iSixSigma

Minitab capability when spec limits are indexed

Six Sigma – iSixSigma Forums Old Forums Europe Minitab capability when spec limits are indexed

This topic contains 3 replies, has 4 voices, and was last updated by  Adam Bowden 14 years, 7 months ago.

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
    Posts
  • #23582

    Bill Kraus
    Participant

    I am looking at test data across a broad current range (5A – 100A). I am mometarily stumped for how to assess capablity across that range because the spec. limits are a sliding scale of +/- 5 percent (i.e. at 5A the limits are 4.75-5.25, but at 75A they are 71.25-78.75). The Mintab capability section seems to want fixed spec. limits to assess capability. Is there a slick solution that will allow me to look at the whole set of data without breaking it into buckets with fixed spec. limits? Thanks for the help.
    Bill

    0
    #56274

    Arne Buthmann
    Participant

    Hey Bill!
    Do you use Minitab 13 or 14? If you are working with Minitab 14 (you can get a free, 30-day demo of MINITAB 14 on http://www.minitab.com) use the capability analysis for multiple variables (Stat>Quality Tools>Capability Analysis>Multiple Variables).
    This allows you to enter different spec limits for each group being analyzed. You don’t have to enter these upper and lower spec limits manually, but you can get them from 2 seperate worksheet columns (one for LSL, one for USL). These columns can be quickly generated by
    – inserting an additional column which contains (in your case) the values 5-100, and then
    – calculating the LSL column and the USL column, resp., as +/- 5 % of the additionally inserted column (with the help of the Minitab Calculator).
    Hope that works! For any further questions don’t hesitate to ask me!
    Arne
     

    0
    #56459

    six sigma Guru
    Member

    1)Subtract the Targets of 5A and 75A from the measurements ofthe two sets
    2)Plot the control chart for individuals. Ensure that the process is stable.
    3)Plot the histogram and normal probability graph and ensure normality.
    4)If process is stable and Normal, then you can calculate CP and CPk using Xbar,s and the specifications of +/- 10%.
    For some unknown reason, most of the six-sigma practioners forget steps 2 & 3 and calculate meaningless Cp and CPks

    0
    #56460

    Adam Bowden
    Participant

    Hi   Mr Guru,
    Most people don’t follow the steps because they don’t appreciate “normality” or do now wish, or know how to, transfrom the data.  Another hypothesis is that is you collect data to infinity it becomes normal any way.
    You do have a Good Point that I think most BBs need to understand and practice.
    Best regards,
    Adam
     
     

    0
Viewing 4 posts - 1 through 4 (of 4 total)

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