N43 with strange behavior

One car with N43 engine got lucky – it happened to get to qualified mechanic. It turned out, that valve seals had lost flexibility, exhaust valves were swimming in oil. Oil in large amounts was found also in exhaust.

After engine repair (replacement of valve seals, prophylactic maintenance of engine head, new seals for exhaust) came round for management system maintenance. Some “improvements” has been made by previous maintenance “specialists”, but let it be a little intrigue – for a bit later time.

Status, when the engine control inspection was started: several as if successful driving sessions after deleting of long-term trims, heated till work temperature. No active error messages in MSD80 memory. The engine performs in Homogeneous; Homogeneous lean and Stratified charge are not used. Strange – that would be the correct description of this current engine’s performance.


First step: inspection of long-term trim menu. Already here first oddities are spotted:

a) for a long time after starting the engine (more than 5 minutes) the PWM of control probe heating was 5 .. 7% only, which shows, that the probe is not heated normally. This mode shows “before working mode” and is typical for first/second driving session after full deleting of long-term trims. This mode is kept, if by some reason (for example, incorrect registration of injectors, errors in basic operation modes of engine) MSD80 can not go to normal working mode.

In the image: PWM indication of control probe is marked (only after waiting more than 5 minutes, MSD80 started to heat up the probe):

BMW INPA adaptations menu

b) both banks has 0,00 mg/stk offset type LTFT in idle run. Perfect engine? Unfortunately, unbelievable. It look’s that (for unclear reasons for now) long-term trims for banks are not created.

BMW INPA adaptations menu

In the same time: integrators performs their duty, the values of integrators: in the limits, LTFT of mid/high load range gives adequate values, it means: closed loop is active.


Inspecting status bits:

BMW INPA learnbits

We see, that (for unclear reasons for now) application of individual Lambda corrections has not performed (successfully completed). In normal situation, these long-term trims has to be completed during second/third driving session after deleting long-term trims, in around 10 cycles of homogeneous mixture (with length: half-minute .. minute) evening the cylinder operation. Later these data will be used also for Stratified charge mode.


We continue the inspection.

BMW INPA injectors adaptations

In this image you can see individual adjustments for working range (factory coded adjustments + additional adjustments applied by MSD80). Current numbers indicate, that no additional adjustments are applied! It means, the adjustment data of injectors complies with factory coding. Individual adjustments are not incomplete, they are not even started!

Next menu:

BMW INPA injectors adaptations

Direct confirmation – no individual adjustments nor for warm, nor cold engine.

BMW INPA cold engine adaptations

No Lambda additional corrections of banks for different engine temperatures also existing! It would be understandable, that no additional adjustments for -20 oC existing – the outside temperature now, in summer, don’t goes below +10 .. +15 oC, but there are no data nor for 0 oC, nor +20 oC, nor working temperature. No data at all.

And now we check performance of Lambda probes.

BMW INPA Nernst resistance check

As we see in two last rows, dynamic resistance of Nernst cells is in the normal range for all three probes (two wide-band and common control probe).

Unfortunately the voltage test shows some oddities: the voltage of control probe is only 0,05 V (or 50 mV) by correct (Lambda = 1,00) data of regulating probes. 0,7 .. 0,8 V would be normal value (in the upper part of linear region – MSD80 keeps control probes in such values in Homogeneous mode with Lambda 1,00). Reason – the defect of control probe?

The data of Lambda control check (using the Lambda measuring tool) shows data 1,06 .. 1,08! It means, that the fuel mixture actually is lean (not normal), and the check-up of control probe performance has to be continued (because the voltage is appropriate according this Lambda value). During the test drive, when accelerating (when the fuel mixture is enriched for short time with Lambda till 0,85), the control probe shows voltage till 1,0 V, which means, that it can generate EDS without problems.

Conclusion: one (or both) wide-band probe shows incorrect values. But no adjustment of setpoint for these (wide-band) probes is performed according to values of control probe! One more mystical fault.

In current situation, the control probe does not performs at it is supposed to – it don’t calibrates reference values of wide-band probes and is not able to perform the control of CO catalytic converters. One more sign of problem and defect. Unfortunately, it’s was not possible to perform the test sequence of Lambda probes (asymmetric data of regulation probes would allow guess the damaged probe and evaluate the deviation of wide-band probe from norm). And no error messages recorded in MSD80 memory.

According to very strange situation, that engine works only in Homogeneous mixture mode, we continue the inspection with check-up of performance of NOx sensor (specific damages of NOx sensor could be the reason, that engine don’t switches to Homogeneous lean and Stratified charge modes). To achieve that NOx sensor turns on, we perform one more driving session (one of the conditions for NOx sensor to switch on – driving speed at least 20 km/h). Unfortunately, the sensor don’t performs the necessary performance in necessary moment. Indications shows, that it don’t sends live data, and still there are no error messages in MSD80 memory. It means, that sensor responds (sends status, ID and other basic information to MSD80), but refuses to turn on (go to active mode). When all hopes were almost gone, the sensor however turned on. It happened with very significant delay (more than 5 minutes), which suggests, that the sensor is aging and it has troubles to perform self-test (balancing the current between oxygen and NOx cameras).

BMW INPA NOx sensor menu

The wide-band data of NOx sensor gives more or less correct value or Lambda, but narrow-band data shows lean mixture (correct value with Lambda 1,00: 450 .. 550 mV)! Existing data of NOx sensor gives the signal, that fuel mixture is leaner (Lambda around 1,03) than wide-band data of Nox sensor shows in first moment. Yes, after a moment also wide-band data gives Lambda 1,02 .. 1,03. The mixture is lean, it’s obvious.

Additional questions arise regarding reference values of NOx sensor in overrun mode.


Elevated value of reference can signal regarding:

a) huge drift of reference value after self-leveling process of sensor, when it starts to work (which signals of “drifting” parameters of sensors – result of damage or aging)


b) slow reaction of sensor to changes in content of NOx: fall (it signals, that diffusion barrier of sensor is clogged up – wear of sensor because of high level of oil ad mixture or other fuel by-products of increased temperature of exhaust gases).


Attempt to re-write the reference values was made. Unfortunately with no results, because the sensor showed very slow reaction to overrun mode (taking in account, that level of NOx in exhaust was decreasing much faster than in case of “regular” exhaust, see the description below).

Conclusion: already second sign, that the sensor aging. Existing reference value is not yet so high to reach the level of error message, for this reason there are no error messages in memory of MSD80.

While performing this stage of diagnostics, the permanent engine shivering is detected, but Rough run menu also in this moment shows almost perfect balance of cylinders (difference prom ideal value is not higher than 0,003 .. 0,004). Obviously, long-term trims of individual efficiency are turned off! In the same time – the status of long-term trims of flywheel shows, that they are completed. One more problem, cause of which is not directly indicated (no error messages, no abnormal values of sensors).

Desulfating menu of NOx sensor shows, that 650 .. 700 km are driven since regeneration (or registering a new one) of NOx catalytic converter, which is more than enough for long-term error messages to appear. Error log shows error messages regarding VANOS (this could be related to oil pressure, with which, thou quite very rare, was some problems) and also trim & misfire error messages for separate cylinders. Last ones just increase suspicions, that there are fundamental problems with basic operation modes of engine (obviously, also LTFT are not overwritten successfully).


Now it’s time to reveal the intrigue, mentioned in the beginning of this post – the engine was taken care by “professional” diagnostics specialist, which cut out the NOx catalytic converter! This is a extraordinary technique, and reasons for such choice are not clear. Such “solution” causes losses of around EUR 1000 to customer (price of used NOx catalytic converter + price of damaged NOx sensor in future) and don’t solves the possible problems with engine in any way. Obviously, “the masters” had no understanding about N43/N53 series engines and basic principles of NOx system performance.

But there are no error messages regarding NOx catalytic converter, the engine stubbornly performs in Homogeneous mode. It turns out, that the engine is “left” in forced (diagnostics) mode of Homogeneous mixture. I don’t suppose, that it’s done with intent, obviously it’s a result of “playing” with diagnostics…


When Stratified charge is restored, the NOx values confirm, that the NOx catalytic converter has been removed:


The NOx level in exhaust is around 1300 ppm (1500 ppm is max value, which can be reported by INPA; according to manufacturers information – 3200 ppm is max value, which can be measured by the sensor) in Homogeneous lean mode, right after this mode turns on.

After restoring Stratified charge (of course, not for long – only till moment, when MSD80 will detect the fact, that performance of NOx catalytic converter is 0), immediate formation of individual Lambda adjustments for performance mode and idle run was observed. These adjustments will allow to normalize the performance of engine till moment, when normal performance of NOx system will be restored (new catalytic converter or NOx emulator).

Performance of engine in “healthy” Stratified charge with Lambda 2,75.

BMW INPA NOx sensor menu


Brief summary:

totally damaged NOx system (expensive NOx catalytic converter has been removed, huge load created to NOx sensor because of elevated temperature) as a result of “professional” maintenance plus normal performance or engine management system has been paralyzed:

a) no offset LTFT are created;

b) no individual adjustments of cylinders for idle run are created;

c) no individual adjustments of cylinders for working mode are created;

d) no adjustments for banks depending of temperatures are created;

e) no long-term trims of mechanical efficiency are performed;

f) long-term trims of setpoints for Lambda trims are not performed correctly;

g) the Lambda control probe don’t performs correctly etc.

Of course, this is only the question of time, when the engine, performing in such mode, will start to register error messages regarding Lambda probes, trim, misfire etc. And then “the masters” could start the ”maintenance”…


And the simple (additional) problem:

BMW INPA OBD menu engine temperature

Data of engine temperature after test drive. As seen, the temperature is around +84 oC. Temperature of oil + 85 oC, temperature or coolant (outgoing) +77 oC. The data shows, that all three temperature sensors are in working condition, temperature of engine – lowered.

Normal temperature of engine (Normal: in default mode) should be +95 .. +100 o C, which can not decrease, when the engine runs idle. In this case – the temperature of engine is decreasing in idle run mode, dropping down to +75 oC during 10 .. 15 minutes. +75 oC is below lowest treshold of thermostat, obviously it (thermostat) is not closing (the mechanical element of it is pledged). Unfortunately this is quite common problem and no necessary attention is payed to it, because, if only mechanical part of the thermostat is damaged, no error messages are shown in engine’s management system.


Related entries:

How I perform adaptations

Flywheel adaptations

Management of Lambda probes

Stratified charge performance check

Rough run. Trim errors. Hidden cause