iSixSigma

MSA – Hardness Testing – Destructive or Not?

Six Sigma – iSixSigma Forums Old Forums General MSA – Hardness Testing – Destructive or Not?

Viewing 8 posts - 1 through 8 (of 8 total)
  • Author
    Posts
  • #35708

    Chris L
    Participant

    Hi,
    I am setting up a MSA (Gauge R&R test) on hardness testing, and in doing so am trying to decide whether Hardness Testing of Metal Components (either Brinell or Rockwell C) constitutes a Non-destructive or a Destructive test? i.e. is the analysis performed as a Crossed or Nested Study?
    The minitab definition of a destructive test is that each part can only be inspected once by each operator – clearly a single part can be hardness tested many times – therefore it is a non-destructive test (crossed) test. However, it can also be argued that each repeat measurement on a single part must be taken in a slightly different position and hence the test is ‘destructive’ as the exact same position cannot be measured more that once – so it is a destructive (nested) test!
    If anyone has faced this conundrum already please give me your thoughts / advice ASAP.
    Many thanks,
    Chris L

    0
    #100979

    Anonymous
    Guest

    Chris,
    You could also argue that the probe deteriorates from test to test. If I were you I would start off by taking a second measurement nearby and later justify this by studying the sources of variation. For example, if you can show you have an accurate and precise measurement system; how much variation do you observe around the component – was your assumption reasonable.
    Andy

    0
    #100985

    Mike Clayton
    Participant

    I agree that hardness can be treated as non-destructive, IF you make multiple measurements near same site (not too close as past test changes its neighborhood slightly), then use the Shainon Plot to show that the family of variation piece to piece is MUCH less than within piece.  Of course if you have a large piece, then make measurements wherever it matters to find the spatial variation (within piece) with a MAP perhaps for more understanding.  And then if it shows that points in same zone or region are similar, then those “repeated” measures in same zone show your one of the variance components, while the wider within piece, and the piece to piece, and operator to operator are other components of variation.  Then use the new GUM method of stating the UNCERTAINTY of the measurement, rather than the classic %R&R with its strange “rules of thumb” for gage goodness.  TN 1297 is good reference document from NIST.  http://physics.nist.gov/cuu/Uncertainty/basic.html

    0
    #100989

    Mike Carnell
    Participant

    Chris L,
    You can do it either way. Actually try it both ways and see what the difference is.
    If you are worried about hitting the same spot repeatedly fixture the tester/samples so you get the same spot each time. Run it as crossed.
    Don’t worry about fixturing and run it nested.
    You can also select samples and test tehem in various spots and run a hypothesis test and check an assumption of homogeneity.
    Lots of options.
    Good luck.
     

    0
    #100996

    rathinakumar
    Participant

    Chris,
    Metallurgically speaking, hardness tests are non-destructive. Scatter in readings do occur due to material and heat treatment inhomogeneities. To accomodate this, most specifications ask either for a minimum (API) or a maximum(NACE) or a range(user specs in oil field eqpt.). Seldom is a target alone specified. Within piece as well as Part to part variation is something we encounter daily.
    Scatter can be minimised by eliminating human error in reading the indentation dia. in Brinell by use of Brintronic/Foundrax computerised equipment.
    you may email me at [email protected]  
    with more details of your part and spec and I’ll see if I can help.
     Regards

    0
    #100997

    M Tuesday
    Participant

    It was defined as non-destructive and unable for replicatation. See AIAG MSA 3rd manual that has some explanation.
    DOE book by Dr.Montgomery has some good example that most nearly your area.

    0
    #101010

    Chris L
    Participant

    All,
    Thanks for your input, all of if valuable.
    I am comfortable with the different in hardness values I will find across a single component due to metallurgical inconsistencies – I can cope with these by creating a ‘map’ of the component using hardness results as a means of certifying the raw material / component. I am quite experience in gauge R&R testing , having completed more than I care to remember. I suppose my query was, do I process the results as ‘crossed’ or ‘nested’?, as this can have quite an effect on the GR&R total StDev, thereby affecting the number of repeat measures (CLT) I have to take to establish a ‘capable’ measurement system before I create a hardness map of the component.
    So, Crossed or Nested?
    Thanks,
    Chris.

    0
    #101035

    Dillon
    Participant

    Chris,
    You can perform your study as “Crossed” if you can prove that you have a consistent product from piece to piece. 
    We have done this in the past by control charting various samples down the length of a coil of material.  Using this study, we could state that samples from the material were statistically the same which let us treat the material as the “same” part.
    If you cannot prove this, nested it is.  Hope this helps.

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

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