RPM steady, but idle very rough - need help

[wbondsteele]

My first post on this great forum. Hoping to get some help from my friends on the other side of the pond. So here goes. I recently purchased a 1990 325i 4DR auto with 54,000 original miles, all stock. First BMW I've ever owned, beautiful car, no mechanical issues ... except for an idle problem that is very frustrating. I have read a lot of posts regarding this issue but none that seem to address the symptoms I am having. Upon morning start up the car starts and idles fine, not super smooth, but what I guess to be a fairly good idle for this car from what I've read. The tach jumps up a bit and then quickly settles down to a stead 700 r.p.m. But as soon as the engine warms up, usually within a minute or so, the engine starts getting rough, though the idle speed never varies. As I drive around town and reach full operating temperature, the idle worsens and the body shakes. The roughness, however, is fairly controlled--though erratic--if that makes any sense. Idle speed never varies though and as soon as I touch the throttle all roughness vanishes. I should note that the car idles noticeably better when the A/C is on. Idle speed rises to a steady 750 (versus 700). Nevertheless, the idle is still lumpy. Another thing I should note: my gas mileage is poor, possibly indicating that the car is constantly running rich? As far as troubleshooting this, here is what I do know. It is probably not a vacuum issue. I haven't had the car smoke tested (yet) but I have thoroughly checked everything for leaks and found none. Next was the Idle Control Valve. I removed and cleaned it, and it appears to be functioning normally. I also don't think this is an ignition issue. I installed new wires and plugs and gapped them correctly. I also tested the coolant temp sensor and it was okay. The TPS also appears to be working normally. About the only thing I haven't checked is the oxygen sensor and the AFM. I suppose that could be the culprit, but before I go messing with that I would like to know if anyone here has experienced this specific problem or can troubleshoot based on this information. I would very, very grateful. It is frustrating. If I ever fix this problem I promise to post the solution for the benefit of forum members. Thank you from Lawrence, Kansas.

[Brianmoooore]

How/where did you test the coolant sensor? At the sensor itself, or at the ECU plug?

[wbondsteele]

I tested at the output of the sensor itself, hot and cold. It was within the normal range.

[Brianmoooore]

You need to repeat the test at the ECU plug to test the wiring as well. Details are in the wiki.
As a general rule, you should always try not to disturb a suspect component when testing. In the case of this sensor, it's not inconceivable that the pressure from the test meter probes on its socket would cause a temporary cure.

[wbondsteele]

Hmmm, I see. I will check the wires from the ECU this weekend report back. I now have a check engine light, so something is going on somewhere.