1HCL not passing CAN bus thru out connector, evil dust ball?

[PuddleOfPolyglot]

I'm daisy chaining the CAN bus on a set of 3 1HCL expansion boards off a 6HC main board. One of the 1HCL's has a bad exit CAN bus RJ-11 port, and won't send. I've switched cables around, it's not a bad cable, it's not the termination jumpers, I measure zero ohms between the gold wires on the (same position) tops of the connectors, and I've looked at the schematic and verified everything I can see that isn't covered up by the connector bodies. It's either a very very determined invisible chunk of dust sitting on the inside of the connector or something sneaky I'm not seeing yet.

I don't seem to find any trouble exactly like this mentioned in the forums here. Before I clean off my desk and set up an oscilloscope, is there anything obvious I forgot to look for (other than trying to file the inside of the connector) ? Maybe this doesn't happen very often and I'm lucky? I'm thinking I'll make a special cable with a short tap on it for that particular 1HCL board.

[dc42]

@PuddleOfPolyglot does the soldering of the CAN socket pins on the underside of the PCB look OK?

Have you tried a different RJ11 cable in that socket?

[PuddleOfPolyglot]

Hey, thanks for the reply!

I haven't dug out my old network tester but all the cables I've made so far work fine when rotated around. Any of the 3 1HCLs that follow the affected one blink the status LED fast, any between the 6HC and the affected blink slow and show up in the list on the machine-specific config web page.

I might be hallucinating. All of the solder joints look really good, nice processing. I'ma do all the tests again, yeah, it seems like I must be missing something easier.

[dc42]

@PuddleOfPolyglot I presume you've looked into the two CAN sockets to check that the gold plated fingers look OK and are not bent?

If you swap over the cables going into that board, does that board also fail to get CAN sync? If it does then that would confirm that it's just the CAN socket that is faulty, or not properly connected to the electronics and the other connector.

Please confirm that you have the termination jumpers set correctly, i.e. fitted on the last board in the bus and not fitted on the other boards.

Also please confirm that you have set the 1HCL boards to different addresses, by connecting them t the main board via CAN one at a time and changing the address to a unique value from the default.

Unless we can resolve this, we'll offer you a replacement under warranty; but first I'd like to see a photo of the back of the PCB where the connectors are soldered, and one or two photos looking into the CAN sockets.

[PuddleOfPolyglot]

So I spent a few hours over a couple of days trying to figure this problem out, then spent a day reading the internet to see if anyone else had it, followed by posting a description of the problem.

Within a day, I get one of the best answers I've gotten on any forum in the last ten years. So guess what happens when I go check the problem again?

Evil invisible dust bunny has fled. Computer acts like nothing was ever wrong. Thank you for the magic blue smoke engineer voodoo.

The 6HC and three 1HCL boards CAN bus links up within a second of booting. It all works perfectly and is a modest complement to my building skills (no comment on my cleaning skills).

"replacement under warranty" - that's what I was trying to avoid. I do not like returning goods to well-meaning manufacturers and frankly... returning open-source hardware is against my religion. Everything else about the Duet3D hardware/firmware/software is working perfectly by the docs. I'm very impressed!

Thank you again for the assistance, I have so many many questions.

[dc42]

@PuddleOfPolyglot I'm glad you got it working! I'll mark this as solved, but please revert it to unsolved if the problems occurs again.