Regarding CAN buses. What do you do when you have devices all over the machine? Should you split the bus like a star topology, or does the cable have to visit every node in a linear pattern?
A line with 2 termination resistors, at the beginning and at the end.
So it’s always a single line with no forks or star topology?
It can be star too. You just have to consider the lengths and propagation delays.
Is a can hub a thing? Do branches need 120 ohm resistors as well?
Sorry for the late reply. No, only 2 are required. One at either end of the furthest points on the network. We use CANbus on a tree topology for an ROV control system with no issues at all. But speed is 250Kbps. Bus lengths are no more than 4-5 meters from the main CAN controller. The surface to ROV Is via RS232 and gets converted to CAN in the ROV. We do that as almost all ROV multiplexers are RS232 or RS485. Some of the newer ones support CAN, but as this is an upgrade to an existing older ROV, so they don’t have this feature. Like Modbus, CAN just works and makes for a drop in new module so easy.
1 Like