Measuring thickness of a "fish bowl"

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Measuring thickness of a "fish bowl"

    I have a part that is basically a 3D printed metal fish bowl. (See attached pic, where I used clip planes to "cut" the CAD in half) I have been tasked with using the CMM to measure the thickness of the wall of this fish bowl. What I have done now is measure a plane on top. Level and set 'Z' to the plane. Then I go down set distances -0.5", -1.0", etc and take opposing auto vector points on the outside and inside at the same Z height. I am using the smallest probe I can to eliminate any probe cosine errors. However, the numbers I am getting are larger than expected which makes me think this is a bad measurement technique. Does anybody have any suggestions to get this right? I basically need a caliper/mic type measurement but we don't have any hand gages that will work.
    Last edited by Parker; 02-19-2020, 12:44 PM.

  • #2
    If you measure outside and inside at the *same* Z, the measurement will only be correct on the 'equator', all other heights will give too large values.

    If the bowl is spherical, you should use spherical coordinates, same angles, just different radii (and direction), for the opposing points. If it is not spherical I don't know how to get the right points, maybe have them generated from the CAD system?
    AndersI
    SW support - Hexagon Metrology Nordic AB

    Comment


    • #3
      Put the INSIDE cad on one layer. Put the OUTSIDE cad on a different layer. Pick your point on the INSIDE data (with only the inside showing) and create it. Close auto feature, hide the INSIDE, show the OUTSIDE, open auto feature, click FIND. It will then put the point on the outside data, directly outward from the inside point. You will probably then have to flip the vector.
      sigpic
      Originally posted by AndersI
      I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

      Comment


      • Parker
        Parker commented
        Editing a comment
        I will try to separate the CAD. If I can't do that can I use the same X,Y,Z of an inside point, then flip the vector for an outside point? The bowl is not a sphere unfortunately, I already looked into that method. Thanks guys. The nominal thickness is only .02".

      • Mike Ruff
        Mike Ruff commented
        Editing a comment
        This is a slick way to do it! Last time I had to do something like this, I did all the math by hand for one side based off the vector from the other side and the nominal thickness. This would have saved me some time

    • #4
      Yeah, you could do that as well, your XYZ nominals wouldn't be correct, HOWEVER, you could just create the same point twice, flip the vectors of one of them, THEN put in a STOCK THICKNESS value to get to the other side.
      sigpic
      Originally posted by AndersI
      I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

      Comment


      • Parker
        Parker commented
        Editing a comment
        I'm going to try that and re-run. Thank you.

    • #5
      Shameless plug for this suggestion over at the idea center:
      https://hexagonmi.userecho.com/commu...irrored-points
      PC-DMIS CAD++ 2o19 R1 SP11

      Comment


      • thomas pilc
        thomas pilc commented
        Editing a comment
        lol just as i was reading the issue i went o man we just went over this same issue but in 2d last week, please vote people this would be such a nice feature to have

    • #6
      It looks like a torus, so measuring as it in /out, and compare radii ?
      Or measure some circles.
      What Matt said seems to be a good solution.

      Comment


      • #7
        Originally posted by Matthew D. Hoedeman View Post
        Put the INSIDE cad on one layer. Put the OUTSIDE cad on a different layer. Pick your point on the INSIDE data (with only the inside showing) and create it. Close auto feature, hide the INSIDE, show the OUTSIDE, open auto feature, click FIND. It will then put the point on the outside data, directly outward from the inside point. You will probably then have to flip the vector.
        +1

        Comment


        • #8
          Originally posted by Matthew D. Hoedeman View Post
          Yeah, you could do that as well, your XYZ nominals wouldn't be correct, HOWEVER, you could just create the same point twice, flip the vectors of one of them, THEN put in a STOCK THICKNESS value to get to the other side.
          +2

          Comment


          • #9
            There is a reason why I received PC-DMIS total consciousness.......
            sigpic
            Originally posted by AndersI
            I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

            Comment


            • #10
              Originally posted by Matthew D. Hoedeman View Post
              There is a reason why I received PC-DMIS total consciousness.......
              http://images.clipartpanda.com/kudos...part_image.jpg

              Comment


              • #11
                I did a combination of Matt's suggestions and it worked perfectly. Took the inside CAD "layer" out, then created two OD points, flipped the vector on one and added the .02" thickness. It does change the Nominal XYZ when you adjust the thickness also. Thanks a bunch guys for the help!!

                Comment


                • #12
                  Originally posted by Parker View Post
                  I did a combination of Matt's suggestions and it worked perfectly. Took the inside CAD "layer" out, then created two OD points, flipped the vector on one and added the .02" thickness. It does change the Nominal XYZ when you adjust the thickness also. Thanks a bunch guys for the help!!
                  You used THEO THICKNESS to get the nominals to change, and that is probably what you want. HOWEVER, there is also ACTL THICKNESS, and when using an actual thickness value, the nominals won't change, but Pcdmis will still measure correctly, and still be looking thickness away, and the reported values will show the cad model nominals, but the deviations will have the thickness 'removed' from them.

                  Example: X0 Y0 Z0 I0 J0 K1

                  Use THEO THICKNESS of 0.2

                  NEW nominals of X0 Y0 Z0.2 I0 J0 K1

                  If part is 0.21, you will see X0 Y0 Z.21 with a ZDEV of 0.01

                  Use ACTL THICKNESS of 0.2

                  Same part will show X0 Y0 Z0 (noms), report will be X0 Y0 Z0.01 with a ZDEV of 0.01

                  Just additional info to keep in mind. The ACTL THICKNESS can also be a defined variable. Say you are making a program that will check an in-process part as well as a finished part. The in-process process is leaving 0.005" of stock on the part to be removed after heat treat. Why make TWO programs? Make ONE program, with ACTL THICKNESS being a variable, then using operator inputs or various other coding methods, have the variable be either the 0.005" (in-process) or 0.000" (finished part). One program will check both parts.
                  sigpic
                  Originally posted by AndersI
                  I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

                  Comment

                  Related Topics

                  Collapse

                  Working...
                  X