W203 220 CDI Limp home - overboost problem

[JonnieBeeb]

Hi All,

My first post, and I'm hoping someone can help with the problem I am having. I have searched the forum and have not been able to find a definitive answer to what I am seeing.

A couple of weeks ago my 2006 220CDI (120,000 miles) went into Limp Home Mode. I scanned for codes using an iCarSoft scanner and I saw "2632-8 Check System Charge pressure control Charge Pressure too high" and a "2359-2 Check charge air system, Boost pressure too high". My first thought was the turbo vanes had seized, or the actuator had failed. I had heard this is a common problem. So I took off the air box to have a look. When I did I found the air pipe on the inlet to the turbo had come adrift, so I thought OK, this is probably what is causing the problem, air is going into the turbo that has not been metered. While I had the air box off I did check the turbo vanes. I disconnected the control rod and the vanes moved freely, no sticking or roughness at all. The turbo spun freely with no play in the bearings.

I reconnected the air pipe and turbo control rod and started the car up. I could see the turbo actuator moving up and down as I revved the car, so I feel happy the actuator is working correctly.

I took the car for a short drive. I went for a mile and stopped. Car revved freely and was happy to accelerate. On way back, it went into Limp Home again. I scanned for codes and I had the same overboost codes as before, so obviously the inlet pipe coming off was a red herring.

I have taken all the turbo pipes off the car and there are no splits or leaks in any of them. The intercooler looks sound too. I took off the EGR and a little bit sooty, but it was able to move freely.

I took the car for another drive while logging with my scanner. The same happened again, A mile drive with no problems and no codes flagged. On the way back, it went into Limp Home, and I had the same overboost codes. In the logging data I am seeing some strange numbers, I have pasted four frames, where I believe Limp Home was triggered. I am seeing some high Boost pressures, but also some very high boost air temperatures. I really can't believe those temperatures are real ! The Charge pressure positioner goes to 10%, which I presume is part of the Limp Home mode. When I stop and restart the car the boost pressure positioner goes back to 95%, so I don' t think it had stuck at 10% for a mechanical reason.

Frame 12/33:
4540 Engine speed ----- 779rpm
4547 Boost air temperature ----- 16.36°C
4592 Atmospheric pressure ----- 1003mbar
4609 Boost pressure ----- 2422mbar
4611 Intake air pressure ----- 974mbar
230 Charge pressure positioner ----- 95%
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Frame 13/33:
4540 Engine speed ----- 2975rpm
4547 Boost air temperature ----- 47.76°C
4592 Atmospheric pressure ----- 1003mbar
4609 Boost pressure ----- 2963mbar
4611 Intake air pressure ----- 976mbar
230 Charge pressure positioner ----- 10%
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Frame 14/33:
4540 Engine speed ----- 1962rpm
4547 Boost air temperature ----- 53.56°C
4592 Atmospheric pressure ----- 1002mbar
4609 Boost pressure ----- 2200mbar
4611 Intake air pressure ----- 990mbar
230 Charge pressure positioner ----- 10%
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Frame 15/33:
4540 Engine speed ----- 1567rpm
4547 Boost air temperature ----- 35.16°C
4592 Atmospheric pressure ----- 1001mbar
4609 Boost pressure ----- 1721mbar
4611 Intake air pressure ----- 989mbar
230 Charge pressure positioner ----- 10%

Does anyone have a idea what is causing this? Is it a boost pressure sensor, or could it be an EGR problem? Is that where the high boost temperatures are coming from? Can I drive the car with the EGR disconnected to see if that is the problem? Or will that cause the car to go into Limp Home mode anyway.

Thanks in advance for any advice.
J

 

[alexanderfoti]

Charge air temps look fine.

On your last freeze frame data, 1.7bar of boost with a charge positioner of 10% (or 2.2 bar on the one above!) seems implausible.

Either the charge actuator is sticking and this is the commanded position or your boost pressure sensor is failing. They do fail so I'd check this sensors readings ignition on engine off and see if it's realistic?

 

[mersum1es]

as above... plus are swirl flaps working?

 

[JonnieBeeb]

Thanks for the suggestions. The issue has the feel of a sensor giving spurious readings to the ECU.

@alexanderfoti I agree, the boost pressures look scarily high, surely if that were the real pressure something would have gone pop? I have looked at the boost pressure readings with engine off (around 1000mbar) and running at idle (again around 1000mbar), and when I rev the car I see boost pressure go to around 1500mbar).

I might just replace the pressure sensor as they are not "that" expensive.

@mersum1es I have looked for the swirl flaps and I just can't see the control actuator and control rod anywhere on the manifold. I have googled for pictures of what it should look like and I have had an inspection mirror to look at the underside of the manifold. I can't see anything that looks like the pictures on my car. Where should I be looking for the control rod?

 

[mersum1es]

I am not sure if all OM646 has flap fitted from factory, but Flaps might be removed too.

 

[alexanderfoti]

If its being read as absolute pressure (which it is as it reads 1000mb with engine off) it means 2200mb is 1.2bar of relative pressure, which is not that high. Peak stock pressure is probably 1.5bar so it's probably a sticking actuator causing the issue.

 

[JonnieBeeb]

@mersum1es The specs of the OM646 read as though they all had swirl flaps, but my car has been in the family since new, and I know it has never had swirl flaps removed, so I think it must mean some OM646 didn't have swirl flaps. It is one less thing that could be causing the problem !

@alexanderfoti I will take off the turbo actuator at the weekend, and see if there is anything obviously wrong with it. It must be intermittent if there is something wrong with it, as I can see it happily moving when I start the car and apply some revs.

 

[JonnieBeeb]

A quick update.

I fitted a new boost pressure sensor. A quick drive proved that hadn't fixed the problem, as the car went into Limp Mode after a couple of miles.

I have since taken off the turbo actuator, and opened it up. Clean as a whistle inside and nothing obvious wrong. No broken teeth on the gears and no dry joints on the connectors.

Kind of at a loss what else to look at. I'm reluctant to take it to a dealer, as they will happily empty my pockets and no guarantee of a fix.

 

[bennymechanical]

Hey Jonnie, I'm from Australia and I'm having the exact same issue with my 05 203 c220. The issue came on suddenly not too long after replacing the high pressure pump.

My issue normally triggers at a boost reading of 1700mbar so barely 8-9 psi of boost.

I have replaced. Turbo actuator, map sensor, maf sensor, baro sensor, EGR valve, cam position sensor, pcv. I have removed the intake manifold for cleaning.

I am yet to do a physical boost pressure test to confirm that the engine is making the boost as read by the ECM,

I'm going on three months now and I'm about to scrap the car entirely.

Let me know if you have any other insights.

Regards, Ben

 

[JonnieBeeb]

Hi Ben,

I still have not got to the bottom of what is causing this, and it sounds like you have replaced a lot of stuff and still not solved the problem.

I assume if you stop the car and restart the "limp home" clears and then the car runs fine for some time? I am working on the theory that it is caused by a "transient", something is causing the boost pressure to very briefly go out of range from what the ECU is expecting for a particular RPM/engine load.

I had stopped using the car as I have a few steep hills where I live. There are a couple of steep hills on my way to work, both 10% gradient and one is twisty and narrow with nowhere to stop if the car went into limp mode. I just couldn't risk using the car if it was going to let me down. I was then forced to use the car as my other car was off the road and I have found that I can take it for a long drive (30 miles) and the problem (usually) doesn't appear. What I have found is that if I stop and restart the problem often appears, so I am working on the theory that it seems to occur after a hot start.

I have realised the short test drive I was using, is (almost) guaranteed to cause the problem. The test drive is a 3 mile drive, I stop in a lay-by (engine off and restart) then drive back. Coming back, as I pull out of the lay-by I have to pull out onto a reasonably fast road, but it is also up a gentle hill, so I have to give the car a bit of welly. Nearly every time the car will go into limp mode. There is something about the "profile" of that short drive that causes the issue.

One idea I had was could the turbo vanes be sticking or grabbing when the turbo is hot and the engine has been stopped. When I checked the movement of the vanes they moved very freely, no roughness or anything to suggest they are seizing, but that was on a cold turbo. I'd be surprised if the vanes could behave so differntly on a hot turbo, after all the tubo is intended to get very hot, and the limp mode clears after an engine stop and restart.

My other idea is that there could be a momentary intake leak. I know that I am seeing an overboost, and you'd expect a leak to cause an underboost but I had read that a leak can also cause an overboost. The idea being that a small leak might not be enough to cause the ECU to log a code but it will command the turbo to provide more boost, but if the leak is suddenly closed it can cause a pressure overshoot and register an overboost. I feel like this is clutching at straws a little bit, but as I can't see anything else wrong I'm not sure what else it could be. I'm trying to think what is it about my short test drive that could cause this. The roads where I am are in a fairly poor state, with pot holes and bumps including on my short test drive when I pull out of the lay-by. Am I hitting one of those bumps that is causing a leak to briefly open up in the intake ...?

I visually checked all the hoses and joints on my car and they all looked sound. I also visually checked the intercooler and it too looked sound. I didn't take off the manifold, but what I'd like to do is pressure test the intake system. I need to get a large bung with an air connection on it that I can put into the intake pipe from the air filter box. If you do this you also need to block the crank breather hose that goes to the intake pipe and use a pressure of about 10psi. I would firstly check there are no leaks, then give all the hoses and joints a small bit of movement to see if there are any leaks. If you are also able to connect an accurate pressure guage, you could cross check the pressure sensors on the car.

Sorry I can't give you a definitve answer, but hopefully I have given you some things to think about.

I'm keen to know if you do get to the bottom of what is casuing this.

Regards
Jonnie

 

[bennymechanical]

Hey, thanks for the informative write up. Yes mine resets after every key cycle even if it's super quick. I have found that it doesn't matter the temp or length of drive it always occurs as I cross 1700 mbar on the boost sensor. I have driven the car with my scan tool tracking the data for 30 minutes without a hint of trouble until I add enough accelerator input to push the car past 1700mbar then it instantly jumps to 2200 or 2500 on the reading and the turbo actuator goes to 10 percent and it's all over.

What I have noticed is that even though the boost jumps the car does not feel any more powerful nor can you hear boost sound from the engine.

Today I did another round of data collection and notice that as the boost spiked so did the charge air temp. It went from 25c to 45c almost instantly.

These are the best leads I have
1. Mine played up for 2 months exactly like this without a single fault code at all until I tried the new turbo actuator, only then did I occasionally get high and low boost codes.

2. Before changing any parts I took the intake manifold off and cleaned out all of the build up. When I pulled the air filter to turbo inlet pipe off I saw that the turbo end of the hose that has the orange seal was all mangled. I carefully glued and taped the end to the point where it was well sealed. The car must have gone a full week without issue I regularly so a mountain pass. I thought the car was fixed the suddenly the issue returned. Unfortunately I touched many things so cannot be sure why it was fixed briefly

 

[bennymechanical]

The only part I changed that affected the car at all was the turbo actuator and the new part made the boost less predictable and a little bit of surging under boost. So the aftermarket actuator was actually worse than the original Merc one.

I agree that the issue is either air metering pre turbo or turbo boost scaling vs predicted boost curve.

Next step is to confirm that the ECM boost reading is the same as an actual mechanical boost gauge and the other is to get a smoke test done on the intake.

I am a Mercedes systems tech and a Ford master tech and I have seen many curve balls in my career but this is one of the trickiest.

 

[JonnieBeeb]

Hey,

After reading this forum and a few others, this overboost problem seems fairly common and seems to stump a lot of garages. As you probably know, the om646 is also used on MB Sprinters and I have read a lot of owners have replaced turbos etc when trying to fix this and still not solved the problem.

My feeling for my car is this is not a turbo issue, but as you say a control loop problem, the ECU is commanding a boost pressure, and it's seeing a value out of range. From what I have read, the tubo boost is just controlled by the turbo actuator - the engine wants a particular boost pressure for a particular RPM and engine load and it uses the turbo actuator to achieve that, and uses the intake pressure sensor to monitor the pressure. So you can ignore EGR, PCV etc. Not sure if this is correct, someone else may confirm if this is the case.

I also noticed on my car that the turbo inlet pipe has a habit of coming off, I'm surprised there is no lip on the metal to clamp the hose around. I was able to tighten the clamp sufficiently to prevent the hose from coming off. I have also bought a replacement orange seal, but have not fitted it yet. My seal hasn't disintegrated but it has hardened so might not be providing enough grip.

If you are seeing a pressure jump and a temperature jump in your data, then I would say the pressure jump is real and not a spurious/faulty sensor reading. As I'm sure you know, when you increase the pressure of a gas, the temperature also increases, so the fact you see a temperature jump is significant.

Regarding the aftermarket turbo actuator, I have read these need to be calibrated to the turbo, so that might be why you found that made matters worse. But I have also read that they don't need to be calibrated. It really depends if the actuator works in open-loop or closed-loop, and I have not been able to figure out which. If it's closed loop, then it shouldn't need to be calibrated, the actuator would report back to the ECM its position and the ECM would nudge the actuator either open or closed to achieve the required boost. If it's open-loop, then yes I can see the actuator would need to be calibrated. The ECM does know the position of the actuator as it can be seen in the logs, but is what the ECM has asked for, or is it the real measured position...

I think a smoke test would at least rule out leaks in the intake as this seems to be very common on older vehicles. It's often obvious as there is visible oil, but not always. I can't do a smoke test, but just a pressure test would at least tell me if I have leaks and if there is a leak I might be able to hear it without using smoke.

One other thing to consider is worn or damaged wiring. It's another thing to think about on an older vehicle. The wiring going to the turbo actuator would be a likely candidate.

 

[JonnieBeeb]

I might try a simple pressure test on the intake system which is easy to do. A spray bottle with water and washing up liquid, spray over the intercooler, hoses and joints etc. Run the engine with a few revs and look for bubbles.

 

[bennymechanical]

Hi, cheers for that. Unfortunately I don't have access to an good data to understand the full description of operations on these turbos. Based on looking at the internals of the actuator I belive that the ecu is in full control.

The data I'm seeing when the car is behaving sugguest that the turbo is given maximum attack angle from first start and that only changes if you back off the accelerator quickly, it appears to jump down in actuator percentage consistent with how harshly you back off engine demand.

Or it will or should curve off as you are achiving max boost, which I have not seen other than the aggressive drop from 95 to 10 percent

Intake pressure test will work fine. The smoke machine I used to use did both so in my mind those are the same thing. Only issue with the pressure test is that the pressure always gets into the crankcase no matter what you block off so make sure your oil cap is off, I have seen and done somd weird things experimenting with pressurising.

Can't rule out ECM and in terms of wiring I might just do a pin to pin load test of the turbo wires just for laughs if I can get hold of a wiring diagram.

 

[bennymechanical]

Hello,
I have some interesting news. I managed to get hold of the factory manual. I did a lot of reading and I understand a lot more now.

The turbo actuator is only 3 wires. Common power and ground and a single wire to the ECM.
These wires are reasonably straightforward to test though you will need to remove a few things to access.

I load tested all three wires and did a wiggle test while they were holding current. All came up strong.

So today I removed the plugs off of the rail and pump, and the cam sensor and map sensor. I did a relative compression test with my oscilloscope. The result was great. Kudos to my son for being the key turner.

I then removed the airbox ect to do the wiring tests on the turbo actuator. Then refit the original factory actuator. I reassembled and cleared fault codes.

I went for a drive and the car is perfect. The actuator is acting dynamically through its various percentages. The boost tops out at 2200mbar but I did hit 2400 once.

 

[bennymechanical]

What does all this mean for you. Well I know that the only 2 sensors that were not replaced until I had installed the aftermarket turbo actuator were the cam position sensor and the MAP sensor.

It is entirely possible that the map sensor was the issue and I created a seperate issue with the aftermarket turbo actuator. This feels most likely.

I may have disturbed the wiring to the actuator enough to remove a short or open circuit.

Perhaps my car is not actually fixed and I just had a weird lucky run for 20min over a mountain.

I will update if the car plays up again. I will also answer any questions or give you information from the manual if you need.

Best of luck to you and fingers crossed for me.

 

[JonnieBeeb]

Thanks for the update - fingers crossed you have fixed your car.

Interesting that the turbo actuator is only three wires. I had seen a youtube video where someone had shown them testing the three wires, and the third wire is a PWM (pulse width modulation) control line, which kind of suggests the ECM is not able to read back the position. All it can do is drive the actuator to a position and use the manifold pressure as feedback. That does suggest that the actuator does need to be calibrated and maybe that is why your problem became worse when you used the aftermarket actuator.

When you say you changed the MAP sensor, you mean the air flow sensor in the intake pipe between the filter box and turbo?

When I get an overboost fault, I was seeing 2963mbar ! The sensor that reads that pressure tops out at 3000 mbar :-/

My car went into limp mode on my drive to work this morning. I was able to stop and restart at some lights. I can tell when it has done it (before even trying to accelerate) as the engine note changes, and there is a real roughness to the sound.

I know this sounds weird but I have a feeling it goes into limp mode when accelerating and doing a hard right turn. It makes me wonder if there is a leaky gasket somewhere and the engine shifting on its mounts opens up the leak.

 

[bennymechanical]

Hey, so I thought it was a PWM or it could just be a generated voltage from the ECM. Yes I believe that there is no feedback for actuator position the data provided under turbo actuator is just what the ECM is commanding. The workshop manual says it uses the MAP (boost sensor with the little hose on the intake manifold) and the charge air temp sensor (located in the boost pipe near the fan directly in front of the intake manifold) to control desired boost levels.

As far as I can tell that's all the whole loop is just the two sensors and a PWM output from the pcm.

I just went on another decent drive over the mountains at night and everything was fine.

I will look for another genuine turbo actuator incase the problem returns.

 

[JonnieBeeb]

So it could have been your MAP sensor that was giving spurious outputs. I changed mine quite early on as it was a relatively inexpensive part, but sadly that didn't fix my problem.

I have bought an endoscope for looking into hard to reach spaces - the kind that plugs into a phone, with a small camera on the end. I will try and find some time to use it to see if I can see any potential leaks.