reorientation error tcp

    Hello everyone!
    I ask for help in my question.
    positioning my tool?
    On the robot irb4600 installed welding torch. when reorientation is used, tcp shifts within 2 mm. This is normal? where do i find a tcp positioning accuracy document.
    Tcp is fully configured using Bullay. maximum and minimum error within 0.5 mm.

    2mm, while not horrible isn't ideal, you should get a better result than that.


    Do a manual 5-point (4+elongate) TCP and see what your mean/max is.


    Also worth checking since it can play tricks on you is the welding wire:
    Depending on the routing of the liner the wire will "wiggle" or rotate as you re-orient and it's in reality not your TCP drifting but your wire moving.... so what I used to do is to cut off 10-15cm of wire and make the top part s/z-shaped (to hold it in place inside the torch), then release the feeder rolls and pull back the wire from the spool/drum, insert your 15cm piece from the contact tip backwards (push the wire into the torch) and then redo the orientation test.
    This way you know you have a "dead" wire that can't / shouldn't move around on you.

    Ok . Yes, I will try. but it seems I did it during the first calibration. And in what documents I can see the data on the reorientation error tcp?

    There isn't a number for TCP-reorientation accuracy.


    There's position repeatability, path repeatability / accuracy but those numbers are measured during the ISO cube test and isn't representative to a reorientation test.


    If you're not programming a lot offline (robotstudio) then I wouldn't really worry about it too much.

    of course. but I recently started programming and learning to not create a lot of points. therefore this question bothers me because the robot has been re-calibrated. Can I see anywhere, does my robot have absolute calibration accuracy? I want to understand how accurately I can use it


    To add to this, make sure that there are no external forces pulling on the back of the torch. I.e., the whip which contains the pas, wire liner. Such would make it look like the tcp is drifting. Finally, which version of RW is your robot? When 6.0 came out, there was an issue with the tcp drifting during reorientation which was related to the jog mode. ABB responded by saying that was in Responsive? mode and that it should be changed to the other mode. Or the other way around, I cannot remember for sure right now, I would have to look it up.

    There is usually no single root cause to these errors and the normal reasons varies depending on robot modell.


    The 4600 is a well known bastard child off ABB and puts shame on their long legacy of excellent path performance.


    Most likely your tool is very light compared to the max tool payload of the robot. This will result in your dynamic factor being all wrong.
    There are many things you need to fix to get the perfect motion but many of them are complicated and require secret knowledge about the robots.


    Best advice i can give you is to download tunemasters, it comes with the tools and utilities package and is a free software tool from ABB.
    https://robotstudiocdn.azureed…And_Utilities_6.08.01.zip
    Depending on your RobotWare version you should let the software Tune you DF (RW5.15) or your Foundation stiffens(RW5,6x and up).
    If your RobotWare is 5.6x or newer the system will save the data to your motion process modes, otherwise you need to write down the results and add TuneServo commands manually.


    You should Increase your "Finite Impulse Response filter" with the command TuneServo.


    Code
    TuneServo ROB_ID,1,60\Type:=TUNE_DH;


    Tune_DH can not break the robot so play around with it, lower value then 100 means more filter, This is described in depth in the Rapid reference manual. Also make sure your arm payload and tool weight are 100% correct.


    If you fix all this your drift should be around 0,2mm.

    The problem did not dare. The service support answered that for the irb 4600 series - up to 3 mm. but I am still tormented by another question. when moving the positioner together with the robot. Those. the robot is given a point on the positioner, and it follows it. As a result, the robot moves away from the point by 5-6mm. This is normal? What do you thinck about it? ???

    How much do you move the positioner and which part / how does the robot rotate with it ?


    You already stated that the robot itself drifts ~2mm on it's own so out of your 5-6mm how much is the robot ?


    If it's a 90 degree rotation on a single axis positioner then it's a bit excessive, if it's 180 on both axis on a 2-axis positioner, then I'd say it's not too bad.


    Also keep in mind that these are more or less "theoretical" exercises that really doesn't matter all that much unless you're doing offline programming and had hoped you'd get away with no touch-up, and/or if you for some weird reason need to search in one position and then rotate everything 90-180 degrees before welding.

    This is a dual axis positioner. I turn the first axis 130 degrees. and the robot deviates by 5.5 mm from the point. the robot moves with all axes except 6.
    I need it for welding complex parts.
    detail accurate enough. I want to install this stitch through one point. then it is easier to edit the program.
    the robot makes a circular motion together with the movement of the positioner. but the welding torch deviates to 5.5 mm. this may give bad for welding.
    in addition, he had a knocked down calibration on 5 or 6 axes. it has been calibrated.
    I thought the robot has precise positioning. as on coordinate machine. it is bad..
    although the offset on the uniaxial positioner reaches 1.5 mm. there may be a problem in the positioner...
    On your 4600 series robots ,are you also have?

    Quote from Selivanov


    I need it for welding complex parts.
    detail accurate enough. I want to install this stitch through one point. then it is easier to edit the program.
    the robot makes a circular motion together with the movement of the positioner. but the welding torch deviates to 5.5 mm. this may give bad for welding.


    As I mentioned previously, as long as you program your parts manually and/or touch them up after offline programming, this is not as bad as it seems.


    Normally you'd split a 13 degree move into 4-5 positions, each position taught / re-aligned into the correct position.
    The real error / deviation would be the drift you see when executing that program.


    Did you ever do a manual TCP ? I've seen instances where Bullseye TCP would pass and give a "reasonable" TCP with a robot that still was poorly calibrated and the easiest way to check that is to do a manual TCP. (it'll give you a large max error if it's bad)

    OK. All the same, it is not clear to everyone whether such an error is present.
    anyway. Thank you all so much for your help.
    If I can solve the problem, I will write how I solved it.

Advertising from our partners