Probe Calibration and Form Resutls

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

  • Probe Calibration and Form Resutls

    Can probe calibration affect form results?

    Using TP20, 3 way (T-shaped) build. Taking about 50 vector points around a cylinder and constructing a cylinder (least squared). Output cylindricity and it measured 0.0024" yesterday afternoon. Calibrated the probe this morning, re ran the same part and cylindercity measured 0.0011". All the other diameters and distances measured were different by 0.0001"-0.0002".
    Last edited by acgarcia; 11-17-2022, 11:12 AM.

  • #2
    As far as I am concerned the calibration only sets the position and the diameter of the sphere. It should not affect the cylindricity especially by that much.

    For me it is best to have a graphical analysis of the cylindricity and compare. A small impurity on the cylinder can cause this if you are not using any outlier filter.

    Comment


    • #3
      My first thought is: Did the part move between yesterday afternoon and this morning causing this mornings points to be taken in a different place than yesterday afternoon?
      Sheffield Endeavor3 9.20.8, Tesastar-SM, Leitz LSP-X1s & LSP-X1M, PCDMIS 2011 MR1

      Comment


      • #4
        constadin, I don't have much experience using graphical analysis but i'll play with it. I was thinning of reporting t values of each of the points as a test program.

        Hi-Tech Redneck , yes, the part did move out of the CMM and was put back in. We are in the upper 40s low 50s right now.

        Comment


        • #5
          Does your alignment make sure that you hit the same bit of material for each point on the cylinder, or is it just aligned on the cylinder without a clocking feature?
          3.7mr3CAD++ / 2011mr1CAD++/2012mr1CAD++/QUINDOS7

          Comment


          • #6
            When you constructed your cylinder did you use BF or BFRE? You should use BF.

            Comment


            • #7
              If you are using a 3 way probe and taking hits with each probe then probe calibration could have an effect on the measurement 100%. It is unlikely to have a .0013" impact on the results unless the probe (probes) have changed since the last calibration. I don't think it is the reason.....but it does have the potential to change the readings.
              https://www.pcdmisforum.com/forum/of...lignment-kinda

              Comment


              • #8
                Liambo734, I skip the manual alignment and go straight to dcc with those points and then use that cylinder in my alignment. That could be the culprit right there. I use a cylinder that is perpendicular to my problem cylinder for the alignment.

                BKulpa I think my construction is BF .

                Comment


                • #9
                  Originally posted by acgarcia View Post
                  constadin, I don't have much experience using graphical analysis but i'll play with it. I was thinning of reporting t values of each of the points as a test program.

                  Hi-Tech Redneck , yes, the part did move out of the CMM and was put back in. We are in the upper 40s low 50s right now.
                  Go the insert, report command, analysis. Choose the cylindricity reporting (the mouse cursor on your code needs to be after the cylindricity callout) change the multiplier to 100 (the bigger the value the bigger the arrows that showup the deviation) and press view window. You will have a visual representation of the points, the limit lines etc. Do this for the of them and see if there is 1 point that is causing this difference (most probably a burr or impurity on the part) or there are more. Also repeat with a different tip. Maybe the tip has a micro crack somewhere and is randomly causing spikes.

                  Comment


                  • #10
                    acgarcia your machine and head probe can guarantee that repeatibility?

                    I don't know your industry, but in a 3 way build on TP20, measuring whit 50 points, it's completely normal and acceptable get 0.0013 difference between measurements.

                    Comment


                    • constadin
                      constadin commented
                      Editing a comment
                      the 0.0013 is in inch not millimeters. The 0.0013" is in fact 0.033mm which is a lot to have as variation for cylindricity.

                    • Jfcampos
                      Jfcampos commented
                      Editing a comment
                      Ok I understood now your pov.

                      Thank you.

                    • BIGWIG7
                      BIGWIG7 commented
                      Editing a comment
                      @constadin
                      yes but not for the TP20. in my experience when it comes to checking diameters with the TP20 , the less hits the better. The TP20's spring just doesn't allow it to get great measurements when checking anything over 8 hits.

                  • #11
                    1) If you're using more than one tip of the star probe for the cylinder, the calibration could indeed cause that (small discrepancies on one tip, could put it out of sync with the other - check your cal results for high StdDEv and/or tip diameter). As someone else mentioned, a graphical analysis might show hits with one tip to be causing the issue.

                    2) How big (diameter and length) is the cylinder? 50 hits often seems a lot, but when looking at form you might be getting an incomplete picture. Especially if you're not taking the hits in the exact same locations.

                    3) The TP20 will exhibit lobing anyway, depending on the length of the build I'd expect 0.005-0.015mm so some of your error might come from that.

                    4) The normal (Std/Med/Ext) force module aren't really designed for star / T configurations. You can get 6way modules that are (but really I try avoid star probes on TP20s)


                    In actuality, I'd say it's probably a combination of the above. Have you checked the repeatability (not moving the part) and without recalibrating?

                    Applications Engineer
                    Hexagon UK

                    Comment


                    • #12
                      In the past, without removing the part from the CMM, the results are repeatable. The diameter of the cylinder is around 8 inches and is about 7 inches long. So the axis of my cylinder lays on my 454s Y axis. I can't fulling reach all the way around the cylinder so I move as low as I can go to maximize the spread of points. Using an extended force module and cal results are usually always under 0.0004" std dev.

                      Comment


                      • #13
                        Calibration can absolutely affect the form when using multiple styli to determine one shape.
                        As can residue, dust, oil, or any other FOD on the parts, cal sphere, or any of the 3 styli tips.
                        Depending on your probe length and weight, TP20's SF should be able to reproduce with 1.6um variation per the brochure
                        (these are from Renishaw "Data sheet_TP20-MCR20_EN.pdf", Issued 0505 Part no. H-1000-2064-02-B)

                        Haaha notice TP20 is being probed at 8mm/sec probe speed, THAT'S FAST!!! I was always told 2mm/sec for TP20.
                        image.pngimage.pngimage.png​​
                        Last edited by louisd; 11-18-2022, 12:30 PM.

                        Comment


                        • #14
                          This is from Renishaw's TP20 User's guide (H-1000-5008-04-C)
                          image.png

                          Comment


                          • #15
                            Originally posted by NinjaBadger View Post

                            4) The normal (Std/Med/Ext) force module aren't really designed for star / T configurations. You can get 6way modules that are (but really I try avoid star probes on TP20s)


                            In actuality, I'd say it's probably a combination of the above. Have you checked the repeatability (not moving the part) and without recalibrating?


                            ^ all this. the TP20 is a limited probe. you've got to find the sweet spot.
                            Last edited by BIGWIG7; 11-18-2022, 01:50 PM.
                            Che Guevara is a communist scumbag.

                            Comment

                            Related Topics

                            Collapse

                            Working...
                            X