Stall Detection for Delta Calibration



  • Hey Everyone,

    Was wondering if anyone's done any work to investigate stall detection as a mode of calibrating a delta printer? Just thinking how wonderful it would be to not have any end stops or hot end probe. I have a few ideas on how to reduce the errors mentioned in the sensorless homing page and possibly dial in stall detection to the point within calibration bounds.

    Anyone done any work here or have any suggestions?



  • On the individual axes for homing, it should be perfectly doable. As it says in the wiki page you linked:

    On a delta printer, the towers home to the top and a variation of 1 full step in the homing position would cause a significant calibration error. This error can be removed with just one cycle of delta auto calibration. Therefore, if using sensorless homing, you should always auto-calibrate after homing and before printing. For example, if you normally include G28 in your slicer start script, add G32 after the G28. If you want to use the resume-after-power-cycle facility, then in your resurrect-prologue.g file you can home and then probe the bed at 3 or more points at the periphery while keeping the head clear of the print, and use S3 on the final G30 command to adjust the endstop positions only. Or you can just home and hope that the motor stall positions didn't change.

    However, it would be much more difficult homing Z to the bed, as all three motor would be moving, so the stalling 'torque' would be split three ways. It would also make delta calibration difficult. Feel free to experiment and prove me wrong!

    Ian



  • Everyone, including me, thinks that Stall Detect is not anywhere near reliable and repeatable enough to home a delta to the point it can print without further calibration. Neither at the top of the towers or with the nozzle touching the bed.

    As mentioned above, this could potentially be proven wrong. But it will take a working physical demonstration... go for it!


Log in to reply