Hi all,

1999 XK8 4.0. I made a mistake during AC diagnostics — grounded Pin 1 of the AC compressor clutch relay socket with the engine running. Engine shut off immediately and won't restart.

What happened:

• Engine Compartment Passenger Side Fuse 14 (10A) was blown — covers ignition coils relay, throttle motor relay, AC compressor clutch relay, HO2S heaters, EGR and EVAP
• All other fuses good
• Fuel pressure confirmed present
• No spark — confirmed by pulling coil pack

What I've ruled out:

• Fuel system — pressure confirmed
• All fuses — checked with ohm meter

Interesting observations:

• Ignition coil relay receiving correct signals from ECM but not switching — tried 3 relays
• "Stability Control" warning on dash — new since the incident

Could the ABS/TCCM be sending a false fuel cut-off or ignition retard via CAN bus? Could the ECM be damaged? Could it be the immobilizer? Has anyone seen this before?

Any help appreciated — keen to sort this myself before going to a specialist.

Thanks

 

Hello,
I would start by disconnecting the battery (to reset the control units), in case you haven't done that yet.
I hope you find a simple solution.
Best regards, Willi

 

I think it's unlikely you damaged your ECU - you grounded an output pin via the relay coil. The pin itself turns the relay on by grounding it, so it can handle whatever current that might produce, albeit in the wrong direction.

This is where I would be looking:-

• Ignition coil relay receiving correct signals from ECM but not switching — tried 3 relays

Are you saying the relay is being grounded correctly by the ECM (tested with a bulb going to +12v and the relay pin) but not coming on? In which case it can only be the +12v from the fuse box - maybe something got fried in the fuse box itself. Anyway that is what you need to confirm first before suspecting the electronics.

 

Have you tried reading the OBD codes with a decent scanner?

Richard

 

I am working with Claude AI to help me. this is the results of todays diagnostics based you your all posts, Claude thinks the

Next step is to: Unplug modules one at a time and recheck CAN bus voltage after each to identify which module is pulling the bus down.Any advice on which module to start with on the 1999 XK8 would be appreciated.

This is what i did today:

Ignition coil relay testing:

• Relay socket has 12V on Pin 3 (power supply) and ECM correctly grounding Pin 1 (trigger signal)
• Pin 5 shows 0V — relay not switching despite correct signals
• Headlight bulb test confirmed full current on Pin 3 — fuse box internal damage ruled out
• Relay socket visually inspected — no burn marks or damage visible
• Different relays tried — same result every time

OBD Scanner testing:

• Tried iCarsoft LR V3.0 in Jaguar specific diagnostic mode
• Scanner locked up completely when trying to communicate with vehicle modules
• Could not read any module specific codes — scanner just froze

CAN Bus voltage test at OBD port:

• Pin 6 (CAN High) = 9.4V — should be 2.5-3.5V ❌
• Pin 14 (CAN Low) = 9.4V — should be 1.5-2.5V ❌
• Both pins reading identical voltage confirms CAN bus is shorted or severely faulted

 

I don’t know if I’d trust an A-I to diagnose a 27 year old Jaguar.

Z

 

One of us is looking at the wrong wiring diagram. On mine it shows the ignition relay as:-

So pins 1 and 3 should be Bat+ with ignition on. Pin 2 should be gnd when the ECM wants to energise the relay. Pin 5 will be at Bat+ when the relay is on.

For the can bus, with the ignition off, measure the resistance across CAN H and CAN L - it should be 60 ohms. Also measure resistance to ground.

If you want to unplug modules, start with each end of the CAN bus, which is the ABS module and the instrument cluster.

By the way, Claude is a compulsive liar and then pretends you were the one who got it wrong.

 

I think you're on the right track. It's not a job that I have ever had to do as I've always known where the short/disconnection is but you presumably need to start with anything that you touched or moved since the car was running properly and check the wiring for continuity and the modules for damage by unplugging or, ideally, back-probing the connecters whilst still connected and live.

If I were to have to do something similar I would probably purchase an OBD2 breakout box as used by people like DiagnoseDan on YouTube. The boxes are not expensive but Dan uses one to diagnose problems with the OBD systems.
Richard

 

Its worth all readers noting that the ECU switches in the earth not the live . So each of the outputs is backed by a MOSFET so if you were to join an output to 12v live the MOSFET would be blown and rendered open circuit. I had this happen with a ecu earth for my heater pump and had to get inventive with other outputs.. just be aware .
The ignition relay is not switched by the ECU but by the ignition switch in the old fashioned way

 

@dibbit I think yours is the wrong diagram - for the SC version.

For the NA 99 XK8 my diagram is showing



So looks like there should be 12v on pins 1 and 3 with the ignition switched on and as long as pin 2 is grounded the relay should switch giving 12v on pin 5 and all coils should be powered.
As @scottatl says there is 0v on pin 5 there is obviously something wrong with the relay or socket

Just re-read post #5 and if it is correct that pin 1 is at 0v this is wrong and should be 12v, This should be switched by the EMS relay, controlled by the ECM, via fuse #14 in the engine management fuse box

 

In this instance, the ECM doesn't switch ground for the ignition coil relay. Pin #2 is permanently grounded by the black (or blue/black depending on which diag you're looking at - the pinouts are identical)

The ECM supplies B+ to pin 1 via the 10A fuse in the engine management fusebox to energise the relay when the ignition is on.

If you have B+ on pins 1 & 3* and 0V on pin 2, there should be B+ on pin 5. If not, the issue is in the immediate vicinity of that relay.


The CAN voltages are awry - they should be, IIRC, ~2.4V for CAN low and ~2.7V CAN high. Is it possible you measured against B+ rather than ground? They should always be similar in any case as the resistance between the two wires is low. If those are the true voltages to ground, definitely a problem there.


* use a test bulb in conjunction with the multimeter

 

I have a nice ending to the story:I read on another post that back probing fuses is not reliable. Back probing a fuse can give you a false "good" reading even when the fuse is damaged internally. A fuse can have enough contact to show voltage but not enough to carry the current needed to actually switch a relay. Always pull the fuse and check it directly.

So to summarize what actually happened:

1. You grounded the AC relay socket which caused a surge
2. Fuse 14 in the engine management fuse box was partially damaged — not completely blown but not fully conducting either
3. This meant the ignition coil relay never got enough power to switch properly
4. Everything else — ECM, CAN bus, wiring — was completely fine all along

Thanks everyone!!!!!!!!

What we all just did was "Human in the Loop" with AI. With AI I had a great time testing things in ways i just had no idea how to do, and it helped me read the complex documentation, but without human in the loop the problem would not have been solved. AI blew the fuse and did not think about the back probing issue, even though that is how i told it that is how I was testing. I also put this thread in the AI project file and it helped design test from all the good input. A lot of fun had!!!