CAN bus backbone vs daisy chaining

This thread is a spinoff from the below thread since it is departing from the original topic of that thread.
http://www.chiefdelphi.com/forums/showthread.php?p=1121087#post1121087

The four-port hub is used in our test bot for a while and it doesn’t seem to give us any trouble. Regarding termination, the whole backbone is terminated in two places. One is built-in to the 2CAN, the other is the 100-ohm resister I soldered at the end of the bus. If a port is unused, it is not terminated either. Each port is connected to a Jaguar. There is no termination on the other port of the Jaguar. At first I was thinking of terminating it but then I considered that the two ports on the Jaguar are probably a passive connection. If that’s the case, terminating each Jaguar would lower the impedance of the bus. So I decided I don’t have to. Let me know if I am wrong. The overall length of the backbone is really short, probably within 2 to 3 feet including the pig-tail that plugs into the 2CAN. The cable length from the hub to each Jaguar depends on how far the Jaguar is from the hub. Typically, it’s about 1 to 2 feet.

It’s been over a few decades for me, so I can’t remember any of the wave reflection math I’ve learned :). To make things worse, I am now in the software industry. So I am really rusty on this. But my impression is that since the cable length is really short and the CAN bus is running only 1Mbps, so if there is any reflection, it would be electrically insignificant. Correct me if I am way off :slight_smile:

I will when I have some time. This is very interesting.