iSixSigma

Sample Size Required for Batch Testing

Six Sigma – iSixSigma Forums General Forums Methodology Sample Size Required for Batch Testing

Viewing 5 posts - 1 through 5 (of 5 total)
  • Author
    Posts
  • #53826

    Buckle
    Participant

    Morning all,

    Looking for some advice on sample size required for testing a batch. We manafacture 10,000 items and have a spec limit of no more than 1% of items within a batch can be over 5.3mm

    How would i work out the sample siz e required for testing of each batch?

    Any help would be appreciated.

    Thanks
    Mike

    0
    #191578

    Robert Butler
    Participant

    You will have to provide more information before anyone can offer much in the way of suggestions.

    1. How big is a batch?

    2. 10,000 items makes it sound like you have a large number of very short runs (perhaps as few as one run a year) is this the case?

    3. If #2 is true then what are the statistics on the runs in terms of items per run – minimum, maximum, mode, median, standard deviation etc.?

    4. Again, if #2 is true are what does a pareto chart of item importance tell you with respect to item ranking (vital few – trivial many).

    There are other questions that need to be asked and answered but if you can provide the answer to the above perhaps I or someone else could offer some thoughts.

    P.S. Did you try the .135, 99.865 approach to your earlier problem? If so, how did the result compare to whatever it was that you chose to do?

    0
    #191579

    Buckle
    Participant

    Hi,

    A batch is 10,000. We actually run every week on this product (its tabletting). The issue i have is that to test each tablet takes 4 minutes and is a destructive test on a very expensive product. So obviuosly the less tests the better…… Do i still need to provide other detail? as it isnt short runs?
    Thanks

    Ref the earlier problem i managed to sort this, i just used the function in minitab to identify which distribution it was, i then re-ran the tests and got the PPK values. i had a few special causes (data errors) which were skewing the data too!

    0
    #191580

    Buckle
    Participant

    I think im on the righ track using Minitabs acceptance sampling plan by attributes…..

    For AQL i put in my specification limit ie 1% but for RQL im stuck? I have read that i should just put it down as 5% (CI @ 95%). Is this correct?

    Im happy with the producers risk and consumers risk areas but just confused with what to include for the RQL (LTPD)

    Can anyone please help??

    Thanks
    Mike

    0
    #194991

    I have a batch of 4000 items that I want to be 99% confident have no defects (defect is either yes or no for a visual defect). How many of the 4000 items would I randomly inspect to get 99% ?

    0
Viewing 5 posts - 1 through 5 (of 5 total)

You must be logged in to reply to this topic.