Audi 3B w/ 5 coils forcing idle help

[jrbe]

I'm new to VEMS but not new to tuning.  Working on a friends car and having a hard time with it.  The ecu/settings seems to be stuck thinking it should be idling no matter what throttle position and i can't figure  out why. Its a v3 PNP into an audi 3B with 5 coils.  It was supposedly dyno tuned but the ve map is too flat and moves in multiples of 5.  It supposedly ran better than this previously but i havent seen it run well.  I know the previous owner also. He got rid of it in part because of how not well it runs.

Things that have happened so far:
It popped a fuel pump driver.  Harness was hacked pretty bad.  Thinking a clipped wire shorted it out and took it out.  Was fixed and harness all cleaned up. Maybe the map got screwed up when it was repaired? or maybe something is still wrong in the ecu.

I fixed a poorly calibrated tps sensor.  There was a delay where tps displayed 0% to maybe 1/8" of throttle tip in then it went to 1%.  After adjusting that the same amount of tip in is ~5% and tps display is smooth through the full throttle swing.

It idles around 13' at 1k rpm.  Previously if you tried to give it gas you could watch it continuously pull timing all the way down to 2'.  To me this says it is thinking it should be idling. 
The ignition based idle control had an advance change max advance of 13.0 and an advance change max retard of 13.  I adjusted those both to 3.0 and it isnt pulling timing down below 10' anymore. After this  change i can baby the throttle up to about 2200 rpm, past that it goes super lean and rpms fall.  If i keep my foot there it bounces between 1300 and 2200 rpm slowly.  Its not a launch control type thing.  If i stab the gas it burbles up but hits the same cut.
Tps value for idle threshold was set at 2, i also tried 1 and 0.  Doesn't seem to have changed the seeming forced idle situation.

It starts up poorly but idles ok.  If you step on the throttle you can watch the tps slowly go up to about 5-6% (after fixing tps calibration) where it seems to fuel cut at ~2200.

It seems like the ecu/tune is trying to force it to idle.  I have been through all of the menus that have tabs on the tuning screen in vems.  Im not seeing what is wrong.  I did try adjusting the VE table where it falls flat with no response at all it seemed. I cant post a map yet.

Anyone have any ideas or suggestions on where to look or what to try?

Writing this i thought that the plugs may have been fouled.  I'll check those and update this when i can check them.

[mattias]

Yes, like you said if you posted an error report within VemsTune and the link to the web page for the report here it would help immensly.

My tips :
1. Get a new TPS, the old one is likely worn out near closed throttle.
2. You need to set a natural Idle TPS threshold, which is not possible, see 1.
3. Overrun fuel cut and resume is not working due to 2 and 1.

[jrbe]

Tps is smooth through its swing. The low end was adjusted too low on the a/d which gave the "dead" 0 tps area. It's perfectly smooth after adjusting the a/d # on the closed side of the tps calibration.

Thanks for the reply. I'm not sure how many posts I need until I can post an attachment...

[mattias]

No need for attachments. Open vemslog/vemscfg in VemsTune, make an error report (Help menu), post the URL (address) you are given.

[jrbe]

http://vems.hu/vemstune/bugreports/reports.php?cmd=view&key=FgcxVG

As you can see tps is smooth now.  Idled at first and slowly gave it more throttle input.  gets to 11% tps and 2200rpm then falls flat. 

After letting off as it seemed to die i crept throttle up slowly then held it steady as the engine bounced between about idle and 2200 rpm at ~11% tps.

Thanks for looking.

[VEMS]

Hello Jbe,

I have reviewed your report linked above, for starters i would suggest the following:

* To correct your idle stangeness, please correct your Idle Settings -> TPS Value for idle threshold  to a reasonable value of 1.5-2% instead of 0%.
* There are constant secondary trigger errors flagged, perhaps it is not connected or sensor broken ?
* You are using a very old firmware (1.1.95), if the setup is already tuned it should be okay, but i would not recommend it for a new install (1.2.31 firmware will be your best bet).

I would definitely check and correct the first two steps, take special care about changing firmware it will require expert knowledge and re-tune might be needed after upgrade only proceed if you feel confident about those steps.

Best regards, Dave

[jrbe]

I changed the idle tps value for idle threshold back to 2, no change.

I did record triggerlog.  It shows the 2nd trigger is dropping out over idle rpms.  Ignition events are dropping out as well which goes with the engine cutting out. Weird thing is the hall sensor in the distributor should be pretty robust. Only time i've seen a hall or optical sensor signal drop out was when a pull up/down resistor wasn't strong enough.  I'm not sure how that is configured with this ecu, software or hardware...
I'll try to borrow an oscilloscope to verify what the triggers are actually doing.

Primary trigger settings are Auditrigger divby3.  I have read of people having issues and switching to divide by 27...  Any chance there were issues with the current firmware in the ecu? Its currently 1.1.95

I am willing to update firmware but i dont know enough about this ecu yet. I can redo the VE map if it shifts for the new firmware and re fine tune.  I just dont know what to expect when updating firmware, whats going to change or what to expect to change or have to change and what if anything will be incompatible / problematic with the new firmware. 

I have read that config info has shifted position/location over different revs, i have no idea how to fix or adjust for that currently.

I'm not looking for someone to do my homework, just to hopefully point me in the right direction.

[Kamuto]

have you swapped crank home pin wires when installing ecu?

[jrbe]

Ecu was installed 4? Years ago and dyno tuned then. I assume they did this already. It didn't run this bad previously (supposedly.) I haven't checked though. I'll look this up.

It looks to be the distributor signal that's dropping out. I see 5 ignition events between 2nd trigger pulses. But it goes away. I can't zoom in to see what's going on.
Does this ecu reference the tdc pin/sensor (3rd oddball) trigger on older Audi's?

Does the distributor (sync/cam) pulse need to be aligned with a crank tooth signal on this funky Audi setup?

[jrbe]

I put an oscilloscope on the hall sender in the distributor. It is indeed dropping out. Going to install a new one and see if the issue goes away.

[jrbe]

2nd distributor doing exactly the same thing.



The oscilliscope i'm using is crap, voltage (power and signal) at the hall sender is indeed 5v.

[VEMS]

Hello Jre,

I put an oscilloscope on the hall sender in the distributor. It is indeed dropping out. Going to install a new one and see if the issue goes away.

If two different hall sensors give the same result it is more likely that some other common denominator is causing some problem, i suspect there might be a problem with a pullup strength on your hall sensor signal on your ecu or some intermittent wiring connection. It could also be a measurement error but less likely since both the Vems v3 and the scope confirm intermittent cam pulse signal.

You scopeshot lists the cam signal max as ~2.1V which seems to confirm the case for the improper pullup resistor on the cam-hall signal. Could you verify with DVM (digital volt meter) what the voltage of the signal on ec36-pin13 is with cam-hall inside the window and outside of the window ?

Best regards, Dave

[jrbe]

Thanks for the reply Dave.  I was thinking along the same lines as the too weak of a pull up resistor or a bad wire.  I am getting some occasional help, he recommended shorting the hall signal to ground. If it starts (luck of where the engine is) it takes the hall sender out of the equation.

Doing that, i got it to start but it seemed very sensitive to noise and didnt run great.  It only would get to ~2000 rpm like this still.

I measured between the hall sensor wires with the sensor unplugged. 
Between 5v power and -:
5.07v

Between signal and - :
1.85ma (shorting the 2 with a meter, figure 2.7k pullup resistor in there)
4.2v    weird.

Between signal and +5v :
.39ma (shorting the 2 with a meter, 13k ohm resistor or a strange voltage divider in there???)
.86v 5v to signal

Im not sure what to think of the weird floating voltage between the signal wire and 5v.  Cant see a good reason why this isnt at 5v.  Not sure if this points to something being wrong with the ecu or if theres some strange circuitry going on in there.

I have another crank sensor to try.  Ran out of daylight/time to try a new one. If the signal is at all what the oscilloscope is showing its probably bad. Though its pretty strange that hall and a mag trigger are both dropping out...

I have seen an extremely noisy alternator freak out these fairly noise sensitive sensors.  I'll probably pull the belt and see if it changes anything running just off of the battery.

[jrbe]

The idle issue was fixed by recalibrating the tps and fixing the settings.  Timing increases as it should now.  Still wont rev past 2200rpm unless you stab the gas.

Shorting the hall signal to ground did not change the way it runs.  Still broke up around 2000 rpm.  It was hard to start and had to try a few times.  It also seemed sensitive to noise (ran rough) like this.

Kamuto, there are short plug in harnesses to swap the trigger wires around (male / female connectors that swap position of the wires.)

I also took the belt off of the alternator, I've seen alternators that dumped tons of noise in the system.  There was no change when i took the belt off.

I just logged for 2 hours and could not catch it dropping out with this better oscilloscope.  I used both the hall signal ground and the chassis ground for the scope reference to check if the ecu wasnt grounding these well.  Signals were exactly the same.





The sensor that watches flywheell teeth had a spike in the signal.  The spike was double the normal voltage that the 135 teeth normally had.  Im working near the coils though too, may have been the coil firing.  I dont have a pic of it.

Every once in a while you see noise spikes in there.  This was a long record rate on the scope that i later zoomed in on.  The resolution / quality suffers.



Strange thing to note is the hall pullup voltage is only at 4.2 - 4.3v. Not sure why its not at 5v.  I did measure voltages and amperage to pull the pullup voltage down.  I measured between the hall sensor wires with the sensor unplugged.

I also tried to add a 4k ohm pullup resistor between the +5v and the signal wire.  This didnt change the way it ran.  Not sure it did much to the voltage on the signal wire.  I forgot to measure it.

Im not sure what to think of the weird floating voltage between the signal wire and 5v.  Cant see a good reason why this isnt at 5v.  Not sure if this points to something being wrong with the ecu or if there is some strange circuitry going on in there.

Most hall sensors have a minimum rated voltage of 4.5v (but not sure what the specs are for the bosch? hall sensor in the distributor...) this could be an issue at some point.

The distributor signal seems to be inverted in the ecu or maybe externally.  My assumption is something is wrong with that circuit.

I attached a trigger log and another vemslog log file.

Please let me know if the 4.2v at the hall sender is normal.  I'm thinking there is something wrong in the ecu, whether its trigger setup or hardware that processes / inverts the hall signal.

Some log files here, http://vems.hu/vemstune/bugreports/reports.php?cmd=view&key=FgcxVG

[jrbe]

Got a response from the guy helping.  Sent a divide by 9 config.  It works as it should now. No more sec trig error and it revs up well.