iSixSigma

MiniTab Control Charts

Six Sigma – iSixSigma Forums Old Forums General MiniTab Control Charts

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
    Posts
  • #49562

    uhlan
    Member

    I need to set-up a control chart. It will be for continous numerical data. All from one part type, same assembly line, same process, same suppliers, operators, etc. The process/part test results all need to be below a maxium value. In general, with the scant data I have provided you with, what control chart option within Mini Tab should i probably be selecting?
    I’m sure that the answer may very well be, “well it depends” but any thoughts would be very approciated.
     
     

    0
    #169562

    sickpig
    Member

    Use control charts to see X-Bar and I-MR charts. Assuming your process is in control and you have historical data you can see what Minitab offers as the UCL/LCL compared to the Customer USl/LSL

    0
    #169564

    Sloan
    Participant

    It isn’t as simple as just picking a control chart type and jumping im. I suggest you start first with the basic questions about your process. How often will you sample the data from the parts? How easy is it to measure? What should your sampling strategy be? Do you measure every part or just spot check periodically? If you decide to subgroup your measurements, what is the right subgroup size? Do you produce these parts in batches or in lots? Once you have some of those questions answered, (I suggest building a measurement plan) then you can decide whether to use X-bar/R or I-MR or something else.

    0
    #169786

    JC
    Participant

    I like using the I/MR.  Though my data are discrete and not continuous.  At first glance I would use I/MR and select parts randomly across process points;day, time of day, etc.

    0
    #169906

    Jen
    Participant

    Uhlan,
    Keep in mind that I&MR charts are for homogenous data such as sampling from a waste water influent tank, paint vat, gases etc., meaning if you were to take five samples from something like this there probably wouldn’t be much difference in the range and it would be very similar to having taken just one sample.  However, taking samples in different time chunks may show differences and therefore the “range” is then determined by subtracting the previous sample from the current sample.
     Xbar&R charts are for those processes where when you take information from a sub group of items there is more opportunity for them to have differences in the measurements and therefore give you a range.  Examples are diameters on a cylinder bore of an engine block, flatness on the surface of cylinder head decks etc, with each part being its own measurement and the subgroup is made up of how every many parts you pulled off the line (2, 3, 4, 5, …).
    Once you have looked at your process, such as Outlier suggested you will know better which chart technique is correct for you.

    0
    #169926

    uhlan
    Member

    Thanks Jen.
    Precisely (well lets say well within tolerance of) the answer that I was looking for!!!!

    0
Viewing 6 posts - 1 through 6 (of 6 total)

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