iRVision - Child Locator Tool (Use as Reference Position)

    Hello guys!

    I was wondering if someone could help me please...

    I have a 2D Single View Process to perform a Frame Offset with a GPM Locator but I do not want to use its position as the Reference Pos. Instead, I want to use the Edge Pair Tool position, which is a Child of the GPM, as the Reference Position.

    In the attached picture, I made an exaggeration just to exemplify what I am trying to do. A taught a model which is intentionally rotated, so the R angle you see is almost 0, but what I really want is the position of the Edge Pair Locator, which is centred on the workpiece and actually shows the angle of rotation of it.

    How can I use this information as my Reference Position so I can move the robot to pick this up?


    I cannot simply rely on jogging the robot manually to the pick position because for the next workpiece that I have to pick I will have less than 150 microns of tolerance, so jogging the robot by hand is basically impossible.

    Thanks a lot for your help guys!

    Quote from Nation

    You could setup some measurement outputs to spit out the X Y and R, and then build the position in the TP program. I think it would be the same as using found position.

    Thanks for your help! The thing is, this tools only outputs pixel values (Row and Column). Converting these values to milimeters is not much of a hustle because is just multiplying by the calibration resolution. However, corelating the measurements with the User Frame would require more work...

    There is got to be a way to output this information easily.

    Quote from HawkME

    You shouldn't need to use any child tools just to offset the position, as long as it is always the same relative offset.

    On your GPM tool you can just click 'set origin' and move it where you want.

    The thing is, I cannot rely on manually clicking to set the reference position, because if I miss for 2 pixels my entire process goes sideways...Assuming that I can set my TCP as accurate as I need, I wish to get the exact position of the center point of my part, so I can move the robot to pick it up. My error cannot be more than 100 micrometers (my calibration gave me 49microns/pixel) setting the center point manually and jogging the robot does not give me the precision that I need to.

    That is why I am doing this way...I am using the GPM to locate the part and other child tools to get the center point.

    The robot itself is not accurate to 100 microns, so good luck with that.

    You will just need to jog it in very fine mode, until you get it dead center, or use direct entry to shift the position. But even then, you will have some variation due to both the camera accuracy and robot accuracy.

    If you really need better accuracy then you need to design a mechanical process to fixture the part.

    Quote from HawkME

    The robot itself is not accurate to 100 microns, so good luck with that.

    You will just need to jog it in very fine mode, until you get it dead center, or use direct entry to shift the position. But even then, you will have some variation due to both the camera accuracy and robot accuracy.

    If you really need better accuracy then you need to design a mechanical process to fixture the part.

    Yeah, I have not yet discarded a possibility to have some sort of handling station. But the repeatability of the LR-Mate 200iD is +/- 10 microns, so I would think that the camera error would be the most significant in this application.

    Anyways, for now, my main concern is to get the position from other child tool other than the parent GPM Locator.

    FANUC support asked me to perform a firmware update but it didn't worked as well...

    Repeatability is not accuracy. That spec only applies to taught fixed positions. You are going to be dynamically offsetting that with the vision system so repeatability doesn't apply. Instead it is the accuracy of the robot. Robotics companies typically don't publish accuracy data, because it isn't very good.

    If you are in a small confined area, then your accuracy will be better, but still, may not be good enough. You will need to get your user frame and tool frame perfect, as well as your z height value that you set in the vision process. Even then 100 microns is dicey.


    I still don't think the child tool is necessary. You can dial it in by adjusting your points. Use dial indicators if necessary.

    It you are successful, or not, please post your results. I'm sure others will find it interesting.

    Quote from HawkME

    Repeatability is not accuracy. That spec only applies to taught fixed positions. You are going to be dynamically offsetting that with the vision system so repeatability doesn't apply. Instead it is the accuracy of the robot. Robotics companies typically don't publish accuracy data, because it isn't very good.

    Typically, in my experience, Accuracy is at least Repeatability*10. I've found this to be a reasonable rule of thumb across brands.

Advertising from our partners