Still having high RPM and idle problems...

[paanta]

I feel like a damn broken record here. Sorry to keep posting about this problem... Car in question is a '91 535i. I'm an engineer by training, have owned a dozen or more Bosch fuel injected cars, and haven't visited a pro mechanic in 10 years....and still this is leaving me completely stumped. I need input from the Gurus. :)

Symtoms:

Idles rough after warm start (car sits for 10 minutes to ~3 hours). Cold start idle is great. Restarting a hot car immediately is great.
No power for ~30 seconds after warm start. Throttle input causes engine to bog down. Car will barely move. Revving over 2500-3000 rpm clears it up a little.
Won't rev past 5800 rpm or so. Cutoff is immediate and feels exactly like a rev limiter. Holding WOT in 2nd gear the car will sit right on 5800 rpm and not go over. Doesn't feel like wheezing out of breath as with a fuel pump problem.
Normal otherwise. Tons of power between 3K-5K, sounds good, idles perfectly, good throttle response and it starts on the first crank every time.
Intermittent Error 1222. Lambda rich/lean control stop. CEL flashes on sometimes during bad idle. Seems better after O2 replacement, but persists nonetheless.

Tests:

Crankshaft Position Sensor: Replaced with non-OEM, but no difference. Resistance is fine. Air gap on loose side, but within, spec.
Coil: Replaced. Did nothing. Returned part.
Plug wires: Replaced
Distributer+rotor: Look fine
Plugs: Haven't checked...seems unlikely culprit. Still, check this weekend
Alternator+Battery: 12.3v car off, 13.78v measured between jump start terminal and engine ground. Constant with RPM and only slight dip with larger power draw. Borderline low?
Throttle Position Sensor Tested and good. Unplugging at idle causes instant bad idle.
AFM: Cleaned (plate was sticking a bit), resistance of wiper smoothly goes from 500->1100 ohm and then back down to 800 with door fully open.
Idle Stabilization Valve: Cleaned. Makes a sound when shaken. 20/20/40 ohms. Not much difference when connector unplugged though...
Air temp sensor: Air temp sensor reads normal.
Coolant temp sensor: Tested OK
Oxygen sensor: Tested OK. Replaced anyway. Old one had whitish deposits on it (previous owner replaced cyl head after gasket failure)
Fuel pressure : 3.0 bar with pump running, engine off. 4.5 bar if return line pinched. 2.5 bar with engine running, up to 3 bar with throttle input.
Fuel flow: 0.9 quarts after 30 seconds. Bentley calls for 0.93. 3% low enough to matter if pressure tests OK?
Fuel Pressure Regulator: Replaced. No difference.
DME: Replaced. Symptoms persist. New DME seems to cause car to idle high (1100 rpm at times) though it has stock chip vs. EAT.
Vacuum leaks: Cracking oil cap causes instant bad idle. Spraying carb cleaner all over causes no idle fluctuation
Tachometer: Tested against hand held tach. Works fine.
Fuel filter: Replaced


Thoughts....

Are the two problems (rev limit and bad warm starts) related?
The car runs fine most of the time, and the cutoff at high RPM is VERY sudden, so I don't think fuel pressure/flow is a problem.
Bentley's Bosch fuel injection book says Bosch defines the warmup stage as from the end of cranking for about 30 seconds. This coincides with by bad running. Maybe it's mistakenly thinking the car is starting cold when it's really starting warm and giving _way_ too much enrichment. Or maybe it thinks it's starting hot and is giving not enough enrichment. Air temp sensors are fine at the sensor itself, but perhaps the wiring to the DME is bad, causing higher resistance, which makes the sensor read low?
The stomp test works, but it seems intermittent. I have to try 5 or six times to get it to cause the CEL to flash. Throttle switch works when tested at the switch, but again, maybe the wiring to the DME is bad?
If the car wasn't seeing WOT because of a flaky throttle switch, would it still pull to red line?
How sensitive is the CPS to the air gap? Mine is on the loose side of spec.
Mark D'Sylva said a bad wheel speed sensor can keep the car from revving past 6K. Is this true on my old 535i with no traction control?


Next steps...
Start checking sensor values from the DME side?
Check and gap plugs, just for the hell of it.
Double check all grounds.
Others?...

[Tiger]

Intake manifold leak or major vacuum leak. Pull out every hose and make sure they hold tight... if not cut back a little and reinsert.

[paanta]

Would a vacuum leak cause bad idle all the time? The car idles perfect 99% of the time.

Adding a tiny vacuum leak causes instant bad idle. I've always been told that indicates you don't have any really huge ones.

I'll triple check the vacuum hoses, though.

[Tiger]

Not really... everything expands when hot... so better sealing when hot. Your vacuum hoses are 20 years old! and you still think they don't leak?

[paanta]

Not really... everything expands when hot... so better sealing when hot. Your vacuum hoses are 20 years old! and you still think they don't leak?

All, or almost all, the vacuum hoses are new from when the PO did the head. :( or maybe I should say :)

Possible one of the hidden ones slipped off, of course...

[Ross]

On cold, or is it warm, starts isn't the vapor canister purged? I am similarly stumped but this system may be worth investigating.
Even though you say it feels like a hard cut off as a rev limiter would be all the ignition variables have been addressed so I'm in the fuel/mixture camp.

[Jeff N.]

Wow. Really strange. It sure sounds like the idle problem is fueling related. Seems to have all the symptoms of being lean - no power, CEL on.

Curious that when you unplug the TPS on idle you get bad idle. I can't recall if it should be that way or not. I would tend to expect a slight increase in RPM but should generally run OK.

Do you have the ability to swap in another AFM? I'm still thinking that sounds like the culprit. Where are you? I have a remanufactured AFM sitting in my parts bin that has < 10k miles on it.

[paanta]

I'll try to find a local AFM and swap and see. If I can't find one, I may hit you up. :)

It seems like the post-warm start problem results from the motronic warmup sequence. I can't overemphasis how well the car runs except for those 30 seconds after you start a warm engine. Starts on the first crank, idles like a rock, great power (except for the pre-redline cutoff) and throttle response. All the major components pass the tests and if any were really bad I'd expect other symptoms...especially since normally problems tend to show up on a hot engine or a cold engine but not on a lukewarm one. So I keep coming back to more esoteric theories...

Does the EVAP purge happen just after warm startup, or is it continually dumped into the intake as on older cars? If it's just after startup, that could definitely explain things if there were a big leak in the line near the valve making the car go waaaaay lean until the valve gets closed again.

Other possibility: there's a shared ground for TPS, Coolant temp sensor, and shielding on the ignition coil+o2 sensor+CPS. I'll give that a look. Dunno if a bad ground on the shielding could cause the CPS to crap out at high rpm..

Does anyone else have trouble with the stomp test? It always takes several attempts for it to take. Maybe that TPS wiring is bad?

[Jeff N.]

This is a manual, right? On a stick, the car will run OK at WOT without the TPS. Might not make max power but it shouldn't cut out.

Re the evap purge, you should be able to disconnect the harness wire that opens the valve and see if anything changes.

Ah - just had a thought. The DME has a safety function that cuts out the engine if it sees more than like 4.9v on the signal wire. Hard cut. I've had this occur with my MAF conversion if I get the translation tables wrong. Mark D. has confirmed this "feature" of the DME. Makes me wonder all the more if your AFM is fubar. I'd really push to swap in another unit and see what happens.


I'll try to find a local AFM and swap and see. If I can't find one, I may hit you up. :)



It seems like the post-warm start problem results from the motronic warmup sequence. I can't overemphasis how well the car runs except for those 30 seconds after you start a warm engine. Starts on the first crank, idles like a rock, great power (except for the pre-redline cutoff) and throttle response. All the major components pass the tests and if any were really bad I'd expect other symptoms...especially since normally problems tend to show up on a hot engine or a cold engine but not on a lukewarm one. So I keep coming back to more esoteric theories...

Does the EVAP purge happen just after warm startup, or is it continually dumped into the intake as on older cars? If it's just after startup, that could definitely explain things if there were a big leak in the line near the valve making the car go waaaaay lean until the valve gets closed again.

Other possibility: there's a shared ground for TPS, Coolant temp sensor, and shielding on the ignition coil+o2 sensor+CPS. I'll give that a look. Dunno if a bad ground on the shielding could cause the CPS to crap out at high rpm..

Does anyone else have trouble with the stomp test? It always takes several attempts for it to take. Maybe that TPS wiring is bad?

[paanta]

Yeah, the WOT thing is odd, since it should be open loop at that point.

The reference voltage is 5v, right? So if the AFM has any sort of short it'll kill the party? I'll see if I can find a spare AFM. Without the AFM the car goes into limp home or does it just not run at all? Maybe I'll try disconnecting the AFM completely and see if it'll pull to redline. You never know...

[Jeff N.]

Recalling my motronic a bit. Mark can swoop in and fix any of my mistakes.

There are 3 basic maps in the Motronic 1.3. An idle map. A mid throttle map. A WOT throttle map. The TPS triggers the switch. Car runs closed loop during idle and mid, goes open loop at WOT. It's never been 100% clear to me if/when on the 1.3 the AFM signal is ignored. I have heard reports it's ignored at upper RPMs at WOT but am not sure. I know that on my porsche 951 (similar vintage motronic) the AFM signal is read at all throttle openings regardless of the TPS signal. The BMW TPS triggers the WOT setting about about 2/3 opening of the throttle plate.

Car will idle without the AFM. Never tried anything outside of that.

I think another AFM is definately a good thing to try at this point. Guessing it will fix your high RPM problem but may not your startup problem.

[paanta]

Just went and tried to drive the car w/o AFM plugged in. It'll idle (which I guess is independent of the AFM anyhow) but it won't drive. Ah well. It may very well be ignored at high rpm, but on the off chance that the car decides to run way lean as a result of my disconnecting the AFM, I don't want to burn up any engine bits trying to get it to rev to 6200. :)

[Bill R.]

Its fairly sensitive to this. Close it up to even less than the specified gap of 1 mm, say .035 thousandths instead.. At higher rpm it might be not getting clean signals and might be responsible for the cutting out.




Just went and tried to drive the car w/o AFM plugged in. It'll idle (which I guess is independent of the AFM anyhow) but it won't drive. Ah well. It may very well be ignored at high rpm, but on the off chance that the car decides to run way lean as a result of my disconnecting the AFM, I don't want to burn up any engine bits trying to get it to rev to 6200. :)

[Bill R.]

Its fairly sensitive to this. Close it up to even less than the specified gap of 1 mm, say .035 thousandths instead.. At higher rpm it might be not getting clean signals and might be responsible for the cutting out.


Also do a full sweep test on the gauges.. .I think 91 has that capability to see if the tach reads all the way.

[paanta]

What gap? The AFM or the CPS? Both of 'em are gapped around 1 mm, right? :)


Anyone know how to do the full sweep test? I haven't seen it before.

[Jeff N.]

CPS gap.

Sweep test. Oh my, this like a memory test. I think you hold down the trip reset button and then turn the ignition on. The cluster does a full test; sort of like a fireworks show.

[paanta]

Better invest in those memory enhancement pills....no luck on that sweep test. :)

[paanta]

Ah, found it.
http://www.m5board.com/vbulletin/e34-m5-discussion/69581-speedometer-took-crap.html

[paanta]

And of course the sweep test just makes the tach and speedo go part way up. ~3450 rpm and 75mph. Fun light show, though. :)

[Jeff N.]

hey...I was close. What's a button or two amongst friends?

[Bill R.]

the e39/38/46 and newer do it.





And of course the sweep test just makes the tach and speedo go part way up. ~3450 rpm and 75mph. Fun light show, though. :)

[Ross]

I recall E34s do the full sweep. Will check mine later.

[paanta]

Gapped CPS to .030, still no luck.

TPS WOT connection is a bit weird after all. Tested it before and it was OK, but now it's reading anywhere between 5 and 30 ohms at the DME harness. Is that enough that it wouldn't register as being continuous?

AFM when warm does, indeed, have a flat spot where it's reading ~100 ohms from the DME connector. I can get a used AFM here locally, but if you've got a pretty new one Jeff, how much do you want for it?

And when it rains it pours.....I forgot to remove a jack stand from the crossmember behind the engine before I backed off of ramps. Now I'm dripping oil. Hopefully it's 'just' an oil pan.

[paanta]

For the permanent record: I think the new AFM fixed the high rpm cutout AND the idle.

[632 Regal]

wow success! What happened to the oil pan?

[paanta]

Coincidence, I think. I really slammed the oil pan on that jack stand, but a leak from the valve cover gasket got bad enough to start dripping that day. Hence my post about dropping the stud down into the bowels of the engine...also resolved.

Amazing to have the car running right. I almost don't know what to do with my time.

[Jeff N.]

Hey Patrick! I have a spare M30 oilpan too! You just let me know :D