Mercedes-Benz Forum banner

M137 Misfire saga (blue smoke and misfire at cold start)

5431 Views 108 Replies 4 Participants Last post by  tim687
I have got issues with blue smoke coming out of the exhausts together with misfires upon cold start.
This morning I undid the oil filler cap and cranked the engine; the oil filler cap started to dance around for a brief moment. By the time the smoke was stopped, the cap settled.
I suspect that I might have a blow-by issue, which is quite common for this engine.

The car also starts to misfire when the temperature is rather low; at around 7 degrees C. When the engine warms up a little, I can restart the engine and the misfires are gone.
I'm not sure if the misfires are caused by the oxygen sensors getting fooled from all the oil burning.

When the engine cold starts, the RPMs go high, as you'd expect, but then it is barely able to keep itself running. Pressing the accelerator does not rise the RPMs by the amount you'd expect. Sometimes the RPMs do respond to the accelerator pedal getting pressed and then the engine runs ok.

I used catalytic converter cleaner to clean my cats, as I do have some cat efficiency codes.

In document GF07.61-P-3011F
There is an section stating:
"Post-start enrichment compensates for the fuel which precipitates on the cylinder walls. The fuel-air mixture is enriched sufficiently until a certain engine speed has stabilized (e.g. 1500 rpm with no gear engaged at +30 °C coolant temperature for about 30 s)."

Post-start enrichment is not happening as the engine does not rise to 1500 rpm. It did do that, but then I resetted the ECU and it stopped doing that.
The cold-start checks on DAS look ok, apart from the ignition angle being to low and the post-start enrichtment and TWC heating not being active.
The retarted ignition angle is probably due to the incorrect mixture.

Does the ECU shut-off procedures (post-start enrichment and TWC heating), if it measures that they are not working?

I've cleaned the PCV system venturi valve about 10k~15k km ago and it stopped to produce smoke upon cold start after that, I've replaced it about 2k km ago, as the old cover was broken and oil was seeping down the engine.

Last november, one of the coil packs went and the car started to gradually produce more and more blue smoke upon startup. I guessed this was due to fuel and oil mixing or something.
It took me about 1 to 1.5 months to order/receive/install the new coil pack and I was driving around with half an engine until that point. I did not disconnect the injectors, so the cylinders probably got washed everytime on startup.

I verified, with an oscilloscope, that the voltage transformer is still working properly upon cold start.

I did notice that the oil was had a nice metalic cloud, during the last oil change. I thought that those where the ZAS actuators since they are ratteling like crazy when ZAS is turned on.

I tested compression on all cylinders and there where a few lower ones.

Values are in bar; update: my method of measurement is probably not right and I'll have to do these over again
1: 9.6
2: 8.2
3: 6
4: 9
5: 10
6: 10.1
7: 11
8: 11
9: 10.5
10: 9
11: 10.1

This weekend I did a smoke vacuum test and the results where quite confusing. With the PCV system connected, alot of smoke came from underneath the intake manifold.
With the PCV disconnected and my pinky plugged in the hose that comes from the oil water cooler inside the valley of the V12, there is no smoke.
My guess is that we are lifting a gasket of somesort, by using positive pressure (around 1 bar) at the crank case when it is actually designed to handle negative pressure.

At this point, I think I either still have a vacuum leak, my fuel filter is clogged/fuel pump is not working properly or my injectors are shot.

Any suggestions? I'm kind of lost at this moment as there where a lot of problems summing up at this point. (Ignition coils, voltage transformers, oxygen sensors)
See less See more
1 - 20 of 109 Posts
Any suggestions on tracking down the vacuum leak? I see alot of smoke coming from underneath the intake manifold (from my smoke tester) but I can't really see where it comes from.

Please do keep in mind that it only smokes big time, when the PCV tubing is connected!

I've replaced that gasket 6 maybe 7 times now, so you'd think that I'd be good at this point.

Also, the LTFT's are not ridiculously high or something

@Kraut56 do you have any additional insights on this issue?
It's the fuel pressure.
2.5 bars when it should be 3.4.
I've replaced the fuel filter, runs much better, but the fuel rail pressure only goes up to 1 bar without starting. When the car runs it will slowly go up to 3.6 bars. I'm still suspecting the fuel pump to be bad too, as it made a strange noise during the fuel pump test when DAS was connected.

How quickly does your fuel rail pressure rise?

I also redid the smoke test, this time with me blowing on the regulator. Took quite a breather to get it done, but I saw no smoke.
So, I've attempted to decarbonize the engine a little, by using water. This morning I cold started the car and NO MISFIRES and blue smoke! The car ran better than ever, but can still run a bit better.
It doesn't increase it's RPMs in the first 10/15 seconds of the start procedure and for the first 10 minutes of driving, the fuel trims are still all over the place (+20% and -8%)

I did get an electrical issue fault code on the lambda of cylinder 1-3, just wondering if that might be due to the water ingestion.
In the name of god, what is going on here xD.
The STFT also peaks right after the O2 sensor going to 0.0v
Line Font Parallel Screenshot Electric blue
See less See more
Driving the car hard seems to bring the sensor back to life somewhat.
Since this sensor is roughly 10k km old, I do not expect it to go bad.

Yesterday I resetted the fuel mixture adaptations and throttle body end stops and it performed like a beats. Engine was really happy when I ran it to redline a couple of times, during the 3 pulls that I did.
Today, well not all was well. Sometimes I'm doubting if I should start my car in the middle of the day to keep the engine somewhat warm, to make it run better.

So to be clear about the problems that I'm having right now;
  • Misfire at cold start seems to be gone (might be related to the high temperatures for the last month or so), but there is alot of engine knock at cold start and driving until partially hot
    • The DAS oxygen sensor graphs are not crossing the 0.45V mark exactly every second, sometimes they skip a peak or two. I'm not sure if this is due to sampling speeds, as the updates of other values are not particularly fast either
  • The RPMs do not go up to 1500 for the first 30 seconds until coolant is +30C degrees, which to me is an indication that something it still wrong
    • I think I've read somewhere that this doesn't happen when there are fault codes, but can't remember where
  • The ECU sometimes tries to protect the cats by shutting down cylinders 2 and 3, I'm wondering why it doesn't shut down cylinder 1 too, because that's on the same lambda
  • CEL comes on when driving on the highway and when driving a bit aggressive
  • It takes ~30kms for the engine to run like a proper smoothed out V12 after a cold start and I only drive 40km one way
  • I'm still hearing a hissing sound from a mysterious vacuum leak (or not)? M137 Hissing when running
I'm taking a look at the freeze frame data quite often, but that doesn't show the exact reason why it has shutted down a cylinder. So I'm having a hard time figuring out what is happening.

During a cold start test, there are a few processes that are not getting operated/controlled correctly
  • Heating TWC doesn't always seem to turn on
  • Post start enrichment doesn't always seem to work
  • SAI solendoids do not get operated
Please find a video here: Google Photos
See less See more
I've done some research and document gf07.61-p-3032f and gf30.22-p-0003e both state that the rise in engine RPMs are there to heat up the TWCs. In the video you are able to see that the idle speed detection is working, that the SAI is turned on and that both check valves should be opened. I've tested these valves and they seem to work fine.

So; is this not working because there where TWC inefficiency codes?
I was investigating lambda graphs and the car ran trash. I had misfire fault numbers of 50 to 60 on loads of cylinders. I turned off the engine to reset everything, coolant was around 70 degrees.

I resetted the ECU using the following procedure:
  1. Resetted the mixture adaptation and throttle body stops
  2. Loaded the correct implicit coding in DAS developer
  3. Turned off ignition and waited 10 seconds before turning it back on
  4. Verified if all the settings where still as I'd like them
  5. Waited 10 minutes before I started the car
The car started and ran flawlessly.

I only got the TWC efficiency codes when I drove the car.
I'll see what happens when I cold start it on monday.

Just a quick question for you to wonder on; was I to quick during my previous resets? I started the car back up immediately after step 1.


The STFT and LTFT are off massively. This leads me to suspect the fuel injectors, since one cylinder might be overcorrecting, leading to a low LTFT for one bank, which causes problems on the other cylinders that are working correctly.
See less See more
I recently spoke with an Mercedes expert and his guess might be that the variable valve timing system is not operating as expected.

For reference: Variable timing adjustment issues - M137 V12
@Dave2302 could you explain the throttle pedal issue your referencing to in this post?

@Kraut56 I've just seen your post regarding blocked cats and I've been having a lot of TWC Damage codes and I've got the feeling that the cats are blocked.
Could you explain what the O2 sensors do when the cats are blocked?
12 posts and no replies.
Thanks for your honesty, 12 posts and no replies and a V12 that doesn't misfire now, a numbering coincidence?

I've figured out the main problem: clogged up cats. I've disabled the diagnosis of SAI and post cat lambda's and the car did not misfire at all this morning.
I've been WOT multiple times and that usually resulted in the ECU shutting down cylinders, but not this time.
No check engine light, no shaking after WOT, nothing, only a sliky smooth V12 humming away

I did notice that my generic OBD-II scanner reported an LTFT of -25%, which never happened before replacing the fuel pump.
I'm still guessing that there are some injectors not working as they should.

Could somebody hint towards the fuel trim levels in SDS? I've only managed to find the mixture adaptation, which lists injection times, dwell times and such, but those are all well within spec so I can't seem to find any -25% LTFT.

In the mean time, I'll replace the boost sensor, camshaft sensors and the throttle body gasket to hopefully get rid of the vacuum leak.
See less See more
I'll just continue to update this thread with my findings, and hoping that somebody who has a hint can chime in.

I've replaced the camshaft position sensor on bank 2 (because bank 1 requires you to take off the intake manifold :mad:), replaced the 'boost sensor' of the throttle body, replaced the throttle body gasket, replaced the two PCV hoses that run from the oil water heat exchanger and I've reinstalled the new O2 sensor that I removed.

The car runs with a LTFT of -6%, which is quite a bit better than the -25% that I was getting, but the car is misfiring all the time now, which wasn't there before.

I did notice that when I idled the car, that the intake temperature peaked to 90 deg C, a vacuum leak after the MAF but before the throttle body?
Having chased many misfires on an M137, I feel your pain.
It's a great engine, but it isn't so great it it doesn't run properly

By the way, have you ever taken off the intake manifold? Those gaskets are super easy to have slip out of position and cause leak-related misfires too, and the gaskets themselves can fall apart due to corrosion if you've ever moved it. I think you saw my pics from the other thread.
I have! Multiple times. You can visually check if the gaskets are in place by looking down next to the manifold ports. The studs that hold the gaskets in place are visible with the manifold on and you can feel if the other ends are in place properly.

Also wanted to note the last time I had a cat issue (on an M112), it had literally cracked and broken, requiring it to be outright replaced. Unless you're able to inspect them visually and confirm they're still holding together, its possible no amount of cleaning will help fix the flow there.
Hmm, I do have an endoscope, but I don't think it's able to make the bend into the exhaust pipe.

Now, as discussed in another thread (M137 V12 oil burning), I think it might just be oil getting sucked in through the intake manifold causing misfires and blue smoke upon startup. When the engine gets warmer, the rough running goes away.

The engine isn't as clean as I'd like it from the insides, cam adjusters and ZAS might not be working properly due to oil blockage. The engine load is also on the edge of the acceptable values at idle, according to DAS.
See less See more
So, I've confirmed that the intake manifold does not leak anything.
So this makes the mystery of where the smoke from the smoke tester is coming from even bigger.

However, I did a few exhaust backpressure tests (by using the compression tester) and I did not find anything strange. The pressure was close to 0 all the time, but when I unplugged the O2 sensor of cylinders 10-12, the engine fired up and shutted down immediately after the startup procedure was completed. This was a strange behavior since that did not happen with the O2's of 1-3 and 4-6 unplugged.
I reinstalled an old O2 sensor and the fuel trims where back to 0%, unfortunately not for long. They are now -7% again.

This mornings startup was rather messy, but I've never had such a smooth highway experience in a few months as this mornings was.
I'll try to unstick the piston rings as the car sat for a year before I bought it, just to get the compression of the cylinders up by a few bars, and hopefully cylinder 3 will not sit at 6 bar, but at 10 bar afterwards

Oh yeah, I had marvel mystery oil in my oil and in my fuel tank
See less See more
The car does seem to run a bit better. I've had marvel mystery oil in my oil and fuel tank for half a week and during my 40km commutes I do feel like the engine vibrates a lot less.
Just a quick notice, my engine runs roughish when cold (little bit of knocking), but purs like a happy kitten when hot.
A second quick notice, keep in mind that I've gotten multiple codes on the camshaft position sensor of bank 1. I'm not sure if that sensor is very reliable, but I'll replace it soon

During my daily commutes I noticed that the long fuel trims where quite low, around -20% to -15%, so to diagnose things, I've made the following changes
  1. Reset the mixture and throttle body end stops
  2. Change the ignition correction to use 93-95 octane (which corresponds to 98 or higher) -> this will be a permanent setting
  3. Disconnect the MAF sensor to prevent the LTFTs from being learned
Again, due to the MAF being disconnected, the LTFTs stay at 0%, which is fine for now.

The STFTs hover between -10%~10%, which is perfectly fine. When under power one bank's STFT goes down to -20%, which is strange. Idling can be a bit of an issue, since the MAF is disconnected, it won't have all the feedback on the engine's performance, so it starts to misfire now and then, because of an incorrect mixture, I guess.
Another thing I noticed, the engine knocking when cold has been reduced greatly, especially under load.

The engine runs great with these changes, mpg not so (it went up to 12l/100km, instead of 10l/100km or lower), but the big question still is:
Why does this improve the engines running?

At this stage I'm thinking about one of the following things
  • Is the MAF faulty? I've only replaced the sensor, not the tube that the MAF is attached to. I think I've replaced it with Pierburg, which is the OEM of the intake manifold itself.
  • Is the MAF working fine, but is the engine compensating for the decreased air intake caused by bad compression? During idle tests in DAS, the air intake value was around the upper limit, so I think that this can be excluded.
  • Are the O2 sensors faulty? During exhaust backpressure testing, after I disconnected the O2 sensor of cylinder 10-12, the engine didn't run. This engine is known for being a cat and o2 sensor killer.

I would invite anybody, not just the M137 geeks, to shoot me with common sense things at this stage, as in the end, it's just another ICE.
See less See more
I just reminded that the fuel trims went very negative after I replaced the fuel pump and filter
Allow me to attach an video of some values during diagnostics. Please note the value of the lambdas, they are not jumping at exactly the same time

Also the self adaptation at idle speed. I just resetten the adaptations so I haven't driven the car, thus those values are not adjusting yet.

@Kraut56 @Astro14 @wallyp @Dave2302 @calgary
@Deplore could you please look at these values, like you don't know anything about the history of my car?

I know, 3 of the 4 cats are pretty much dead.
1 - 20 of 109 Posts