Audi A8 | Injector fault

Vehicle details

• 2003 Audi A8
• 4.2 Litre V8
• Engine Code BFM
• LPG conversion

The problem

A 2003 Audi A8 was having a problem with rough running which in turn was causing the engine management light to illuminate.

The first stage of the investigation was to use a diagnostic tool and obtain any ‘DTC’ (Diagnostic Trouble Code) information and relevant data from the ECU. The scan tool showed the following codes stored in the DTC memory:

• 17546 Fuel Trim Bank 2 (Add) System too Lean P1138 Upper Limit Reached Intermittent.
• 17538 Fuel Trim Bank 2 (Mult) System too Lean P1130 Upper Limit Reached Intermittent.
• 16684 Random/multiple Cylinder Misfire Detected P0300 Upper Limit Reached Intermittent.
• 16588 Injector Cylinder 4 (N33) Circuit Malfunction P0204 No Signal/No Communication Intermittent.
• 16688 Cylinder 4 Misfire Detected P0304 upper Limit Reached Intermittent.

The first 3 codes concern general misfire and fuelling. These are probably due to the engine switching from Petrol to LPG and back.

The diagnosis

As this was an LPG conversion, we didn’t trust the DTCs so we checked with PicoDiagnostics. This would help to identify the misfire, highlight any signs of rough running, and more importantly show whether the misfire was a mechanical error or related to a fuel or electronics issue. Figure 1 shows the results from PicoDiagnostics:

Figure 1

With the ‘Variation’ data enabled, PicoDiagnostics shows that each cylinder can produce a high amount of mechanical power (represented by the blue line). The above screen capture indicates that while cylinder H is mechanically OK, it does have an intermittent misfire. Further investigation is required into the surrounding components which supply the cylinder.

Figure 2

We assumed that the first 3 DTCs were due to the engine switching from Petrol to LPG and back. This is because these codes are general misfire and fuelling codes, which aren’t relevant to the fault under investigation. The PicoDiagnostics software backs up our theory.

From the ‘16588’ error code, there seems to be an issue with one of the injectors in the circuit on cylinder 4. At this point we cannot determine whether this is due to the LPG switching, so we carried out a quick test. Using the Fuse Adaptor and a current clamp we can see if all the injectors are drawing even current, if any were poor or even absent.

The screen capture to the right shows that there are only 7 current draws from the 8 injectors. This indicates that there is an issue with one of the injectors, or possibly the wiring harness.

One more check was carried out, which was to switch to the LPG system from the standard fuel system. As the LPG requires the ignition system, if all 8 cylinders fired correctly it would have proved that issue lay with the fuel injector.

The Audi A8 is designed with a sophisticated system that allows one or more cylinders to be deactivated in the event of a continuous misfire.

This is important to remember, as a simple test of resetting the ECU can be performed to show if the problem is continuous or intermittent (this is done by switching the ignition off and then restarting the engine). If the engine misfires intermittently, the ECU will register the misfires a given number of times before it automatically deactivates the fuel into this cylinder to prevent premature catalytic converter damage. This will then lead to a constant misfire.

Figure 3 (front of vehicle)

We knew that we were currently firing on only 7 cylinders and so, before switching to LPG, we must switch off the ignition and restart the engine. Upon restarting the engine and switching to LPG, all 8 cylinders began to fire successfully. This confirmed what we had already suspected: that both mechanical power and spark were present.

The V8 configuration used by Audi (Figure 3) identifies the cylinders in the following order: left side first, numbered front to back, followed by the right side, also numbered front to back (as in the image to the right). With this information and the ‘16588’ error we now need to investigate cylinder 4 further with regards to the injector and circuit.

From here it was obvious that we would need to use the PicoScope Automotive software and begin running tests on the cylinder 4 injector. At this point it made sense to use cylinder 3 as a benchmark for comparison to determine if a difference could be seen, or more importantly if there was a fault with the injector or the wiring harness.

We then connected the PicoScope to the injectors on cylinder 3 and 4, measuring both voltage and current by connecting the channels as follows:

• Channel A to the Injector Cylinder 3 Signal Wire Using the Acupuncture Probe
• Channel B to the Injector Cylinder 4 Signal Wire Using the Acupuncture Probe
• Channel C to the Injector Cylinder 3 Signal Wire Using the Amps Clamp
• Channel D to the Injector Cylinder 4 Signal Wire Using the Amps Clamp

The following waveforms (Figure 5) were captured using the scope. The screen was split into two sections to allow a comparison of all measurements. The sections were set up as follows:

• Left: Injector Cylinder 3 Voltage vs Current
• Right: Injector Cylinder 4 Voltage vs Current

Figure 5

We can see that the cylinder 4 injector is not switching and therefore no current is being drawn. We can also see small voltage spikes, which indicate that there is possibly a high resistance in either the injector or the circuit, which is restricting the full voltage from reaching the injector solenoid.

From this data we could identify that the problem was with cylinder 4’s injector, and that there was some kind of resistance preventing the full voltage from getting to the injector. We decided to strip back and remove the wiring protector to have a closer look at the wiring harness and take note of any breaks, short circuits or obvious problems.

We found slight breaks in the wiring insulation (possibly from previous testing) which might indicate that this was a previously investigated problem. However, these breaks were not sufficient to short to any earth wire or earth point and there were no other breaks within the wiring. From this we were able to determine that the problem was not occurring from these breaks.

At this point the multi-plug was connected back onto the injector and the engine was restarted for further testing. Upon restarting, we noted the misfire had disappeared. We retraced our actions to identify what could have affected the misfire problem. Our only thought was that we had removed the multi-plug and reconnected it with the harness removed. It was decided that a further test was required on the harness. The scope was reconnected to the ‘Fuse Adaptor’ with the current clamp, and a ‘wiggle test’ was performed on the harness. This returned no change in the results, so a slightly more aggressive ‘wiggle test’ was performed on the multi-plug, and at this point the cylinder began to randomly misfire. This can be seen on the waveform captured in Figure 7:

Figure 7

Figure 8

With the multi-plug removed (Figure 8), it could be seen that the terminals inside had pushed apart over time, and therefore were acting as a pivot when moving with engine vibration. This would cause an intermittent misfire as described in the original fault codes.

The repair

The multi-plug was stripped down and the terminal pins were replaced and refitted to the injector. The entire wiring harness protector was built up again and the engine no longer misfired.

The road test

The codes were cleared from the DTC memory and the vehicle was road tested, with no misfire present. To ensure the misfire had been resolved, the LPG was activated and deactivated while on the road test to ensure that the engine could cycle and run on the alternative fuel. No misfire occurred. After the road test, the scan tool was reconnected to check for pending and stored codes. The following codes were retrieved:

• 17546 Fuel Trim Bank 2 (Add) System too Lean P1138 Upper Limit Reached Intermittent.
• 17538 Fuel Trim Bank 2 (Mult) System too Lean P1130 Upper Limit Reached Intermittent.
• 16684 Random/multiple Cylinder Misfire Detected P0300 Upper Limit Reached Intermittent.

The conclusion

The 3 DTCs (see above) returned by the scan tool were due to the LPG running conditions when the fuels were switching over. This confirmed that there were no new DTCs returned that required further investigation.