iSixSigma

Gage RR

Viewing 8 posts - 1 through 8 (of 8 total)
  • Author
    Posts
  • #50445

    Ashman
    Member

    I’m attempting to perform a gage R&R measuring runout on a motor shaft (0.0005″ tolerance) using a test indicator and having a difficult time getting reproducibility even < 20%.Β Β We going to attempt to standardize the process between the three operators but any suggestions on how to further reduce this characteristic?Β  Perhaps a better measuring instrumentation?Β Β  Thanks.

    0
    #173419

    Taylor
    Participant

    Yes a better instrument would help. Also, and just speaking from experience, a controlled environment, one that is free of vibration and air currents. Fixtures used to rotate the shaft or the instrument also can be suspect if not kept in perfect condition.
    Just my 2 cents

    0
    #173420

    Ashman
    Member

    Thanks…any advice on a measurement devices that have worked well for you in the past?

    0
    #173425

    Taylor
    Participant

    Steve,
    Β Some of the new Laser & Optical roll gages are probably the most accurate. Also a few people out there make some air bearing spin fixtures that will do the trick for most applications. I’ve used these for some fuel injector shafts in the past, and they get the job done with a good indicator. I have also seen a Company called Quality Assurance Engineering that has some really nice high end equipment, that will do more than just measure TIR. I would suggest doing a Google and see what you come up with that suits your needs.

    0
    #173428

    Farooq Ch
    Participant

    In addition to investigating a new measurement device, the way the shaft is positioned for measurement is equally critical. If yourΒ set-up is repeatable, then going for a betterΒ measurement device makes sense.
    On a side note aΒ R&R of 20% is not bad, I have seen a R&R of more thanΒ 30%Β  on similar measurements and acceptable per the specs. You may want to check what the R&R specs are for your product.
    Β 
    Β 

    0
    #173430

    Michael Mead
    Participant

    I suggest a roundness testing instrument. I don’t know exactly what to call them, one tradeneame is Tallyround.
    The resolution of a test indicator is probably not good for this. The best you can get is o.00005, or 10% of your tolerance range. With this, operators must interpolate and that can also lead to problems.
    Β 

    0
    #173443

    Michael Schlueter
    Participant

    Hi Steve,
    Just a few stimuli for you. May be one of these can lead you to better results?
    1) How can you eliminate the need for this measurement? Many times a measurement is required to control some other harmful effect. Can you eliminate that one, and as a byproduct, eliminate the need for this measurement?
    2) How can you replace this measurement by detection? I.e. which other effect can you use, which just tells you “no it’s out of tolerances”?
    3) How can you use secondary effects from this tolerance? I.e. once you exceed tolerances which other parameters do change as well, which are easier to detect or to measure?
    4) How can you detect or measure a by-product? Think of noise, resonance, lever effects etc.
    Illustrations:
    1) Noise protection – and noise measurement –Β is needed near urban highways. Eliminating the cause of noise, loud cars, eliminates the need for noise measurement.
    2) Tires have special markers built-in. When they wear-out the markers become visible. This situation is easy to detect. Though it could be measured there is no need for it to do so, because detection is easy to do in this case.
    3 + 4) Can you, for example, correlate the tolerance/deviation with an audio spectrum sampled from your device? Characteristic peaks may come and go or certain plateaus may shift.
    Of course Chads remarks on the measurement environment are more than valid. With the spectrum approach you may even have a way to compensate for such effects.

    0
    #173486

    Mikel
    Member

    You are on the wrong track here, instrument problems don’s show up
    as reproducibility. Reproducibility is technique, training, procedure,
    fixturing type problems. A wrong instrument will show at repeatability or in both repeatability
    and reproducibility.

    0
Viewing 8 posts - 1 through 8 (of 8 total)

The forum ‘General’ is closed to new topics and replies.