The CAN bus error can be generated by devices operating at different speeds on a single bus (not in your case, since you have one device and an interface).
Another source of errors is electrical problems. Missing, wrong, or swapped connections.
The most common mistake is swapping CAN high and CAN low on the DB-9 connector.
I thought the same on CAN Hi/Lo being incorrect, double checked and it was right. Even swapped them around despite being convinced I was right, it made no difference.
I was wrong on the orientation of the CanBus Hi/Lo connections, the words were wrong, but the connection was right… So, Pin 36 is connected to Pin 2 of USBtoCAN for CanLo and Pin 23 is connected to pin 7 of USBtoCan for CanHi. But as I said before, I have tried both.
In the status bar of the PMU Client, apart from saying disconnected it also says “CAN1:passive ack” if that helps.
I did have one small success, I connected a Blink Marine keypad to the USBtoCAN and tried the Light Client, that worked fine so proved the USBtoCAN is working ok.
I am connecting the keypad and PMU seperatly, so they are not interfering with each other, also I do know they are running at different speeds which cannot be mixed on the same bus.
Cables are not very long, the CanBus is a twisted pair and about a meter long.
So still stumped on getting PMU talking. Any ideas anyone ?
Remove the connector and check whether the pins are pushed all the way into it. If they are not pushed all the way, they might not make contact with the device’s pin.
You can also measure voltage on floating CAN wires. Disconnect the PMU CAN wires from the rest of the bus. Measure the voltage between CAN high and ground. Do the same for CAN low. Both should be at around 2.5 V.
In order to be able to measure the voltage at the CANBUS pins on the PMU I figured I would power it up but without the connector in place. So I extracted the 4 contacts from the connector housing and carefully attached contacts to PMU pins directly.
I did not get as far as measuring any voltage, it just worked. It is communicating fine.
What is worrying is the contacts were definitely inserted fully into the connector and took a lot of fiddling to release.
Is there an issue with the plug/connector housing, it was definitely full inserted and the locking lever in place. Makes no sense. Anyone else with similar issues ?
Anyway, will have to experiment, maybe obtain another connector, anyone have a part numbers for the 39 way housing and contacts?
Termination 60 Ω verified (USBtoCAN switch ON + external
120 Ω at PMU end)
Measured floating CAN per your suggestion (USBtoCAN disconnected):
Pin 23 to GND: -4.5 V; reversing probes gives +3 V (not symmetric)
Pin 36 to GND: same behavior
Verified with two different multimeters.
The asymmetric values when reversing probes indicate the CAN pins
are floating, not actively driven by the transceiver.
PMU appears powered and awake (S-LED lit) but transceiver is
not initializing. Could this be a PMU shipped without firmware,
or a defective transceiver?
Try eliminating connector shell, extract contacts and insert on PMU pins directly, BE VERY CAREFUL OF SHORTS maybe some tape around the contacts.
Mine worked first time with this approach, I have been busy since so not had a chance to reintroduce the connector shell/housing. I would like to try another shell if anyone knows a part number, also for some more contacts so I can have a bench set up and a car set up too.
I was able to prove I can get a good CANBUS connection if I remove the contacts from the connector housing and push them onto the PMU pins directly. I ordered another connector set and tried the contacts in the new housing, connection failed. Removed the contacts from the housing and insert on the pins again and it connects.
The contacts are reaching the end of the housing with a satisfying click as they find their retaining tang. The ground and +12v on Pin 7 work as the PMU powers up OK. It is just the CANBUS that fails.
What’s going on here, it seems very strange. Any ideas anyone…
The only things I can imagine is that there is either no electrical connections, or a short circuit.
I would check for short circuits between CAN high, low, power, and ground.
With the device unpowered.
You can also compare the voltage measured on CAN high and low (in reference to ground), when using the connector shell and without to see if there is an actual, electrical difference.
This is a very unique problem.
First time seeing something like this.
Finally sorted this.
One contact, the Can Low one had spread. This means that out of the shell it connected because it was pull off centre by the wire, also it was push onto the pin deeper. When it was in the shell it was held by the shell and must have cleared the pin somehow. Anyway, closing the pin a little by pushing the tangs on its side inwards seems to have solved the issue.
I will not be using that contact in the vehicle, now I know, its getting cut off and binned.
I am guessing that the amount of contact that actually mates with the pin is not the full length of what is available once mounted in the shell.
Wonder if others have had issues with connections…
I know about a couple of cases like this, but they happened after people either pushed the multimeter probe into the terminal and spread the tabs or used a poking tool to push the terminal out.
Did you do something similar that might have spread the contact tabs in the terminal?
