Skip to content
Shop
Explore

INDEPENDENT BMW ENGINE EXPERTISE

B38. Mini. Idle

N43 / N53 · NOXEM · DIAGNOSTICS

Select a technical image to view it in detail.

In my introduction to the MINI with the B38 petrol engine, I mentioned an odd idle behavior. This article follows up on that problem and on what happened after the DME software update.

The problem

The car vibrates strongly and constantly at idle. The whole interior visibly shakes, and the driver’s seat belt hits the door trim audibly. Oddly, the vibration drops sharply when the clutch is pressed.

It did not look mechanical (a gearbox or clutch vibration, for example). It looked electronic, because the vibration faded and returned gradually over several seconds. Diagnostics in ISTA Expert mode supported this. The DME does not balance the mechanical efficiency of the cylinders until the clutch is pressed, although it should do the opposite and start evening them out when the clutch is released. The reaction to the clutch was therefore the reverse of what it should be.

The dealer and BMW AG

The owner took all the technical information to the BMW dealer: ISTA diagnostic screenshots and a description of the problem. The dealer’s specialists confirmed the problem and opened a PUMA (technical assistance) case.

That sounds hopeful, since BMW AG as the manufacturer should help. But after two weeks the answer arrived, and it recommended comparing this behavior with other cars of exactly the same specification. Is the dealer or the owner supposed to go looking for a car with the exact same specification? A fantastic proposal.

The owner updates the software himself

Seeing this “professional” attitude, the owner took the initiative. Using the newest ISTA release, he found a DME software update intended for this specific vehicle. The dealer had at first volunteered to flash it, but eventually declined because the manufacturer had not confirmed (by a technical assignment) that the procedure should be performed.

Communication with the manufacturer is slow: it took several weeks to get an answer of one sentence. The owner had already driven a vibrating car for several months and did not want to wait another month for the sentence “We recommend updating the software”. So he did the update himself. He paid for the online work session, downloaded the data from his car, and used original equipment bought with his own money. When the manufacturer answers one sentence per two weeks, and without much imagination at that, there is little else to do.

The update covered both the bootloader (obviously the manufacturer is fighting “chip tuners”) and the main program. That sounded promising, because a “regular” bug like this should have been removed long ago.

This article describes how the MINI DME behaves with the newest software release. It is also meant as information for the dealer and for the manufacturer, in the hope that somebody finally acts.

Behavior after the update

The owner decided to be careful: no unnecessary steps, and as much information as possible for the manufacturer’s technical support. For that reason he did not re-adapt the engine after the update. No document gave any justification for it, not even a hint that it was necessary.

Subjectively, the car now vibrates both with the clutch pressed and with it released. The vibration changes, but it is hard to tell what is happening. Is the problem “solved” this badly? What is going on?

To get an outside view, I joined the process. Subjectively, the vibration is permanent and strong, and it changes when the clutch is pressed. A nightmare.

ISTA Expert mode shows something different, though. The permanent vibration occurs ONLY with the clutch released. When the clutch is pressed, as in the first test session, the efficiency of all cylinders evens out.

Here the clutch is not pressed, and the strong permanent vibration can be felt:

The 2nd cylinder (green curve) may be slightly less active, but the difference is only a few percent.

Now we press the clutch:

Subjectively the vibration seems to decrease. In the data, however, the 1st cylinder (blue) is clearly less active than the other two, by more than 5%. How is that possible?

Evidence that the flywheel adaptation is wrong

Even in the newest ISTA release (as of the beginning of May 2022), the flywheel adaptation status is not available for the B38, unlike the B48 and B58. The only logical explanation is a problem with the flywheel adaptations, so we looked for it in the data.

Overrun mode at 70 km/h (73rd to 80th second):

This looks very suspicious. The blue line is “lower” than the red one, which should not happen. Zooming in:

The flywheel signal is clearly wrong. One more control test, at 100 km/h:

Overrun starts at the 149th second. There is no doubt that the flywheel adaptation data is incorrect.

At this point I cannot tell what happened. There are two possible scenarios:

  • The manufacturer never adapted the flywheel, and it has worked incorrectly all this time.
  • The flywheel adaptations were deleted or damaged during the software update.

In either case, nothing required a re-adaptation of the engine. The description of the re-adaptation is limited to “drive in different modes”. The flywheel adaptation procedure is not described at all, and the flywheel adaptation status is not available. Is that not an extraordinary situation?

The current owner also likes sporty driving. He regularly brakes with the engine and uses Overrun mode. After the update he drove more than 500 km, with many engine-braking events, and the flywheel is still unadapted, or adapted completely incorrectly. What is going on, BMW?

Idle without vibration is impossible with incorrect flywheel adaptation data, so the logical conclusion is that the flywheel must be re-adapted.

The menu even has a separate subsection for the flywheel adaptations:

Impressive. The only question is where the learnbits are that show whether the flywheel is adapted. Where is the exact procedure, and where is the data with accurate measurements? There is nothing of the kind.

The re-adaptation procedure

Since there is no hint whether the initial adaptation was correct, or whether the update damaged any adaptations, the logical conclusion is that a complete engine re-adaptation should be performed.

The funniest part of the description is its last sentence (you can read it yourself when you run the procedure), which says the report may be incomplete and/or inaccurate. Look at the description on the right side of the image. It does not say how to perform the engine re-adaptation. It describes how to put the car to sleep, and states in bold that this can take up to 30 minutes. That raised several questions for me:

  • Is there any confirmation that the car has gone to sleep? No.
  • “Up to” 30 minutes means what, typically? Unknown. Do you have to wait the full 30 minutes, or would 10 be enough? Why not simply write “wait 30 minutes”?
  • Could BMW not find a faster way to put the DME to sleep? I have seen SI procedures that require an engine re-adaptation after EACH repair. Is the dealer supposed to leave every car resting for 30 minutes just to put the DME to sleep? Then the engine has to be warmed up again, because the adaptations must be performed on a warmed-up engine. BMW, seriously? Is this what you offer in 2022?

Again there is the standard sentence about what to do during the re-adaptation: drive in different modes. Really? Here, it is 2 minutes at idle and then “End service procedure”. Impressive.

Where are the learnbits for the successful completion of the adaptations? There are simply none.

Results after re-adaptation

Here is Overrun mode with the unadapted flywheel, in 4th gear starting at 4000 RPM:

The initial “linearity” of the flywheel is VERY bad. It reports the 1st cylinder as very active, with a deviation from the ideal of around 30 units, which is several tens of percent.

Here is the 5th Overrun cycle:

By 4000 RPM (171st second), the scatter between the cylinders has fallen to +/-10 units, and the 1st cylinder (blue) is already “tightened” to the 3rd (red). That is enormous progress.

This shows the result of the flywheel adaptation in the 3000 to 2000 RPM range. It looks much better than at the beginning of the session.

After the flywheel re-adaptation, the car was put to sleep and the fuel adaptations were generated.

The first image:

The data confirms the uneven cylinder performance until the 300th second, with the clutch released and the vibration noticeable. At the 300th second the clutch is pressed, and the vibration disappears COMPLETELY. It feels as if the engine had been switched off. Perfect. The data also confirms that all cylinders now run evenly.

One more of many identical experiments:

This time the clutch is pressed at the 331st second. The situation is identical, both subjectively and in the data.

The idle balancing stops at 0 km/h

Finally, the owner and I tested one more maneuver.

Years ago, working with an MSD80, I noticed that the DME evened out the idle in Stratified charge mode only once the speed dropped to 0. That is a little silly. With a manual gearbox, the idle can and should be evened out as soon as the gear lever is in neutral, and this MINI even has a gearbox position sensor. BMW, how come? Why such lousy work?

Returning to the topic, we checked how the DME manages the idle evening-out. We rolled in neutral at low speed with the clutch released, and the vibration was ZERO. As soon as the digital speedometer dropped from 1 to 0 km/h, the vibration started immediately, because the DME STOPS evening out the idle. We repeated the experiment more than 5 times, and the effect is stable.

BMW engineers, really? For several years nobody has noticed such a glaring (and simple) defect? How is that possible?

P.S. The MINI was bought brand new at the beginning of 2022 and now has around 8000 km on it.

BEYOND THE FAULT CODE

A clearer picture.
A better next step.

Get in touch
Enlarged technical image
Open original image