iSixSigma

First Pass Yield

Six Sigma – iSixSigma Forums Old Forums General First Pass Yield

Viewing 10 posts - 1 through 10 (of 10 total)
  • Author
    Posts
  • #36522

    Tanya
    Member

    The company, I joined recently constructs in process first pass yield charts (Xm,R) using the following method: 
    First the overall average from all in process first pass yield testing is calculated.  For example, if 1 product took 2 tests to show the first time passed, then the average is taken as 50% for that product. Then, average from all products for the week are taken to get the data point for that week.
    What I’ve studied is absolutley different,
    Yield = #of units that pass on first test(without rework)/Total # of units tested
    Q: What is the correct method?

    0
    #105614

    A.S.
    Participant

    Xm R chart is moving average range chart and it can be used for anything not only for first pass yield.
    Both methods are correct.Usage is depends on your requirement as detailed below:
    1.If you want to assess the process variation impact on a particular product or if you want to compare the different prodcuts,the method followed by your comapany (% for individual product) is ok.
    2.If you want to assess the overall capability of your design and process function and to study the trend over the period,the method you studied is applicable.
     
     

    0
    #105777

    gutierrez
    Participant

    Why I can not use Xm R chart to assess First Pass Yield?
     
    Eduardo

    0
    #105781

    Len
    Participant

    Xm-Rchart is for variable data only,In your case , i suggest you to use P,Pn,chart. This my poor opinion , hope to help you.
     

    0
    #105789

    gutierrez
    Participant

    But if i had to multiply two or more numbers to know the Rolled First Pass Yield, I can not use a P,Pn chart because I have a variable quantity of parts in the differents steps of the process. Because of this, I use the Xm R chart.
     
    Eduardo

    0
    #105823

    vpschroeder
    Member

    we do a combination
    for each visual process, we use your method
    for each machine/functional test process we use # passed/total # tested.
    in some cases, a series of tests may be considered as a single ‘test’ point, in which case if any of them fail, we consider that 0%.  (in your case, it seems, you average the result.  in our field, that doesnt fly)
    for the week, we roll all products by each station and report a single data point

    0
    #173658

    Viren Rami
    Member

    I agree what Tanya has suggested. ie.
    The First Pass Rate Formula should be : Yield = #of units that pass on first test (without rework) / Total # of units tested

    0
    #180346

    Rosa Nunez
    Member

    If I have a unit that did not pass the test because our supplier did somenthing wrong on the PWA this failure on the first pass yield can not count againt my production. I mean Can I still documenting this as a failure?

    0
    #180353

    Anonymous
    Guest

    Dear Rosa,It’s my anniversary today – so you’ll have to forgive me for being a little worse for wear ..Many years ago MOS 8 produced the best Lot of wafers ever seen, but they bombed at final test.Later, we found out they were so fast the test area couldn’t synchronize them ..Unfortunately, we couldn’t change the process because the first time yield is always right!Just kidding …

    0
    #180356

    DLW
    Participant

    Hi, Rosa.
    Yes, of course it is a “failure” in the sense that it is not acceptable
    within your system, or for your customers. Remember, they do not
    care WHY it was not right.it is not about keeping score (although your post sounds like that
    is what you are dealing with). It should be about identifying
    problem areas, their causes, and opportunities to improve. Just
    because the problem source is outside your walls does not mean
    there is nothing you can do to make it better.DLW – BPEX

    0
Viewing 10 posts - 1 through 10 (of 10 total)

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