Cartesian Coordinates are off between VC and RC, even though everything is the same!

bwinter714

Hello again,

I am having an issue where the Cartesian coordinates are no where close to being correct, even though I made sure that everything (all frames and TCP'S) matched in the VC and RC. This obviously presents problems when doing offline programming. For example, this is the real robot at coordinates are [X 76.6,Y 585.9,Z -261.0,EX -116,EY 3,EZ 90.2] (relative to the table work object), second picture.

The VC, however, looks like the third picture when I plug those coordinates into both the linear jog tool, and if I make a specific robtarget for the position. (I made sure that the robtarget referenced my correct work object, so it isn't something simple like that )

This doesn't make any sense to me, because I am very careful about how I set up my systems in order to avoid things like this! But I can't explain what I did wrong here.

The first picture is the same Cartesian coordinate with respect to the world frame, and the VC can't even reach it! Which is why I had to show only the tool at the target (the lance pointing towards +Y). 

Does anyone have any suggestions for me? 

Thanks!

Henrik Berlin

Hi,

The root-cause of your problem is not obvious to me, but perhaps it is for someone else. Anyway, what I would have done is to take a few steps back.

1. Start by jogging the robot to 0 position for all joints and compare the TCP of tool0 with respect to the base (you can set the robot to UCS in RS)

2. Jog one joint at the time away from zero to 30 degrees, say, and compare TCP of tool0 with respect to base.

3. Then compare with respect to world.

4. Third, compare with respect to your work object.

5. Now you can compare TCP of your user-defined tool with respect to work object.

6. The last step is to jog to your robtarget

Somewhere along the chain you should find a deviation. Do you think it can be related to the calibration of the robot? 

Kind regards,

Henrik

bwinter714

Thank you Henrik,

I have compared both at zero, and as long as I am viewing joint angles, everything works correctly! Well, it's not exactly the same, but I haven't done any fine calibrations of the cell and objects inside of it yet

See attached picture of the same position in my first post, except I moved the robot to the joint value of the position rather than the Cartesian coordinate.

I think maybe the disconnect between the Cart. coordinates in my VC and RC has something to do with the tool and how it was created? I will make another tool with a new SAT file and see if that helps/fixes the problem. I will keep everyone updated....

-Brian

bwinter714

Also, here are some pictures of the VC and RC at identical joint targets (J1 is at 30 degrees like you suggested), and everything set to Wobj0 and Tool0. The XYZ is correct, but the rotation not so much. 

IDK about the robot calibration, it was calibrated from ABB, it is a brand-new robot. 

-Brian

bwinter714

This has been fixed.

In case this ever happens to someone else, check the tool file! The tool wasn't created right in the first place (I wasn't fully aware of the proper sequence when making a tool, I assumed that it was OK to move the tool on J6 once attached) and since the TCP was so off in the VC because the rslib wasn't correct (not in numeric data, but in the CAD), that it was screwing everything up in the manner above. 

Does anyone know why this happens? Under normal circumstances with any other offline program, if I create a tool, and attach it to J6 of a robot, if I want to rotate or translate the tool, the TCP follows it. Where in practice, it does not seem to be the case in RobotStudio. It isn't a huge deal, now that I know, but I was just curious if there was a specific reason why this is so.

Thanks!

blind3rr

Make sure you have set Base Frame of your robot and external axis right in Controller/Edit Systems (robot should have zero position and external axis should have values related to robot).

Henrik Berlin

The tooldata in RobotStudio is created when you attach the tool library to the robot. It is created from the 'template' frame that is in the tool mechanism. Moving the tool will not automatically update the tooldata. I will take note of this as a potential future usability improvement (internal ref: WI11996)