Variance between measurements

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

  • Variance between measurements

    I'm getting measurements that vary from 0-0.015 mm if I measure more than once. Is that to be expected? It can affect if parts pass or fail. Any troubleshooting tips here?

  • #2
    We need way more information to help. Code, what type of features....

    Comment


    • #3
      Just send them. I am confident in this and have them call me if you have any issues. The African prince who recently solved his money transfer issues should have my number......

      Comment


      • #4
        Too generic of a question.

        As BKulpa said, what are your features? Number of hits? Could be an alignment issue. Please post code.

        Comment


        • #5
          block steel, sheet metal, plastic, rubber, milled, stamped, injected, ground...... and biggest of all.... MACHINE and environment
          sigpic
          Originally posted by AndersI
          I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

          Comment


          • #6
            You need to know the accuracy of your machine, it is on your cal certs.

            How solid is your alignment?
            a) datum features are in good shape (flat round) or should you use simulators?
            b) datums take up the majority of the part's volume, or they are .01mm apart and you are inspecting 10m from zero-zero?
            c) your hit placement on the datums is representative of the part drawing or you are doing 3-2-1 on a 10m cube?
            d) your alignment is hitting the same places on the part every run, or you are placing it loosely into a fixture, so it is close enough to last run to find the part?
            e) your alignment strategy was designed with the tolerances you are inspecting in mind?

            How is your probe qualification?

            Is the probe shanking anywhere?

            Are there burrs that obstruct measurement?

            If you are inspecting a compound angle hole, or single angle hole, are you able to get full measurement, or you have a head position that is sort of close and you are only hitting 65% of the diameter so you don't shank?

            If you tolerance is super tight, are you even articulating the head (presuming you can)? You probably shouldn't be. Align and check all tight tolerances with the head in a single position then check the looser stuff with the head articulating.

            Is the stem on the probe sufficiently stiff?

            Is the machine moving at a reasonable speed for the probe in use? (I have a .3mm stylus with a .2mm stem. If it moves too fast, the stem deflects before the touch trigger registers the hit)

            On that, what is the design of your probe system? There are a few, most everything better than touch trigger.
            a) if it is touch trigger, where are the 3 directions with the least resistance, can you measure datums and features with tight tolerances only there?
            b) if touch trigger, can you measure holes with the same angles of contact, with respect to the tri-lobe construction of your probe?

            Is the CMM the correct tool for the inspection you are doing?
            a) is this a run-out with a touch trigger probe?
            b) diameter?
            c) Position?

            Have you compared the results to what you get inspecting with other methods? If you are 100% sure the part is .0075mm and the CMM is telling you 0 - .015, then that is one thing, if you are certain the part is .025mm and the CMM is saying 0 - .015 that is something completely different.

            Hows the temperature where you are inspecting?

            Hows the vibration?

            Clean room?

            What span of time is this occuring over? I'm guessing a few minutes, but you don't say. Could be a lights out program with a 14 hour cycle.

            Long story LONG, there are a lot of factors.

            I would start with what the part truly is (presuming the CMM is the correct tool), measure 2 or 3 other ways and figure out of you are already close, then look at refinement. Eliminate head articulation. Cycle (itterate) the alignment so it is very close to hitting the exact same points for alignment on the last itteration prior to inspecting.

            Comment


            • #7
              never have I seen anybody make asking a question a project the way you guys can....

              Comment


              • #8
                Originally posted by Douglas View Post
                never have I seen anybody make asking a question a project the way you guys can....
                I have 2 machines, one is "new" (3-yo) and one in "old" (30+ yo).

                0.015mm (0.0006") is less than the OEM specs (0.0007") for the old machine, so I wouldn't really think about it on that machine. I do sheet metal parts of all kinds for automotive. 0.015" 'thick' steel moves, 0.015mm, wouldn't bat an eye, 12mm THICK steel, on the new machine (0.0002" spec) would make me look twice.

                There really is a lot we need to know to tell the OP if 0.015mm is 'too much'.
                sigpic
                Originally posted by AndersI
                I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

                Comment


                • Douglas
                  Douglas commented
                  Editing a comment
                  yes I know I didn't mean it in a bad way... but I have gotten the impression the questions for more info have made a guy or two learn how to even answer them

                • Douglas
                  Douglas commented
                  Editing a comment
                  Boss at work: Go on PCDMIS forum to ask the question

                  Rookie Operator: I now have more questions than when I went in

              Related Topics

              Collapse

              Working...
              X