Uneven running. Rough run. Smooth run. Part 1.

Never ending story about N43/N53 series petrol engines and troubles with even running.. I believe, each owner of car with any engine mentioned has seen names mentioned above in error message descriptions! There is a legend, that the engine of this series running smooth – it’s mission impossible.

Yes, these engines has troubles with ignition coils. When replacing ignition coils and spark plugs, you have to follow, that the environment is clean and the procedure or replacement is strictly followed, the coils has to be improved with snubbers, to ensure, that coil’s overrun don’t damages the coil repeatedly (please, read the post about ignition coils).

But everything is not so easy…

I had some suspicions already for a while, and… Now I decided to confirm/reject them. Seriously. Well, to do that, I had to strain a bit.

At first, I purchased  ELM327 adapter.


Next – install the appropriate software in my computer. One, which read’s OBD2 06 mode.

Both activities didn’t caused any problems – both adapter and software started to perform as necessary immediately after installation.


Why is it necessary?
In short about basic issue: MSD80 detects misfire from change of speed of flywheel (for example, Bosch uses accelerometer).

First the system of Siemens misfire detection was introduced in MS43 management module, which was used in M54 engines.

Unfortunately both INPA and ISTA D don’t display the count of each cylinder misfire, no online monitoring possible.

The basis of my suspicions was following: what’s happening, if BMW (in this current case: E60 LCI, N53B30, manual gearbox, M suspension, 19″ alloy wheels) drives on quite uneven road. Does the uneven road surface can cause false error messages regarding engine misfire? 

The experiment.

In local city, I turn on the ELM327 adapter, make sure, that misfire counts for all cylinders are 0. Everything is perfect. I drive thru whole city, the engine performs in all modes: Homogeneous, Homogeneous lean (Lambda = 1,5), Stratified charge with Lambda till 3,0 .. 3,2. Then I stop  and check the misfire counts is 0 for all cylinders. Totally 20 km driven.

Then I turn to famous road with “very good” surface, turn in the cruise control in the 6-th gear, speed ~ 75km/h. To drive faster- no, I feel sorry for my car…

My experiment was completed after several minutes. MSD80 has counted in total 17 (!) misfire points (so, uneven running was detected at least 30 .. 40 times – see the description below), turned off Stratified charge, stopped to count misfires. The problem found – let’s drive to service center! But let me to point out, that I didn’t felt any incorrect actions in the engine’s performance (mass misfire for sure not).
BMW misfire counters

As you can see, the engine is broken down totally: regular problems in 1-st, 3-rd and 5-th cylinders. 3-rd and 5-th cylinders (additionally to new ignition coils and spark plugs) has brand new injectors – so it is hard to believe, that these cylinders can be faulty. If you check ISTA D, it will have records for 1-st/3-rd/5-th cylinders, ISTA will detect (performing test for cold injectors), that ignition coils and spark plugs are in working order, and – will suggest to change injectors for these cylinders.

After a while I repeat the experiment on my way back,but now I use different software. Again, in the beginning all cylinders with 0 events, after “crappy” road… See the image:
BMW misfire counters

In this image: misfire counters are 12 bottom lines, 2 pieces for each cylinder. Now “only” 14 legit misfire points (so 20 .. 30 times the unevenness of flywheel movement was detected).


Oh, stupid manufacturer!

No, the manufacturer is not stupid. BMW had foreseen, that road can be uneven. For this purpose the DSC/DTC module send signal “Rough road”, and for a while the misfire detection (as in cases of variable load and/or in the moments, when the throttle angle is changed) is turned off. Obviously, when you touch the brake pedal often and twist the steering wheel (or when wheels slip out), MSD80 receives signal, that the road is uneven. The Germans has not foreseen, that our roads (in all their width) are so uneven, that it makes no sense to brake or change the direction.


Bad news:

a) driving mid/low quality roads, MSD80 records error messages regarding uneven running;

b) the misfire counters for this error message (checking ISTA D misfire test procedures) don’t correspond to real situation;

c) when N43/N53 “catches” this error message, the Stratified charge is turned off for a while (but Homogeneous lean injection continues to perform with Lambda around 1,5), the fuel consumption grows slightly;

d) when the N43/N53 engine starts new session, it starts without Stratified charge mode (it is not very OK – lukewarm engine to force to work in Homogeneous lean: this is the most unpleasant mode, if we look from the point of view of fuel combustion), Stratified charge is restored after a while (5 .. 15 minutes) – that’s the good news.


The good news:

a) N43/N53 restores Stratified charge in a while (if you don’t change driving mode) or after change of driving mode (and after re-calibration of cylinders individual corrections). For notice: much lower misfire event count is needed to turn off the Stratified charge mode repeatedly;

b) this error message (not in active, not passive status) don’t stops the long-term trim formation and don’t causes any noticeable problems in engine’s performance (except slightly increased fuel consumption).


Note: if uneven run (real or fake) is detected also in Homogeneous lean injection mode, engine switches to Homogeneous injection mode and start re-creation of bank and cylinders individual fuel adaptations.


The flywheel is quite massive piece of iron. If the misfire is catched in idle run, the identification of faulty cylinder is quite effective. But by higher RPM, a jumble of one cylinder is not easy determined in time – it’s not easy to recognize the exact moments, when problem starts/stops. Because of this, the possible “guilty” receive coefficients. For example, this cylinder is guilty for 70%, but next one (in time measures) – guilty for 30%. When any of “suspects” cumulatively achieves virtual possibility of 100%, the misfire count grows for 1 unit.

If in the beginning (some cylinder always will be as first even in very chaotic distribution) cylinder happens to be possible “guilty”, then this cylinder has chance to be the dominant “guilty”. And MSD80 records list of possible “guilties” and their coefficients, and this list is saved during sessions with purpose to make life easier in next sessions with finding the possible “guilty”.

Interesting coincidence: the working sequence of cylinders is 1/3/5, it means, that all “defective” cylinders (in time meaning) are consecutive, it means, their working cycle creates full circle (in time meaning).

Siemens monitors misfire events in following way: counts (flywheel uneven running cases) density (for example, MS43: counts events of each cylinder to 200 crankshaft turns, total count of events to 1000 crankshaft turns). Accordingly – if the engine “catches” separate and rare misfire – it is not counted and recorded as error. As understood from MS43 description, that the length of individual memory buffer of each cylinder is 5 .. 10 seconds, the total memory register length – up to 1 minute.

Relatively good news – for misfire error message status become active, the road has to be very uneven for a longer while, short sections of uneven road will not activate the error status. But, as the real situation shows (as you can see in experiment described above), the misfire count reaches several tens of time in few minutes.


Morality of this story:

in our driving conditions these misfire detection systems gives false error messages, and cylinders recognized as damaged are NOT damaged – the system reacts to unevenness of road. Even engines in perfect condition will have misfires in one or several cylinders. Also the method offered by BMW (switching coils/spark plugs/injectors between “good” and “bad” cylinders will give fake results!

After half an hour drive on a normal road surface.
BMW misfire counters


Be careful – some MSV/MSD software releases have problems with diagnostics programs. Usually misfire is shown in cylinder No. 1 (sometimes also with real shivering/rough run).

Below- the image after 10 km driven with plugged-in INPA in live data mode (real shivering was not identified):
BMW misfire counters

As you see, the cylinder No.1 has 86 (!) misfire points counted!


Related entries:

Uneven run. Conclusion