The car in question runs fine for several kilometers but than suddenly starts running roughly. The garage already tried several things, from changing ECU, Hall sensor and because sometimes fault code Oxygen sensor malfunction was set, this sensor as well. The VW dealer who also looked at it, didn’t know what to do and delivered the car back to the garage, not repaired. The car was delivered to GMTO for a critical analysis. As usual we start our investigation by monitoring the most vital control signals of the engine management system, the injector and the ignition control with a scope, and went driving. At first the car behaved like it should do. If one single hitch would occur we could immediately record and analyze it on the scope.

After about 6 km we did encounter the failure and the signals were frozen. This action resulted in scope picture number 1.

Picture 1: Primary ignition coil & Injector control

Evident is that the primary ignition coil control signal shows mistakes and besides missing a spark also displays a varying dwell control. The control signal of the injector shows no problem at all, a stable signal during a long period can be seen. But the engine of course has to get it’s spark every compression stroke. If these misfires happens more often the oxygen sensor sees to much oxygen in the exhaust gases and it’s not impossible that the ECU after sometime thinks the oxygen sensor misbehaves. This could explain the fault code on this sensor.

First conclusion

Back to the unstable varying dwell seen in the ignition control signal. A modern ignition module “looks” at the current that the ignition coil takes. This current always increases curve wise and when the current reaches the right value the transistor is partially out off conduction. This shows as the period after the coil (negative) has been pulled to the ground and the voltage is about half of the battery voltage. A (too) long period (current take-off) can be the result of a low coil resistance. A (too) short period can be the result of a voltage supply problem on the coil (transfer resistance) now it takes to long before the current reaches the right value. Scope pictures 1 shows two different things.

• The absence of a spark, during which the power supply is a stable 12Volt
• An extremely variable dwell

The 12 Volt shows that the power supply on the ignition coil is not the problem because this voltage is measured, through the coil and wires, near the ignition module. What remains is the control of the ignition module because the ECU still dictates the moments of switching on and off the coil current (dwell and timing).

A logic next step is to measure the control signal of the ignition module from the ECU. With the second channel of the scope again the injector control is measured to get a clear picture on the regularity in the system (scope picture 2). The scope picture shows a varying control during the problem.

Picture 2: Ignition module control & Injector control

This signal is “made” by the ECU and because we already checked the camshaft sensor we decided to check the crank angle sensor. Both crank angle and camshaft signal were measured during the problem (which almost continually was present) and can be seen in scope picture 3.

Picture 3: Crank angle sensor & Camshaft sensor (both Hall-type)

The measurement immediately caught our attention because something odd shows up. There was a correct regularity to be seen in the signal and also the missing lobe was visible but there was a strange fluctuation at the top of both signals. Because a Hall-type sensor always switches to the ground part of the signal that is 0 Volt is directly related to the working of the sensor. All above the 0 Volt is been generated from the ECU and if strange things are present there this always is a result of in- or outside the ECU including connections or power supply/ground wires.

Both power supply and ground of the Hall-type sensor originate from the ECU and scope picture 3 shows that the signal is proper switched to the ground (measured against battery negative) so we can say that the ground of the ECU is well. We concentrated on the power supply of the ECU. From the wiring diagram of this system can be seen that the power supply of the ECU is provided through the system relay which is switched on by a wire coming from the ignition lock. There also is a separate power supply wire from ECU to relay with the purpose of holding the relay while the ignition is turned off. This way of controlling the relay is more and more common with the purpose of, example given, keeping the cooling fan control active while the ignition is turned off. Measurements were made on the power supply of the ECU (pin 23 of ECU). When the problem occurred again it was obvious that the power supply was the cause of the problem. This can be seen in scope picture 4.

Picture 4: Power supply of the ECU (pin 23 of ECU)

The power supply fluctuates heavily by so much as 10 Volts!! That’s not normal and the cause has almost certainly been traced.

After this conclusion a few more measurements has been done on the control wire and solid power supply of the relay but no variation could be found there. The conclusion is that the relay itself must be the problem (picture 5).

Picture 5: Relay

After replacing the relay the engine again behaved as it should. Looking back this turned out to be an easy problem of a faulty component but was difficult enough for the mechanics involved in the garage. Because of the complexity of the system and amount of components it’s no use to simply replace all components by trial and error. To measure a few components at the time seems time consuming but in the end pays itself back if the right pattern of fault finding is applied. Knowledge of the system, of course, is crucial.

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