96 XK8 suspected CAN BUS issue (ECU TCU ABS Cluster)

[jwpalfrey]

Good Morning all,

Further to my post in the Engine section I now think I've narrowed down my engine cutting-out issue to a fault somewhere within the CAN network. Or at the very least there is a fault within the CAN network AND on the engine. But I need to fix both if indeed there are two issues.

So far it's shown P1797 P1367 P1368 P0706, and now shows P0335

Initially it only showed P0335 and so I changed the crankshaft sensor.

Anyway I've been getting "Engine Stalled" appear when the engine cuts out, this is usually after "Engine Fault" appears, but within the same time-frame, Trans Fault and Stability control also come up on the dash, the trans goes to limp mode (4th gear) and recently the rev counter and temp gauge have both intermittently stopped working even though the engine is running.

I only have a cheap engine fault code reader so I've only been able to read some of the codes. Doing some reading up the above symptoms seem to point to a fault somewhere on the CAN network.

Anyone have any experience of this and any advice, I'm studying the 1997 electrical guide to locate the pin-out diagrams and module locations to start checking connectors and part numbers (if I can obtain identical parts to swap out to try and locate where the fault lies)

I have a few questions, can the ECU TCU and instrument clusters be swapped without needing re-coding with WDS/IDS? if so which ones?

[mike020150]

In general, if a module is part of the car's security system, it will need to be recoded. The engine ECU and instrument modules definitely are so it's unlikely substitutions will work correctly. Mike.

[jwpalfrey]

Thanks Mike,

I've worked on other VAG group vehicles and it is certainly the same for those. Thanks for the info.

[Astromorg]

There's a whole set of algorithms for checking the CAN system in Jaguar Technical Information System (JTIS) in the Electrical section under Module Communication Network, but broadly it involves starting at the diagnostic socket and checking for open circuits or short circuits of the two CAN wires all the way through to the ABS module at the end of the line. It's not as simple as just disconnecting the wires and checking them; the checks include measuring resistances through the modules too. You really need to get a copy of JTIS. Here's a link if you need it
(I've edited this to a better link)
https://redirect.viglink.com/?format=go ... n_1999.pdf

Just click on the green download button and save it where you want. When you then open it, go to Electrical, 418-00 Module Communications Network and follow the instructions!

As you've got the wiring diagrams, Fig 19.1 is the one showing all you need, but it's going to be a long job by the look of it!

[Astromorg]

Further to my last post and checking on some details, it turns out that the two CAN wires have 120 ohm resistors across them, one in the ECU at one end of the system and the other in the ABS module at the other end. If all is well, measuring the resistance at the diagnostic socket between PINs 6 and 14 should be 60 ohms as the two 120s are connected in parallel. I've checked mine and its 61.6 and as I have no problems, I think the 60 figure is a nominal one. Others have also reported 61.5 ohms. A higher or lower figure points to a poor connection or an open circuit.

Next stage, if the CAN is disconnected at any of the relevant modules, a good system will show 120 ohms across the wires leading upstream and 120 ohms on the module pins downstream. If there's a bad/high resistance connection, one of those 120 figures should be high and if an open circuit then low. Repeating that procedure at each module in turn should allow the problem to be narrowed down.

Hope that all makes sense!

[MaxTheDog]

Seems that this post is dead. However, for future reference for anyone, the two 120 ohm resistors at the ends of the CANBUS are there to ensure that the two digital states, 0 and 1 (roughly 0 to 1.3 volts equates to logic 0, and 3.5 to 5 volts equates to logic 1).

Without the resistors there is the likelihood that the two logic states can ‘float’ about between 1.3 v and 3.5v giving the microprocessors no clue as to how it should interpret the signal - i.e. is it on or off if it’s a switch, or complete gibberish if it a data stream and should be an analogue value.

TLDR:
As Astromorg has stated above, around 120 divided by 2 is the resistance reading that should be obtained if they’re still present and correct. A small amount of difference from this value is ok and it’ll be down to the % tolerance of the resistors and a small amount of resistance from the data cables.
Hope this will be of use to someone.
Regards,
Michael.

[mike020150]

It's good to record this information for posterity even if no-one is looking for it right now. Many thanks. Mike.

[Astromorg]

Hardly seems like eight years ago since this thread was last active and I hope it's another eight before it's needed again . Whenever CANBUS problems arise anywhere online, they usually seem to take a lot of time and effort to get them sorted.

Thanks to Max for adding helpful and interesting info.

[MaxTheDog]

Hi All,
I feel I should post a amendment. It’s been a long time since I’ve worked on this type of equipment and on doing some further research, the information I posted previously was incorrect.
Strictly speaking I was talking about pull up and pull down resistors, whereas I should have been discussing end of line resistors (EOL). EOL’s are required to suppress the reflection of data back down the cables due to the high frequency these systems operate at.
Long story short - it still interferes with data transfer and can confuse the microprocessor.
I hope this puts the record straight and no one has invested any wasted time using the incorrect information.
Regards,
Michael.