In one of my odrive contraptions (odrive pro), I use software torque-limiting to find a hard stop. I’m using trap_traj input mode in position_control while this happens, so there is a destination position set beyond the hardstop (in later cycles this gets modified to be just barely past the hard stop). It all works pretty well, but in the cycles where the motor stops due to torque limits well before the trap_traj target position, there is a weird latency in the NEXT trap_traj trajectory moving away from the hardstop, as it seems to be calculating from the old target position that I never reached.
What’s the most effective/efficient way to get past this? I’ve thought of
- Temporarily switch to passthrough input, apply the current (hardstop) position as the setpoint, then switch back to trap_traj input to apply the position for my next move. Not sure if this will really clear the trap_traj-related history.
- Temporarily switch to velocity_control, and passthrough input, and set to 0. Then switch back to trap_traj input to apply the position for my next move.
- Find out about some magic setting I haven’t yet found to clear the trap traj planner. Or start it from the current location. I’ve tried setting the destination to where I am now, then the new target, but it still seems to have trouble unwinding from where it thought it should be.
Thx