Gage Study When No Appraiser but Several Test Bay (Sockets)

Six Sigma – iSixSigma Forums General Forums Methodology Gage Study When No Appraiser but Several Test Bay (Sockets)

Viewing 3 posts - 1 through 3 (of 3 total)
  • Author
  • #247892


    Hello, at my company we manufacture Automatic Testers for Automotive and I need your help to better define some aspects in order to improve our product Quality by using MSA.

    Important aspects:

    1) A tester is usually made of more than one “Socket” (a kind of Test Bay where Device Under Test il placed).
    2) Each socket is expected to provide the same measurements as the other ones.
    3) There is no human interaction: after the Device is loaded the tester executes automatically by SW an automatic test sequence on Electric/Electronic DUT components.

    I was thinking to evaluate Cg, Cgk with a Type1 Gage Study (Ref. Minitab) and then a GR&R.

    My questions are:

    1) Should I perform Gage Study Type1 (Cg,Cgk) on Each socket first and the on ALL sockets (without filtering by socket) in order to evaluate the correlation?

    2) Since there isn’t any APPRAISER’s contribute (automatic testers) shold I avoid considering APPRAISERS in GR&R and replacing this concept with SOCKETS?

    Thanl you very much!


    Roger Hill

    Hello, I have some questions.

    1 is the socket the interface between the DUT and the testing electronics?  If so, does it contain signal conditioning electronics?

    2 are sockets dedicated to test bays or can they move from test bay to test bay?

    your description sounds like multiple assembly lines each with a final dedicated test bay.  Depending on what is being assembled, a particular socket is used to attract the DUT to the test bay.   Is my understanding correct?



    Mike Carnell

    @Rumbero65 You make the assumption that the person loading the parts has no effect. There is probably a good chance you are correct. Just to be sure I would run a hypothesis test just so you have it in your back pocket in case someone asks.

    We have had situations like this frequently. I am not sure I am exactly clear on how the test is set up but it isn’t that uncommon the substitute a test set, tooling, etc. for the appraiser. If I were doing this and making the assumption the appraiser had no effect I would still use 1 person to insert the parts. I would probably also use 1 appraiser for trial #1, a second appraiser for trial #2, etc. That allows you an additional analysis option.

    The one thing you do want to be very careful of is randomizing the order the parts are tested. Do not take them in the same order for every trial.

    Just my opinion.

Viewing 3 posts - 1 through 3 (of 3 total)

You must be logged in to reply to this topic.