SN65HVD251: SN65HVD251

Christian,

Thank you for bringing this issue to our attention. I just want to make sure I understand the setup completely, if you have any schematics, waveform screenshots, or a block diagram to share that would be great!

• "...in addition D(CAN Transmit) and R(CAN Receive) lines are isolated from the power supply via a dual channel digital isolator..."
  • Do you mean the signals coming from the micro controller are isolated from the transceiver? Or the pull-ups on TXD and RXD are isolated?
• "Both the D(CAN Transmit) signals are tied together and connected to my micro-controller CANTX port, the two R(CAN Receive) signals are gated to the micro-controller CANRX port via a 2-port AND-gate."
  • Do you mean both the D and R signals from nodes B and C are tied together to one Microcontroller that controls both nodes? And that is separate from the controller of node A.

Thanks for bearing with all the questions, I want to make sure I fully understand the system before I approach the problem you are seeing.

Regards,

Hey Eric

Thank you for your swift response.

I have attached a simpel block diagram of the set-up.

Best regardsIsolated CANBUS.pdf

Christian

Hey Eric

I have been in a meeting the whole day and have therefore not had a chance to provide additional data to you.

Do you need more info from me in order to evaluate my design?

Best regards

Christian

Christian,

I wasn't able to get to this today, but I will look tomorrow and have a response.

Regards,

Christian,

From the problem description you've given, it sounds like the correct thresholds aren't being met on the bus. A BIT error is given when the level of the bit that was received doesn't match what was transmitted, and the STUFF error indicates that six bits of the same state were received (think bit stuffing on the controller side). In both cases, it is apparent that the correct voltage levels aren't being interpreted correctly by the transceiver from the bus, or the transceiver driver is having trouble driving the correct levels to the controller. The former seems more likely.

There could be two possible causes of this: since isolation is being used, the GND references could be off between the three transceivers, though it doesn't look like you've connected anything incorrectly. The other cause could be too much capacitance on the bus, but in your application this doesn't look to be the case.

• Can the customer, just for testing purposes, keep the GNDs consistent (don't isolate node C from A and B) and see if the errors go away? If they persist, try connecting the 5V all together and see if they go away.
• Is there any way to get oscilloscope screenshots of the bus while the errors are occurring? And probe before and after the AND gate?
• Can you explain the function of the AND gate? Both nodes B and C are receiving the same information from the main controller, and then the laptop and control panel are supposed to respond with a dominant bit otherwise the response is recessive. What exactly is this application?

Regards,

Hey Eric

Thank you for your inputs, appreciate it.

I will try se if we can do the tests you propose and will let you know.

The AND-gate was put there because we didn't want to tie the two R outputs together in order not to damage the tranceivers.

Best regards

Christian

Hey Max

I think the problem is that the devices can't see each others' messages(collision), and I also believe it can be solved by adding additional 2 AND-gates to model the CAN bus' dominant state such that the TX signals are AND-gated with the other devices transmitted data. I will test it mid next week and will let you know and will describe the modification more detailed.

Best regards

Christian

Hey Max

Thank you very much for your link to the TI reference, it seems to be what I'm looking for, I will take a closer look at it today.

For your info I have attached a schematic of what I had in mind before you sent me the link to the TI reference design, I guess this design could end up in an infinite loop as you explaned earlier.

Bnew isocan.pdfest regards

Christian