We’re considering using CAN this year for the position mode during hybrid and speed mode during teleop.
I noticed that there were several discussions last year on how to use an encoder with a two CIM gearbox. Some suggesting splitting the encoder wires but the ground wires needed to be isolated. Others suggesting connecting the encoder to the master jaguar and then using the voltage output of the master as the input into the slave jaguar but that sometimes created problems. I may have missed other methods. Is there any consensus on the “correct” way to use an encoder with a two CIM gearbox?
The correct way to do this on board the Jaguar is to have the Jaguar firmware modified to include a “follow me” mode so you can implement a master and a slave Jaguar. Since that didn’t happen for this year, the correct answer is to connect the encoder to the cRIO and to run the PID algorithm on the cRIO.
If you use diodes the forward voltage drop will reduce the output voltage from the encoder (0.3VDC - 0.7VDC) and if you look at the encoder under that load you’ll discover that might not be desirable.
More importantly you won’t really get noise isolation from that.
Just to share our experience from last year. We ran into the same encoder splitting issue on our lift and drive train… We were not successful in splitting one encoder to two Jags but were able to get something to work that’s along the lines of what you’re describing as a “follow me” mode.
We set one motor controller up in a speed control mode over CAN. We then queried that “Master” controller’s output voltage. Then commanded the “Slave” motor controller (in VBus mode) to the Masters’ voltage.
This obviously is not ideal as the slave controller will undoubtedly lag the master, however it worked in our testing. That said, we ended up not implementing closed loop control on the Jags for other reasons, and ended up not driving the motors in this configuration during the competition season.
If you are tempted to use this method of Master/Slave control, be conscious of the traffic generated on the CAN bus, especially if you’re using the RS232 bridge
We tried a number of different approaches, and like other teams we ended up running the PID loop on the cRIO. The best results we got was running a master/slave setup, but I didn’t like it so we went back to just running the control loop on the cRIO.
You can successfully split the encoder, and we in fact ran them to both Jaguars and to the cRIO. But as I said, we ended up ignoring the inputs to the Jaguars.
Ideally, it would have been nice to have a firmware update that would have let us do this on the Jaguars, but no such luck. There was some talk that it might happen, and I was hopeful when there was a firmware update. Maybe next year.