Auto Calibrate does not succeed in some arm positions2018-10-30T12:19:22+00:00

Forums Niryo One Troubleshooting Auto Calibrate does not succeed in some arm positions

Viewing 5 posts - 1 through 5 (of 5 total)
  • Author
    Posts
  • FindingNiryo
    Participant
    Post count: 44

    From some arm positions Auto Calibrate is unsuccessful.  For instance when arm is straight up in the air: joints (0, 0, 1.57, 0, 0, 0)  instead of the ending coordinates having the usual values < 0, -0.329, -1.246, 0, 0, 0>, the arm is still backwards facing <3.011, -0.761, -0.473, 0, 0, 0> .  A second Auto Calibrate from this position will succeed.

    Edouard Renard
    Keymaster
    Post count: 110

    Unfortunately we won’t be able to change that 🙁 . Before and when executing the auto calibration, the program has no idea at which position each axis is. So there is no way to know whether the arm is straight up, or at home position, or any other position.

    So usually we recommend to place the robot at approximately the home position so it runs smoothly. In your case I guess that you don’t have a choice because you’re controlling the arm remotely. If a second auto calibration works, then it’s a non blocking issue for you – definitely not perfect solution, but the only one as for now.

    FindingNiryo
    Participant
    Post count: 44

    I appreciate that having the calibration succeed when arm is straight up is a corner case and that the positions are unknown.

    But since the calibration also will first-pass fail any time the arm is laying on the table or any base rotation less than -1.7 radians I propose:

    CanCommunication::autoCalibrationStep2()  increase calibration_timeout to 40 seconds and

    sendCalibrationCommandForOneMotor(&m3, 1300… decrease m3 delay to 850

    Or maybe automatically retry Auto Calibration a second time.

    • This reply was modified 2 weeks, 1 day ago by  FindingNiryo.
    Edouard Renard
    Keymaster
    Post count: 110

    Increasing calibration timeout to 40s seems doable, it should not add too much to the waiting time in case of a calibration problem. I’ll add this to our feature backlog for next update 🙂

    For the axis 3 delay I’ll make some tests on some robots to see how much we can reduce this delay.

     

    Edouard Renard
    Keymaster
    Post count: 110

    You can find here a new commit that changes the following:

    – Calibration timeout is now in a config file so you can easily modify it without having to recompile anything

    – Calibration timeout now default at 40s (tested with axis 1 & 3 at the opposite maximum position when starting calibration)

    – Axis 3 is a little bit faster during the calibration

Viewing 5 posts - 1 through 5 (of 5 total)

You must be logged in to reply to this topic.