Lost CAN

[jon volk]

Anyone lose a CAN connection on an LDU board? Ill probably need to pull my logic board to do some investigating.

I blew a HV fuse getting on the throttle this morning, but after replacing it, everything seems to work as it should in terms of contactors and startup except the LDU CAN output being dead. I think there may be a short in the CAN circuit as if I leave the drive unit on the bus, my charger will not generate current (I connected the internal CAN of the 10kw charger to pins for external comms/custom triggers). Furthermore, for the life of me I could not figure out why I couldnt get my ibooster CAN messages to be received even after logging them without the booster being connected to the rest of the bus, but noticed if I unplugged the LDU connection it would send messages. This was discovered last night, prior to the fuse blowing.

Just curious if anyone else has experienced any CAN issues.

[jon volk]

Only getting .6v on CANL and 1.0v on CANH

[Jack Bauer]

With no messages on the bus you should see 2.5v between ground and Ch and Cl. Looks like a blown transciever but hard to see how...

[jon volk]

Yea, probably just a stroke of bad luck. Im assuming bad transceiver as well due to the interaction with the rest of the bus. I have a few extra laying around so it should be an easy fix, aside from pulling the motor out again. Checked resistance on the leads to my test point and they .2 ohm so that looks good.

[jon volk]

Foresee any issues in using a SN65HVD230DR over the SN65HVD234DR from the BOM? Looks like some less robust max voltages, but generally operable pinouts, 3.3v, etc.

[Jack Bauer]

One pin is different from memory.

[jon volk]

Ill order some 234s then

[jon volk]

Update, pulled the logic board, removed the transceiver and transplanted it onto a board with a Teensy. No CAN activity and same voltage observed, confirming the problem. Easy fix and something to look at if you ever lose CAN coms.

[FJ3422]

A spike on the HV-lines of the motor at the moment the fuse blew could have induced some transient on the CAN. Maybe a TVS diode could have protected te SN65HVD sufficient. Perhaps that's the reason they are recommended in the datasheet
For example the 'PESD1CAN', JLC stocks them. 5-cent insurance for your linedriver

[johu]

Not a bad idea, should be part of the little redesign: viewtopic.php?f=18&t=991

[jon volk]

If I can find a way to import a Designspark project into Kicad, I may try taking a swing at it. I'd like to order a run of 5 boards from JLC to have backups after a capacitor breaking a leg loose and now this.

Im also going to throw a model s later revision "smart" pyro fuse in there instead since it makes sense to me to use the factory matched part and itll drop in my contactor box without any real effort.

[Jack Bauer]

Sadly Kicad don't talk designspark. I tried. I'll jump in and tidy it up:)

[larsrengersen]

And is it perhaps an idea to also add a common mode choke and isolated transceivers (but that means you also need to add an isolated power supply). And in terms of ESD, I would also recommend the PESD*CAN. Myself I use the PESD2CANFD24V-T.

[johu]

Haven't seen isolated transceivers in
- The Leaf BMS
- The Leaf inverter
- VW ECU
- VW CAN gateway
But yes common mode chokes and protection diodes are actually quite standard, just so far we mostly got away without

[jon volk]

New transceivers showed up. Despite my dodgey soldering, we've got CAN.

I also now have iBooster brake position showing up over CAN now to my VCU. Even before the fuse "event" there must have been a problem with the transceiver putting low voltage onto the bus and blocking some messages. All works as expected. Now to finish mapping motor regen to the ibooster output instead of pressure transducer.