+ Reply to Thread
Results 1 to 3 of 3

Thread: Acceptable Ppk in relation to confidence/reliability

  1. #1
    Points: 29, Level: 1
    Level completed: 58%, Points required for next Level: 21

    Thanked 0 Times in 0 Posts

    Acceptable Ppk in relation to confidence/reliability

    Hi All,

    Most of the acceptance criteria that my team works with is based around a given confidence/reliability interval that needs to be met. There are multiple situations where capability analysis is being used (specifically, Ppk), which I'm not entirely familiar with. Is there an equation to relate a given Ppk with a %conf/%rel? All I've been able to find is the general acceptable Ppk of 1.33, but I haven't found anything to relate this number to confidence intervals. Ideally I'd like to be able to calculate a minimum required Ppk value for a range of required confidence/reliability requirements.

    I appreciate the help!

  2. #2
    TS Contributor
    Points: 15,152, Level: 79
    Level completed: 61%, Points required for next Level: 198
    Miner's Avatar
    Greater Milwaukee area
    Thanked 408 Times in 366 Posts

    Re: Acceptable Ppk in relation to confidence/reliability

    Ppk is a quality metric, not a reliability metric.

    Ppk is the smaller of (Upper Spec Limit - mean) or (mean - Lower Spec Limit) divided by the long term process variation (3*StdDev). A Ppk of 1.33 means the closest spec limit is 4 standard deviations from the mean. This provides an allowance for a process shift to be detected and corrected without producing nonconforming product.

    Ppks do have confidence intervals associated with them based on the sample size used to calculate them. However, capability indices have no time element to them and are unrelated to reliability.

  3. #3
    TS Contributor
    Points: 12,501, Level: 73
    Level completed: 13%, Points required for next Level: 349
    rogojel's Avatar
    I work in Europe, live in Hungary
    Thanked 334 Times in 314 Posts

    Re: Acceptable Ppk in relation to confidence/reliability

    to build on that. if you are using Minitab 16 or 17 you can just select the option of including confidence intervals. However, in my experience this rarely done in practice, mainly because of the low sample sizes used to calculate the indices.

+ Reply to Thread


Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts

Advertise on Talk Stats