Last Saturday evening, the HDi Xantia was not behaving itself. The transponder light was flashing for 30 seconds but the car would actually start and run. I narrowed it down to the fact that the ECU wasn't locked which to me was an OK fault to have. Better than the other way around.
By Wednesday, it would start and run for about 30 seconds then cut out. By Thursday, this was crank, start and stop within 1 second. It was intermittent, sometimes it wouldn't even start so I started going about investigating it! Transponder light still the same.
As my Lexia is broken, all I had to go on was wiring diagrams and some prior knowledge.
On Wednesday night, I decided that it might be the in-tank fuel pump and so retired thinking that tomorrow I would bang on the tank a few times whilst it was being cranked and maybe even jump the pump with direct feed and earth from the battery
Thursday came and I banged on the tank. It didn't start. So I wired directly to it and the fuel pump clearly works but it didn't start.
On the advice of a friend (a very good diagnostics guy), he remotely suggested that it would be worth checking the fuel pressure sensor by scoping it which I did. Even though the car didn't start we could see that the fuel pressure regulator and diesel pump were probably working OK because the voltage signal was in range.
I moved onto the next thing and performed waggle tests on various parts of the loom around the double relay, ECU and engine compartment fusebox, all to no avail.
I checked the battery earth points using a test lamp to load the circuit. All good.
I then checked the powers and grounds to the relay using the test lamp. Seemingly all good and capable of performing under load so I committed the cardinal sin and purchased a replacement double relay because there is no way to properly load test a double relay unlike a single relay where tools readily exist for testing those thinking that it must be a faulty double relay.
Then I noticed that one of the earth cables from the battery post terminal clamp was not well secured so I re-secured that. No change. Sometimes it would start and stop and sometimes it wouldn't start.
Spent a good part of Friday checking power and grounds to the ECU and this is where it all got a bit sketchy...
The ground earthing of the ECU (88 pin single plug Bosch EDC15C2 06) at pins 33, 49, 51 and 53 was all good. I had disconnected the ECU connector, clamped the test lamp to battery positive and touched each pin (on disconnected plug) in turn and the lamp shone brightly. GREAT! The earths are fine.
Where it became sketchy was then understanding how the double relay works. It's not at all simple for a layman like me.
I checked continuity using the test lamp for which of the wires (again, disconnected ECU plug) to see if the wire integrity was good. Pins 1, 29, 86 and 87. All good!
So what could be wrong? The relay or the ECU?
What I did next, tentatively, was perhaps stupid as I connected the test lamp to BAT+ and probed the powers to the ECU with it all connected. Now I cannot recall the exact pins which caused the relay to actuate but I think it was all of them - 1, 29, 86 and 87 but it couldn't have been. I know that now.
I had gotten as far as I could go without risking blowing the ECU to bits.
My diagnostic mate called me and asked if I wanted him to pop in and take a look. He arrived at around 7:40 on Friday evening. I went over the symptoms again with him and told him what I had tested and the process of testing. This allowed us to focus on the relay, the inertia switch and the ECU.
He has a habit of teaching me by continually questioning me to see if my understanding is solid. I fall down on many occasions and then question myself and how much I do or don't know.
I show him that there is power to pins 3, 11 (Relay 2) and 15 and 8 (Relay 1) under load with a test lamp.
So we sit down and work through the wiring diagram and write down a plan.
Firstly, we are to test continuity with a test lamp between pin 1 relay and pin 1 ECU as we know the input side of the relay is receiving power.
Secondly, pin 2 relay to pin 29 ECU.
So we perform the test and confirrm that those wires have integrity.
So finally we're left with Pins 7 and 10 at the relay. Looking at the diagram on the relay, to me these look like they are the wires that control the relay but I am not 100% sure.
Pin 10 goes straight to Pin 86 of the relay. We test integrity under load and that proves fine.
Pin 7 goes to the inertia switch. We choose to remove the connector and bridge it. We don't appear to have continuity under load. We test each segment of wire and we have continuity. We test it again whilst bridged and we do have continuity. We strip the wiring protection away and pull test the wires and double check it visually. It's all OK. Maybe I'd not backprobed one of the terminals correctly.
So now we are left with one grisly fact. It must be the ECU!!!
Mark explained to me that the relay must be earth by the ECU when the ECU tells it to earth the relay.
ECU REQUIREMENTS
With ignition on, pins 33, 49, 51 and 53 should be 0V as they are the earths.
The fuel pressure sensor is Pin 74 and at idle should be approx. 1.3V. However, this is of no use unless you have a digital multimeter that is able to register min/max values and if your car will not start. It should show 0.5V with ignition on so that is of some use.
Pins 1 and 29 should show battery voltage with ignition on.
Pin 86 should show battery voltage with ignition OFF but 0-1V with ignition on.
Pin 87 doesn't actually appear to act on the ECU until the relay is fully energised and only then if the inertia switch has not tripped. I guess this is quite critical.
So I backprobed Pin 86 under Mark's direction and we were showing 0V instead of battery voltage with the ignition off. I switched the ignition off and there was NO change.
This proved the ECU had an issue internally.
So yesterday morning, I took the ECU apart and traced Pin 86.
I could see slight corrosion when the pin goes through the moulding and then checked for continuity from that pin to various parts of the track.
I could clearly see oxidation of the copper track even though it was minute in places. So I wiped the PCB over with Isopropyl Alcohol and gently scrubbed with an old toothbrush to clean what I could of the oxidation.
I could then see under 10x magnification that the track to pin 86 was corroded through. I verified this with a basic continuity test on the DMM.
The adjacent track which I didn't bother tracing had also suffered. I suspect that that track is something to do with the immobiliser circuit.
I stripped some wire out of a broken iPhone Lightning USB cable. I could have done with a 30x stereo microscope because the copper tracks cannot be 0.5mm wide!!!
I carefully scraped back the protective conformal coating on the copper tracks and repaired the two tracks. I decided that given the soldering equipment I had that it would be too risky to try and repair both tracks in the same location so on the non Pin 86 track I scraped the coating back further along and decided to then solder back to a test point further up the board.
I then coated the repairs with pink nail varnish (which coincidentally had blue glitter flecks in it which could short the ECU). I also decided to liberally coat all the areas of oxidation I could see where oxidation has occurred to hopefully starve the oxidation of any more oxygen and hopefully prevent further degradation of the ECU PCB.
There was also one semi-conducter which could do with replacement just from a visual check. That's a good for another day.
Hope you enjoyed reading about the battle to keep War Horse on the road!
