Crankshaft sensor

Patient: E39, M52B25


The engine had error messages regarding crankshaft sensor and problems with starting.

The owner of the car replaced the sensor with one, manufactured in China – the situation didn’t get better. The error messages regarding the crankshaft sensor and problems with starting continued.

At the next stage, a new Siemens VDO sensor was installed, the error messages regarding crankshaft sensor disappeared, but suspicions regarding its incorrect performance and problems with starting – continued.

Checking the INPA live data, most time of starting moment INPA indicated 0 RPM. It’s not correct – obviously, there are problems!


The oscilloscope was called in for help. The crankshaft sensor MS41 was connected to pin 83.


When connecting to the sensor, at the starting moment the following picture was observed:

In the picture, you see 5 .. 6 full turns of the crankshaft.


When zooming the picture, we had to conclude – the sensor “sees” 15 .. 20 “teeths”, then a longer pause follows.

In the picture – one full turn of the crankshaft.


In addition, the fulfillment of “teeth” is incorrect.

In case of the correct signal, the pauses between “teeth” have to be of the same length as “teeth” itself. Exception – the “empty” teeth (at this time the signal level has to be low), which allows identifies DTC of the first cylinder.


At the image: flywheel with marked “empty” tooth:


Why does the sensor see only half of the flywheel? Obviously, the pulley itself is slightly eccentric – displaced. It’s nothing unusual in a range or norm. The sensitivity of a correctly working sensor has to be in such level, that slight shifts from the ideal would not affect the performance of the sensor – the manufacturer of the sensor has thought about that!

Here, in this oscillogram you can see – the first impulse (from “pack” of the impulses, which are seen by the sensor) is “thinner” than others.


And also in this picture – the last impulse is “thinner”.

There is no doubt – the pulley is eccentric. When visually evaluating the pulley, it was concluded the offset from the ideal cannot be observed visually. In a set with incorrect fulfillment of impulses – unfortunately, it turns out the VDO sensor was defective.


An interesting nuance – as the “ruptures” of the sensor signal are short, DME could not identify problems of the sensor itself (it was “thinking” the engine regularly stops for a moment). Yes, newer DME (equipped with more powerful MCU) won’t have troubles to identify such defect, but MS41 – could not handle it.


After installing another sensor, we see: the sensor “reads” all disc! In the picture, 9 .. 10 full turns of the crankshaft.


Fulfillment of the impulse – correct:


And here – also the “empty” tooth:

The sensor works correctly – the problem is solved!


In the end – several nuances regarding the process of oscilloscope data analysis (review). For example, if we have a brief look to the correct (as I state above) working sensor:


You can see the empty “teeth” (in the picture below marked with red marks):


But, at the same time – you see as if other, short data interruptions:

This is a peculiarity of digital oscilloscope data display. The resolution of the monitor of the exact oscilloscope is 800*600 pixels. It is not a little, but in this image, 14 million (!!!) of measuring points are compressed. It’s clear in 800 pixels it’s not possible to display 14 million measurements – the oscilloscope “averages” the displayed information.


To make sure the signal is correct (or directly opposite – to find mistakes), after stopping data frame of all 14 million measurements (storing them in the memory of the oscilloscope) it has to be zoomed in and each stage of it.


Here, the zoomed signal. Looking at a previously inserted image with an empty “tooth”:

We see short impulses “not in place” – where they should not be. Or the sensor still sends an incorrect signal?

No, this is interference. Interference from other hubs of the car – actuators (their moment of switching on/off) or the mains network – during measurements, the car was connected to the booster, also the battery of the oscilloscope was charged from the mains.


What confirms interference, not incorrect signal of the sensor:

a) these peaks of the false signal are very short – much shorter than for “normal” impulses;

b) these impulses are with “positive” and “negative” polarity at the same time. Pay attention – these impulses “drop” below marking of 0.0 V. The real signal can in no way drop below 0.0 V!


Do these disturbances do harm to the performance of the engine? No, in a normal situation – don’t do. At first, DME software filters such very short signals and ignore them. Second – possibly these impulses got in the oscilloscope, inducing in its measurement connections. As you see at the first image of the entry, the feeler is used to measure the signal. For around 10 cm it is not shielded. In this – unshielded area – all types of interference can be induced. Not without a reason the wires from the crankshaft, camshaft sensors are shielded! It is very important to avoid damaging it – then the density and amplitude of induced signals can reach above the ability of DME to filter them, and then problems are guaranteed!

The second obstacle – the Ground connection of the oscilloscope during the measuring was not made correctly – not by appropriate input pin of the DME, but by the chassis of the car. If you have reasonable suspicion, that the level of interference has overreached allowed, the correct connection of the oscilloscope – first thing, with what to start! In the current situation, taking into account the experience, was clear – everything is fine with the sensor, so I wasn’t worried about the interference I saw. But I had the chance to show you, how the real signal looks alike, not the one from books.


The sentence of this entry – the sample of the real signal of the real sensor. Even more – even for such, as if simple things, there is very little exact information.