Greetings, everyone!
I finally got around to probing my car with a digital multimeter in an effort to diagnose the rough running issue.
Symptom:
Rough idle, especially under load. Virtually undriveable. Engine smooths out when revved to about 4K RPM.

In addition to the EHA valve, I just replaced the O2 sensor.

Here's what I found:

Current to EHA, Key On/Engine Off: 19.6mA
Current to EHA, engine at idle: -1.5 - 3.5mA (varies)
Current to EHA, idle & O2 sensor unplugged: 0.4mA

X11 port, Pin3-Pin2 (measured in DC volts, since I don't have a duty cycle function)
Key On, Engine Off: 3.8V (about 72%, assuming 13.7V battery)
Engine idle: 6.5V-8.2V (rises steadily from 6.5 to 8.2, then drops back to 6.5 every 5 seconds or so)
Engine Revving (varying speeds): 4V - 11V; varies
Engine idle (warm): typical range 6.2-7.2V

I discovered that the tamper-resistant cover on the mixture adjustment tower had already been removed, so I tried adjusting the mixture a bit. Using a 3mm allen wrench, I gently pressed down on the screw until I could feel it engage and I could feel resisance to turning. I adjusted the the mixture screw about 1/2 turn in either direction with the engine running and while monitoring Pin3 of the X11 port. Oddly, the voltage at Pin 3 remained virtually unchanged at 6.8V, and the engine did not seem to run any differently when he screw was turned. It seemed to have no effect.

I also tried unplugging he EHA and O2 sensor with the engine running. Neither seemed to have any effect; same rough idle.

Can anyone make heads or tails of this? Your input is sincerely appreciated!

Mika

What are your motor mounts like?

I have tampered with the 3-mm allen before - half a turn in either direction is quite an adjustment. On my car just a small incremental turn made a change. Regarding your comment about the engine not responding to your fiddling - You don't feel this change instantly. I know it takes (the computer ) time to react. Were you patient enough ? Hopes this helps.

The motor mounts are pretty shot. That aside, the motor runs rough. At idle, I can hear occasional thumps. It sounds almost like someone is kicking the engine every so often. On pressing the accelerator, the engine stumbles badly, but doesn't stall anymore.
Mika

How long should I have waited? I probably waited about 15 seconds... is that too little?
Mika

After each small turn, I mean a hair, you need to blip the throttle and wait about 30 seconds or more for the engine to settle again. That is how you do it. Quick and dirty way is to measure the voltage at the battery with the engine running. Cut that voltage in half and there is your %50 duty cycle value. Basically try and get close to that. You may need a new Lambda tower if your value is swinging that far. They are only a few bucks and fairly simple to install. The lower the voltage the richer the engine and higher voltage is lean. Clockwise to en-richen and counter clockwise to lean!

__________________
~Jamie
_________________
2003 Pewter C230K SC C1, C4, C5, C7, heated seats, CD Changer, and 6 Speed. ContiExtremes on the C7's.

1986 190E 2.3 Black, Auto, Mods to come soon.....

Can anyone describe how I can test the O2 sensor I installed?
The sheath around the 1-pin connector makes it difficult to probe when connected.
If I probe the singnal pin when it is disconnected, will it tell me anything useful?
Thanks,
Mika

Pull the 1-pin connector apart. Ground your multimeter to the chassis, and read from the 1-pin line from the O2 sensor. You should be seeing fluctuations, mid-point being around 0.45 volts

__________________
1990 190e 2.6 - The only one I can really call "My Car".

1987 190e 2.3 - The donor car's up and running, only mods are Euros and a Sony headunit. My Dad's runabout now.

1990 300e 2.6 - The parents' ride.

Thanks for your reply. Should I see fluctuations when the sensor is cold, or only once warmed?
I belive I measured as you described, and got about 0V....
Mika

Sorry about that.....yes, you do need to wait until it gets warm. Basically the engine needs to be @ normal operating temp, and you want to have it there for several minutes before testing the signal.

__________________
1990 190e 2.6 - The only one I can really call "My Car".

1987 190e 2.3 - The donor car's up and running, only mods are Euros and a Sony headunit. My Dad's runabout now.

1990 300e 2.6 - The parents' ride.

Thanks for the reply. The O2 sensor checked out OK - roughly .55V when warm, changing slighly.
I got a hold of a duty cycle meter, and rechecked the mixture. I adjusted the screw until it averaged 50%, and verified that it remained there for about 1/2 hour. The value did bounce around a bit, occasionally climing up to 65%.
I also noticed that the idle would occasionally rise up to about 1500 RPM.
Any more thoughts?
Mika

No, you won't see any fluctuation with the sensor disconnected because that disables the entire lambda system.

If you disconnect the O2 sensor the system will run on the basic mechanical mixture only, which should be slightly rich, so the sensor voltage will be constant and high.

It's best to leave the O2 sensor in the circuit by pulling the connector just enough to attach a small test lead to the exposed pin. Now you have a parallel connection and the circuit is undisturbed. With a scope you should see the sensor voltage jump back and forth between about 0.15 and 0.85, and you can evaluate its response time. It should complete about one to three cycles per second at idle. With a voltmeter it should read the average output which is about 0.45 volt.

The first rule of testing an operating circuit is that the test set up CANNOT effect circuit function or your results are garbage.

Duke