iSixSigma

Capability

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

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
    Posts
  • July 23, 2008 at 2:35 pm #50589

    Schuette
    ★ 10 Years ★
    Participant

    Hello – does anyone know if it is possible to do a capability study with attribute (pass/fail) data?  I realise it is normallly with continuous data but was wondering if there was some sort of “fudge factor” that could be used.
    cheers
    Jim

    0
    July 23, 2008 at 2:39 pm #174082

    Robert Butler
    ★ 20 Years ★
    Participant

    Yes, check Chapter 9 of Measuring Process Capability by Bothe – the chapter title is Measuring Capability with Attribute Data

    0
    July 23, 2008 at 2:52 pm #174085

    Schuette
    ★ 10 Years ★
    Participant

    Thanjs for this – Unfortunately I dont have the book but need to be able to do the test fairly quickly.  Could you possibly give me an abbreviated version please?
    cheers
    Jim

    0
    July 23, 2008 at 2:52 pm #174086

    Ken Feldman
    ★ 10 Years ★
    Participant

    If you have access to Minitab, check it out and you will see that it allows you to do capability on about any type of data and any type of distribution.

    0
    July 23, 2008 at 3:25 pm #174089

    Schuette
    ★ 10 Years ★
    Participant

    Great – thanks Darth.

    0
    July 24, 2008 at 1:34 am #174116

    Michael Mead
    ★ 10 Years ★
    Participant

    I just had some sort of action in my brain. Several years ago, I set up a control chart for attributes rather thatn variables using narrow limit gaging. I wonder if that could apply to a capability study as well.  I might need to do sum thankin’ on that.

    0
    July 24, 2008 at 11:52 am #174129

    Chris Whittaker
    ★ 10 Years ★
    Participant

    Hello,
    The best way to do a capability analysis for discrete data is to do DPMO. (defects per million opportunities).
    If the data is non- normal use DPMO method  to convert continuous Y into discrete X by formulae Defect = Y/N and this will give you capability for discrete non normal data
    DPMO
    Unit (U) = number of parts or systems
    Opportunity (op) = a test characteristic or inspection part
    Defect (D) = anything that does not meet customers specs or is non conforming.
    DPU = D/U
    Total op (top) = U*OP
    DPO = D/top
    DPMO = DPO*1,000,000
    Hope that helps.
     

    0
  • Author
    Posts
Viewing 7 posts - 1 through 7 (of 7 total)

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