KRC4 Position monitoring doesn`t match with real world

    Hello everyone!


    I have a problem that is puzzling me. I am trying to palp a jig to build the Baseframe, but after I go to the virtual coordinates they don`t match with the robot position as I go further from the origin (Also the milling trajectory is not good).


    I have tried to measure the probe(with automatic robot macro) and the error is insignificant.


    One thing that I observed is that if I am, for example, at the coordinate X:2000 Y: 2500 and I jog it to Y0. my X coordinate changes gradually to 1996.9. Can this cause the measurement error, or there is any logic explanation to it?

    I am not sure if I have the "Absolute accuracy option".


    "To deactivate KUKA’s Absolute Accuracy you should set variable DEACTIVATE_ABS_ACCUR to TRUE. You can find this variable in the file KRC:\STEU\MADA\$custom.dat (around line 73)."


    I don`t have the $custom.dat file so maybe I don`t have the Absolute accuracy option.

    Is this error (approx-3 MM on y at coordinates X:841.22, Y:1130.18) normal for a robot marked HA?


    The load data is set correctly for each tool and spindle.




    Quote from yootommy

    I am not sure if I have the "Absolute accuracy option".

    You should have. You mentioned your robot is an EXTRA HA. The HA types all use absolute accuracy. Use the Info tab on the HMI. You should read something like "Position accurate robot" if you have the abs accuracy option.


    Fubini

    Yes, it is active. just researched a bit more on the topic of HA.


    The robot has a rail. We are cutting carbon fiber parts and we implemented 4 jigs.


    When we wanted to implement the 5'th jig it was a collision.

    The collision (god knows why) affected all the other 4 cutting trajectories. The robot has no calibration error but everyone else is insisting to recalibrate.

    We measured the errors on the TCP and they were insignificant, also they have been auto-corrected by a macro.

    Trying to palp the jig to build the base frame again, we noticed this offset between virtual and real coordinates (some colleagues tell me that this might have been the case before collision too).

    Besides the fact that the 3d data doesn`t match with reality, considering the extra HA option, I have no explanation for this big error.


    KUKA and a 3d control laser are coming tomorrow and I will keep you updated.

    Wow, what a ride...


    The laser tracker measured multiple points in the air one meter apart on X Y and Z. All of them had different errors from the value that was shown in the teach pendant ranging from +1mm to -1 mm from point to point. (We are waiting for the 3d report). -The robot was jogged manually for the measurement to exact "x000" X,Y, and Z values. -


    The KUKA representative had nothing to say than "pretty odd". -Me too my friend, me too- :loudly_crying_face:


    The robot has been recently moved from another location so the KUKA representative suggested that the robot should be recalibrated "after any major collision or relocation".


    As the TCP rotates PERFECTLY along the measurement pin, I am not sure the recalibration will do any good. Any thoughts about why the robot can have variable errors in different locations? As I mentioned early, the payload is set correctly.


    Can recalibration erase some of the HA functionality?

    Quote from yootommy

    Can recalibration erase some of the HA functionality?

    No. But effects of a bad mastering can be bigger than on non HA robots. As stated before on a HA robot do not use jogging to exactly position robot to a defined position coordinates because in that case abs accuracy is offline (but $POS_ACT the value behind the current position menu on the HMI is calculated including absolute accuracy). Better to use a program here and run it in T1 to avoid trouble. Probably not a problem in what you did but this system behaviour often leads to confusion and using a program helps to minimize this.

    This is usually not an issue when teaching because you visually position yourself to a reference spot and you are not intrested in instead of moving to defined given coodinates.


    Fubini

    What means bad mastering?


    As far as I know, you plug the EMD and the axes disappear as they are mastered by the tool.


    I took the head of the robot off and remastered it. For a HA robot, do I need to add "load correction" and recalibrate for each tool? Meaning 8 times with the same tool as it is a milling robot?

Advertising from our partners