iSixSigma

MSA with 30 Plus Operators Measuring

Six Sigma – iSixSigma Forums General Forums Tools & Templates MSA with 30 Plus Operators Measuring

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
    Posts
  • #53864

    Rappuhn
    Participant

    I have a situation where a supplier has over 30 people using calipers to measure a specific feature. They used 3 of the operators to do a MSA which passed so I can assume that they can adequately measure the feature. However can I assume that all of the other operators are capable of doing the measurment or should I ask them for some other type of information?

    0
    #191686

    Strayer
    Participant

    Every assumption is an acceptance of risk. If you aren’t comfortable with the risk then, yes, you should ask for more information. You might want to know, for instance, whether the 3 operators were randomly selected, is that an acceptable sample size, do all operators get the same training and skill certification, have they done a Gage R&R on the calipers, are all of the calipers the same model, from the same supplier, and of the same age.

    0
    #191687

    MBBinWI
    Participant

    While Straydog is correct, and you need to assess the risks, I would also suggest that you also think about the purpose of the MSA. What you are looking to understand is whether the measurement system is “good enough” to make the decisions necessary based on the data measured. While some advocate including all operators in a GR&R study (which is the safest approach) I would instead look to include the operators with the worst capability. Just observing the various operators perform the measurements in their regular manner should quickly point out which of them are going to have the most variation/least accuracy. If you include these as the 3 operators (3 so you can get a better estimate of variation) should provide the info required to understand whether the gage system in question is “good enough.” If the worst operators can pass the criteria for adequate accuracy/stability/bias/linearity/etc. then the rest will be even better.
    Of course, just my humble opinion.

    0
    #192035

    Robert Ballard
    Participant

    I concur with MBBinWI. The “safest” approach is to include all operators in the Gage R/R study. The next best option to obtain a statistical evaluation of measurement error is to conduct a Range study. This method can be found in the AIAG manual. The range method will allow you to evaluate either Repeatability OR Reproduce-ability. You could use the original Gage R/R study to determine if there is more Repeatability error or more Reproduce-ability error and use this knowledge to design the study. The Range method does not provide an evaluation for Operator*Part interactions so if an interaction is present in the original study you should understand the risk if you go the Range Method route.

    Good luck.

    0
    #192070

    Stan Mikel
    Member

    @Ballard That sounds like a good approach, who taught you that?

    0
    #192133

    Afsar Ahmed Choudhury
    Guest

    Please go to minitab.. read the help file and do the tests

    0
    #192136

    Mike Carnell
    Participant

    Which part of the Minitab help menue do you think answers the question?

    0
Viewing 7 posts - 1 through 7 (of 7 total)

You must be logged in to reply to this topic.