Intermittent roughness/power loss

[t_marat]

I had this thread on this issue before
http://www.bimmer.info/forum/showthr...ght=idle+rough

Well I changed the O2 sensor to a new original BMW one (sensor is a siemens one, my car has an MS40 siemens ECU, that wasn't exported to USA). Unfortunately I still have this issue.

When I first start the engine it runs normal. After a while it starts running rough with black smoke coming out of the exhaust. Taking for a drive and keeping revs at about 2500-3000 level makes engine run normal again. If I stop and idle for say a minute or so, then roughness and black smoke are back. Holding the revs at 2500-3000 for a minute is enough to bring the engine back to normal. I can also say that this behavior seems pretty much irrespective of engine temp, be it cold or warmed up. All this made me think the heating element in O2 sensor is gone bad, so I replaced it. But the problem is still there, exactly as before.
I thought maybe the heater voltage wasn't coming to the O2 sensor. checked at the connector and a 5V signal and 12V heater voltage are there, so an O2 sensor is now ruled out.

What else could be causing this problem? Injectors? Coils? TPS?

[t_marat]

Tried disconnecting TPS. Issue is still there.
Tried disconnecting MAF. At first engine runs normal, but then it starts wobling, rpm rising and falling, coming to near stall. The moment I reconnect it returns to more or less normal (black smoke).

[Jeff N.]

Black smoke is too much fuel. You may have a bad ECU engine temp sensor causing the ECU to keep the car on the cold run cycle. Or, a stuck thermostat so it never warms up.

Does this car have codes? Do a stomp test and get the codes. Work it from there.

[t_marat]

Thermostat is working, as the engine warms up.
The stomp test won't work on my car.

It could be the temp sensor, but would it depend on rpm? As long as I keep driving/rpm higher than 2000, then no black smoke.

[Morgenster]

Thermostat is working, as the engine warms up.
The stomp test won't work on my car.

It could be the temp sensor, but would it depend on rpm? As long as I keep driving/rpm higher than 2000, then no black smoke.

Sounds logical: higher RPM means higher temperatures and better combustion as well as better cat functioning. It's probably still running very rich. Only way to know is have the exhaust gases analysed and look at condition of plugs. I'd have thought that the fixed O2 sensor would correct for this when warmed up though. Apparently not. If the resistance of the temp sensor is higher than normal that means the DME gets the idea that the engine is cold, but If you really have that much smoke I'd also look at injectors or FPR?
Looking at your plugs and seeing one or two fouled plugs could quickly point to the injectors.

[Ross]

X2 what Jeff N says.

[t_marat]

I disconnected the intake manifold temp sensor. Still had the same problem. After a while engine would start running rich.
When I just start it, it runs normal. It takes a minute or so (shorter if fully warmed up) to start running rich. Raise the rpm to about 2500 for about 30 seconds, and its all normal. Leave idling, and it goes rich again.

So now things left to check are injectors and FPR? How can I check the FPR is working correctly? If the pressure in the fuel system is normal, then FPR is OK?

Could this be injectors? Or coil packs?

Another thing. I read that Siemens ECU can control the O2 heater. I understand this as its ability to turn off the O2 heater. Also there seems to be no heater relay. Could it be the ECU is going mad?

[Morgenster]

I disconnected the intake manifold temp sensor. Still had the same problem. After a while engine would start running rich.
When I just start it, it runs normal. It takes a minute or so (shorter if fully warmed up) to start running rich. Raise the rpm to about 2500 for about 30 seconds, and its all normal. Leave idling, and it goes rich again.

So now things left to check are injectors and FPR? How can I check the FPR is working correctly? If the pressure in the fuel system is normal, then FPR is OK?

Could this be injectors? Or coil packs?

Another thing. I read that Siemens ECU can control the O2 heater. I understand this as its ability to turn off the O2 heater. Also there seems to be no heater relay. Could it be the ECU is going mad?

I don't think the manifold temperature sensor has the same impact as the engine temp sensor and I don't know if disconnecting these sensors one by one will yield good diagnostic info.
Any ECU can go mad.
I'm just thinking logically here and not speaking out of experience, but if the FPR is failing It would let fuel pressure build up in the fuel rail in idle since the FPR is operated by the level of vacuum in the manifold. Normally more vacuum means more fuel is bypassed to recirculation and less pressure. What I don't know is if that would ultimately result in an over-rich condition because injector timing is pretty sharp and fluids can only be injected up to a certain limit. If the injectors were designed to be dependent of the FPR for metering the amount of fuel I'd definitely check it out, if not I'd look at the injectors.
But I could just as well be talking BS since I have only my logic and an idea of how these engines manage themselves. It's best to listen to the real engine gurus here.

[t_marat]

How engine determines timing for the injectors? Using the crankshaft sensor?

[t_marat]

Had the car scanned. At first the following errors were read:
98: Idle Speed Control Valve
161: Unkown. Absolutely no info anywhere.
100: Self diagnostics error of ECU
Errors were erased and after 5 minutes of driving only error 98 reappeared.
As I wrote above my main problem is that if I idle a couple of minutes engine starts running really rich with black smoke from exhaust. This actually happened during the second scan. Right at the same time as engine started running rough and rich another error code popped up
97: Tank Ventilation
After a pressing the gas couple of times engine returned to normal condition and ran for about 5-6 minutes fine. Final third scan revealed only error 98 (ICV), error 97 dissapeared by itself (no reset after second scan).

I don't think such a rich running condition can be caused by some vacuum leak. Is it possible? Engine runs really rich, you can smell it.
What is tank ventilation? Evap purge valve?