Hello everyone,
I've had my car for a little more than 2 years and it has been mostly trouble free, however this season it has failed to start (does not even turn over) multiple times. Sometimes I get the gearbox fault, limphome mode, ASC not avail etc... on the display. Sometimes it just does nothing. Everything seems to point to a weak 12v battery causing electrical ghosts, but voltage looks good each I check it and when it does turn over it cranks very fast. I agree that there are electrical ghosts in the system and there does seem to be some voltage sensitivity, but it seems to me that replacing the battery is a bandage that does not address the underlying issue that so many of these cars have. I've been trying to capture the issue when it occurs, but have been so far unable to collect good data during the fault. I'm attempting to capture the raw CAN data from the diagnostic connector (Pin 6 and 14) to see if I can identify which controller is not responding correctly. It looks like the CAN system is very basic and contains a fairly small amount of data. So far this is what I see:


MSG ID; Bytes; Cycle time[ms]
0x64h; 3; 8
0x96h; 6; 4
0xC8h; 6; 10
0xFAh; 7; 8
0x12Ch; 8; 4
0x1F4h; 3; 10
0x3E8h; 6; 10
0x44Ch; 8; 100
0x4B0h; 8; 8
0x7D0h; 1; 200
0x7D1h; 1; 215
0x7D2h; 1; 200
0x7D3h; 1; 200


Curious if anyone else has attempted to decode the raw data or knows of a reference that details the message structure. What I'm primarily looking for at this point is an indication as to which controller is transmitting which message.


Thanks

 

You're way over my head trying to decode the CAN. But just to be sure it's not an electrical problem check the voltage at multiple points while trying to turn the engine over. The bulkhead posts and grounds, the mains fuses in front of the battery, even the battery cables themselves, not the posts.

Do you observe the red P being lit, and the other letters when you shift? That's another indication of trouble.

You say the battery checks okay, do you mean 12.5 V or more?

 

That part is documented in the electrical manual (at least it is for the '03). There are a number of tables in the appendix that document the receiving module(s) to each of the messages and their source.

 

+1

JTIS lists the CAN messages, source and destination(s) but unfortunately not the associated IDs. I don't know where you would get that information: much if not all of it is proprietary.

An OBD2 reader would at least provide codes that we can decipher.

Replacing the battery isn't a silver bullet, but the underlying issue is that the car's electronics are sensitive to voltage.

Some info here:
CAN Hacking: The In-vehicle Network | Hackaday

Mike

 

I'll keep my eyes open for that, but I can tell you that I could not shift the car out of park and that the HVAC control also beeped and said Er on the display which leads me to believe there is a communications problem between all of the controllers.

 

I've seen the CAN lists in JTIS, helpful but still need to determine ID. Short of disconnecting individual controllers (and bridging the CAN bus across the connector), it will be difficult to determine which is the transmitter of each message. The JTIS list is really a list of signals and those signals will be grouped within a message. There are only 13 messages and of the 5 nodes only 4 transmit. I also assume that the last 4 messages are status tokens of each transmitter, so we really only have 9 messages with data. Based on the JTIS list, I can get an idea of which types of signals are grouped together and then try decoding the raw data looking for signals in groups similar to the JTIS list. Things like throttle pedal, brake pedal, engine speed, vehicle speed, and temperature should be pretty easy to spot in the raw data. It should also be pretty easy to find the instrument pack because we know that CAN message will contain information about the status of the headlights which we can manipulate when looking at the raw data. I'll be poking around at this for the next couple of weeks, but would love some feedback from the experts who have gotten to the bottom of the voltage sensitivity issue. Does it get worse with time? Is it just a matter of cleaning connection, replacing fuses and worn relays, or is it more sinister like failing caps within the individual controllers?

 

Cranks very fast does not prove the volts are OK.

CAN messages are interesting but tough to decode and I reckon unlikely to help you figure out what's wrong.

Comms faults are usually bad battery/power/ground or chafed wiring / bad connector.

You can check CAN circuits if you wish but nothing so far strikes me as it being relevant.

 

Here it is again for reference.
Please check the INST PK.

bob

 

Mine is out of the VIN range as far as I can tell, but this seems to be a good place to look. Battery voltage today after sitting 24 hours was 12.43 at the big fuse block in the back as well as the main terminal under the hood. Some success with decoding CAN today. 0x12Ch is certainly from the ECM. Bytes 4 and 5 make a 16 bit signal for Engine RPM, Byte 1 is engine throttle position and Byte 2 is accel pedal position (but not exactly since this is still moving around during cruise control operation). Byte 6[Bit 0] is brake pedal active. Byte 6 [Bit 3] looks like the cruise control switch and Byte 6 [Bits 4 and 5] seem to be related to cruise control engaged.


I probably won't be able to look at this for the next week or so, but as I find new info I will post it here just in case anyone else finds this useful.


Also in case anyone is interested I'm using this CAN adapter to view this information.


https://gridconnect.com/pcan/can-adapters/can-usb.html

 

While you are viewing the CAN information, take a minute to spray some electronic cleaner in the ignition barrel, followed by 'canned air', and a lock lubricant, paying particular attention to the little 'door' or 'flap' that partially conceals the ignition hole.

This 'door' or 'flap' gets gunked up with crap and fails to fully close. When this happens the vehicle does not go into 'sleep' mode, and continues to draw power from the battery. Over time a daily driver's battery voltage will dip to the point that erroneous error messages are displayed and disappear leaving no trace, once alternator voltage kicks in, as well as doing other weird things, i.e., windows loosing their travel limit memory is a dead giveaway. My top started it's open cycle upon inserting the key without being near the top down button one evening, the next night the battery was dead. Cleaned the cylinder 'door', charged the battery, and all has been fine.

 

Failing electrolytics seem to mainly affect the ECM, particularly in the very early units (97). Doubtless the engine bay environment is a factor here.

Remember that even the youngest examples of our cars are now 10 years old; unlike good wine they don't improve with age without some TLC. Connections can work loose and corrode, and stuff just plain wears out.

These cars are tetchy about voltage, and it's not confined to just a single module, e.g. window memory and, IIRC, adaptive cruise are both susceptible to voltage drop. Quiescent battery voltage may be OK, but what does it drop to under load when cranking?

You're obviously keen to progress with the CAN message decoding, but unless you can get hold of the DBC you've got a big task, and I'm not sure it will help with a diagnosis. It might be embedded in the Jaguar software?

Mike

 

Been looking for an excuse to hack the CAN bus anyway. I have a lot of experience with CAN bus analysis and debug so if it helps with the diagnosis great, but if not we'll benefit from a best guess DBC file for our cars. I'm away from my computer with JTIS installed (only have my iPad for vacation). Does anyone have PDF links to the 2000 xk8 schematics? Looking for something to keep me busy.

 

Go to jagrepair.com

 

Thanks. Should keep me busy.

 

Well that explains why you are doing troubleshooting the hard way!

As you may know, generic codes can be decoded with a $10 Chinese ELM327 scanner.

Jag proprietary codes can be docoded with a real or Chinese clone Mongoose JLR scanner (see https://www.jaguarforums.com/forum/g...an-tool-66558/), or an AutoEnguity scanner with the Jaguar package.

If you really get into this, and add extended Jaguar proprietary PIDs to the widely used TORQUE app (for use with the $10 Bluetooth ELM327) you would certain have many grateful fellow members!

Cheers

 

Hi 03 XKR,

Please enlighten me with the part you referred to as "the ignition barrel"? Can you post a photo or describe the exact location of that part?

TIA

 

He's referring to the lock cylinder where you insert the key

 

Double check the brake switch. Pressing the brake normally allows the activation of a solenoid that releases the shifter out of 'P'. In a quiet environment, you should hear it.

There are many posts on this, this is a common failure.

Best of luck, keep us posted.

 

Quick update:
Thanks everyone for comments and suggestions.
Car was sitting for 1 week while I was on vacation. I measured the voltage at the battery and at the OBDII connector. 12.19 volts. I also checked the door in the lock cylinder - I looked like it was operating smoothly. I also checked the CAN bus to ensure that there was no short to ground. CAN bus looked good with the correct 60 ohm resistance across CAN high and low.
I was expecting a fault with such low battery voltage, so I connected my laptop and did a CAN trace to look for issues, but the car turned over and started perfectly. Brake switch seems to be working fine and I hear the click at the shifter. I'm going to attempt to make a measurement of voltage during start, but I need to build some hardware first so that I can measure and record the voltage at several points simultaneously when I key on and start. Any suggestions of the 4 or 5 most problematic points to check?

 

A totally uninformed guess would be to start with the main battery positive feeds to each of the fuse boxes which all come from the "High Power Protection Fuses" in the trunk (boot).