1991 300SL Idle Speed Puzzle
 

I am consulting with a good local shop that has run out of time and ideas on this car. It is a 1991 300SL five-speed with appox. 78K miles. Generally good condition. The car came in with a variety of hard starting, poor acceleration, and other running problems, all of which have been fixed.

The following parts have been replaced:

- fuel filter and pump
- coolant temp sensor
- throttle position switch
- decel micro switch
- cold start valve
- idle air tubes
- plugs, wires, rotor, cap
- EHA
- air sensor potentiometer
- OVP relay

The remaining problems are that when the car is started hot or cold, it will idle at 1100 -1200 rpm until you press and release the throttle, at with point the idle drops to around 650 RPM. If you turn on the AC, the idle goes very low and if you turn the wheel, it almost dies.

When the fast idle occurs, the voltage to the idle control valve is around 7 V. When the idle drops, there is no voltage. You can unplug the connector and there is no difference in the idle.

The ECU was swapped for a known good unit. A functioning idle valve from another car was tried, but made no difference.

I have checked the verified the operation of the decel and throttle position switches at the ECU connector. I drove the car extensively with my meter connected to the EHA, and it is functioning perfectly.

So, any ideas why the ECU refuses to control the idle?

Thanks,

Sounds like a bad idle valve. If you disconnect the valve it should default to wide open giving a very high idle. If it doesn't then that is an indication of it sticking mechanicaly.

Just forget it and sell the car to me.

:-)

I asked the shop about the idle valve. They said that they tried another one, but it made no difference.

You raise a good point. I assumed that on this model, with no power the valve was closed and that's the way it was designed. I will look into this some more.

[Added 9/9: per Bob Terry and David Poole, and the MB documentation, with no power, the valve is closed]

Additional Information
 

I took the car back to the shop yesterday because they had a similar car, but an automatic, which means the ECU is different. We swapped the MAS modules with no difference. The MAS module receives tach signal from the ignition module and compressor on and passes them to the ECU. So there went that theory.

The we dug around in the cabinets and found the MAS testing harness and a breakout box. So I was able to test a couple of things with the engine running:

Voltage from ignition module (tach) to MAS: 6.4V, spec is 7.5-9V.
Voltage from MAS to ECU (tach): 6.6V, spec is 5-7V.
AC signal from MAS to ECU 9.5V, spec is 5-12V.

I also tested the vehicle speed input from the speedo. At stop it was .77V. It went to 3.3V when the car started to move, then gradually increased with speed, like 4V at 50 mph. Spec is >1V at 13 mph.

So it looks to me like the ECU is getting good input on tach and speed. I do wonder about the low reading from the ignition unit to the MAS, though.

Thoughts ???

Found the diagnostic connector ...
 

This car actually has a 16 pin diagnostic connector, cleverly concealed under a cover on the right side of the firewall with the hood light switch sticking out of it. So I dug out my impulse counter and pulled the following codes:

CIS:

10 - Idle speed and throttle switch implausibel
4 - Air flow pot current implausible
18 - Current to idle speed valve implausible

Ignition

17 - Crankshaft position sensor defective

I cleared all codes and drove the car. The only one that returned was the 17. Not sure what this would have to do with the idle problem, but it's possible that this has screwed up the RPM signal.

chuck,is this the 104 engine?

Yes.

Put a new crank sensor in, cleared code from the ignition module. Same issue - high idle, then low idle, no response when AC is turned on, idle drops even further.

Code 17 from ignition returns if engine is started. No codes in CIS module.

Thoughts, ideas?

O2 Sensor in mine made a big difference in fluctuating idle. Also, does this have an air idle control valve? I took mine off and cleaned it out. Found a cracked hose coming from fuel assembly. Replaced it and now my '91 idles like brand new. I also put some cheap 1.99 copper plugs in. I did all of the other things you have listed above to no avail.

Just to re-state the problem. The car starts and idles at about 1500 rpm, hot or cold. Then at some random time, it will drop to 650-700. It will almost die when the AC compressor engages or the steering wheel is turned.

Today's update:

The DTC 17 from the ignition module turned out to be my bad. This needs to be checked with the engine at idle. Thanks to Bob Terry for pointing that out. So did that, and no DTC.

Yesterday, I re-checked all continuity and operation of the various sensors and switches that connect to the ECU, along with the operation of the idle and decel switches, all of which were OK. Also checked all power and ground to the ECU - all good.

So then I farted around with the throttle linkage. If you carefully move the linkage with the air cleaner removed, you can see that there is maybe an inch of movement in the cable before the throttle plate actually moves and cancels the "idle" signal to the ECU. But the way the linkage is arranged, the decel microswitch releases immediately.

I discovered that if I moved the linkage just enough to release the decel switch but not move the throttle, the idle would drop. After a bit of experimenting with the accelerator pedal, I was able to start the car like this. When I did this, the car would start and idle at 650-700.

So something about both switches closed is telling the ECU to raise the idle after starting. When the decel alone opens, the ECU decides to cut all power to the idle valve. David Poole tells me that this is a "limp" mode and suggests there is a rev limiter not working correctly.

Remember, doing this with two different ECUs.

His suggestion is to check the O2 sensor output to make sure that the lambda reading reflects the mixture, and to check the engine and chassis grounds.

Anybody got anything else?

Some background info
 

From the "Model Year 1990 Model 129 Introduction into Service"

#1
Should there be an open or short circuit of the idle speed air valve, or an interruption in the data exchange between the CIS-E and the EZL/AKR units, the power to the idle speed air valve in shut off. At that point the ignition timing is extremely retarded, so the idle speed increases only insignificantly.

#2
If the idle speed contact is closed and there is a malfunction in the idle speed air valve (increased idle RPM), a fixed operating mode will be chosen by the EZL/AKR unit. Thus the most retarded ignition timing possible will occur, preventing an increase in idle speed.
========================================

Based on this and the experience with the car, I think #2 is happening. But I am at a loss explain why. The car is shooting itself in the foot. The ECU is causing the increase in idle speed by opening the idle valve. Then it reaches a point where it notices that the idle speed is too high. So then it kills the power to the valve. WHY?

Solved
 

Per Steve Brotherton, if the voltage from the air sensor potentiometer at idle is out of spec, the ECU will not understand that the car is at idle and thus decide to go into a sort of limp mode, cutting the power to the idle valve and retarding the timing.

The spec is .5-1.2V output at idle measured across pins 1 and 2 at the pot. The easiest way to measure is to remove the back of the connector. I tested at the pot and found .07V. Adjusted it up to .8V by rotating the pot very slightly, maybe .1 mm. Now have perfect idle including AC compressor signal - no more drop when the AC comes on.

I tested the Key On values at the ECU harness, but did not test at idle because of the good readings from the EHA current. Lesson learned.