Need T/S help - E36 M50

[m20power]

Troubleshooting quandry on our '92 325is. CEL illuminated showing codes for both bad AFM and bad O2 sensor. After CEL reset, car barely ran and wouldn't rev. Once the CEL lit, car ran better but still rough and wouldn't pull over 4k on the highway. We repeated this test several times - reset CEL, ran lousy...after CEL lights, runs better. While running unplugging/plugging the AFM had no effect.

Thinking that the AFM was bad, but not convinced, we sourced a known-good AFM for this odd-ball model year M50 (thanks Brian). Installed good AFM, reset CEL, car starts and ran good, a little lumpy but OK. Idled for 15min, no CEL - problem identifed!(?) Installed old AFM to verify, but no difference - no CEL and runs the same. Hmmm...maybe not AFM. test drives with either AFM was the same - ran fine, pulled hard to redline but stumbles when dropping to idle and hesitates from stop - no CEL. Smoking more than usual also.

Decided to look at O2 sensor - although I don't believe a fault here can have such a profound effect. First noticed no voltage to heater wires - relay OK, socket voltage OK, not getting ground signal from ECU to energize heater. I think that the O2 sensor heater on a 4-wire sensor always turns on when car is started simply to expedite warm-up, but the temp is not modulated so it's not important to warm running as, say, in a WB. O2 sensor signal voltage steady at 450mV indicating, I believe, a sensor not yet up to temp but not dead.

My next step is to run at high-ish RPM to see if the O2 sensor comes up to temp and into play. If so, I'll move on for now, worry about the heater circuit later.

Any help/suggestions to pinpoint this problem would be appreciated (Vic, Ray, Don?). Since the car is OBD1 I can only get the codes my Peake reader can access. AFAIK my AutoEnginuty can not interrogate an OBD1 ECU.

TIA, Bob

[alken]

Bob -

Any signs of a vacuum leak? Frequently, leaks -> lean conditions will trigger an O2 sensor code. FYI, the codes you're seeing from the Peake tool are factory, though you can't delve as deep as you can with the PC of course...

Ken (not a mechanic...)

[jnscaldwell]

Bob,
You already probably checked it, but what about the cam position sensor?
#11 in this diagram:
http://www.realoem.com/bmw/showparts.do ... g=11&fg=15

I made the assumption you have a non-vanos M50 in there. I saw a NV M50 throw AFM codes and it turns out it was this part that was failing...nothing to do with the AFM at all.

Jeff

[RAYADAM]

Hey Bob,
did this just happen all of a sudden? this problem can go in several directions...going to throw some stuff out to you:

*check fuel pressure?
*did you actually look at the ECU,the early E36 could get water into the
ECU compartment if the drains got clogged..we used to add another
drain port behind the valve cover to fix this.
*did you test running condition with O2 unplugged?,a good healthy engine
with proper fuel press,cam timing,etc will run fine..
*ICV could cause issues similar to your case
*Throttle potentiometer
*camshaft and crank sensors?

how many miles on this car Bob?

[m20power]

Thanks for the reply guys, good suggestions all. To answer Ray, car has ~250k and this did just happen. Car was running fine, parked for a week or two and then wouldn't start until after a lot of cranking - then lots of smoke and very rough. Greg drove it home from Philly and couldn't pull over 4k on highway.

I didn't check any of the other sensors or FP. It is a non-vanos motor. Wouldn't a failure for many of these other sensors set appropriate fault codes?

What confuses me is the behavior observed last weekend (not now) with the bad AFM code, and the way we could reset the CEL and trigger another CEL time after time. Also, the way it ran better when the CEL was lit. What's related to AFM function closely enough to throw that code?

I was hoping to shortcut a long day of troubleshooting...I'm not too hopeful as all these are good suggestions. Any other ideas? Thanks, Bob

[victory1auto]

Get the DME running closed loop before doing any more troubleshooting. It is fairly common for the DME output that pulls in the O2 heater relay to fail. You can rewire such that the relay pulls in with the ignition turned on. Indeed, this is how the original O2 heaters were controlled.

Vic

[m20power]

Thanks Vic. Is it possible that a non-heating O2 sensor could cause such poor running and throw an AFM code? Is the O2 sensor heater always on when engine is running?

TIA, Bob

[victory1auto]

Despite its being controlled by the DME, I believe the O2 heater on that vintage is on virtually all of the time the car is running. Given how far back in the exhaust system the sensor is, I would think the heater would have to be on, except perhaps for protracted high-load operation.

I think it is unlikely that O2 heater problems would set an AFM code.

A cold sensor can definitely cause driveability problems. Or, more precisely, an operating sensor can mask them.

Vic

[RAYADAM]

Bob,
It could even be a bad tank of gas ,all depending if you filled up recently.

sometimes bad sensors can't be detected by older ECUs..

Like I said earlier ,if you unplug the O2 sensor ,and the engine still runs badly ,that will eliminate the possiblility the O2 sensor is the problem.ECU will operate at default settings if the O2 is not present and of course obviously set a CEL .. will not cause significant power loss, but indeed compromise emmissions and fuel economy.

[jnscaldwell]

I still say it's the cam position sensor (or a much deeper DME issue). I just don't think the O2 sensor could cause such bad running problems, even if it's not heating up. If the DME is not getting the proper signal to fire the injectors/plugs etc then this will cause all or most of the problems mentioned. On the car I saw this happen to we disconnected the AFM and connected it repeatedly with NO change in the performance of the engine. Connecting and disconnecting the O2 sensor did nothing either. The DME threw no codes for cam position or misfire, only AFM codes. We basically threw everything but the kitchen sink at the problem and replacing the CPS is what fixed it.

Jeff

[m20power]

Thanks all - I've got my marching orders. I'll try the next pass at t/s and report back. Thanks much for all the good advice. Bob