82 SC with 92 3.6L and bouncing idle

[yelcab1]

I have an '82 911SC with a 3.6L engine from a '92 964. Pretty well done except that it has a wildly bouncing idle when stone cold for about 2 minutes, then it settles out at 880 rpm. It does have a Steve Wong chip in it. So far, what I have done to diagnose are:

1. Smoke test the intake circuit, no vacuum leaks.
2. Measured fuel pressure when idling: 50 psi and steady. Pulled off the vacuum line and no fuel is dripping out. It seems fuel pressure regulator is good.
3. Took out the Idle Valve, cleaned it, applied 0-5V to it and saw that the valve continuously actuates.
4. Looked at the PWM voltage at the two pin connector and saw the right voltage, 0-12V with about 40% duty cycle at 100 Hz. Duty cycle increases a bit when headlight is turned on to add a load to the engine.
5. For good measure, swapped the ICV with a known good ICV from a 89-C4. Bouncing idle remains
6. Swapped out the vacuum switch next to the ICV, bouncing idle remains.
7. Swapped out the entire ECU for a known good 89-C4 ECU, bouncing idle remains
8. Durametric scanner shows the O2 sensor signal to be 800-900 mV. Changed out O2 sensor and bouncing idle is the same. In fact, the Durametric reads 800-900mV O2 sensor signals on BOTH 964 cars. But the actual sensor is putting out 100-800mV (measured at the sensor). Of course, Durametric says they never saw that problem.

So far, I have eliminated the ICV, vacuum leaks, vacuum switch, ECU box, Fuel regulator, O2 sensor to be the problem. Don't quite know where to go next.

I did look at the wiring harness on the SC-3.6L car and found that it has a wiring harness for a Carrera car. The harness has a DME relay which is different from the DME relay for a 1992 964.

My next question is: Are the two DME relays functionally the same? They are physically different and wired a little differently.

Does anyone know the functions of the Carrera DME double relay? What does it do when the ignition key is off, and when the key is on?

It seems like ECU is making the idle bounce around in order to look for something at cold start-up. Then after a couple of minutes, it either finds it or gives up and steadies the idle at 850 rpm. Could it be the idle adaptation (on a 964 engine) parameters are lost every time the key is turned off?

[Solamar]

Have you checked the mixture at cold idle?

Does the 3.6 use the same socket as the 3.2 that you bridge two contacts to set idle speed and mixture? If so, make sure you go through that procedure correctly.

[jpnovak]

Didn't see on your list.

Did you verify that the idle switch is engaged properly? I have seen trouble with this on a few 3.6 swaps where the adapted throttle cable does not allow full engagement of the idle switch.

You can manually close the switch to test or pull back on the linkage to engage switch.

Make sure this is adjusted properly going forward.

[yelcab1]

Originally Posted by jpnovak

Didn't see on your list.

Did you verify that the idle switch is engaged properly? I have seen trouble with this on a few 3.6 swaps where the adapted throttle cable does not allow full engagement of the idle switch.

You can manually close the switch to test or pull back on the linkage to engage switch.

Make sure this is adjusted properly going forward.

Yes I checked the idle position switch and it clicks as I move it just a little bit from its rest stop position. The impedance changes state before and after the click.

[yelcab1]

And the answer is (this time, anyway):

The MAF sweeping arm and its resistive bands are contaminated with carbon. A good cleaning and the idle bounce disappeared.

There is still a closed-throttle stalling, but that is another issue.