iSixSigma

Capability Rollup

Six Sigma – iSixSigma Forums Old Forums General Capability Rollup

Viewing 12 posts - 1 through 12 (of 12 total)
  • Author
    Posts
  • #47666

    Savage
    Participant

    I was talking with a co-worker the other day and he mentioned that in a previous company they used a method of capability rollup. BAsically calculating the capability of subassemblies and them combining them to result in a rolled up, composite capability metric. Kind of like RTY, but using CpK instead. My friend didn’t have much detail about the method. Has anyone done this? Is this a valid method? Any resources for me to research? Just curious – thanks.Matt

    0
    #159148

    M. Salim MSA
    Participant

    I have done that using Z scores to rollup scores from parts, process, product, software etc into one score for the company ! This is done using scorecards for Z-values and rolled up into one final score ….

    0
    #159149

    Chad Taylor
    Participant

    Why would you want to? Whats the value behind it?
     

    0
    #159151

    Savage
    Participant

    That’s what I’m wondering.  I can’t imagine a reason for doing this.  I am trying to understand if there is some statistical justification for doing this.

    0
    #159152

    Chad Taylor
    Participant

    Matt, there is known. Its stupid

    0
    #159153

    M. Salim MSA
    Participant

    Stupid for some ….. overall six sigma level for others !

    0
    #159154

    M. Salim MSA
    Participant

    Key is to have the understanding of a capability score rollup !

    0
    #159164

    Chad Taylor
    Participant

    OK, then explain to me how 3 sigma * 3 sigma = 4.1 sigma?????????
    Then Explain Why this useful. I have a complete and open mind.

    0
    #159223

    Savage
    Participant

    I don’t understand, that’s why I asked.  If you do, please share your insight.  Thanks.

    0
    #159225

    M. Salim MSA
    Participant

    I can send you some examples if that will help ….. I was referring to the other gentleman when I mentioned “understanding” ! some people are curious which is good …. but some people don’t have the understanding nor do they have the ability to understand !

    0
    #159226

    Jim Ace
    Participant

    Matt:
     
    This is how I learned to do this task.  From my example you will see why you can not just average the Cpk values or Z scores
     
     
    Step 1: Find the yield of X independent items.
    Example: 5 indpendent items were found to have the following yields;
    X1=90%, X2=95%, X3=99%, X4=84% and X5=73%.  The 5 items could be the number of steps in a process or the parts in a subassembly.
     
    Step 2: Multiply the 5 independent yields.
    Example: .90 x .95 x .99 x .84 x .73 = 0.51904314.  This is the rolled throughput yield
     
    Step 3: Find the average yield
    Example: 0.519043^(1/5) = 0.877082995.  Notice that 0.51904314 = (0.877082995)^5, thus confirming that the average yield is NOT the mean of the item yields.
     
    Step 4: Convert average yield to a Z score
    Example: Excel equation =normsinv(0.877083) = 1.160527733
     
    Step 5: Add 1.5 sigma to estimate the average short term capability
    Example: 1.16 + 1.50 = 2.66.  This is your overall Sigma score for the 5 independent items.  Notice that the overall Sigma of 2.66 is NOT the mean of the subordinate sigma scores.
     
    Summary:
     
    Item                  Yield            Sigma (Z)
    X1                      0.900                2.78
    X2                      0.950                3.14
    X3                      0.990                3.83
    X4                      0.840                2.49
    X5                      0.730                2.11
    Overall              0.877                2.66
     
    Jim Ace

    0
    #159264

    Savage
    Participant

    Thanks Jim.  Just before I read your post, I found the same method in one of my reference books.  Thanks again.

    0
Viewing 12 posts - 1 through 12 (of 12 total)

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