iSixSigma

CpK determination

Six Sigma – iSixSigma › Forums › Old Forums › General › CpK determination

Viewing 14 posts - 1 through 14 (of 14 total)
  • Author
    Posts
  • November 4, 2008 at 6:14 pm #51257

    Lost it!
    ★ 10 Years ★
    Participant

    I’ve lost it.
    Maybe somebody can help me out.
    First of all I’m not a star in statistics, so maybe I’m not making sense at all.
    In the calculation of CpK. What standard deviation do I use? Do I use the standard deviation from the measurement system alone (determined by repeatedly measuring a standard) or do I use the total standard deviation that includes the variability of the property I am measuring plus the deviation caused by the measurement itself?
     
    Thanks

    0
    November 4, 2008 at 7:08 pm #177385

    Found it
    ★ 10 Years ★
    Participant

    Neither. Use the estimate of the standard deviation from Rbar/d2 obtained from the control chart.

    0
    November 4, 2008 at 9:06 pm #177387

    Ron
    ★ 10 Years ★
    Member

    First the formula for Cpk assumes that you know the actual process Cp (otherwise known as Pp) if you don’t have long term data take a sample of thriy to 100 pieces and measure them. Calculate the standard deviation of this large sample size.
    Once you have establish the proces capability you can calculate your capability ratio.
    If you use only a samll sample you are risking accuracy and inference erros.
     

    0
    November 4, 2008 at 9:21 pm #177388

    mcintosh
    ★ 10 Years ★
    Participant

    Found it is correct.  Ron is not.  See the three part explanation that starts with this one:  http://www.spcforexcel.com/ezine/oct2004/oct_2004.htm

    0
    November 5, 2008 at 6:50 am #177393

    Lost it!
    ★ 10 Years ★
    Participant

    But which variation?
    I have two.
    1) variation originating from the measurement system
    2) variation in the object I’m measuring
    (and ofcourse total variation)

    0
    November 5, 2008 at 8:50 am #177396

    Szentannai
    ★ 10 Years ★
    Member

    Hi,
    in principle you would want to use the process variation (variability of the product) and this variability will unfortunately include the variability of the measurement system. How you estimate the product variability is another story – but it will always include the measurement noise as well.
    Regards
    Sandor

    0
    November 5, 2008 at 12:14 pm #177400

    Ron
    ★ 10 Years ★
    Member

    first do a gage R&R if your measurement system is okay you always measure the part to part variation of the process.

    0
    November 5, 2008 at 2:26 pm #177405

    Derya
    ★ 10 Years ★
    Participant

    Hi;
    Before calculatin cpk, first of all you must make sure that %contribution of mesaurement system is <%1. Ýf not, you must improve it and then calculate cpk.
    If measurement system is ok, then you must use pooled standard deviation.
    Derya

    0
    November 5, 2008 at 3:52 pm #177411

    Lost it!
    ★ 10 Years ★
    Participant

    Thanks you all!!!
     
    I was on the right track afterall. I was already in the process of evaluating my measurement system in a Gage R&R but somehow got completely confused in all the statistics. It cleared up after a good night sleep and your input.
     
    Thanks!!!

    0
    November 5, 2008 at 4:18 pm #177414

    Mikel
    ★ 10 Years ★
    Member

    There is no requirement to have % contribution < 1%.Where do you come up with this number?

    0
    November 5, 2008 at 5:37 pm #177418

    Ron
    ★ 10 Years ★
    Member

    Wrong!
     

    0
    November 6, 2008 at 6:55 am #177433

    Derya
    ★ 10 Years ★
    Participant

    Hi Stan,
    According to you, what is the requirement? and is it possible to mesaure process capability rightly without an acceptable measurement system.
    Thanks for your opinion

    0
    November 6, 2008 at 10:33 am #177434

    Craig
    ★ 10 Years ★
    Participant

    I think Stan’s issue was with 1%.  The statement should have said less than 10% is preferred. (R&R/Tolerance, or P/T ratio)
    Before you release a measurement system to a production area, you should do a 5 part MSA. (Repeatability, reproducibility, linearity, bias, and stability).
    When you estimate cpk, the measurement variation (which as been deemed to be low I hope) is included in the estimate of your standard deviation.

    0
    November 6, 2008 at 11:42 am #177438

    Derya
    ★ 10 Years ★
    Participant

    % contribution ( % PD, or variance of MS/variance of process)

    0
  • Author
    Posts
Viewing 14 posts - 1 through 14 (of 14 total)

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