TIDA-01487: 5 MBaud CAN-FD repeater design

Hi Thomas,

thanks for the quick replay.

If i understood correct, the overall delay TX-RK-Loop must be optimized. 

I see that there are some new devices  (e.g. TCAN3404) with low loop cycle time. 

do you think it could work without using a isolation device?

best regards

Martin

Hi Martin,

This is correct, you need to reduce the time of the TX/RX loop as much as possible. If you can remove the isolation device, this would for sure reduce the time of the loop in the benefits of achieving a higher baud rate. It would need to be tested if you then an achieve 5Mbaud - but I would think that this might be achievable then.

I also looked at the TCAN3404 device, but don't see the tPROP(LOOP1) & tPROP(LOOP2) time getting improved much. It is more or less a bit higher.

In addition, start with the arbitration logic as provided in TIDA-01487 design guide. Use logic family that is fast (I think we already used a fast one), and play around with the timings of the arbitration logic (the R/C constant, that is feed into the buffer with the Schmitt-trigger inputs.

Note that I cannot test this improved method on a real TIDA-01487 board here at TI Freising as we don't have access the original board design anymore. 

I hope this helps. Let me know if you have additional questions.

Regards,

 Thomas