Select a technical image to view it in detail.
My experience with BMW N-series engines (MSV/MSD engine management units) is that in at least 80 to 90% of cases, replacing the injectors is unjustified.
After such an unnecessary replacement, one of two things usually happens:
- The replacement is done together with other repairs, such as preventively replacing spark plugs, coils or the CCV. The adaptations are deleted and re-learned, and the engine really does run better.
- The replacement itself is done incorrectly, and there is no improvement.
An example of how not to code injectors
Here is one example of how NOT to code injectors:
This video has more than 6000 views, 30 likes and 0 dislikes. The first dislike is mine. Sorry, but I couldn’t resist showing my attitude.
The diagnostician made two essential mistakes.
1. Incorrect injector coding data was stored
As described here, the last digits of the injector coding in the DME differ after the data is stored. If you try to enter the stored data again:
- ISTA D will not confirm the new data.
- INPA will accept the new data but will not use it, and the creation of adaptations will be disturbed.
If you use the stored values to check whether the correct data was entered, you will wrongly conclude that the data was stored incorrectly.
This is the data that was entered:
And this is the data stored by the diagnostics “specialist”:
2. The old adaptations were not deleted and new ones were not created
If the old adaptations are not deleted and new ones are not created, the engine keeps running in a limited-functionality mode. The adaptations concerned are the flywheel adaptations and all types of LTFT: bank offset and multiple, cylinder offset and multiple. The consequences are:
- No individual cylinder adaptations are performed. The engine keeps vibrating, the misfire count rises on individual cylinders, fuel trim and catalytic converter (CO) errors can appear, and the engine runs unevenly.
- The creation of the other adaptations (such as bank fuel trim) is disturbed. All the problems above can result, up to and including the EML indicator lighting up in the KOMBI.
Unfortunately, none of the 6000 viewers pointed this out. Only one person asked why the entered data differs from the stored data, and the “specialist” who did the coding never replied.
How to replace injectors correctly
I regularly receive e-mails complaining that, after injectors were replaced, the desired result was not achieved and nothing improved. So, once again, briefly, here is how the replacement must be done.
1. Code the injector flow-rate data
- Using ISTA D is recommended.
- If you use INPA, check the entered data very carefully, because INPA does not perform an integrity check on it. When coding injectors for a 4-cylinder engine, use this link to determine the cylinder order.
- Store the data printed on the injectors, not the data ISTA D or INPA displays after coding.
2. Delete the adaptations
- In ISTA D, follow the Test plan sub-task that suggests deleting the adaptations. If for some reason the sub-task is not created, use INPA to delete them.
- In INPA, delete all adaptations via ../F8/F2/Shift+F9 (status: ignition ON, engine OFF).*
\*On N43/N53 engines, I recommend performing a desulphation session of the NOx catalytic converter before replacing the injectors.
3. Create the new adaptations
- Flywheel adaptations.
- Bank offset and multiple adaptations.
- Cylinder offset and multiple adaptations. On N43/N53 engines, these are done for both the Homogeneous and Stratified charge modes.
How to create the adaptations correctly is described here.

