Interim checks

  • Filter
  • Time
  • Show
Clear All
new posts

  • Interim checks


    I read about this in the NPL Good Practice guide No. 80 - Fundamental good practice in dimensional metrology.

    Interim checks

    The regular performance verification check of a CMM should be a key component of

    the company quality system and records should be kept of the results obtained.

    Unfortunately these tests may only be performed once or twice a year – so what do

    you do to ensure you have confidence in the machine on a daily basis? The answer is

    to perform interim checks. These checks are also described in the ISO 10360-2, but

    essentially involve regularly measuring an object that is similar to the regular

    components inspected by the machine – this may be a special test piece, a calibrated

    master part, ball plate, hole plate, machine checking gauge, or ball bar. By measuring

    a series of features on this artefact each day, or perhaps first thing on a Monday

    morning and recording the sizes on a control chart, one builds up a visible history of

    the machine’s capability.

    Trends may be observed on the control chart indicating that the results are within

    some specified limit and likely to stay within the bounds for some time, but the real

    benefits are observed when sudden changes are highlighted. Sudden changes are often

    indicators of nasty things happening to the machine – damage to an air bearing, a

    collision with a forklift or crane, or simply dropping a heavy object - sometimes

    people are embarrassed to admit they have damaged an expensive item and may not

    say anything about it. It is up to the operators to ensure that they are confident in the

    results from the CMM and getting the result they expect from a daily interim check.

    If the CMM fails the interim check then the operator is encouraged to undertake a

    performance verification test to see if the system really is out of specification.
    And I found it interesting, so I decided to give it a try. I just wanted to see how you guys would go about doing it?
    Which measurements would you take, and how would you evaluate them? What evaluations would be most relevant?

  • #2
    Make a “golden part”. Measure and record all dimensions manually, (surface plate checks)

    Then use CMM and measure part at various locations on the CMM table.


    There are no bugs, only "UNDOCUMENTED ENHANCEMENTS!"



    • #3
      One can buy or have made a ball-bar, a ball plate, a step gauge, a Glastonbury gage, a Renishaw MCG, or as dph51 said, just use a "golden part". I had a golden part I repurposed from the scrap bin. An ideal golden part would be one that is of similar size to most of your components and has the types of features you check most often. It doesn't matter if the part is 'in tolerance' or not, you just want to establish a baseline of what it checks now, and then remeasure it periodically and compare the new results to the historical record.

      The advantage of using a 'precision' artifact of known 'Calibrated' size is a verification of traceable accuracy, but at the cost of a precision gage and sending said gage out for calibration/certification.

      There is a most excellent case study of the benefits of a golden part, how it indicated a subtle issue with a CMM system and process of problem solving based on that clue written by one of the retired greats of this forum Mr. Hilton Roberts. I recommend searching for it.

      Happy Friday!
      sigpic"Hated by Many, Loved by Few" _ A.B. - Stone brewery


      • #4
        If you have some cmm's, you can measure the same part with the same program on different cmms, then calculate the average, the std dev, and do it once a year, for example. (It's only a n intercomparison !).
        If there are intercomparisons organized, ou can suscribe to them, so you can follow your cmms and your methods, if there are not imposed.


        • #5
          Invariably, our Production Manager doubts the dimensions I tell him (when I have the audacity to tell him that a part checks OOT).

          I've just gotten into the habit of verifying specific dimensions via Height Stand, comparator, pins, etc. If the two dims corroborate (which they generally do), all is good, and I can reject the part with confidence (doesn't mean that they still won't send it...), and I can tell him that I double checked it via alternative method, and shoot his argument down before it even gets out of his mouth.

          However... There have been times where the numbers don't jive, and the alternative method checks the part good... Which is right? Which is wrong? Why is it wrong?

          When it's wrong, (which isn't that often, thank God) I usually learn something about why my setup was faulty, or find out that somebody bent a probe (JOY!) or some other such nonsense...


          • #6
            I've found an old part that was used a few years ago to calibrate a measurement equipment for flatness in the production line. It's pretty large and got lots of holes, pockets etc. It doesn't come with a CAD model though but I'm pretty good at programming without CAD so it'll be fun.

            This will be my last project before going on vacation.

            Thanks for your inputs, it's always interesting to read. Now I'm going to try find that study by Mr. Hilton Roberts.


            Related Topics