iSixSigma

Required Ppk higher than Cpk

Six Sigma – iSixSigma Forums General Forums General Required Ppk higher than Cpk

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
    Posts
  • #53715

    Pastore
    Participant

    In reviewing a customer’s Quality Manual, they require a Cpk>=1.33 and Ppk>=1.67. Due to long term variation being greater than short term, I would think the required Cpk should be greater than the Ppk.

    Should the required Ppk be greater than Cpk? If so, can anyone explain why?

    0
    #191200

    Mikel
    Member

    Your customer has their definitions backwards, something not that uncommon a decade ago.

    Get them a copy of the AIAG SPC manual for their birthday.

    0
    #191216

    Klerx
    Participant

    Cpk is based on within variation, Ppk is based on total variation and both have nothing (!!) to do long term or short term. Ppk is worst case (not necessarily in statistical control). If worst case is better than 1.66, probably in the stable situation the Cpk is better than 1.33

    0
    #191218

    Mikel
    Member

    Uh.

    Duh.

    If worse case is better than 1.66, wouldn’t Cpk be better than 1.66.

    If you really don’t understand the question, you shouldn’t try to answer it.

    Duh.

    0
    #191251

    Jeff Barnhill
    Participant

    Great question!

    Cpk uses an “estimated sigma” value (Rbar/d2) in its formula. It is commonly used to measure whether the process is capable of meeting a customer’s requirements.

    Ppk uses the “actual sigma” value ( S – Google it) in its formula so it is commonly used to measure whether the process is actually meeting a customer’s requirements.

    This is why Cpk is typically known as a “short term Capability” indicator and Ppk a long term.

    Remember, on Isixsigma.com, there are no stupid questions!

    0
    #191252

    Mikel
    Member

    Eddie,

    There are a lot of stupid questions. Pay attention.

    0
    #191260

    Severino
    Participant

    Are you sure that’s an “and” and not an “or”? At one point in time (not sure if it’s fixed) the AIAG manuals had requirements that you must have a Cpk>=1.33 if your process was statistically in control and a Ppk>=1.33 if your process was not statistically in control. Montgomery et. all likened this practice into statistical terrorism since your capability studies analyze history to predict the future and we cannot have faith in a prediction of a process that is not stable (in-control). I could see someone trying to compensate for a lack of faith by stating that if you are in control you need to achieve the Cpk>=1.33 or if you are not in control then Ppk>=1.67 under the incorrect assumption that the higher capability requirement safeguards you against your lack of stability. The key his is that is an “or” condition and not an “and” condition.

    0
Viewing 7 posts - 1 through 7 (of 7 total)

You must be logged in to reply to this topic.