04 Cadillac XLS CAN Bus drop out

About to start working on one of the only three XLRs this side of the pond, has a few issues but my main concern is that it had a random full shutdown, disco lights on the dash moment. Took about 20 mins to get it re-started. I was not there for that as on my test run just prior to check the ABS and HO2S it had run pretty much flawlessly for half an hour!

As far as I can ascertain from the documentation I have (when the job is confirmed I'll buy Mitchell access for a month as no UK data suppliers have anything) the HS-CAN network is ->

DLC
BCM (terminating resistor and class 2 gateway)
--> Branch to RTD (active suspension)
TCM
ECM (terminating resistor?)
ABS
ACC (Adaptive cruise)

EDIT - found a pic


Subsequent to the shutdown I had a mountain of U codes (Autel 906BT), as far as I can see the everything except the RTD on the HS network was complaining about lost comms with each other. Everything on the Class 2 network was complaining about loss of comms with everything except the RTD on the HS network. In fact the RTD was suspiciously absent from the DTC's.
The RTD is the no.1 suspect as it is in a very damp area of the trunk and there is evidence of previous work, but if it died and took the network down, why did the other modules not miss it and if it was alive why did it not miss the other modules?

EDIT - This Topoligy is wrong see post a few down www.scannerdanner.com/forum/post-your-re...-drop-out.html#47628

Are you sure the vehicle has a CAN network? The 'vette didn't go to CAN until '05.

yes, well I think so, the terminology is the same



though this shows it differently to the reference from the Systems book above

EDIT - This Topology is also wrong see post a few down www.scannerdanner.com/forum/post-your-re...-drop-out.html#47628

That's more believable than the earlier pic. At least the Class 2 lands on the DLC. I've attached network diagrams for the '05 'vette which might be close enough to help.

Thanks very much for that, very useful! - It does make more sense like that. Still don't understand the lack of RTD/ESC related DTCs. Might be a scan tool error.

Aftermarket scan tool error is definitely a possibility on something that rare. Another possibility is that RTD/ESC has been deleted with a Tech 2.

corvettesalvage.com/disabling-f55-on-a-2003-to-2013-corvette/

Unfortunately as the ESC was never not installed on the XLR (it was optional on the 'vette) the F55 delete does not work. - At least that is my understanding from failures posted by people trying exactly that. That non optional status is also why the cheap vette modules don't work and the XLR ones are $2500-$3000 for 04/05.

Anyway, I need to get the scope on the CAN lines and hopefully find that the signal is already degraded just not enough to drop out. Then I have something to fix! If it is all good then I will move on to fixing the top and trunk as that's at least broken!

In the middle of fixing the trunk/top on this and it has a network hissy fit entering a limbo power state, not on , not accessory and not off.
Battery off overnight, working fine next morning
I posted a bunch of pics and a few more details on the Pico forum www.picoauto.com/support/topic22552.html but in summary

This is the Topology


CAN from DLC/BOB during limbo event :S



Split the network at the red slash in the Topology pic, CAN from BOB = great!



Next day however when all is well again, however something is pulling down the dominant state voltage when nothing transmitting - this was on both sides of the network split with the ESC module disconnected as that was a suspect and easy to get too.





Any ideas what could be pulling that voltage down even when the network is split?

At the time of the event I was running the hydraulic roof pump, but the last time the fault occurred it was just driving, tests done just in acc mode.

Wightscope wrote: Any ideas what could be pulling that voltage down even when the network is split?

At the time of the event I was running the hydraulic roof pump, but the last time the fault occurred it was just driving, tests done just in acc mode.

Since you were testing in ACC mode the network might have been trying to go to sleep. I'd retest with the key in RUN as a first step.

OK Will do.
When I was testing on both sides of the split I deliberately waited until the modules shut down after a few minutes of ACC so I could see the voltage it dropped to when not communicating the - BCM side dropped to 1.3v and ECM side dropped to 1v

This was with the engine running : -


It's higher but still low- but the B+ would have been higher

Looking through the 155 DTC's set they are mostly to be set by modules on the Class 2 network complaining about no-com with modules on the HS CAN. However the ABS complained about everything on the HS (except the ESC), Cruise control the same and oddly the ECM had a current code for no com with BCM. I am wondering if Autel have the ABS and the ESC mixed up. If nothing else it tells me that the Class 2 network did not go down!

Wightscope wrote: In the middle of fixing the trunk/top on this and it has a network hissy fit entering a limbo power state, not on , not accessory and not off.

Take a look at Power_Mode attachment.

Wightscope wrote: This is the Topology

Does that diagram also show the Class 2 landing on the BCM? Also if the vehicle is an export build it probably doesn't have a VCIM, AKA OnStar, module.

Wightscope wrote: Looking through the 155 DTC's set they are mostly to be set by modules on the Class 2 network complaining about no-com with modules on the HS CAN. However the ABS complained about everything on the HS (except the ESC), Cruise control the same and oddly the ECM had a current code for no com with BCM. I am wondering if Autel have the ABS and the ESC mixed up. If nothing else it tells me that the Class 2 network did not go down!

155 might be a record :woohoo: Some of them may have been set when you were testing with the ESC unplugged though. Save a report, if your scan tool can, and clear them out. See what comes back next time.

If you want to do further testing with the ESC unplugged you can jump the CAN wires at its connector to maintain network integrity. Tan to Tan and Tan/Black to Tan/Black. You could also try scanning with the ESC jumped to see if the ABS module is on the network. Or unplug a wheel speed sensor and see which module sets a code for it.

Also have you seen any comm codes related to the ESC anywhere?

Thanks for the info Matt - there were 328 DTCs when I first scanned the car, I cleared the lot then scanned after it's first shutdown (158) and I have cleared it since.

I paid for Mitchell access so I have all the wiring and principals of operation - reads exactly like your attachments.. I noticed that the RCDLR is the back up power mode sender - and none of the remote fob operations work - though it does pick up the key fob as being in/out of the car.

Yes the full dwg shows the BCM as the gateway and the class 2 layout- but mindful of posting too much info.

I assumed the limbo mode was just a symptom of the network issues - for example if I pressed the start button I would get "put foot on brake" but then no crank.
I don't think it was an export model, I think it was exported (it is LHD no fog light as one was hacked in). It did have an Onstar system as apparently it was removed to stop it draining the battery (I think they just pulled the fuse!) Also the VCIM seem to deal with a lot more stuff like the radio and nav?
It clearly had/has a CAN issue from the waveform on the day and the day after.

Report attached, note no mention of ESC, also not sure what "Engine Electrical" is - seems to cover a whole bunch of stuff.

Wightscope wrote: Also the VCIM seem to deal with a lot more stuff like the radio and nav?

It interfaces with the radio, and possibly nav. Might be worth checking to see if it's still in the vehicle though. If it was physically disconnected the CAN wiring might've been hacked to bypass it. And while you're in there doing a proper spliced bypass would remove the module as a potential problem.

Wightscope wrote: Report attached, note no mention of ESC, also not sure what "Engine Electrical" is - seems to cover a whole bunch of stuff.

That report is bizarre!! On page 3 ABS has lost comm with ABS :silly: :blink:

The U0140 being both current and passed looks suspect too. GMs of that era tend to move U codes to history immediately when the fault clears.

And yeah I did "note" that the ESC isn't mentioned by any other module, or reporting any codes itself. I'm not convinced it hasn't been deleted..........

Since the scan tool isn't giving any direction all you can really do is start taking modules off the CAN network and jumpering to maintain network connectivity.

Matt T wrote: That report is bizarre!! On page 3 ABS has lost comm with ABS :silly: :blink:

Yea I noticed that too. U2108 when I look it up it says only the BCM, ESC or DSCC sets it.

I am pretty sure it can not be deleted, lots of people have tried as it is no longer available and the easiest option is just to swap out the shocks for normal ones, but it the module goes bad it sets speed limiters. Difference with the Corvette was it was never an option. I left it plugged in but pulled the ESC power plug and I got "80mph restriction" - but it still did not set an identifiable code. I'll see if I can talk to it.

Anyway as the voltage drop seems to be common even when the network is split I'm not sure it's a module problem. OTOH the network came back when the main fault occurred when I split it so that means it is a module fault!

Going back to work on the roof tomorrow, if it faults I'll look again at the network issue. One thing from my notes is a lower voltage on the power mode switch than battery and I see codes set for that too. "device 1"

Looked at it again yesterday and almost back to normal, except now I look closely I notice that about 20% of the packets are at a higher voltage. This was ign on engine off, but its the same running and acc. I did however have a battery charger on it so when I get the next chance I will scope it before I hook up the charger. It has an intermittent parasitic drain as well.



Didn't do any more work on the network as it isn't that broken so I can't fix t!

Learned something interesting today, in that the "limbo" that the car was locked in that started this was "Sleep mode" I found this out from here

www.vehicleservicepros.com/service-repai...-tips-and-techniques

In Sleep mode the can signals are 1.4-3.6 v which is exactly what I have in the first pic. So something is locking the car into sleep mode randomly, back to looking at the Power Mode Managers Matt attached earlier.
There were many more lights on than would be expected if the car was asleep but it gives me a bit more direction

After some more research (which TBH there is not a lot I can find on this) Matt was on the right track about the modes. I found this pic in an article about Allison truck transmissions (!) on the same site as above


I have a theory that the sleep mode bias voltage is present in all modes when it faults. The original fault (Sunday) was the car being locked in sleep/low power mode, then the following day (after battery disconnect) the sleep mode bias voltage was still present on the CAN bus in acc/ign and running.

The day after, it is gone.

Acc mode on Monday


Acc mode on Tuesday


Now I have too look and see what and how sleep mode is commanded and if there is a wake up signal, or if the power monitor thinks the battery is low (the car has faulted like this whilst driving!), are all modules commanded into sleep/low power mode which would explain the bias low on all the modules even when the network was split.

Just to add more thoughts to this (I'm on a roll!)
There are 2 Power Mode Managers, the BCM as the main one and the Remote door lock receiver (RCDLR) as the back up. The remote doors do not work - if that module is down combined with a faulty ignition button or it's circuit then then that's a feasible scenario.

Wightscope wrote: I found this pic in an article about Allison truck transmissions (!) on the same site as above.

Guess who used to own Allison...............

Wightscope wrote: Now I have too look and see what and how sleep mode is commanded and if there is a wake up signal, or if the power monitor thinks the battery is low (the car has faulted like this whilst driving!), are all modules commanded into sleep/low power mode which would explain the bias low on all the modules even when the network was split.

That low bias might be the sleep mode command to modules that don't have a hard wired run input??

Wightscope wrote: Just to add more thoughts to this (I'm on a roll!)
There are 2 Power Mode Managers, the BCM as the main one and the Remote door lock receiver (RCDLR) as the back up. The remote doors do not work - if that module is down combined with a faulty ignition button or it's circuit then then that's a feasible scenario.

If the back-up is down everything, other than the doors, will work normally unless the primary drops out. Focus on the primary, which is also the Class 2 to CAN gateway.

Trying to walk a tightrope here between helping, or at least giving a little moral support, and sending you down a rabbit hole........................