expl0rer

Mine? Don't really have one, tbh. Just adding my limited empirical experience/anecdotal evidence and what I've attempted to tackle carbon issues.

Considering all of the steps Eric has taken already, I'm not sure I could add any helpful tips about faulty hardware or software.


Edit: Actually, just after I posted this message, I thought, what are the chances that this might be pedal or throttle body related? I've seen AMG E55/E63 guys on a FB page post about driveability issues and some solutions point to pedal sensor or TB. Obviously, the E55/63 are not equipped with the same motor as the GL, but a TB problem or pedal sensor sending the wrong data to the TB causing the TB to not respond as expected could exhibit as Eric's driveability description. I would think there would be codes related to the sensor or TB, though.

eric_in_sd

Thanks gents for chiming in. I was beginning to feel a bit lonely in my misery.

Expl0rer, to your question, I have been running Mobil 1 for over the last 100k miles. Early on I was using 10W-30 because San Diego has such a mild climate, but noting that the viscosity at temperatures was enough different, I switched to 0W-40 at about 60k miles ago. I have been recently changing the oil filter every 5K; I used to not change the oil at all because enough was leaking down one or more of the intake valve stem seals, but when a Liqui Moly additive seemed to swell the seals, and dropped the leakage considerably, I started sucking out oil when it came time to top it up. Pulled two quarts last time. This brought the average age of the oil in the crankcase to 8K.

I surely understand why good oil would clean passageway deposits, but don't know how it would affect the combustion chamber.

Regarding the piston carbon: That is the only thing that (a) is visibly wrong and (b) fits. I am guessing that the carbon lends itself to preignition, causing the ECU to grossly retard the timing, and (maybe) the spark gap voltage at that retarded timing is low enough that the ECU registers it as too low of a primary coil current. I have noticed that it was always the same six cylinders, always skipping 7 and 8, and now one of the cylinders seems to have healed somewhat, with the coil current error only happening on five of the previous six. I am guessing the ECU simply shuts down both the coil and fuel injector when that error happens, retrying the coil only after restart. This would fit with the observation that the problem disappears almost every time the motor is restarted.

This all hinges on the software being stupid enough to not report what it has done. Given it is Mercedes, I would not put it past them.

When I tried dripping water into the front of the intake manifold, I noticed very little happening until late in the game; when I revved the motor, huge clouds of condensation came out the exhaust. I think the pooled water got sucked up in the action of the tumble flaps. It is fortunate not enough water went in to hydro lock cylinders 4 and 8; the risk of this is reason enough not to attempt this.

All the water seemed to do a little bit of cleaning. Some of the spark plugs look better now.

The motor stalled and would not restart; I opted to not push it and took out the spark plugs to let the cylinders dry. I tried to peer in with the borescope but the lens promptly fogged. I hope these adventures did not rust the piston rings.

Since I have nothing else to try, I will tap 1/16" NPT holes in the intake manifold, plugging with these:



I'll take one, or perhaps also get a brass set, to put a small hole in the center, allowing water to be dripped in. If right over the intake valves, there should be no risk of it going in the cavity below. Could be a useful thing to have as therapy for regular cold short drives. These are nice and small; their outside dimension is just 5/16". I was really proud of myself setting up a dripper siphoning from a gallon jug overhead; pity the setup did not work out.

Any other hardware to consider? MAF sensor? Pedal position sensor? A problem with the MAF sensor should manifest as a fuel ratio trim, while a pedal sensor problem should show up as erratic behavior, with the motor simply taking the erratic input and attempting to follow its orders.

eric_in_sd

Here are the plugs, with 1 at the upper left.


6 and 8 have gotten visibly cleaned. When I first pulled the plugs, all were equally blackened.

Note 7 looks no better than the others, which raises the question why it is not suffering the coil low current error.

expl0rer

The oil would not do much for carbon build up on the pistons' flat surfaces or valves but will help with the cylinder walls. The gasoline additives will be your main carbon defense for the valves and pistons, plus the cylinder walls during the combustion cycle. I don't think that you have a bad enough carbon build up on the pistons. I would not worry about cleaning them up, but let the fuel/additives do their thing gradually. It would take more than just a couple of tanks, though. I realize that I was not clear about the part the fuel plays in the carbon fight when I added that bit in my post. Apologies about that.

I think that inadvertently some amount of the gasses find their way into the oil, so having good quality oil to perform "clean up" only helps. I think we're both in agreement there.

The spark plugs don't look that bad, either, but you might want to either clean them or replace. Not sure how many miles are on them, but in any case if you measure them and they are in spec, I would clean them and reuse. At least in the short term.

About the water drip, I'd rather see water mist being injected simultaneously with fuel. Something that you could connect to the injectors which would sync the fuel and water entering/combusting once cyl at a time. That would require some engineering, though. If it were me, I'd leave the water out of the equation for the time being and use fuel with techron additive (Mobil/Esso, Chevron, Texaco, there are others too, not sure what's available in your area) and dump Techron in the fuel tank too. I think that trying to clean the carbon quickly carries risks which I would not be willing to take. I'd rather let the additives work slowly. If the carbon is in bigger chunks and does not burn, it is abrasive and could scratch the walls. Which reminds me... about rust in the cylinders, check with the borescope. You should be ok, but who knows. You can drop a teaspoon or two of clean oil in the cylinders of concern (through the sparkplug holes) and turn the motor by hand several cycles to spread the oil on the cylinder walls on the combustion side. This should remove some, if not all the rust from the walls. You can try vacuuming the rest of the oil and rust deposits (tube and syringe, nothing fancy), then run the motor to burn off what's left.

I've gotten in the habit (with my diesel before the GL, and after a while started doing this with the GL too) to take scenic routes around town to give the motor ample time to run up to op temp. Sometimes it takes me 4 times longer than going straight to the destination, but the fuel and time I am burning is cheap insurance against problems I will inadvertently run into if I did not let the motor warm up. And at least once a week I give the GL a run for an extended time and higher speeds to keep gremlins away.

eric_in_sd

I tried an extended idle (about one hour) with a concentrated (12 oz / gallon) Gumout Regane Complete (which is 30% PEA, 30% naphtha) but not much happened. They say chemical cleaners do work but are slow. What I have read is that the only remotely quick way to clean combustion chambers is water injection, though people do raise the question of how safe it is to dislodge chunks of carbon junk. One fellow on BobIsTheOilGuy said fuel additives wrecked his fuel system o-rings and another muttered about "dissolving the tank liner", both notions which I took with a grain of salt.

While I'm waiting for the 1/16 tap to arrive, I suppose I'll resume the idling with Regane (PEA is what's in Techron); it can't hurt at this point. I diluted it with about 5 gallons of gas. PEA is mostly good for cleaning the intake, as it burns up in the combustion chamber - but hey, why not try?

The spark plugs are nearly new. I changed them early on in this saga.

I believe the only steel in the combustion chamber is the valves and the piston rings, all else being aluminum. And aluminum+silicon for the cylinder liners.

Regarding engineering: I thought about checking on how long water + alcohol stays suspended in gas, to see if a recirculating system might keep it suspended long enough to go through the injectors. That certainly could be simpler than pulling the manifold, although it could turn out a whole lot more complicated! Part of the problem would be getting enough water in to clean the combustion chambers but not so much that the motor misfires. It could turn into a huge mess!

It's a real good question how the carbon got in there. I have not ruled out bad gas; I had done a partial fill with a kinda sketchy station in town, and once in the distant past I similarly filled up at a dodgy station and ended up with, according to the dealership, "carbon encrusted valves" which they graciously replaced for me under warranty. I think they were so generous because that vintage motor was prone to leaky valves - MB knew it was a problem.

I might try burning off a cup of the old gas to see if burnt snot remains behind. Or maybe it was the month or so of only short trip driving in the first cold days of winter, October - November. Or both.

Thanks for all the comments.

expl0rer

You would need several Regane treatments to see results. I doubt one tank of sketchy fuel was the cause of carbon buildup. The short trips might have contributed, but doubtful that 2 months worth is the sole cause. Carbon buildup is just something that happens overtime and if not negated by detergents/chemical treatments, will just happen. If you drive the car over longer distances, it will happen less, but it will, still.

I don't think that the carbon buildup you see on the pistons is that much, and am skeptical it is your problem's source. I'd look at the pedal sensor. Can DAS manipulate it? Some recommend to reset the position sensor with a procedure involving the key and pressing the pedal for so long, which also resets the transmission, I think. Not sure if the same procedure will reset both the sensor and tranny on the 164 platform. Also worth looking at the TB. It could need a clean up. Or did you blow seafoam through it already? I don't recall where you pushed the seafoam onto the system.

Max Blast

I also don’t think carbons your root cause. Even if you have a generic scan tool it should give you live data where are you can get a TPS read out to at least rule that out. If it’s acting nonlinear or erratic then you could rule it in.

with regards your plugs they just look like normal 200 K mile plugs…… Nothing out of the ordinary there.

eric_in_sd

The tool I have is the higher end Autel. It allows ECU programming, push commands to Airmatic, that sort of thing. I have not for the life of me been able to find anything that would specifically tell me whether the carbon is the culprit. The only thing is the freeze frame data when the coil undercurrent errors appear: It shows heavily retarded (positive number; I assume that is retarded? After TDC?) spark. So the ECU is retarding the heck out of the spark when the errors actually manifest.

Consistent with the sooty plugs is a sooty exhaust and rich mixture smell. That makes me think of preignition, not spark ignition.

When I have pulled the plugs in the past, even when high mileage, they look normal, with white-tan insulator. This is different.

I replaced the throttle body, which includes the position sensor. No help.

I hate to try swapping the accelerator pedal, as it is not cheap, and it seems to me that erratic readings from the pedal sensor would manifest in erratic throttle response, not misfire type failures in the motor. It also would not explain why the mixture seems rich.

Note if the combustion is incomplete, the oxygen sensors would pick it up as a lean burn - still excess oxygen in the exhaust.

Max Blast

Yes carbon in access may cause preignition and or dieseling, and in any case it will cause knock which the Knock sensor will detect and retard timing accordingly.

But now that would have to be so much carbon and it would be at higher RPM for significant knock to occur than what you’re seeing so I’m just not sold on that being the root cause.

In any case carbon bad. Legit streetcars has a new YouTube out where he runs 20 gallons of water through the motor to get carbon cleaned out but it doesn’t end well.

eric_in_sd

Water injection works to clean carbon but is not without its risks, that's for sure. As I figured out, it is not a simple matter on the GL's manifold.

I wonder how many weeks of continuous idling it would take to go through a hundred gallons of PEA-rich gasoline. I would drive it and rev the snot out of it, but it won't comply.

One of the problems with the vacuum hose approach is how the water is distributed to the cylinders. Even in the legit streetcars video you can see the vacuum port probably favors the front cylinders, on the left in the video, because that's where the vacuum line is.

expl0rer

I'm not sure I'd want to idle through a whole tank. Idling for long periods contributes to carbon buildup. While we're still talking about carbon, how about your MAF, oxygen an coolant temp sensors?

Disconnect the MAF and take a drive. You'll run rich with the MAF off (also contributes to carbon buildup), but if your problem is solved, you have your culprit.

If the MAF is not it, test the oxygen sensors and the coolant temp sensor, if you have not. I'd also monitor the coolant temp with your Autel. I wonder if you're running a little hotter than normal.

eric_in_sd

The motor will run with the MAF sensor disconnected?!?

Max Blast

It should run in open loop I think, only one way to find out. Other cars I’ve had run with the maf disconnected because basically that’s the failure mode when the maf fails.

expl0rer

It has to. Imagine the MAF failing while you're on the highway. If the motor dies, it's a disaster and a lawsuit in the making.

eric_in_sd

You're using your common sense. That has never quite been a guiding force for MB, now has it.

Thanks for the suggestions, guys. Looks like my tap shows up today, so the schedule got pulled in a little.

I forgot to note another symptom: During idle, the motor misfires intermittently and when it does so, burps out the intake. The burp is subtle, but clearly audible. Partial ignition of the intake charge on one cylinder, that is, extreme preignition?

Max Blast

That sounds like a ignition timing issue - or could be a timing issue in general. Have you ruled out skipping a tooth?

could also be as you say pre-det but I’d rule out the engine jumping time or a ignition module giving it wayyyy advance.

expl0rer

Injectors? Do the misfires happen on the same cylinder(s)?

eric_in_sd

How would I know? There are no DTCs besides those I mentioned. Camshaft timing adaptation?

eric_in_sd

No clue. I pulled the injectors and cleaned them with my fuel injector cleaner doohickey, which gives you a channel through which to spray carburetor cleaner while the injector is activated. All looked reasonable, spraying nicely, with no contamination around the injector tips. Cold stalling stopped after I did this, by the way. It is possible they were somewhat plugged and poor fuel spray led to the pistons carbonizing - or whatever jammed the injectors to the extent it did also carbonized the pistons.

I could disable fuel to each cylinder one at a time, checking to see if one only stops the hiccup noise and misfire.

Max Blast

i think cam adaptation values would be way out to lunch on the bank that is affected.

Has the reluctor wheel code you menstioned early up in this thread persisted? That’s timing related and could cause this.

expl0rer

You may have a leaky injector. I'm not sure if your injector cleaning doohickey could be used to diagnose one or not for leaks.

To disable them, would you use the Autel? Disconnecting the injector from the harness should give you similar results, but more painful on the knuckles. Although you may see codes thrown then.

eric_in_sd

I have not seen the P0336 DTC again.
I believe the cam adaptations were similar between the two banks. I'm pretty sure I would have noticed significant differences.

I believe the ECU is disabling the fuel injectors on five or six out of the eight cylinders (to protect the precious TWC) and so when the motor is limping it throws arbitrary errors.

The fuel injectors can be individually disabled electronically.
I am skeptical that there is a leaky injector because the soot on the spark plugs was uniform across all cylinders.

I might do the injector disabling diagnosis before pulling the manifold.

eric_in_sd

I finished the port water injection project and reinstalled the manifold. It works great.

I also got a much better endoscope. Here are the cylinder pictures, in order. I only figured out after the fact that it is best to have the piston at BDC for the picture, sorry.









Having spent some time idling and revving the motor while dripping water in - with a needle valve you can literally drip in one drop at a time, it's just enough to make the idle stumble when it hits.

The plugs look better, all except for #2 which is sooty. Could be that #2 is getting the most oil leakage in. Note the carbon in that cylinder is wet. The vertical streaks on #7 could just be carbon dragging up and down with the piston, or it could be scoring. Note also the flakes of carbon on #8.

I'm going to try forcing a higher idle - propping a 2x4 in there, I guess - and moving water in faster. This takes too long otherwise. I'll focus on cleaning #1, #6, and #8. Plus maybe #2; possibly that 3-D lump of carbon in the center is causing preignition.

I did notice that one of the cylinders had evidence of preignition inside the intake manifold. There was soot at the port. Being a dummy I did not note which cylinder that was.

At this point I am near throwing in the towel on this motor. The oil leaking on the intake valves is troubling. I could do the valve stem seals, but how much labor do I want to throw into this thing.

Mite put in a 550 motor, idk. Want to wait til spring anyway, so in the meantime I'll just tinker I guess.

Max Blast

I would look at one more thing before you throw in the towel… If you can at least idle it till it gets warm and then get somebody with a X entry scan tool in to do a compression check that’ll tell you. Nothing heinous noted in these nice pics though.
Could be resold as a core!

Although that 550 motor does sound enticing.

eric_in_sd

Oh, it idles just fine. It refuses to take stabs at the throttle.

I ran the "compression test". It's on the Autel. It returned satisfactory, with a (max minus min?) rpm difference of 12. Thanks for the prompt to try it; I was honestly a little nervous wondering how violent the test would be.

I figured out how to prop the accelerator pedal. Wouldn't you know it, the computer intervenes after a little while and assumes the pedal must be malfunctioning. You have to switch the motor off and restart. Anyway, I got it to idle at 3k rpm. This sped up the water injection. I think I might be on to something; when I started doing the injection on #2, it stopped the light misfiring. #2 is the cylinder with the sooty plug.

I over did the water on #2 a bit; I couldn't take a picture of the cylinder because the camera was fogging up; when it warmed a bit I could see boiling water on the top of the piston. Ha!

Here are the piston pictures after cleaning 1, 2, 6, and 8 (before is on the left). This time I backed each piston down to BDC. I was not, however smart enough to do them in firing order (even smarter: opposing cylinders, 1&6 5&3 4&7 2&8), so I spent a bunch of time turning the crankshaft!










Lookit all that oil on 3, 4, 6, and 7.

Note the before on #1: The piston top indents (for valve clearance) were filled with carbon.

I'm seriously wondering if that 3-D junk on #2 was causing preignition, because the motor definitely runs more smoothly. #2 plug came out still sooty though.

I will probably try more extensive cleaning tomorrow before finishing reassembly and trying a road test.