iSixSigma

Including rework processes in yield calculations

Six Sigma – iSixSigma Forums Old Forums General Including rework processes in yield calculations

Viewing 8 posts - 1 through 8 (of 8 total)
  • Author
    Posts
  • #34819

    mjtoday
    Participant

    This is not a question about the difference between FTY and RPY.  When I am calculating yield for a manufacturing multi-step process, with well-defined rework processes, do I multiply the yield of these rework processes into the calculation for yield of the manufacturing line?
    In other words, assume that a nominal process includes “Process A”, “Process B”, “Process C”, and “Process D”.  I also have a “Rework Process E” that is sometimes entered from one of the main processes.  Which way should I calculate my FTY or RTY:
    1) Yield = (A-Yield) x (B-Yield) x (C-Yield) x (D-Yield)
    2) Yield = (A-Yield) x (B-Yield) x (C-Yield) x (D-Yield) x (Rwk-E Yield)
    It seems that ignoring the rework process yield could result in misleading overall yield numbers if the success rate of that rework < 100%.
    Thanks.

    0
    #96424

    Mikel
    Member

    Your assumptions are correct and it is #2.
    Furthermore, you can look at the yield of the rework and see if its predictable and a quick gain from improving poor analysis tools.

    0
    #96429

    Mu Joe
    Participant

    One of the purposes of six sigma and Lean is to eliminate re-work entirely by eliminating the defects that make the rework necessary. The only way to get a true picture of your first time throughput efficiency is to do the RTY calculation without the rework step. Doing the calculation doesn’t mean that you stop doing the actual rework in the short-run, it just means you don’t count the output of that step when calculating RTY.
    Make sense?

    0
    #96434

    Mikel
    Member

    Makes non-sense.
    Count all defects.
    Period.

    0
    #96441

    Mu Joe
    Participant

    Wait a minute.
    Your goal is to eliminate defects in all steps of your process and thereby eliminate even the need for a re-work step. In determining your initial RTY (the true performance of your process) you want to know where your defects are originating the first time, not what additioanl defects might be coming out of our rework step.
    Yes, of course you do want to know if your rework step is creating defects and you want to correct that, but only as an interim step. Your ultimate goal should be to completely eliminate the rework step by eliminating defects that occur earlier in the process.
    Rework is waste, find a way to eliminate the rework step.

    0
    #96442

    jediblackbelt
    Participant

    The question should be is rework a part of your normal process?  If for some off the wall reason it is then count your defects.  Rework is a NVA operation and should be avoided.  Adding additional rework defects to the normal process defects will skew the data and does not make sense with regards to lean.  Eliminate the rework is the goal and by reducing the parts going to defect is the way to do that.  Counting the efficiency of rework seems like an oxymoron.  Would your customer want to know how efficient your rework operation is or would they rather know how many parts come in to rework and what the plans are to stop this NVA operation.  If parts coming from rework are failing then your DPMO will raise accordingly anyway.  You should want to know if any parts reintroduced into the line are coming from rework because if they are then you are obviously having problems with rework to begin with.

    0
    #96443

    mjtoday
    Participant

    My main concern with approach #2 is that while it does give a quantitative value that also reflects unsuccessful rework processes, the  resulting yield number loses some of its physical meaning (because not every part that starts the line actually goes through one of the processes included in the calculation).
    For example,
    Process A: 20 in / 20 outProcess B: 20 in / 20 outProcess C: 20 in / 20 outProcess D: 20 in / 14 out + 2 scrap + 4 reworkRework E: 4 in (from “D”) / 3 out + 1 scrap
    Calculating yield with each process treated individually:
    FTY = (1 x 1 x 1) x (14/16 **) x (3/4) = 0.66RTY = (1 x 1 x 1) x (14/20) x (3/4) = 0.53
    ** (I doubt it’s correct, but I’m treating rework parts as if they didn’t exist here)
    Regarding FTY, the actual FTY (% of starts that finished the line) is 17/20 = 0.85. By including the rework process in the multiplication product, we would get an artificially low yield.  It actually seems that FTY calculations only work if Rework E is “embedded” into Process D, effectively treating it as if it didn’t exist–just as the examples in the dictionary on this site.
    Similarly, the actual RTY (% of starts that went through the line without a defect) is really 14/20 = 0.7.  By including the rework process in the multiplication product, though, we can’t duplicate this value. 
    Is this basically a philosophical difference in whether to make rework yield loss highly visible by degrading our overall yield numbers with this loss, versus reporting an overall yield number that represents something with more physical meaning?
    Of course once our yields all hit 100%, there will be no more debate…
    Thanks again.

    0
    #96444

    mjtoday
    Participant

    Thanks for your input.  I agree with your statements that the desire is to try to eliminate reworks by fixing the primary process.  However, if it’s going to take significant time for an engineer to eliminate a rework, we still need to monitor the effectiveness of the reworks until we can eliminate them (to highlight if there’s a sudden spike in failure of a particular rework, etc…)
    It sounds like the best approach would be to include only the primary processes in the RTY calculation, and also monitor a separate report listing each rework and its yield, at least until they can be eliminated.
     
    On a side note, I can picture times when a rework is not done on 100% of parts, but won’t ever be eliminated–for example in a typical photolithography process done during semiconductor manufacture, an operator may take a measurement of a critical dimension in the photoresist before an irreversible process (etch, plating, etc…) is done to the part.  If an out-of-spec CD is measured, the resist can be stripped off and the part reprocessed after a minor tool adjustment.  Due to the nature of the process, I think these reworks (and their impact on RTY) can be minimized, but probably won’t ever be completely eliminated. 

    0
Viewing 8 posts - 1 through 8 (of 8 total)

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