Very strange idle behavior

[J.DeFeo]

My car will intermittantly idle very fast, anywhere from 1000-1500 RPM. When the engine is cold, it usually idles around 1000-1100 RPM and then slowly increases it's speed as the engine warms up. Then, after maybe an hour of driving, occationally the idle will "bounce" up and down between 1200-1400 rpm by itself. It sounds like I'm tapping the gas pedal to initiate a street race, which has disappointed many honda civic owners thus far ;). This also seems to affect the fuel cutoff behavior; it will cut off with much less travel in the gas pedal, it feels like I'm losing a great deal of "sensitivity".

Even more strangely, about 80-90% of the time I can fix this behavior by flooring it in first or second, getting the RPMs about 4000/5000 RPMs will bring the idle back down to a perfectly smooth 750 RPMs.

I recently adjusted the TPS, since previously it didn't have continuity at idle. It now reads <1 ohm at idle (on the idle side of it). I believe the TPS works perfectly now since adjusting it fixed my starting problem. The ICV measurements are all within the +/- 2 or 4 ohms that the bentley manual specifies. The coolant temperature sensor read about 100 ohms high for the 176 F measurment but I have no idea if the coolant was actually at 176 degrees. The metal part that the sensor is connected to was nearly too hot to touch so I estimated it was pretty close.

The vacuum hose from the fuel pressure regulator was basically dust, so I replaced it; no change in idle behavior. I'm going to replace all of the vacuum lines with siliconne but I don't think this is a vacuum issue. Any ideas on where to start diagnosing this? Any suggestions at all would be appreciated, especially further tests for the ICV which I'm suspect of even though the resistances are within spec. Thanks in advance, and sorry for making so many new threads!

'92 535i, M30, manual

[Ferret]

My car will intermittantly idle very fast, anywhere from 1000-1500 RPM. When the engine is cold, it usually idles around 1000-1100 RPM and then slowly increases it's speed as the engine warms up. Then, after maybe an hour of driving, occationally the idle will "bounce" up and down between 1200-1400 rpm by itself. It sounds like I'm tapping the gas pedal to initiate a street race, which has disappointed many honda civic owners thus far ;). This also seems to affect the fuel cutoff behavior; it will cut off with much less travel in the gas pedal, it feels like I'm losing a great deal of "sensitivity".

Even more strangely, about 80-90% of the time I can fix this behavior by flooring it in first or second, getting the RPMs about 4000/5000 RPMs will bring the idle back down to a perfectly smooth 750 RPMs.

I recently adjusted the TPS, since previously it didn't have continuity at idle. It now reads <1 ohm at idle (on the idle side of it). I believe the TPS works perfectly now since adjusting it fixed my starting problem. The ICV measurements are all within the +/- 2 or 4 ohms that the bentley manual specifies. The coolant temperature sensor read about 100 ohms high for the 176 F measurment but I have no idea if the coolant was actually at 176 degrees. The metal part that the sensor is connected to was nearly too hot to touch so I estimated it was pretty close.

The vacuum hose from the fuel pressure regulator was basically dust, so I replaced it; no change in idle behavior. I'm going to replace all of the vacuum lines with siliconne but I don't think this is a vacuum issue. Any ideas on where to start diagnosing this? Any suggestions at all would be appreciated, especially further tests for the ICV which I'm suspect of even though the resistances are within spec. Thanks in advance, and sorry for making so many new threads!

'92 535i, M30, manual

Pull the plug on your air flow sensor while the engine's running and doing this bouncing - and see if the idle settles down. You've got classic symptoms of a vaccuum leak, and this will tell you exactly whether you have or not.

[Morgenster]

Pull the plug on your air flow sensor while the engine's running and doing this bouncing - and see if the idle settles down. You've got classic symptoms of a vaccuum leak, and this will tell you exactly whether you have or not.

A pretty big one too because I would only get this if my ICV to manifold elbow hose was disconnected leaving a serious air leak.

[J.DeFeo]

A pretty big one too because I would only get this if my ICV to manifold elbow hose was disconnected leaving a serious air leak.

This almost exactly what it was. The small hose going from the ICV to the plastic L-joint in front of the intake boot had a huge hole it in it right at the base. Replaced hose, idles perfectly now.

Thank you everyone, my car now runs perfectly!

[J.DeFeo]

This almost exactly what it was. The small hose going from the ICV to the plastic L-joint in front of the intake boot had a huge hole it in it right at the base. Replaced hose, idles perfectly now.

Thank you everyone, my car now runs perfectly!

This was only part of the problem, as it turns out. The ICV was also broken. I cleaned the everlasting hell out of it with 4 different solvents and it still idled too high. Here's a fun test I learned out for that:

1.) When the car is idling high/strangely, take a screwdriver and lightly tap on the top of the ICV. When I did this, after maybe 3 taps the idle came down to normal.

After the idle came back down I re checked the vacuum and it was perfect. Got a new valve coming in from AutohausAz, $112 which was the lowest price I could find.