Can't get D-Jet to run right at 2% CO

[ctaylor738]

I am hopefully on my last round of adjustments on my "barn find" 350SL. The details of its wake-up and what I did to it are in this thread:

Waking up D-Jet

Basically, I am trying to undo all of the adjustments that were made by a succession of mechanics to get the car to run with high fuel pressure caused by a kinked hose in the fuel return circuit.

My initial adjustment was to use the MAP to set the CO at 2% at 3000 RPM, and then use the thumbwheel to set the idle at 2%. This resulted in a rough idle and a stumble under light load, worse when cold. I could have a smooth idle in gear and a rough idle in Park by turning the idle air down a bit. Turn it up, and the idle would go rough in gear and OK in Park.

Then I ran into a very respected senior tech at local dealer at a German car event. His technique was to set the thumbwheel 5-6 clicks from the right and adjust the MAP to "where it feels good." So it did that, and I basically needed to go about 1/8 turn richer on the MAP. But then I was back to my lumpy idle in Park. But now it's intermittently lumpy. (I am not making this up.) It would be lumpy, go smooth, lumpy, smooth. I would say it was 2:1 lumpy:smooth.

So then I went one click rich on the thumbwheel, and the above ratio went to 1:2 more or less. One more click, and I have a smooth idle almost all the time. Drove the car. More responsive, smoother pull. Life is good.

But a subsequent CO check showed around 5%. I pulled a couple of plugs, and they were clean.

One other note, the car does not have a good transition to idle coming to a stop. It burbles a bit as the RPM comes down through 1200-1500. And it's a little lumpy at the same RPM leaving a stop on light load.

Any ideas on what's going on?

TIA

[Gurunutkins]

are you sure you have no vacuum leaks? I know in the first set of posts you said no but are you sure? is the air slider working correctly or has the previous owner compensated by blocking the air supply tube or closing down the idle screw. often the up and down at idle is caused by extra air getting in and the engine speeds up and slows down as it burns off with the set fuel. what does the vacuum look like at idle? does it fluctuate or if steady how low is it? as I remember you said you had a gunstone? meter, does that do CO or CO2 and does it show the ratio as that would help
cheers
Barri

[GaryF]

In your position I would start over with all the basics, check every engine signal feed to the ECU was within spec.
- I'd take the hose off the MPS and check its output while applying vacum
- Check TPS signal
- All temp sensors
- Signal from trigger points while cranking
- Good spark in all cylinders.
etc.

Once all of these are eliminated I'd start looking for a known good ECU and try swapping them.
I'm sure Bosch do a diag unit that plugs into the ECU harness but whether you can hire it from somewhere is another story.

[sjefke]

Also, how reliable is that CO meter? I remember you saying something about a certain protocol you followed in order for it to work properly (but I may be wrong). Personally, I would be happy if the car ran fine. 5% is high, but it is an old car after all. I would first run it for a couple of weeks and see whether CO changes behavior.

Bert

[ctaylor738]

All measurements at app. 60 deg F

The air intake sensor = 289 ohms, should be 260-340.

Coolant temp sensor = 33 ohms, should be 2100-3100 (!).

Throttle position sensor OK.

Engine vacuum at idle slightly over 19" and steady

The vacuum reading was taken with a Mityvac teed off the MAP sensor. Reducing or releasing the vacuum with the Mityvac had no effect on the engine's running or on the CO reading.

So I would say I need to get a new coolant temp sensor, and check some more on the MAP sensor. There are some resistance readings to test.

Thanks for the good advice, guys. I should have done this a long time ago.

[Gurunutkins]

yep the coolant temp sensor is toast. 19" of vacuum is good, the MAP sensor is slow to respond. the easiest test is to put the mity vac on it and pull as much vacuum as you can then leave it. it shouldn't drop noticeably for at least 5 mins.
cheers
Barri

[Gurunutkins]

yep the coolant temp sensor is toast. 19" of vacuum is good, the MAP sensor is slow to respond. the easiest test is to put the mity vac on it and pull as much vacuum as you can then leave it. it shouldn't drop noticeably for at least 5 mins. I have a good ecu I can post to you to try if you need to that I have just pulled off my megasquirted car but I will want it back in case the MS fails
cheers
Barri

[350SL4spd]

All very good advice here.

Another place to check for vac leaks is the intake manifold. The propane method is quite a bit safer than the starter fluid method. After how long she sat, the gasket could have dry-rotted a bit...

And you never know with the d-jet ecu's. I replaced tons of parts & adjusted pretty much everything on mine, before replacing my (obviously molested) ecu, and that smoothed out a good part of my lumpy idle. I'd still test all of the wires coming into the ecu to make sure you don't have any bad wiring first.

Good Luck.

[cth350]

Chuck, it's slow and tedious, but putting each wire in the harness through a resistance check is important. There are something like 26 wires to check on an SL.

While on the subject of coolant sensors. Don't forget the purely mechanical wax finger that controls the idle air bypass. Not that I can remember how it's supposed to behave after the warmup phase.

-CTH

[ctaylor738]

After looking at a WorldPac picture of the coolant sensor, I became suspicious what I actually checked. Turned out to be the temp gauge sensor. The injection coolant sensor is right next to it, and the resistance measure is 2490 ohms, which is right where it should be.

I put the Mityvac on the MAP and pumped it up to 19". After 20 minutes, it had dropped to 17".

On the MAP wiring

Terms 7-15 = 88.3 ohms, spec is 90.
Terms 8-10 = 339 ohms, spec is 350

So I am going to call all of these Close Enough.

I did a more thorough test of the throttle switch. It has the "switching" from zero to infinite resistance as the throttle is opened, but it's not like it's a clean switch. At idle, it shows 1.2 ohms resistance, spec is zero. I am suspicious.

Barri, a question. You seemed to say that the MAP is slow to react by design. Is this a correct interpretation? But shouldn't it be quick to react to a drop in vacuum like for acceleration?

[Gurunutkins]

Hi Chuck
the temp sensor is one of the two in the front of the manifold with 2 pins, the bigger one is the thermo time switch and the ohms readings are wrong for that so assume you have the right one which is the smaller of the 2. the temp sender for the gauge is right at the back of the engine and the auxilliary fan switch is at the very front on the thermostat housing.

the TPS should read no continuity at idle so if you have a reading its probably not set right. I will pull the pin numbers if you want this evening. I have included a picture of the insides of the TPS for you to see what you are working with. its the midle one for the mercedes.

When I said its slow to respond, I meant yours not generally. I have included some pics of a few i repaired for spares. the anaerobic bellows is the part that moves and often breaks along the fold lines. if you have a leak in the casing then the bellows cannot move as you wont build up vacuum, holding for 20 mins and dropping to 17" is good. as are the readings. you might just want to lubricate up the spindle a little to let the rod move in the coil a little faster. if you look at the pictures I have included when you turn the screw in the outside you are essentially lenghtening or shortening the rod which the bellows moves and thus changing the response to vacuum
there is also a whole section in the CD that tells you how to test the inputs to the ECU at the ecu plug just using an ohnmeter and volt meter if you dont have the grundig test machine. I can copy to PDF for you if your CD has not arrived yet, very usefull
hope that makes sense

I also have a question how do you guys get to post such big pictures, I just cannot figure it out, help appreciated
cheers
barri

[ctaylor738]

Thanks for the clarification and the pictures. I do have the CD, and used it today, but thanks for the offer.

According to my documentation on the TPS, terminals 12 and 17 should show 0 ohms at idle position, that's where I get the 1.2 ohms.

So now I guess the question is, whether the MAP is sending good readings to a dead ECU, or is the MAP dead? The ECU is at least alive enough to change the CO by turning the wheel.

[Gurunutkins]

have you tried resetting the TPS to get it down to zero by twisting it on the screws there is about 6 mm of movement there. the big thing is if the ecu is expecting a zero signal and is getting 1.2 ohms then it thinks the throttle os open and will over ride the MAP reading and richen up the mixture to account for the acceleration it thinks is going to happen (TPS over rides the map as the tps moves first before the vacuum changes). so untill you get a zero signal to the ecu at idle it wont really rely on the other readings.

try resetting the TPS and then retune her. if you need another MAP I have 4 good rebuilt ones and you are welcome to try one as a test

remember too that the throttle plate has to be 100% closed before you set the TPS as any idle air goes through the idle air bypass and not through the throttle butterfly. check the cold idle air valve as well to make sure it is shutting when the engine is up to temp. then set the idle to get the highest vacuum you can on your vacuum gauge by twitching down the idle air screw at the front. when you get the revs approx. right with the maximum vacuum you can get at those revs then the idle is right and you can set the TPS to zero
hope that makes some sense
cheers
barri

[ctaylor738]

I remeasured between 12 and 17 on the TPS, and today it's 3.6 ohms. I don't think there's an adjustment to be made because the throttle plate is fully closed and the linkage is against the stop.

So then I hooked the Mityvac and my meter up to the MAP sensor. One would think that the resistance would change with vacuum. Otherwise, how would this information get transmitted to the brain. So I would pump it up, let down, in general try to imitate changing vacuum on an dengine. There was no (none, nada, zip) change in either of the windings' resistance. But then there aren't any specs in the manual either.

This is consistent with the results of my test with the CO tester -- no change in CO with vacuum changes.

So does anyone know what should happen to the resistance with changes in vacuum?

[Gurunutkins]

Chuck there are 2 screws or bolts on the TPS itself and enough space to adjust it against the throttle body to get the 0 at idle. you just loosen both and then twist the TPS body. You have to set the throttle plate closed and then hold it closed and turn the TPS itself. In the pictures above you can see the 2 "ears" on the center TPS have long slots to allow it to twist.
I think the rod that rests against the anaerobic bellows inside your MAP may be sticking. the resistances you are measuring are just telling you the windings are intact. try splitting the 2 halves of the MAP and lubricating the rod thats inside the MAP
cheers
Barri