Replace fuel injectors and air intake boot, cannot get stable idle.
 

Hello,

For my 1989 W124 300E M103, I replaced all 6 fuel injectors (along with the holders, rubber o-rings that fit around the holders and the rubber cap) and the large rubber intake boot this past Friday.

I knew I would have to re-tune the mixture as during the first test drive, all prior issues of a slight hesitation were gone, but now my car would hunt for idle slightly when letting off the throttle to brake for a stop sign. (Repeatedly did this during this first test drive.) It even stalled twice, once when switching from R to D and once at a stop sign.

I brought the car back into the driveway, connected my DMM to ports 2 and 3 of the X11(?) connector and tried to adjust the mixture using the duty cycle feature.

CCW -> Leans out the mixture
CW -> Enrichens the mixture

Key in, ignition at position 2, engine not running the reading was 69.2%. (Federal controller = 70%, California controller = 85%)

Upon the first start-up without adjusting anything, my car started in open-loop with a reading of 49.4% and was rock solid during this time. Then shortly after start-up the car switched to closed-loop operation.
The number climbed from 49.4% all the way to 92.3% over the course of about 1-2 minutes and remained at 92.3%.

I started a process of turning the hex screw at the mixture tower CCW to lean out the mixture slightly, about 1/4 turn at a time.

I could never get the mixture number to stop climbing to 92.3%. After a few iterations of turning the mixture screw CCW (leaner) in 1/4 increments, my idle quality started to suffer. (It was starting to get to the point where my idle would be about 350-400 in R or D, the engine was on the brink of stalling.) I have a few things to attend to this afternoon so I just enrichened the mixture (CW) about 1/2 turn from my last setting to restore my idle quality and keep the car from stalling.

A short test drive and the car runs well enough, no stalling or low power, but it definitely running rich. (i think....)

What should I test next? Or am I not being sensitive enough with the adjustments? Does the EHA need to be adjusted now?

I think you are confusing the lambda signal with the O2 signal. When lambda is over 50% the engine is running lean and the ecu is enrichening the mixture. Reading under 50% the engine is running rich and the ECU is pulling fuel out. Reading ~90% lambda indicates the engine is very lean either due to the adjustment being way out or possibly a vacuum leak inadvertantly caused during the previous repairs. After ruling out vacuum leaks you'll be fine to enrichen it to approx 45-55% lambda.

I would probe the O2 sensor wire under passenger seat and ensure that you are getting a good reading first before adjusting the EHA (unless you know it has been tampered with before).

So what you are saying is that my understanding of the duty cycle reading is wrong?

What I did think:
A duty cycle reading of near 100% indicates overly rich conditions.
A duty cycle reading of near 0% indicates overly lean conditions.

What I need to think:
A duty cycle reading of near 0% indicates overly rich conditions.
A duty cycle reading of near 100% indicates overly lean conditions.
AND
Duty Cycle = Lambda?



I am fairly certain that all possible vacuum leaks are taken care of. In the last 3 months I have replaced every single rubber elbow and connector from the firewall to the radiator, and the one that connects to the transmission. Both ICV hoses have been done, and now the intake boot and fuel injector seals. I guess the only possible leak is from the mixture tower itself since the small factory ball has been removed. I JB welded the rough port where someone had drilled into the side of the mixture tower to remove the taper-proofing ball. I can't imagine there being a vacuum leak somewhere but I will have to double check everything, including the seating of the new rubber intake boot to the throttle body.

I replaced the O2 sensor at about 206,000 miles, but I suppose it does not hurt to check, maybe I got a slow or defective sensor... :confused:

Yes, your understanding is wrong. High % = lean (i.e., most of the time spent at 0 volts); low % = rich (most of the time spent at 1 volt). Mercedes is opposite what most others are in how it defines duty cycle.

Ok, I understand it now, thank you all for clearing that up.

I just need to turn the hex wrench the other way! :D

(After testing the O2 sensor and checking for vacuum leaks.)

Okay, so I checked for vacuum leaks using slow burning incense sticks (lots of smoke, should be able to tell where it goes... tests areas behind firewall as well as in engine bay, nothing appears to be leaking...

Tested the O2 sensor, it appears to be giving quick updates within specified range...

After testing for vacuum leaks and making sure the O2 sensor was okay, I started to adjust the 3mm hex screw at the mixture tower again.

DMM in duty cycle mode (%) probed at 2 and 3.
Key in, reading is 69.2%
Start car, Park engaged, reading holding at 49.3%

After a short while, percentage starts to creep up. I stop the engine when it hits near 70%, I turn the hex screw CW 1/4 turn and then restart...

I do this for about 4 iterations, but it gets to the point where the percentage still rises, engine appears to be running worse than when I started... on my fifth one, car barely starts runs pretty rough, yet percentage rises again from about 50%, working its way to 70%ish before I turn it off...

On the 6th iteration, CW again 1/4, but the engine no longer starts...

I then turn the hex screw 1 full 360 degree turn CCW. Engine starts, but again percentage keeps rising at idle in P...

The problem...

If I turn the hex screw CCW, the percentage keeps creeping up until it gets to the point of no start...

If I turn the hex screw CW, the percentage keeps creeping up until it gets to the point of no start...

This is all at idle, in park, no input to the throttle... I would think that one direction would give me a number that increments to 0% and another that increments to 100%....

Any ideas? :o

Using the Static Test procedures found here:
http://landiss.com/mixture.htm
(2nd to last section "Static Tests")

I found that when I pushed down on the air plenum, the duty cycle reading became: 9.3% (10% is desired)

I found that when I actuated the throttle linkage to WOT, the duty cycle reading became: 19.2% (20% is desired)

So it seems that my car passed the static tests.

Also, strange... Port 2 of the X11 diagnostic connector is Ground... yet if I put the red probe of my DMM into this port and the black probe of my DMM into port number 3, I get inverted readings... meaning that at ignition position II before engine start, the reading is 30.6% instead of 69.4%...

I have to put the red probe in port 2 and the black probe in port 3 to get the correct readings... yet this is backwards? :confused:

Does anyone have a picture of which vacuum line needs to be unplugged as per this article: http://landiss.com/mixture.htm

I am not able to get the duty cycle to stop working its way to 92.3%.

Nothing I do with the 3mm hex screw at the tower is changing anything.
(I push down with the hex key, feel it hit the screw underneath, engage it and then turn.)

The car runs decent now, but the idle quality is crappy.

I retested the oxygen sensor.
Results:

Heat coil resistance: 6.1 ohms.
(Tested by plugging in the red probe of DMM into one side of heater wire connector and black probe into other side of heater wire connector.)

Sensor signal at key position 2: 0.498V
(Tested by putting red probe into oxygen sensor signal wire connector, black probe to lightly sanded seat bolt for ground.)

Sensor signal at engine start: starts at around 0.5, works its way to near 0V.
I started the car, engine was in open loop as it was cold, oxygen sensor almost immediately started working its way lower in voltage reading. Took about 1 full minute to hit about 0V. (0.05ish) What I found to be weird is that even if I revved the engine, it would not budge much from 0V, maybe went to 0.1V once. Either this sensor is bad already (10K miles) or my car is running EXTREMELY lean.

Ah ha! It was the oxygen sensor all along!
I swapped in a Bosch 13942 unit (1990 Ford Mustang 5.0), did some splicing and wah lah...

Kinda pissed my initial replacement only lasted 10K, that is the last time I buy a Denso unit.

Duty cycle now jumps between 45% and 55%ish. :vbac47679

Next Problem: Why aren't the aux fans coming on anymore... :rolleyes: