Runout Issues

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

  • Runout Issues

    I'm having issues with runout output. I have a .0005 runout that when I run it at 0 deg I get .00007 in. When I rotate the part 180 deg I get .00054.
    I have been struggling with this for several weeks now. We can open inspect it and get .0001. I'm running version 2013 Mr1. In the next week or so I will be going to Version 2014 Mr1
    The CMM's that I have are Global 7-10-7 with Leitz scanning heads.

    I have 5 linnear closed scans that I construct a cylinder from. The length of this cylinder is 2"... Then I created an alignment leveling on that cylinder -A- X&Z origin. Then I scan the other diameter with a linnear closed scan. This scan is on the oppsoite end of part so I have to rotate the probe 180 deg. Then I construct a circle from that 2nd scan. Output the runout of circle back to cyl -A-

    I also did these same scans using a star probe with the same results. What is funny is the results will repeat within .00003 every time. Please note that I'm not a rookie when it comes to programing and know several CMM softwares.
    I'm thinking that PCDMIS is taking the constructed cylinder and not looking at all of the scanned points. Is there a better way to do what I'm trying to acomplish??? Or is this something with the software??

    I do not know at this point. I have asked several other programmers and had one of our other plants run the same test and they get the same thing. they are using ver 2012 Mr1

    Any ideas or help is much appreciated. I'm not able to add any attachements for some reason. If you think that you can help you can email me at [email protected] or call me at 763-576-6395 and I can send you some snapshots of the results and some of my program code.

    Thanks

    Time for the Trolls to leave.

  • #2
    Check the Cylindricity of the constructed cylinder to ensure is round enough and straight enough.
    sigpicIt's corona time!
    737 Xcel Cad++ v2009MR1....SE HABLA ESPAÑOL

    Comment


    • #3
      Based off of what you said here, sounds like you have a serious form issue with your part. I agree with Roberto, measure your cylindricity before you even attempt to get an accurate reading for runout especially at that tight of a tolerance
      SF7107(PCD), SF454(PCD), 152614(Quindos), 9159(Quindos), 7107(Quindos), B&S Manual, M&M Gear Checker

      Comment


      • #4
        The cylindridity is .00025 and the circularity of each cross section is .00005/.00007.
        Also I'm doing 2 DCC alignments before I measure the features used in the runout.
        Time for the Trolls to leave.

        Comment


        • #5
          Have you done a repeatability check with the two probe angles being used. Make a quick program that measures the qual sphere with the first angle and rotate the head to the second angle and measure it again, then check the distance between the two spheres in each axis. Create a loop that does this several times.
          PC-DMIS 2016.0 SP8

          Jeff

          Comment


          • #6
            When scanning close diameters we always set the scan to cover 450 degrees this ensures that any start finish jumps caused by the initial hit/acceleration an deceleration can be filtered out. This may help if you are not already doing it.

            Comment


            • #7
              Unlike you, i do consider myself a rookie cause I've only been fighting the demon about 3 years now. My question is, How long is the part? I had the same issue with a part I was trying to measure runout of opposing counterbores....Turns out, PCDMIS doesn't do well levelling to a short cylinder then looking at a circle 6" away....Apparently, error is compounded
              Sheffield Endeavor3 9.20.8, Tesastar-SM, Leitz LSP-X1s & LSP-X1M, PCDMIS 2011 MR1

              Comment


              • #8
                Shhrocknroll I will try that next. Thanks
                UKCMM I will also try to cover 450 deg with the scans.
                Hi-Tech the primary datum is ID 2" long and the opposing ID diameter us within .200 of the bottom of the primary datum.

                Thanks for all the ideas everyone. I will continue to solve the PCDMIS Mystery
                Time for the Trolls to leave.

                Comment


                • DAN_M
                  DAN_M commented
                  Editing a comment
                  Solve the PC DMIS mystery? HA that made me laugh =P

              • #9
                With a 2 inch datum and the target surface .2 away I would say Hi-Tech comment would not be an issue. I agree 100% with what he is saying and it was my initial thought as well. If the part was longer and had a shorter datum surface and had this issue I would try and use an OD or Bore on both ends for rotation or level and compare the results.

                Comment


                • #10
                  1. Have you ruled out all part movement during inspection?
                  2. Can you elaborate more on the "open inspection" method/technique? Is it an apples-to-apples (at least as close as can be) measurement?
                  3. When you say "rotate the part 180 degrees", I am guessing you mean as in rotating the part in its fixture (v-block, etc) and rerunning the program?
                  sigpic

                  Comment


                  • #11
                    Schlag: If I understand what you said. If I used both the datum and the feature to align to that would be forcing it to be perfect. Again PCDMIS should not be using any type of active alignement to calculate this output.
                    ir a Programer: 1. The part is not moving. Positive of that as the results I get are very repeatable between the 0 deg and 180 deg rotations. 2. Not sure what you are refering to with the open inspection method. but it is apples to apples. 3. the part is rotated 180 about the centerline of the datum axis. Then running program again.
                    UKCMM: I tried meausring 450 deg and my results got much worse at both positions.


                    Still have not solved the problem. I measure all features 2x in DCC and then measure them again and same results. I'm wondering if it is in the form of the datum that is causing the problem with only one of the rotations. The bottom of datum -A- is at .00025 on roundness and .00002 at top of datum -A-..... Weird how repeatable my results are. I can open inspect it on a surface plate with a pin in datum -A- and get about .0001 which is very close to what I get at the 0 deg position. The pin eliminates the form error at the bottom of datum -A-...
                    Time for the Trolls to leave.

                    Comment


                    • #12
                      Originally posted by William Johnson View Post
                      Schlag: The bottom of datum -A- is at .00025 on roundness and .00002 at top of datum -A-.....
                      You don't say what diameter the probe tip is, so I can't help but wonder if it is shanking at the bottom of the datum -A-. Since your are measuring at 0 and 180 degree positions I'm assuming these are flat locating surfaces and possibly one surface deviates more than the other. Just a thought.

                      Comment


                      • #13
                        So is the diameter of -A- smaller than the diameter being inspected? If not, is it not possible to inspect using the probe angle? This would eliminate the variance between the two tips.
                        sigpic

                        Comment


                        • #14
                          using a 5mm dia ball
                          Time for the Trolls to leave.

                          Comment


                          • #15
                            Datum -A- and feature of runout are concentric circles. I'm using a 5mm probe on datum -A- and 1.5mm probe on the output feature. I have done the measurements with a star probe with no rotations and also with individual probes of same size with Z90B0 & A90 B180 rotations. In both cases I get the same results at 0 deg and 180 deg. What is funny is the numbers are very repeatable at both angles. I have done it with scans and as autocircles both. The results between the scans and circle are also the same.

                            I do not understand how 2 concentric features can give these kind of results. I have tested everything several times on different days with different calibrations done and all the reports look the same. Everything I have tried has not improved my outcome. I just keep getting the same results at both rotations. Again I'm rotating about the datum -A- axis. Not fliping the part.
                            Time for the Trolls to leave.

                            Comment

                            Related Topics

                            Collapse

                            Working...
                            X