Rough Idle - Diagnosis needed! - PeachParts Mercedes-Benz Forum

Bob Albrecht · #1 03-15-2014, 09:01 AM

After driving to work I noticed the idle on my 350SD was very rough. It happened suddenly. There was no gradual degradation in idle.

That morning I drove over some rough ice chunks in my driveway that shook the car up pretty good. I thought it might be some dirt in the fuel tank washing off the sides and clogging the filters. So, I replaced both fuel filters. They were dirty, but replacing them didn't help.

Here are the symptoms.

1. On start the idle is very rough. If I give it a little fuel to bring the RPM up the idle smooths-out.
2. Under normal driving conditions the car is just fine and the engine is smooth.
3. When I come to a stop the idle is smooth at first, but when the RPM drops below about 750 (to about 600) the idle gets rough again.
4. If I give it a little bit of fuel to get the RPM to 750 the idle smooths-out.

Here is what I am thinking:
What is the normal RPM for idle? Is it somehow adjusted too low?
Are my fuel injectors dirty and not as well-balanced, which is causing the rough idle? (It has been about 60K miles since I rebuilt them). I should probably clean and re-balance the injectors, but that is a lot of work if they are not the likely cause.

What are your thoughts?

barry12345 · #2 03-15-2014, 10:20 AM

Disable one injector at a time when the engine is running rough at idle. It may be a cylinder is dropping out. Or on a long prolonged rough idle read the temperature of the glow plugs looking for one that is much cooler.

A milli volt reading of the voltages at the harness connection will find it if there as well and is more sensitive at locating it than a temperature scan. It is just one other possibility. It is also a very fast test.

Since this is a 3.5 engine this should be checked. I hope it is not what it may be. That still does not mean it should not be checked. You may be describing the onset of a bent rod symptom. .

Bob I have limited ability so I tend to resort to simple tests. If when running rough you provide 12 volts to the glow plug system and the engine smooths out it would be semi conclusive to me that a bending rod onset is a probable scenario. You have to add the voltage in some way that enables a heavy normal current flow to the glow plugs. Technically you could do them one at a time. This may have the benefit of locating the offender as well if there. The nice thing about this test if it does not stop the issue it may mean it is something else.

I do not know where the air intake is for the air filter on this model. Is there the slightest possibility you could have experienced some ingestion of water? These engines do not have strong connecting rods. Well they do but not to deal with a design situation Mercedes introduced in my opinion. Actually there may be two of them but that is just my opinion.

Nobody ever asked me why they bend so I just observed all the speculation. I have held my own opinion for a long time but rather than get into a situation of what I think the possible cause is and the resultant derision initially have kept it to myself.

I assume normal idle was in the 600 area before and the idle remained good. If it went lower than it was earlier is another scenario in itself. I added this as an afterthought when reading the following post. I just assumed and that is dangerous that the idle speed was as it always was until the roughness shows up.

mach4 · #3 03-15-2014, 10:33 AM

Edit: misread the IP. Solutions suggested don't apply.

Bob Albrecht · #4 03-15-2014, 11:16 AM

Quote:

Originally Posted by barry12345

semi conclusive to me that a bending rod onset is a probable scenario.

Before you all go off on another rod-bending thread. This engine has the stronger rods from the later 350 engines. I know because I rebult it. There is _no_ sign of "rod bending", and believe me, I know all about it.

barry12345 · #5 03-15-2014, 11:26 AM

That is good to know. Initially I wondered if you were aware of this issue with these engines. I have no interest in a rod bending thread actually. The stronger rods are not considered bulletproof.

It also is good to know you have a fair amount of experience. I still would do the test on that engine. As stated earlier I really hope it is not the issue. On the otherhand it is also a very easy test to do with your experience.

Bob Albrecht · #6 03-15-2014, 02:59 PM

I put a full tank of fresh fuel in and the idle is somewhat better. I still think it might be a dirty injector.

barry12345 · #7 03-15-2014, 05:07 PM

There may be less air being introduced if an air leak where the culprit. The fuller fuel tank is much easier to draw on by the system.

I would also intently have a look at the primary filter for any noticeable air bubbles in transit.

If you have overflow on the return line from the injection pump. If that line can be temporarily stopped to see if the rough idle is essentially the same.

Does that model have the somewhat notorious fuel heater that develops leaks in a lot of cases? Or is that a later or only other model exclusive?

Bob Albrecht · #8 03-16-2014, 08:12 AM

Perhaps it is the lift pump. I am going to have a look at it.

Does anyone think it could be a bad injector?

When I replaced the primary filter there there were no bubbles. I probably should bleed the injector lines after replacing the filters, but the last time I did that it made no difference. The system is supposed to be self-priming.

There is an overflow return line. How would blocking it help?

Yes, it does have a goofy design for the fuel heater. There is a piece of plastic that can crack over time. The seals can also go bad, which is what happed to me. I replaced it not too long ago, so I checked that. If it is bad it leaks slowly and can drain the IP, especially if it has any weak seals.

barry12345 · #9 03-16-2014, 12:36 PM

Quote:

Originally Posted by Bob Albrecht

Perhaps it is the lift pump. I am going to have a look at it.

Does anyone think it could be a bad injector?

When I replaced the primary filter there there were no bubbles. I probably should bleed the injector lines after replacing the filters, but the last time I did that it made no difference. The system is supposed to be self-priming.

There is an overflow return line. How would blocking it help?

Yes, it does have a goofy design for the fuel heater. There is a piece of plastic that can crack over time. The seals can also go bad, which is what happed to me. I replaced it not too long ago, so I checked that. If it is bad it leaks slowly and can drain the IP, especially if it has any weak seals.

Blocking the output temporarily tests to make sure the relief valve has not gotten so weak the low internal base injection pump fuel pressure is causing idle issues.

If this does not correct the idle and the return line you are blocking was flowing fuel prior to the closing off. Then an assumption can be made the fuel pressure is in a workable range or not a problem. Actually this test does have a few issues but in general can indicate things. The ideal test is more complex.

Anyways the easier test only really clears the relief valve of fault. It does not clear your lift pump of any suspicion totally in your case as well. How the engine acts with the line closed does indicate things though but you have to think about the reaction you are seeing.

Since your model does have that fuel heater. Submerging the fuel return line in a container and watching for any air being processed with the fuel is the best test for air in your system I can think of. There is no chance of error in interpretation. That heater is well past the primary filter so there is no other way to see if air is getting into it. Unless the return line is pretty clear.

I cannot dispute your thoughts on a dirty injector. Reading the milli volts from each glow plug at the harness when the idle is very rough would locate the injector if questionable.

The milli volt reading would be very low on that glow plug. The ground for the meter must be to the engine itself. Or very high if the injector was staying totally open I suspect.

A guide to the suitability of your existing glow plugs being used for testing is the range of readings obtained from all the other glow plugs. Idealy they should be a newer matched production batch.

The real world does not always allow this though. So you extrapolate what you have. This is a good way to check injectors for a dud. Since it is a dynamic test. It will catch an intermittent injector where a static or straight test probably would not.

I have to use tests as I am not smart enough to determine what the problem is by guesswork. When combined with buying expensive parts that are not needed or shot gunning problems. I resort to testing instead as I find it less frustrating.

Not being a working mechanic is a big disadvantage. I have no intuitive feel developed over the years that they have developed. So I compensate for it the best I can when required. It seems to work out but again I have endless years troubleshooting things other than mechanical under my belt.

Given that you have several choices at this time. When the idle is rough I would want to examine those milli volts readings. As it is so easy to do and can give useful information or eliminate things.

That is where you save money and time. Only takes a five dollar meter to do it. You may even have a meter now. If you want to get technical two readings would be vastly superior. One reading with the normal idle and another after the roughness has been established for a minute. The meter ground must be to the engine.

barry12345 · #10 03-16-2014, 12:39 PM

Quote:

Originally Posted by Bob Albrecht

Perhaps it is the lift pump. I am going to have a look at it.

Does anyone think it could be a bad injector?

When I replaced the primary filter there there were no bubbles. I probably should bleed the injector lines after replacing the filters, but the last time I did that it made no difference. The system is supposed to be self-priming.

There is an overflow return line. How would blocking it help?

Yes, it does have a goofy design for the fuel heater. There is a piece of plastic that can crack over time. The seals can also go bad, which is what happed to me. I replaced it not too long ago, so I checked that. If it is bad it leaks slowly and can drain the IP, especially if it has any weak seals.

Blocking the output temporarily tests to make sure the relief valve has not gotten so weak the low internal base injection pump fuel pressure is causing idle issues.

If this does not correct the idle and the return line you are blocking was flowing fuel prior to the closing off. Then an assumption can be made the fuel pressure is in a workable range or not a problem. Actually this test does have a few issues but in general can indicate things. The ideal test is more complex.

Anyways the easier test only really clears the relief valve of fault. It does not clear your lift pump of any suspicion totally in your case as well. How the engine acts with the line closed does indicate things though but you have to think about the reaction you are seeing.

Since your model does have that fuel heater. Submerging the fuel return line in a container and watching for any air being processed with the fuel is the best test for air in your system I can think of. There is no chance of error in interpretation. That heater is well past the primary filter so there is no other way to see if air is getting into it. Unless the return line is pretty clear.

I cannot dispute your thoughts on a dirty injector. Reading the milli volts from each glow plug at the harness when the idle is very rough would locate the injector if questionable.

The milli volt reading would be very low on that glow plug. The ground for the meter must be to the engine itself. Or very high if the injector was staying totally open I suspect.

A guide to the suitability of your existing glow plugs being used for testing is the range of readings obtained from all the other glow plugs. Idealy they should be a newer matched production batch.

The real world does not always allow this though. So you extrapolate what you have. This is a good way to check injectors for a dud. Since it is a dynamic test. It will catch an intermittent injector where a static or straight test probably would not.

I have to use tests as I am not smart enough to determine what the problem is by guesswork. When combined with buying expensive parts that are not needed or shot gunning problems. I resort to testing instead as I find it less frustrating.

Not being a working mechanic is a big disadvantage. I have no intuitive feel developed over the years that they have developed. So I compensate for it the best I can when required. It seems to work out but again I have endless years troubleshooting things other than mechanical under my belt.

Given that you have several choices at this time. When the idle is rough I would want to examine those milli volts readings. As it is so easy to do and can give useful information or eliminate things.

That is where you save money and time. Only takes a five dollar meter to do it. You may even have a meter now. If you want to get technical two readings would be vastly superior. One reading with the normal idle and another after the roughness has been established for a minute. The meter ground must be to the engine again remember. If the readings remain the same rough or smooth then is the time to look elsewhere in my opinion.