BEW intermittent rough idle (not cam)

[mountain lion]

The last 6 months I've noticed an intermittent stumble at idle only after lightly reving (after a small reverse or slow move forward especially) and it has been slowly getting worse/more frequent. The RPM drops down between 830-870 and the engine kinda chugs then jumps back up to 903 by itself. It first started only when engine was cool, but now seems to do it regardless of operating temp. Before we think worn out cam, see below:

2004 BEW - 258k miles

Installed at 238k miles:
Franko6 stage 2 cam (original cam still looked pretty darn good)
Bosio 1043 Nozzles (original pd100 injectors)
VNT-17
Darkside full SMIC upgrade kit
DieselGeek timing belt kit
RC 2.5 tune

Cam, nozzles and TB kit were all meticulously installed by a pro on this forum(with some assistance from me).


Things I've checked:

Lift pump is working properly
Tandem pump is the original LUK, but no leaks and car starts the instant the key is turned
No air bubbles seen in fuel return after injectors
Cam/lifters are basically new, but verified no signs of wear
Doubled checked injector connections and harness
Injector health is good: group 23 switch time deviation is around -25 to -30 on each. Group 13 injection deviation is all balanced to 0 and highest quantities are around +/- 0.30.
I had an oil analysis done back in February and everything checked out good, no fuel or excessive metals in the oil.
VCDS has no errors, MAF checks out ok, MAP and boost are good, coolant temp sensor seems to be working fine.
Timing is spot on based on TDC refence marks we put on the camshaft hub and cover after the job.


The torsion has been at -0.5 since the TB job and other upgrades. The last two weeks I decided to start adjusting it. First I went up to +0.5. Car lost FE and the rough idle got worse compared to -0.5. I went up to +1.5 and the idle was even worse there. Went the opposite direction and went down to -1.5. Things got a lot better here, but not perfect. Ended doing a lot of tests at .5 increments all the way down to -4.4. Car idled perfect at -4.4, but lost a LOT of low end. Went up to -3.4 and it also ran great. -2.5 is also pretty good. I've since settled at -2.9. FE is higher, but have lost some low end grunt (as expected). I would say it's pretty much running like a top here.

I'm at a loss as to what could have changed or be wrong at this point for the torsion to have to be adjusted. There were no issues for 16-17k miles since all the new upgrades were put in.

Is it possible the tandem pump has an issue? I did all the tests from the fuel system write up (except pressure readings), and everything pretty much checked out. I was able to pull a bit more fuel through with engine off but lift pump on (85ml instead of 25ml, but someone else had this exact amount on their bosch pump in that thread). There are no high RPM/load problems, car pulls strong all the way through.

I've checked everything I know, not sure if adjusting the torsion is the fix due to upgrades that have settle in or just masking another issue.

 

[oilhammer]

Well I would have never thought "worn out cam" since a mechanical item like that cannot come and go.

My guess would be possibly a flaky CKP sensor, but aside from putting a very fast scope on it, you may not be able to ever detect it. Luckily they are cheap and easy to install. Just make sure the mating surface of the block is clean and smooth before you put the new one in.

Less likely, and these are more apt to flag a DTC, is the CMP sensor.

CMP = cam position

CKP = crank position (sometimes called an engine speed sensor)

 

[mountain lion]

Oilhammer, thanks for the advice! I just ordered a new CKP and will have it in and installed tomorrow morning.

Any chance this sensor would potentially impact fuel economy as well? I have noticed from my logs a steady decline in FE.

I will get the torsion back to -0.5 / 0 and also try advancing it tomorrow. If I set it up at +1.5 I should be able to tell pretty quickly if it is fixed based on how poorly it ran previously.

Will report back. Thanks again!

 

[totalloser]

The symptoms sound so much like an air issue, I'd double check hose routing. If the hoses are routed wrong, air might not be purging. And even a tiny amount of air can cause an intermittent stall if it's not being routed back to the tank via return.

I had this *exact* problem due to previous owner getting the hoses mixed up at the filter. Theoretically air would purge through the injectors eventually, but when it only happens with a high flow rev pushing bubbles downhill, it could take a very long time. And the way the PO set it up, bubbles were returning from the filter back to the tandem pump.

Also, hoses might pull air only when revved due to increased vacuum. And if routed correctly, it might be hard to catch it happening. IDK. Just sounds *so* much like air. But *water* could cause something similar if the filter was totally full. :O

 

[mountain lion]

Thanks for the reply - I'm open to investigate anything further, but I had replaced the fuel filter when this first started about 8 months ago thinking it was something in the fueling department, but no improvement. The fuel hoses are all correct and I temporarily ran clear hose from and to the filter looking for air recently, but saw none. I've checked everything pre and post tandem pump that I know to at this point.

We'll see if the new CKP works tomorrow, if not it was worth a shot, and it'll be back to the drawing board.

 

[Ash03]

My car used to have the same issue (Pd130 Asz), the service happened to be due at the same time. I purged the injectors at the service using liqui moly diesel purge and the problem disappeared.

U can also double check the injector to rocker clearance..?

Sent from my SM-G531H using Tapatalk

 

[Ash03]

Or the injector loom, i read through quite a few posts of loose injector plugs causing similar issues.

Sent from my SM-G531H using Tapatalk

 

[mountain lion]

Well I would have never thought "worn out cam" since a mechanical item like that cannot come and go.

My guess would be possibly a flaky CKP sensor, but aside from putting a very fast scope on it, you may not be able to ever detect it. Luckily they are cheap and easy to install. Just make sure the mating surface of the block is clean and smooth before you put the new one in.

Less likely, and these are more apt to flag a DTC, is the CMP sensor.

CMP = cam position

CKP = crank position (sometimes called an engine speed sensor)

I got the new CKP installed today and it looks like Oilhammer nailed it.

Interestingly, my reported torsion value was a full point higher with the old sensor. I was at -2.9 and as soon as I installed the new sensor, VCDS reported torsion at -3.9! I'm guessing the torsion value is calculated in some way with the CKP and CMP together? I've advanced torsion reading up to 0.0 and everything seems to be running perfect now. Took a 40 mile trip this afternoon and fuel economy on the scan gauge is significantly higher. I'm hoping this is it once and for all.

Should I bother putting a new CMP sensor in proactively or let things go until I get a code thrown? I figured these things only last so long, and if the CKP was skewing measurements, maybe this thing is not far behind.

Thanks again.