Why does my RoboClaw seem to be ignoring the PID gain settings?

Question

I'm seeing a behavior in my RoboClaw 2x7 that I can't explain. I've been trying to manually tune the velocity PID settings (I don't have a windows box so I can't use Ionmc's tuning tool) by using command 28 to set the velocity PID gains, then command 55 to verify that they're set correctly, then 35 to spin the wheel at half of its maximum speed. The problem is that no combination of PID gains seems to make any difference at all. I've set it to 0,0,0 and the motor still spins at roughly the set point.

I must be doing something wrong, but I'm pouring over the datasheet and I just don't see what it is. By all rights the motor shouldn't spin when I use 0,0,0! Any ideas?

Answer 1

Got it! The problem was that there's a feedforward term in the control loop that the datasheet doesn't talk about. The wheel and motor and I'm calibrating for are pretty light so the feedforward was enough to get pretty good tracking on its own (ie little error so little reliance on PID).

To solve it I sent in known-to-be-wrong QPPS settings to induce a small error, tuned velocity PID per the directions in the datasheet, and then reset QPPS to where it should be.