FANUC robot repeatability issues

    Hello All,


    I have some questions about robot repeatability.

    My application is for camera calibration and testing, type of GRnR.


    Essentially, the robot moves from PR to PR and takes a pic.

    The deviation observed is much higher than the repeatability of the robot.


    Currently, the move command looks like -

    L PR[n] 100mm/s FINE ACC20


    Anything I can do to improve the repeatability in this case - will a termination of CNT0 improve the situation?

    Any other parameters I can change to improve repeatabilty?

    Are you testing repeatability or accuracy? Repeatability is not just the same endpoint, but also the same starting point. Moving to the same PR from multiple different starting locations takes this test from repeatability to accuracy, and the error range of accuracy is likely to be 10x the repeatability.


    Do you have your load data set correctly?


    How much settling time are you allowing for the robot to stop vibrating before triggering the camera?


    What steps are you taking to ensure the triggering does not take place before the robot comes to a complete stop?


    For what you're trying to do, FINE should be better than CNT0.

    Hi SyeFire,


    Are you testing repeatability or accuracy? Repeatability is not just the same endpoint, but also the same starting point. Moving to the same PR from multiple different starting locations takes this test from repeatability to accuracy, and the error range of accuracy is likely to be 10x the repeatability.

    I think its repeatability since the robot has to move to different PRs but always the same 2 PRs for start and end. The repeatability discrepancy is seen between different DUTs and even the same DUT after running the cycle multiple times


    Do you have your load data set correctly?

    I believe so


    How much settling time are you allowing for the robot to stop vibrating before triggering the camera?

    There is no settling time as such after the move


    What steps are you taking to ensure the triggering does not take place before the robot comes to a complete stop?

    I can hear the break after each move is done. Programatically nothing

    is there any statement that I can add to the TP code to ensure a complete stop after the linear move statement?


    For what you're trying to do, FINE should be better than CNT0.


    Thanks for you advise!

    Have checked for the robot mounting? Whether the bolts are torque tight? I had similar issue with Repeatability in pick and place and found one of the mounting bolt was loose as the robot was installed at shop floor for testing purpose and not at End-Customer.


    Also check for Payload calibration of the tool. If its not done properly or entered approx. values then also you might face this issue.

    Quote from ROBOTAK

    I think its repeatability since the robot has to move to different PRs but always the same 2 PRs for start and end.

    The point is that the path leading to the image-taking position needs to be the same every time.

    Quote from ROBOTAK

    Do you have your load data set correctly?

    I believe so

    How have you confirmed this?

    Quote from ROBOTAK

    There is no settling time as such after the move

    That is self-defeating. After the motion command, add a short WAIT before triggering the camera. Start with 1sec, and experiment with progressively smaller time delays. If the settling time is contributing to the repeatability error, you will see the effects in the data.

    Quote from ROBOTAK

    I can hear the break after each move is done. Programatically nothing

    is there any statement that I can add to the TP code to ensure a complete stop after the linear move statement?

    The brakes should not normally engage immediately at the end of each motion. There should be a time delay. The brake engagement also introduces vibration and can cause small amounts of uncontrolled motion.

Advertising from our partners