iSixSigma

Zvalue vs Cp

Viewing 8 posts - 1 through 8 (of 8 total)
  • Author
    Posts
  • #53573

    verma
    Member

    Dioes Z value deliver similar result as Process Cabality if not then how do we distinguish between them and how they are important.

    0
    #190721

    Mikel
    Member

    yes

    z values are not important.

    Cp, Cpk (sometimes), Ppk, and Cpm is all you ever need.

    0
    #190730

    Rohatgi
    Participant

    NO !
    Z-Value is different than Cp

    For Cp you require Specificatin Limits

    Statistically

    Z-Value is calculated on N(0,1) Normal Distribution which has a mean=0 and std. deviation = 1
    Z-Value gives the Area under the Curve from Mean to the point selected on X-axis. Generally USL or LSL
    It may be translated into Process Capability based on Spec Limits (USL or LSL )

    Process Capability is a Ratio

    | USL-LSL | / 6*StdDev

    Anupam

    0
    #190738

    Kotval
    Participant

    I agree with Anupam. Z value represents the standard normal distribution. Using the normal distribution tables, the Z value is used to determine the probability of occurrence above or below that value.

    Cp is a capability measure which provides a value which is a measure of how good a process distribution is relative to the specification. Cp is a capability index only for a distribution centered on the target value. Cpk is a better measure for distributions which are shifted on either side of the target.

    0
    #190749

    MBBinWI
    Participant

    Z=3*Cpk and is directly analogous to sigma level, whereas Cpk of 2=6 sigma (no, I don’t want to get into the arguments about shift) is always a confusing issue with those not intimately familiar with the topic.

    Cpk of 1 (3 sigma) was chosen when that was considerred “good”, and 1 was a nice round number to target. Now that that isn’t the case, as well as all the focus on “six sigma” methodology requires a metric that relates to that paradigm (in my humble opinion).

    Also, if you want to be even more specific on the actual capability, a Z-Bench is more appropriate and has no comparable Cpk type metric, particularly when the tails of the distribution being evaluated extends beyond both spec limits – a situation not adequately handled via Cpk.

    Sorry, Stan, you’ve not convinced me yet. I’ll take the ability to analagyze to sigma level and capture true performance for lower capability processes over mere legacy.

    0
    #190750

    MBBinWI
    Participant

    vijay1977 wrote:

    Dioes Z value deliver similar result as Process Cabality if not then how do we distinguish between them and how they are important.

    I guess to be precise, Z does relate to Cpk (actual process capability), but neither Cpk nor Z directly relate to Cp (potential capability). This is because Cp does not depend upon actual mean level, but both Cpk and Z do.

    0
    #190752

    Mikel
    Member

    No problem MBB

    Instead of z-bench, I just call it crappy when I have to worry about stuff in both tails

    0
    #190761

    MBBinWI
    Participant

    Stan wrote:

    No problem MBB

    Instead of z-bench, I just call it crappy when I have to worry about stuff in both tails

    Absolutely agree. Unfortunately, XL doesn’t have a way to plot “crappy” and track to a “non-crappy” level.

    0
Viewing 8 posts - 1 through 8 (of 8 total)

You must be logged in to reply to this topic.