iSixSigma

CPK

Six Sigma – iSixSigma › Forums › Old Forums › General › CPK

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
    Posts
  • November 10, 2005 at 6:28 pm #41367

    Pipkin
    ★ 10 Years ★
    Participant

    Our process involves tearing a plastice blister from a .021pt. sbs (solid bleached stock) that is coated with a heat reactive solvent coating. The machine used to perform this function is a tensile tester that pulls the blister at a 90 degree angle. We measure the force in grams. Because of the nature of the paper making process we get a very large range on our Xmr charts. Then even when our charts show we are in statistical control we get a very low CPK number when we run a capability. Is it possible that we cannot achieve a CPK of at least 1.33?
    Incidentally, what is really being measured in this process is the strength of clay coating that is applied to the substrate of the .021pt sbs board by the paper manufacturer. The plastic blister is attached to the heat reactive coating through the use of pressure and heat.
    Any help would be appreciated. Thanks.

    0
    November 10, 2005 at 7:11 pm #129574

    thevillageidiot
    ★ 10 Years ★
    Member

    Absolutely (I think)….stability and capability are not the same thing….you can have a highly stable process that is 100% incapable of meeting the needs of your customer, and you can have a highly unstable process that is 100% conforming (at least for now). 
    Your control chart serves to tell you how stable your process is over time and to what degree your process aim varies.  Without stability, your Cpk is less than helpful for inferential purposes.  Since you have a stable process that is incapable, I would compare the Cpk measures with the quality ratings given to you by your customer for said output and the real-time yield of the process….these give you a more concrete understanding of your current quality levels than a theoretical Cpk…..Only my opinion….triangulate this off the kids wearing the pocket protectors and thick glasses….I have yet to earn mine.
     

    0
    November 10, 2005 at 7:27 pm #129575

    Pipkin
    ★ 10 Years ★
    Participant

    Thank you to thevillageidiot – your practical approach makes. I have gotten in trouble with customers before when using theoretical concepts.

    0
    November 10, 2005 at 8:06 pm #129576

    CT
    ★ 10 Years ★
    Participant

    Jack,
    Just curious is your data continuous over a long period of time?
    Snap shots of data sets over short runs may prove to be more of an intelectual importance in regard to CpK.
    Just my thoughts
    CT

    0
    November 10, 2005 at 8:51 pm #129577

    thevillageidiot
    ★ 10 Years ★
    Member

    You bet…In my limited experience, the Cpk is useful as a benchmark measure to which I compare the effects of improvement efforts on a process…I find it too unwieldy and quite frankly,  prone to interpretation and maniplulation, to use as a strict prognosticator…especially if using it for external customers…quite effective as an in-house metric if calculation and interpretation methods are standardized and fully understood by all parties…..good luck.

    0
    November 10, 2005 at 8:55 pm #129578

    Pipkin
    ★ 10 Years ★
    Participant

    Ct,
    Yes the data sets are over a short time period. Thanks for you thoughts.

    0
  • Author
    Posts
Viewing 6 posts - 1 through 6 (of 6 total)

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