iSixSigma

Andrew

Forum Replies Created

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
    Posts
  • #209376

    Andrew
    Participant

    @mikecarnell would this mean that the measurement system resolution needs to be improved by a factor of 10?

    @yaseminin I am struggling to understand the example – how many different parts are used, how many times is each measurement repeated? Could you please share the Gage R&R graph if available?

    0
    #209076

    Andrew
    Participant

    You say you can run 60 experiments which is good, that gives enough data points to be confident.
    The problem is that you have too many levels to get a manageable number of runs.
    5 levels could be an overkill if the effect of a factor isn’t linear or quadratic.

    Some options which include reducing the number of levels are:

    1) 2 level Full factorial design: 16 runs to which you can add a few center-points. You may want to do more replicates as it is good to have more data points to take a decision.

    2) 3 level Full factorial design: 81 runs (which is over your limit of 60) BUT if you reduce one of the factors to 2 levels and keep the other 3 at 3 levels it is only 54 runs, to which you can add a few centre-points.

    0
    #209072

    Andrew
    Participant

    Thank you very much for the data set, what you said makes sense.
    From what I see it is not possible to get contradicting Pearson coefficients and P-Values.
    If the Pearson number is high, the P-Value will be low (-> there is a limear correlation)
    If the Pearson number is low, the P-Value will be high (-> there is no linear correlation).

    After giving it some further thought I would say that the Rsquared adjusted is a better predictor of correlation than Pearson’s correlation factor – as it can be calculated for linear, quadratic and cubic regression models.

    0
    #203139

    Andrew
    Participant

    @truong, if we think about an output Y:

    -The Upper Specification Limit (USL) is the highest level of Y the customer will accept, based only on customer requirements (here, your customer seems indeed to be the next process step). It is not obtained through calculation.
    -The Upper Control Limit (UCL) is the highest level of Y at which we consider that the process will still be in control. It is a calculation based only on the values of a set of consecutive data points and the moving range.

    Your UCL needs to be lower than your USL to make sure the output Y is within specification.

    It looks like your current approach is that you are looking at your UCL to define your USL – you should instead be defining your USL based on customer requirements and making sure your UCL is lower than it.

    0
Viewing 4 posts - 1 through 4 (of 4 total)