These two posts were written real time with the progress of troubleshooting as the event unfolded. By mid morning today I knew I was onto the scent to the cause of the problem. It is not the ignition hall effect sensor in the distributor. It is a component in the ECU module that failed the moment when the engine stalled when I was ready to leave for the city. I count myself very lucky that it occur right in the garage rather than on the road. Also have I rely on even the best of Vanagon mechanic I have good reasons to believe he would not perform troubleshooting down to the component level as I did.
Yesterday while debugging the problem on Isolde, I finally had enough of leaving this Pace soldering iron on accidentally for prolong period of time as it has no power indicator. The only way to tell that it is on is the poorly design LCD display which has narrow viewing angle. You eyes have to be perpendicular to the display to read it but most of the time you eyes are at much higher level. I decided to hack in a LED power indicator.
It is a commercial quality programmable temperature soldering iron that was quite expensive at one time.
the hardest part is to drill a hole on the aluminum faceplate as the membrane keys are built right into it, and I could not remove it without damaging; so I had to drill the hole by holding the faceplate by hand
the stepper drill bit comes in so handy to enlarge the hole to the common imperial dimensions
even without schematic I had no problem finding 5Vdc and ground on the microprocessor circuit board and get the power for the LED
a vexing shortcoming rectified
both soldering irons now have power indicator; the Weller on the left is temperature regulated as well but you have to select the soldering tip that has a fixed temperature for the job
A pesky nuisance out of the way, I resumed my fault finding of the ignition problem. The hall sensor not just is used by the ECU to generate the spark ignition switching signal, but also the injector timing as well. I monitor the hall sensor signal at the ECU's main connector pin with the oscilloscope while sitting in the driver's seat cranking the engine. There was no hall sensor pulses.
the group of parts on the left side are the parts diagram for the distributor and the hall sensor, but mine ferrous rotor looks nothing like as depicted
my ferrous rotor that triggers the hall sensor is made of stamped steel
I used the scope to monitor the hall sensor, ignition switching, and fuel injector pulse but they were all static
the hall sensor has three pins - ground (1), power (3) , and timing signal (2)
it turns out it is possible to probe these signals while the connector is connected to the distributor
When I monitor the three pins at the distributor connector, I was able to determine that the hall sensor was getting no power, so the most likely culprit is inside the ECU. I already have the ECU open and were probing the signals as well as trying to find anything that looks abnormal visually. I thought I smelled something hot, but it was very faint so I couldn't be sure.
As I inspect the components and solder joints very carefully I notice this wire wound power resistor protective coating has some cracks. I was not ready to condemn it and jump for joy. Instead I try to determine if it is open, or even intermittent without physically stressing it. It takes some experience to know how to do this. Anyway I deemed it likely not the cause of the problem but an effect.
As I turn the ignition on to check again I can see a wisp of smoke something between the two PCBs. I knew then it must be the 22 ohm power resistor is being overloaded.
this 22 ohm wire wound power resistor has cracks on the protective coating; one end of it connects to the pin that supplies power to the hall sensor
I didn't think the problem is the power resistor, or the wire harness between the ECU and the hall sensor. I started looking for a short, and sure enough the hall effect sensor power is shorted to to chassis ground. I further confirm the short is not on the chassis harness but within the ECU. For me it is quite easy to debug boards like these even without schematic, as most are discrete components. I just need to find components that bridge that node and ground and in no time I identified the likely culprit is the 100uF 25V electrolytic capacitor.
the suspect 100uF 25V electrolytic capacitor
the only way to confirm it is to pull it off the PCB and sure enough it failed in a dead short
As the 100uF capacitor is down stream of the 22 ohm current limit resistor for the hall sensor, its being shorted caused the resistor to dissipate 6.54W of power which is a lot for its size especially in an enclosed case. Note that the resistor is raised 5mm above the PCB precisely for cooling reason. Otherwise it would had charred the FR4 PCB.
I don't have the exact value capacitor so I decided that a 47uF 25V is good enough for the job
I was so sure I have the problem licked that I removed all the monitor signal jumper wires an proceed to start the engine. It fired right up. Only then that I reassembled the ECU module back together and reinstall it under the bench seat. My assessment from the outset that the hall sensor is very unlikely to fail was spot on. The problem turned out lies in the ECU module. Fortunately for me I have open it multiple times and feel quite at easy to do it again. All the electronic engine management signals on the Vanagon can be accessed directly or indirectly on the ECU's 25-pin connector for one stop shopping.
the ECU all buttoned back up
I went for a short test drive in the ranch and all is well now; I am going to give her a good wash at sunset
Another difficult debug done, and I am pleased with the outcome. Most mechanics would shotgun with expensive parts and charge you for them, and if you are lucky they are able to "repair" it. For me I fault-diagnosed down to the failed component inside the ECU, with pin-point accuracy. My cost is $0 except my time. The only sad part is this debug and repair does not improve my track record as I am already at 100% success rate. 😂
Update:
I dissected the failed 100uF 25V electrolytic capacitor hoping to see the failure point. I was expecting a visible short inside due to dielectric barrier failure. This capacitor is of very high quality. It is an EKM made by United Chemi-Con. It was not easy to cut open and remove the rubber plug that main job is to prevent the electrolyte from leaking or vaporizing. The electrolyte typically is of harmful chemical so I don't expect to have children after.
I couldn't perform a tidy dismantling, nor could I see a clear breach of the dielectric membrane. It is quite impressive that this vehicle is over 30 years old now.
No comments:
Post a Comment