Jumpy Jeep

This article is a true description of an AECS technical help desk problem, and how it was solved.

By H.P. Leijen (trainer/research)

er numbered the injection events in firing sequence, not cylinder number. A time measurement between the last two injection events shows that the engine is doing 2,964 rpm just before it dies completely.

This means that injection did not stop as a result of the crank shaft simply stopping with rotating. In my view this may indicate a catastrophic problem with the crank/cam shaft sensor signals.

However, this AECS trained diagnostician had already looked at those signals and saw nothing untoward, I trust his opinion.

Best option is to look at the above pattern again but now in more detail. A much closer look at the same pattern as above shows a problem:

When zoomed in on injection event three and comparing it with event four a big difference in injector activation time can be seen. Injection duration variations should not be jumping around this much.

Please note: injection event one looks different as the scope is connected to that injector. We cover also that in the DMS 1-3 training. We suggested having a look at the adaption values of each injector/ cylinder with the scan tool, as we did not trust that these values were even.

We show a sample picture of an adaption injector value screen. This picture has been made with the scan tool connected to a bare fourcylinder MB ECU (no sensors nor actuators connected). The adaption values in this Launch Pro3 screen shot are shown in mm³ per stroke fuel quantity and are the correction quantities the ECU makes to keep the engine running stable (equal strength power beats per cylinder).

These values are on top of the injector individual delivery variations which are coded into the ECU with the Launch when an injector gets replaced.

When an ECU finds that that adaption values cannot be adjusted any further to keep the engine running smooth, this system will shut down. The injectors are no longer being activated, and the pressure control valve will now do all it can to drop the rail pressure.

The adaption values were 1.8, -2.5, -0.2, -0.6, 2.3! The vehicle data stated that the adaption value limits are -2.0 to +2.0. It was decided to pull the injectors out of the cylinder head and inspect them on a test bench.

The diagnostician was so kind to send us a photo of the number 2 injector, beside a good injector. The injector nozzle retaining nut was split and deformed.

The injector sealing washer allowed combustion pressure to leak past the washer eroding the nozzle retaining nut, weakening it until it split.

The leaking injector washer was letting high pressure Diesel leaking from the split nut into the combustion chamber during intake and exhaust stroke, causing the knock and white smoke. The diagnostician reset the adaption values after five brand new injectors were fitted, and recorded a pattern again.

To analyse and compare the injector activation duration of each injector you have to zoom in down to just a couple of injection pulses.

The measured injection durations do not jump around anymore and showed on the scope: Cyl 5 – 1.763 ms Cyl 4 – 1.763 ms