BMW M5 Forum and M6 Forums banner

Rough idle, worse with A/C off

22K views 63 replies 8 participants last post by  leem5 
#1 ·
My '99 M5 is idling rough at the moment and I noticed that it's a lot smoother if I switch the air conditioning on.

Any idea why that might be?

I'm thinking that the change of load on the engine could just be damping out the roughness or possibly the engine is just operating at a slightly different point in order to drive the A/C.

Could this shed some light on the cause of the roughness?

A little background: I also noticed that B1 and B2 LTFT are ~10% and 17% respectively as reported by Torque Pro, INPA is currently showing 1.11 and 1.18 'Adaptation value multiplicative'. These come down quite a bit after extended cruise on the autobahn. INPA is showing that the B2 cam timing is ~2degrees more advanced at idle than B1 which was also a surprise because I though VANOS should adjust them and then they should be equal?

I've checked for vacuum leaks and checked the O2 bungs for cracks.

Spark plugs, coils, MAF, Cam and Crank sensors, fuel pump, and VANOS solenoid valves are all relatively new. The car passed both the MAF and VANOS tests.

Any ideas appreciated, I will drop it down to the werkstatt soon if I don't figure it out.
 
See less See more
#2 ·
The symptoms you're having are identical to mine a few hundred miles before the water pump failed catastrophically. It's probably unlikely that you're experiencing the same, but figured I'd mention it because it's so similar to what I experienced. I honestly don't have any good theories for why my pump failed that way, particularly since the A/C compressor and water pump aren't on the same drive belt. The idle has been rock steady ever since replacing the water pump.

Anyway, can't hurt to check your water pump pulley for looseness.

Sent from my VS988 using Tapatalk
 
#3 ·
The idle speed is raised when the AC is on, likely why it smooths.
You have all the symptoms of a vac leak or bad mafs, but admittedly there are many other things that can do this. First thing I would do is electrically unplug the MAFs and see how it runs then how it changes over a day or two of driving.
Like the other posters problem anything failing that is driven by the belt could be causing this, might be worth pulling the belt and see if the idle returns to normal. If it does something driven by the belt is bad.
 
#4 ·
Thanks for the help. I looked at the pulleys and belts while idling and they look fine, I haven't run the car without them yet - is it safe to run the engine without the water pump running? not long I guess.

I'd unplugged the MAFs before but only long enough to tell that the idle was still rough. I now ran the car for ~50km with MAFs unplugged and no change. Again, is it safe to run that long without MAFs?
 
#5 ·
If you run with the MAFs unplugged you won't have air temp because it is in the MAFs. There are other things that can compensate. Safe ya sure as long as you don't beat on it, not sure that would be totally unsafe though.
Just watch the water temp with the water pump not being driven. You will know pretty quick if it is something on the belt. The only really hard one is the alt which could have an electrical fault and might try and produce full output at idle.
The rest of the items can be spun by hand and the resistance felt and some like the power steering would likely have symptoms.
For that matter you could watch the voltage on the OBC while driving and see what the ALT is doing. You should see a voltage drop while the rpm is heading for idle and it will recover shortly after idle is reached.
 
#6 ·
I dropped it down to the dealer since I don't have more time to work on this. Dealer says that one of the MAFs is defective and recommends replacing both (~600eur). Current MAFs are BOSCH 13621433566 units which I put in a couple of years ago when I was having issues with starting the car with a warm engine (it fixed the issue at the time).

I'd discounted the MAFs since they are relatively new, the 'MAF test' passed, and - crucially - unplugging them showed no change in behaviour.
 
#8 ·
Just got the M5 back after asking them to replace the MAFs and it does seem better, for sure the 'wobble' between 2-3000rpm seems to have gone and overall it seems much smoother. I'll reserve my full verdict until I've had more time to drive but so far I'm cautiously optimistic.

BMW said their tech had tested the MAFs on the bench and they were giving different readings which is why, even though there were no codes, they recommended replacement.
 
#9 ·
The problem is definitely not fixed. Car feel smoother >2krpm with new MAFs but idle issue is still there, actually it seems to be misfire at anything <1500rpm. WOT from 1000-1500rpm is very jerky, then it smooths out around 1500 and really starts pulling around 2000.

Also noticed that last two days that the throttle was very jerky after about 10 minutes driving - happened at exactly the same spot on my drive to work, and when I arrived at work I pulled 2D and 36 codes with the Peake (In addition to a persistent 8F code that I haven't been able to resolve).

Almost inclined to think that this is an issue with the DME, no idea how to confirm that without a known good unit to swap. Will start at the beginning again and go through vac, fuel, spark, etc again. It's getting real tiring though hence the less than fruitful trip to the dealer.
 
#11 ·
Dude, why did you not mention that before. 8 out of 10 times when the ebox fan fails people get no symptoms because it can fail in many ways. There is one failure that causes your exact symptoms, and that is when it locks up solid.


Clear those codes and see if they come back is step one. Those other two codes are very telling but could be a result of too much power beiig sucked up by the Ebox circuit.
 
#12 ·
Dude, why did you not mention that before
I ruled out the E Box fan code because it's been there forever and I've both replaced the E Box fan and tested the complete circuit related to the E Box fan and it all checks out.

The weird thing with the fan is that my car has the 'old' 3 pin E Box fan but BMW list the 'new' 2 pin fan for my build date. Makes me wonder whether the DME software is expecting and testing for the newer 2 pin fan.

curtissp321 said:
Have you ever replaced your fuel pump / fuel filter? If not - how many miles / Km on the car?
Fuel pump and filter were replaced ~10000km ago ('Preventative maintenance') I did notice that the pump is quite audible when I was under the car a few weeks back but not sure if that's normal or not. Testing the fuel pressure is on my agenda.


I think I need to go back and systematically check some of the basics again. With hindsight I think the dealer was not really motivated to diagnose a 20yo car so that was probably a waste of time. Probably should improve my German language skills and fond some local enthusiasts to help out.
 
#13 · (Edited)
I will take your word for that, but I would check the Ebox fan again just to be sure it is still running. Just in case you did all these tests last year and are assuming it still works.
One or both those codes say you are having voltage issue. It might be simplest to replace the unloader relay than try and test that system but I would also replace the main relay, seems cheap compared to your next test which might be to swap out DME, or have to go to a dealer where they will do a virtual swap.
Which fan did you put in the 3 or 2? has your car had updated software in it?
 
#14 ·
I'll check the fan again but I'm not expecting much, would be a nice surprise if it was that simple though. Note that I never got rid of the 8F code even with the new fan which I confirmed working at the time (old fan was also actually fine).

On the subject of the Ebox fan - the WDS diagrams show the 2Pin fan arrangement as being introduced in Sept'99, Mine is a June car, so I guess my car's not the anomaly and BMW made an error with their ETK.

Afte reading your comments in this thread Sailor (https://www.m5board.com/vbulletin/e...ode-36-battery-voltage-behind-main-relay.html) I'll check battery connections again - battery was out of the car until a month ago while I was on holiday. but this new behaviour doesn't seem to relate to bumps specifically. I took a different route to work yesterday but experienced the same behaviour and codes after about the same amount of driving time and it cleared up around the same time too. Thinking that must be temperature dependent if it's occurring in such a well defined window after cold start. Didn't do it on the way home after sitting for 6 hours. Interested to see what happens today.
 
#16 ·
Crud my failing memory is going to bite me again. Pretty sure it is most often refereed to as terminal 30, but maybe it is terminal 15, which my memory says is the unloader relay. It is one of the two but the best place to find it is on this page way down at the bottom.
https://www.realoem.com/bmw/enUS/partgrp?id=DE93-USA-02-2000-E39-BMW-M5&mg=61
Not going to guess for you because you have a 99 and they are different. Your next question will be is it the DME relay or the one called motor. Take the numbers and reference the wiring diagram the one that feeds the other is term 30 or the main relay refereed by your code.
 
#17 ·
I guess it can't hurt to check the relays since they're cheap.

Car was running much worse today the whole way to work and back, bad enough on the way home that I won't be driving it until I sort this.

The 2D and 36 errors keep coming back, INPA had a bit more information but I forgot to get a screen grab, will do tomorrow. Interestingly, 'Input voltage' was mid-11V to high-12V for the 2D error and ~5V for the 36 error. Assuming this would be the power to the DME.

Something that was odd though was that the 'battery voltage' reported by INPA fluxtuated between ~ 11.8V and 14V at idle. I though this was strange because test Nr 9 in the 'secret menu' shows a pretty constant 13.8V, I'd expected them to be the same. The Torque android app via bluetooth ODB2 adapter shows a constant 14V. Definitely something fishy going on with the electrics.
 
#18 · (Edited)
I wrote all below, then checked it to see if my memory was tricking me and it seems it was not, but I did find an interesting set of notes on a car with a single failed diode in the alt. Then a second failed and he had the car towed to me. I replaced the alt, and it was a beast again. That is the opposite of what I said below, don't know what to tell you, the notes are similar to your car. I have no recollection of the car at all.


It has been a while, I hope my memory holds up. Basically a car has two electrical systems or maybe better described as two sides. That is how the idiot light works, it shows the balance between the two systems. The OBC basically reports the ALT voltage 90% of the time, not really the battery voltage. Yes they are connected, but.... Sorry all I can really say is but.
Without a 5 page essay explaining how things work, take my best guess as you definitely seem to have a bad connection or a shorted battery, something wrong on the battery side of the two systems.



Step one clean your terminals on the battery. Disconnect it and charge it, then read the volts after you remove the surface charge. Easiest would be to connect it again turn the lights on for 30 seconds or so, then disconnect and read the volts. Perfect battery would be 12.9 good battery 12.8, usable battery 12.6 - 12.7. 12.5 might need another charge or charged at a higher rate but likely is not really very good any more.


That is your battery test if it tests good then you need to test voltage at different locations to help find where a connection is not so good. You can't use a volt meter alone because it does not draw enough, you need something that draws power and a volt meter together.


But maybe you just need a short cut, I have lots of them. If you have never noticed any issue starting the car and it turns over strong, we can likely rule out the battery and all the wiring to the starter. But there is a connection in the right side trunk compartment when looking at the battery. That is not ruled out so loosen the red and look at it clean it bolt it back up.


Next you have to get to the starter, even though we have power to the starter there is a second wire that sits on the same bolt, that might be where the problem is. It is not uncommon place for corrosion.Then after that spot you need to check the booster terminal that is where the ALT side of the power joins to the battery side of the power.
 
#19 · (Edited)
I've made a little progress - replaced the DME relay, old one had yellowish burnt looking residue on the contacts coming from the housing, and it pretty much eliminated the throttle issue but not completely and still got the 2d and 36 codes.

Figured it might be the DME so I arranged to swap a known good DME in to test and while doing that noticed that the 'last' plug on the DME was loose. Plugged everything back in properly this time and it feels like a massive improvement when driving. No more jerking, no more 2d and 36 codes (still got 8f though) and it feel generally much smoother when driving and pulls harder <3k. Again, I'll reserve judgement until I've had a bit more time driving because I've gotten overly excited more than once before.

The idle stumble is still there though and actually feels worse (more frequent) now. I reset adaptations with INPA so will wait a bit and see if that helps - idle was _really_ bad initially after resetting the adaptations. Will still try swapping in the other DME since I have it but now thinking I need to more carefully test for intake/exhaust leaks.

Feeling positive though! (also feel like an idiot for not plugging the DME in properly)
 
#20 ·
A good leak test for the whole car is to block the exhaust. I got lucky plumbing temp drain caps fit my Tubi exhaust perfect with a hose clamp. Then two pieces of plywood cut to the size and replacing the air filter finishes sealing up both ends of the engine. Now just add air pressure, dip stick is good if you have the correct fitting, it can be any where. I first added a air fitting to one of the exhaust caps, but since have made a little thing that clamps on the dip stick tube.
 
#21 · (Edited)
Swapped in the spare DME today and It definitely seems to behave differently at idle. It's still a little rough but INPA shows pretty much random peaks in the 'run unrest' screen with what looks like slightly higher on bank 2, which also has higher LTFT (~17% vs ~9% on bank 1). Thinking there could be some issue with the DME then but likely also a vacuum or exhaust leak of some sort. With the original DME there are clear peaks on cylinders 4 and 3.

Also noticed some oil leaking from the sump when I was under the car the other day - not enough to reach the ground but the bottom of the sump is a bit wet. Possible that this could be admitting unmetered air? I'm reluctant to pressurise the system until I fix that though.

The spare DME is actually a US sourced DME so has the wrong software for my European car. I'm getting the same LTFT as with the original DME though and also an '01' code for the fuel pump relay. I've read that the US cars use a different relay so assuming that it's caused by the wrong software and not actually the relay.

I was thinking that I'd try reflowing the solder on the coil drivers of the original DME, and I still might, however I did a bit more hunting and ...

UPDATE:

Just noticed that the 'software number' in the bin file that I pulled from my DME (7837964) doesn't match the software number shown in INPA (7837956). Does the number from INPA mean anything or can it be wrong? Is it possible that I have mismatching software on my DME causing trouble?

Also I compared the bin that I pulled with BMWflash to the 7837964 0DA file and they're basically identical except that mine has 7400rpm rev limit and 299kph speed governor (and checksum of course). Suggests to me that someone has flashed a modified 7837964 software in the past possibly without updating the other software components.
 
#22 ·
Have rough idle when warm, no codes, ICv changed and, no idle issues when AC is on. Car is RK stage 2 with evolve tune which is running rich (evolve need to fix the tune but their rolling road is shanked, or something). Took the car to a specialist in the north. . He said Vanos needs replacing based on Vanos software check .... hmmmm.... I didn’t believe him so I called another reputable specialist and he told me Vanos software checks often fail. The cars with, in my opinion, the most authoritative specialist in the UK. I’ll report what the outcome of his analysis / fixes are once I get the car back.
 
#23 · (Edited)
Well mine is not the DME so I suspect a vacuum leak. I had a code for warmup misfire cyl 6 yesterday. Since the INPA run unrest shows highest peaks on cyl 3 and 4 (right after 5 and 6 in firing order) I'm guessing a leak in or around the cyl 6 throttle body or connector tube with cyl 5. Will smoke test it again at the weekend and hopefully do a better job than last time.

https://youtu.be/k04Y0Esbeag
 
#25 ·
Smoke tested with the plenum removed today. I saw no smoke coming from anywhere other than around the throttle butterflies. It seems to me that most smoke is coming from the cylinders with highest 'run unrest' values in INPA, i.e. 3, 4, 5, and 6.

Cleaned throttles 3 through 8 but ran out of cleaner. Didn't seem to reduce the amount of leakage either, possibly slightly out of sync? I guess a leak here would mean each cylinder getting a different amount of air particularly at idle which would explain the LTFT and rough idle?

Suspecting the leak that I observed when smoke testing with the plenum in place is doe to the throttle body to plenum seal but will test again with the smoke when I have time to see if removing and replacing the plenum helped.

I was only able to take a short drive so will reserve judgement until I've had more time to drive.

Smoke: https://youtu.be/8MGce0d4dNI
 
#27 ·
I drove the car about 20km yesterday without the MAFs but it was the same. However I had reset the adaptations the night before and they didn't change during the drive so I wonder whether that may have affected things. When I later plugged the MAFs back in and drove a bit the LTFT came back up.

I would test more today but the car is booked for fixing the leaking sump. I'll get them to check out the leak too, figure it makes sense to fix the known leak before proceeding with anything else.

Videos of the latest smoke test: bit hard to tell exactly where it's coming from but I think it's plausible that it could be the plenum to throttle body flange.

https://youtu.be/vqI9zyH0OoQ
https://youtu.be/4GaSo_ssT_U
https://youtu.be/sUQFHJahP1Q
 
#28 ·
That does not look like it is a plenum gasket leak, I guess it could be small and defined but that small a leak above the throttles would have little effect on the car. That size of leak below the throttles would have a significant effect. Could it be the way you are testing with the plenum off be missing parts of the engine?
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top