Select a technical image to view it in detail.
Today’s patient is an N53 in its most powerful version (N53B30O0). The engine has lost its surge and torque, especially at high RPM. The task is to find the problem and fix it.
Subjectively, the car accelerates languidly. The torque loss is most noticeable above 3500 to 4000 RPM, and the “push” of the large DISA cannot be felt at all, as if the DISA were not working.
Initial checks
A long time ago I wrote an entry on how to confirm that an engine delivers full torque and, if it doesn’t, how to work out where the problem is. This time I followed my own list.
The short version of the results:
- There are no error messages in the DME or DSC modules, at least none that could cause this problem.
- The throttle opens completely.
- DISA live data show that both valves are being controlled.

The status of the “right-side” DISA first changes to 67%. Above 5000 RPM, the large DISA reading also indicates that the DME has requested its opening.
We have to be clear about what this tells us. The DME sends the control command, and the DISA only confirms that it received the electrical signal and actuated electrically. Whether the valve really opens, and whether its sealing ring is in good condition, can only be judged visually. Still, the DISA valves are connected, there are no error messages for them, and their statuses change. There is also no doubt that the engine runs according to Power_Class 02H, which is the 200 kW version.
Fuel mixture in kickdown
Next I checked the fuel mixture data in kickdown mode. And there was something unexpected:

The mixture is far too rich. By 3000 to 3500 RPM, Lambda on both banks drops to 0.80 to 0.85. When the DME switches to closed-loop mode, which happens 2 to 3 seconds after the throttle opens, we see that this Lambda is reached with integrator values of -15 to -20%. So there is at least 15 to 20% more fuel in the cylinders than necessary. On top of that, this happens after the DME has already reduced the fuel quantity by 15 to 20%. The total fuel “overuse” is therefore as much as 30 to 35%.
Obviously, either too much fuel is reaching the cylinders or the cylinders are short of air.
A fuel problem is doubtful. Fuel pressure is measured and stabilized continuously, and the rail sensor readings are verified and adapted to different driving conditions. The mixture data are otherwise fine: the multiplicative LTFT is correct, and the live data for the mixture are correct. The injectors also behave correctly in all other conditions.
That leaves insufficient air. The throttle, whose position is monitored by two sensors, and the electrical side of the DISA valves have already been checked. What remains is:
- the mechanical part of the DISA valves
- the air filter
The air filter test
I started with the simplest option: I took out the air filter element and repeated the test.

The picture is completely different. Lambda on both banks is 0.95, which corresponds to a correct, slightly rich mixture. The integrators are very close to 0, around -2 to -5%. The mixture is now properly calculated and maintained.
At this point the large DISA could be felt working very clearly.
When I inspected the filter element, I had to admit that visually it was not in terrible shape. It looked like a regular filter. It may well have picked up some water during the winter, which happens in wet weather with air humidity close to 100%, in heavy rain or snow, and of course when driving through a deeper puddle.
Anticipated objections
I can already predict the comments: “you should have started with that” and “what is there to measure, just take a test drive.”
On the second point, the “butt dyno” is a very unreliable measuring tool. It will not help if the problem is more complicated or if there are several problems. Only live data give an objective picture.
On the filter, I have seen mechanics evaluate a filter by looking at it. This test would have failed in exactly that way: the same filter would have gone back in with the note “well, everything looks fine there.”
Filter choice matters
One more nuance must be taken into account. On the E60, for example, the same air filter is used for all 6-cylinder engines, from the detuned M54 (520i) to the 200 kW N53 (530i; N53). The power output of these engines differs by a factor of two, so the amount of air they consume differs by the same factor. The most powerful engine is therefore much more demanding about filter quality and condition. The cheapest filters, which work correctly on an M54, will not suit the most powerful N53. Mechanics who choose budget versions and “analogs” from parts catalogs need to keep this in mind.
In this case the filter number is 13717521033. The available versions are:
- BMW original: EUR 28
- MANN: EUR 22
- Bosch: only EUR 11, which is surprising, since I doubt that a filter at least twice as cheap can be made from quality materials
- Other “analogs” (Mahle, Valeo, Purflux, Meyle, Clean, Denckermann, Mapco and others): EUR 6 to EUR 10
It would be quite interesting to test these analogs for airflow and their ability to filter fine particles, but I don’t expect any miracles. It is doubtful that a filter five times cheaper will give excellent results.