Jeep Compass | Cylinder misfire

I had a customer's 2007 JEEP Compass , European Spec, fitted with the VW 2.0l TDi PD engine (code BLY). Unfortunately standard VAG software (e.g. VCDS) cannot communicate with this ECU (due to it being a JEEP). Communication via standard EOBD displayed "No Codes" although the check engine light was illuminated, and a misfire fault was present.

At this stage I needed to determine 3 things:

1. Are all injectors receiving a signal from the engine ECU?
2. Are all injectors draw an equal current?
3. What cylinder is misfiring?

Firstly I connected Channel A to the alternator output terminal and using the Cylinder Balance test in PicoDiagnostics. I was quickly able to determine that only one cylinder was at fault, as three cylinders read between 98%-100% and one cylinder read at approx. 67% at 850 RPM. PicoDiagnostics was able to determine the correct engine RPM without any difficulty.

I now knew that the problem lay in one cylinder but had no idea which one. In the PD engines the injectors are fed via a multi-plug at the rear of the cylinder head. It is a very simple matter to connect a current probe to check the current waveform of all 4 cylinders via the common supply wire. It is also very easy with a second current probe to get the waveform of cylinder #1. Thus in seconds I was able to see that all 4 injectors were being actuated evenly and drawing very similar currents. This quickly eliminated any possibility of an electrical source being responsible for the misfire. I now needed to see which cylinder was misfiring. Normally I would check the Crankshaft Position Sensor (CKP)  waveform , but:

1. The wire is not very accessible in this engine and,
2. Many of the more modern VAG PD engines use a Hall effect type CKP Sensor, and magnetic coil type sensors give a much clearer indication of a misfire.

As I was already connected on Ch A to the alternator I decided to use the variation in alternator output voltage to find the misfiring cylinder. By setting the scaling of Ch A to x10, I was able to focus very clearly on the slight variations in the line voltage. By setting the low-pass filter to 400 HZ I was able to further enhance the signal and produce a very clear series of voltage readings which varied consistently with engine rotation.

Knowing that the firing order was 1342 I could immediately see without the use of tools apart from my PicoScope that the misfire was in cylinder #2, and that it was a mechanical injection fault rather than an electrical one.