Work Object not saving correctly

Options
Hello

I have a IRB 1600 and I am having a weird issue. I am teaching a work object called J1. When I teach work objects I always save robtargets at the points I am teaching the work object at and teach them in wobj0 so in this case I have a set of 3 points named J1Origin, J1X,and J1Y. I do this so I can verify that my frame is correct if we ever need to trouble shoot. So I teach my work object and then go to reference my part and compare the CAD data I was given. I see that my points in relation to the CAD data are significantly off. My next step is to verify my frame. I go to J1Origin and physically the robot is going back to the same point however when I open up the Jogging screen and change the work object to J1 and I should see X,Y,Z all have zeros, however I am seeing that X shows -12.04 and Y,Z show zero. If I reference my Y  point J1Y is about 20mm off in the Y direction and reads zero in X, and Z, and my X point J1X is about 10mm off in the X direction and reads Zeros in Y and Z. I have done this both in robot studio and on the actual robot and get the same results. I am not sure what I am doing wrong here. Any help or suggestions would be greatly appreciated.  

Comments

  • lemster68
    lemster68 ✭✭✭
    Options
    Some things that I would check:  TCP is good.  Correct robot model is installed.  Base frame of robot in parameters is unaltered.  Calibration offsets and zero position.  Jog in linear, does it drift off the axis in which it is being jogged.  For example, jogging in Y drifting in X.  Pro tip:  Use joystick lock to eliminate accidental deflection.  Finally, does linear jogging look like it is making an arc?  (often incorrect model)
    Lee Justice
  • deltaplex15
    Options
    So I am getting this issue in the real world I only brought it into robot studio to eliminate the robot as the actual problem. My TCP appears to be good however if I delete it and reteach it then update the workobject then I see I am only off about 7.5mm for J1Origin. I am guessing it has something to do with the tool however I am not sure why it is not learning the tool correctly. I am using the 9 point method to be more accurate and it appears to be good when I pivot it and jog it around. 
  • lemster68
    lemster68 ✭✭✭
    Options
    Incorrect calibration offsets can cause this.  So could wrong robot model/type/variant.  Also, check the base frame.  I have never gone more than the 4 points with X and Z elongator.  9 seems overkill, waste of time.  Is the TCP such that you could just measure the X,Y and Z to manually enter those numbers?
    Lee Justice
  • deltaplex15
    Options
    I guess I'm not familiar with your reference to robot model/type/variant. I was assuming you were referring to using robotstudio but is there something I should check on the robot itself?  I have verified the that the calibration offset matches the label on the robot though. 
  • lemster68
    lemster68 ✭✭✭
    Options
    One thing is the robot type, it should have IRB 1600 and not some other type.  There are variants for arm length and payload.  Sometimes, even the same type has a variant and would be marked "Type B" for example.  Those could have a different gearbox for an axis, for another example
    Lee Justice
  • deltaplex15
    Options
    Where would I go to verify this sorry I deal more with Fanuc then I do ABB. 
  • lemster68
    lemster68 ✭✭✭
    Options
    Sticker on the arm will have the robot model and type, should indicate any variant.  Look in a backup for the installed system to compare.  One could also find it in the pendant, but it would be difficult to try to walk you through it from here.  You could try poking around looking for it.
    Lee Justice
  • deltaplex15
    Options
    no worries. I was able to get someone that was very helpful on the line with tech support. Even though our calibration offset looked good we found that J2 at zero degrees was off by about 2 degrees. So we were never able to teach the tool correctly. Once we updated the rev counter for J2 We were able to get the tool taught much better and were able to teach the work object and get more accurate data. Thanks for all your help as always!! 
  • lemster68
    lemster68 ✭✭✭
    Options
    👍
    Lee Justice