Calibrating a Single DCC Rotary Table (Infinite Type)

  • Filter
  • Time
  • Show
Clear All
new posts

  • Calibrating a Single DCC Rotary Table (Infinite Type)


    I operate on an Optiv Performance 1z443, a dual Vision/TTP system.
    This particular unit comes (I believe standard) with a Rotary table, as described in the title.

    For years, the company I work for didn't have a solid understanding of how to calibrate it; as they were the ones that trained me, neither did I.
    Recently, I came across some information that was able to give me some clarity, and I thought I would share with the community.

    Using the "canned" PC-DMIS calibration schtuff for the rotary, I dunno, for some reason I didn't find it adequate.

    After some research, I discovered another method, involving Calibrate Rotary from Features...

    Firstly, I drafted up a standard AutoCalibration program, to calibrate the Camera & All Tips.
    I performed a File>Save As.. & created a separate AutoCalibration program, with the addition of the Rotary table steps, and, more importantly, an IGNOREROTAB/ON command at the top.

    After the initial calibration of the tips, the program switches back to manual mode.
    The operator is prompted to create a 4-point semi-circle on the outside of the rotary itself (which is blue in my case).
    1 point is to be taken on the face, and the program moves to DCC mode.

    Using DCC, a cylinder is drawn on the rotary, to establish a center for rotation. A simple point is used for the face of the rotary, but a plane can be substituted.

    The next part is sort of "****** together" honestly, but it works.
    I don't know the exact nominal of where the sphere is "supposed" to be at 0°; instead I created an Auto Sphere at it's measured location, & measured it again with 17 points using the Master setting (to change the nominals to where it measured).

    I then copied the sphere, set the workplane along the axis of the rotary (in my case, Y-) & Paste-With-Pattern'ed every 30°.
    If you'd like, do it every 45, or heck, every 1; I don't need to rotate less than 30°, so I chose 30.
    In between each sphere, I added a MOVE/ROTAB command to rotate to the desired angle.
    Inserting some clearplanes & move points and what have you, it was allowed to DCC measure all of the spheres at their respective angles.

    Create a plane from all of the spheres.
    Level to this plane, parallel to the axis of the rotary (in my case, Y-)
    Create a circle AFTER YOU ALIGN from all of the spheres
    Insert>Calibrate>Rotary Table from Features
    Use the Plane you levelled to, and the circle of the spheres.

    Voila, it's calibrated.
    Hopefully this helps someone get their rotary going with more confidence.
    I should note though, there are better processes for calibration, such as rotating to every desired angle, then rotating backwards to account for Hysterisis, & constructing a circle from that.

    I just know I was never able to get PC-DMIS to effectively calibrate the rotary, or perhaps I never understood what it was doing, but this custom routine was able to get me going in a way I'm okay with.

  • #2
    Great info! Thanks Jester
    Automettech - Automated Metrology Technology


    • #3
      Looks like this, except the levelling...


      • InspectorJester
        InspectorJester commented
        Editing a comment
        I'm unsure; I read something somewhere (I'm searching for it now actually) that said to level to the plane before you construct the circle.. I will try and provide documentation. I was actually curious about this

      • InspectorJester
        InspectorJester commented
        Editing a comment
        Gotta say, as well, I love how you used variables to calculate the angles of calibration, and just ask the operator how many angles they'd like to check. Very nice!

    • #4
      It's safe to do it, but I didn't think about it.
      I will try with and without, just to check if it's different or not...


      • #5
        In the code, I constructed a 3D circle which has the same vector than the plane, so I think it's the same center than a 2D circle after levelling...


        • InspectorJester
          InspectorJester commented
          Editing a comment
          In your code, you constructed them both from a scan; I believe this would achieve this anyway!
          We're doing the same thing, just different ways. I constructed a 2D circle, so it makes sense to level, but it shouldn't affect anything.
          I'm just glad I got it working!
      widgetinstance 190 (Related Topics) skipped due to lack of content & hide_module_if_empty option.