When the engine control unit (ECU) is constantly making faults then, with knowledge of the system and a simple measurement instrument, it isn't difficult to find the cause of the fault. It's something different when the fault is only present once and while. These types of difficult faults can't mostly be found with the usual faultcode reading instruments, because the ECU can't detect this type of faults. Due to the fact that the fault is only present for a short time and therefore can't be logged in the faultcode memory. Also the transfer rate of information between the ECU and the test instrument is in most cases to slow for detecting fast changing sensor signals (read faults). In the following article we explain the faults that were localized with a 2.5 litre 4-cillinder Chrysler Voyager.

A business acquaintance (an old student of GMTO) called about a Chrysler with some problems. Some facts about the problem. The engine sometimes halts (at most once a day). And, of course, this never happened when then mechanic was driving the car, but always when the customer was driving it. The garage had also done some measurements, but because the fault was not present when they were measuring, there was nothing to measure. After several times having contact by telephone with GMTO it was decided to do some examination with the car using advanced measurement instruments. For this examination the owner of the garage was prepared to spent a day for the examination of the car, due to the fact of the time that already had been spent on this car without any result.

Credibility

Of course the Voyager didn't store any faultcode in the memory. This memory, in Chrysler cars, is easily to read just by turning the ignition key (placed in the ignition lock) tree times on and off. With the use of the faultled on the dashboard the fault can be read. The fault was mostly present when the engine was stationary, at the moment when the engine halted many things occurred simultaneous in just a split second. So it was hard to find out which function halted first. By recording the signals and making them visible it's possible to see the progress of the fault . At the moment when the fault occurred the engine started to halt several times before it completely stopped. The customer already pointed out that the revolution counter went to zero when the fault happened. Because the revolution counter is controlled from the ECU (found in the system schematics) we can enumerate the following points. The signal for the revolution counter depends on the next components:

• number of revolutions-/position sensor
• control device
• Systemrelay
• ignition lock
• Wiring of the above mentioned components

Examination of the revolution counter

Control of the ignition coil

We had reached the point where we should argue where in the system the fault could be localized, (with use of the measurement reports). It was clearly visible that when the ignition skipped (see ignition coil signal) the next contactangle cyclus, with its sparks, came at the right moment. So it definitely couldn't be a problem of the mechanical part of the distributor. On some cars the cage in which the Hall-sensor is mounted can come loose off of the rotor and so turn a little bit. Because the ignition coil is directly controlled from the control device and the signal coming from the Hall-sensor (which has a important function for the right timing of the control signals to the ignition coil and injectors) appeared to been fine. So we started to suspect that there was something wrong with the control device. Replacement of the control device is always pricey, certainly when it isn't the problem. Therefore we decided to examine the HALL-sensor signal to make sure that it was really functioning fine as concluded by the mechanic earlier.

The Hall-sensor signal was also measured using the recordposition of the scope, therefore the signal can be measured over a longer period of time. But then fate stroked and the engine didn't brake down anymore. What to do? We couldn't get any signals of the hall-sensor when the fault occured. We decided to give it a try and so the scope was set up to do its own triggering. The second triggerchannel should be connected to a component that has a feed under normal circumstances and when the engine stops that there shouldn't be any feed. The fuelpump is suiteable for this task, because it shuts down when the engine stops or stands still. With the first measurement channel it is then possible to measure the sensor signal over a longer period of time.

After connecting the scope the car had to be driven until it stalled. A acquaintance of the garage was prepared to drove the car for how long it took for the engine to suddenly stall. It took lots of kilometres before the engine suddenly stalled and so triggering the scope. Finally the Hall-sensor signal was recorded. The acquaintance of the garage managed to start the car again and drove back to the garage and there it was! The Hall-sensor signal showed some problems. The signal of the sensor did not switch correctly to ground. Because all the Hall-sensors use the same principle. The signal wire is switched to ground with the electronics inside the sensor. This creates a square-waved signal. This means that the signalwire, that comes from the control device, must feed the sensor so that the sensor can switch it to ground. Picture 3 shows clearly a gebrekkig signal of this Hall-sensor where the ground switched part of the signal decreases. At a specific moment the controldevice did not see a clear square waved signal and therefore stopped controlling. But it was verified that the signalwire between the ECU and sensor wasn't damaged, due to the fact that the periods of 5 volt were visible on the scope image.

Picture 3: Hall-sensor

Picture 4 shows the engine halting for a moment and then continues functioning (fuel pump went off and on). This triggered the scope.

Picture 4: Hall-sensor

No. That the signal didn't switch to ground could have 2 reasons:

• The wiring (feed) going to the sensor
• The Hall sensor

Before we started replacing the Hall-sensor we attached the scope at the feed wires of the Hall-sensor. Now both the positive and negative pole of the sensor was measured. Also the fuel pump was used for triggering and so starting the recording of the signals. This time the driver had some more luck and after a short period of time the engine stopped. After examining the scope image it was determined that the Hall-sensor was the cause of the fault. The sensor was replaced and the car has never been back again. The bill was written and the garage owner could write the bill for the customer, knowing that the fault was solved.

Picture 4: Hall-sensor

Looking back it is easy to say that if this sensor was replaced directly, the fault was solved quicker. But because of the fact that car systems are still getting more complicated and the components getting more expensive, this is not recommended. Because replacing parts randomly, not knowing if the fault is solved by this, isn't good for the credibility of your garage. Customers have great difficulty with this kind of repairs. Knowing that you are the expert with cars.

R.H.M. Metzelaar
www.gmto.com