atmospheric, I'm seeing other maps show 26-27 degrees vs. ours at ~19. I'm curious if that is due to factory 7A running 10.3:1 compression ratio vs. most others running 8.5:1 or less...does anyone know?

Thanks, Eric

[ejfluhr]

In the previous update/question, I hadn't made a note that we ended up switching to the VW 60-2 flywheel sensor. We suspect our 034 lightweight flywheel is the culprit as, in comparing that to the factory flywheel, we noticed that the cut of the teeth are slightly different. We did not install the factory flywheel to test that theory, just moved to the 60-2 as we had already purchased it. We set up and ran 10-cyl wasted spark using just the cam hall sender for 1 race, since our VEMS ECU was setup for the Audi crank senders on the primary trigger. 10-cyl wasted spark mode ran quite well; I was surprised. However, we don't want to stay there so had Marc S. rework the ECU to take the hall sensor on the primary trigger channel; now moving back to true 5-cyl mode.

The power isn't great, though, only making ~150hp on this na 7A with ~9.3:1 compression. I'm hoping it's the ignition map (see post above) but the guy doing our tuning says the knock sensors don't seem to be working (?) and he cannot safely tune it. So trying to figure that out, also. Help is always appreciated!!