iSixSigma

A Few questions on DPMO

Six Sigma – iSixSigma Forums Old Forums General A Few questions on DPMO

Viewing 28 posts - 1 through 28 (of 28 total)
  • Author
    Posts
  • #39906

    JP
    Participant

    1. Which would be more reasonable definition for a defecta)Product ends up returned in warrantyb)Product doesn’t pass our tests
    At first I was told to use b), but then it wouldn’t make any sence as using an extremely slacky testing system would give our products a superior DPMO. The problem with a) is that ppl in marketing tend not to report all the warranty returns.
    If a), should I exclude errors that are sure to get caught on testing? Again, I can’t see how a good testing system could lower the quality.
    2. Should components coming from our subcontractors be concidered as a potential error for us?
    3. Which of these opportunity groups should be grouped as one or excluded? (Never mind if they are CTQ or not)
    -attaching the screws-wrong/missing screws-attaching the connectors-wrong/missing component-attaching the component-attaching plates etc.-placing the wires-spreading of thermal paste-checking the quality of soldering done by robots-general checking (I suppose this has to be excluded as it doesn’t do any increase in value)-attaching plastic parts-errors in software loading-wrong/missing wires-attaching labels-waterproofing-soldering

    0
    #122556

    Catbyou
    Participant

    It is my belief all defects which occur during the process should be reported.  The purpose is not to get a good score rather to make sure the customers are delighted with the products.  Trying to find ways to make the score look better internally does not pay off in the end.

    0
    #122558

    JP
    Participant

    That just doesn’t sound right, because DPMO would be the same or even lower without any testing and it’s pretty obvious that proper testing improves the quality on products sent to end users.

    0
    #122559

    Tierradentro
    Participant

    JP,
    ‘Proper testing improves the quality’ ??
    are you trying to inspect quality into a product? surely getting it right first time reduces the risk of non conforming product being despatched to the customer?
    for what its worth, the first thing i suspect you need to do is contain defective product, thus reducing warranty claims.Then concentrate on where the defect is happening and try to resolve the issue through whatever means.
    regards
    John

    0
    #122560

    JP
    Participant

    “are you trying to inspect quality into a product? surely getting it right first time reduces the risk of non conforming product being despatched to the customer?”
    Sure, but so does filtering out the defected ones before sending them out. Because of that, I really can’t see how a strict testing system could possibly worsen our quality, as it would seem like that, if we included the defects caught on testing.
    “Then concentrate on where the defect is happening and try to resolve the issue through whatever means.”
    I’m doing the DPMO analysis mainly for ‘benchmarking’ and for enabling comparison between two different assembly lines. We might use the results for straight quality improvements, but it’s going to be someone else’s problem, as I’m just a summer intern.
    So at this point it’s more important to determine the DPMO as kosher as possible than getting straight implications to quality improvement.

    0
    #122562

    Tierradentro
    Participant

    JP,
    A strict testing system (or a more robust containment system) will surely give you less field failures (warranty claims)
    If you are not happy with the DPMO system, why not try rolled throughput yield. This can be done quickly and effectively, although you may need to get extra resources to assit while conducting on the assembly line (s)
    The chances are that if you have a ‘slack’ testing system, the % risk to the customer obtaining defects rises. Once you have done the RTY on each of the 2 assembly lines, you can then compare the data, line V line, process step V process step etc.
    John

    0
    #122563

    JP
    Participant

    “A strict testing system (or a more robust containment system) will surely give you less field failures (warranty claims)”
    Yes, but a looser one would let borderline cases through and some of those would not end up in warranty returns. Thus there would be less defects than in strict testing, although we both agree that the overall field quality is better when the product is tested strictly.
    “If you are not happy with the DPMO system”
    I’m not saying i’m not happy with it, just a little confused about the definitions.
    “why not try rolled throughput yield. This can be done quickly and effectively, although you may need to get extra resources to assit while conducting on the assembly line (s)”
    Well, I’d have to ask my supervisor on that, but if I understanded it at all, it doesn’t seem seem very suitable for this purpose. Both of the assembly lines are essentially quite similar having only one testing stage. Since the products on the other line are bigger and more complex, they will have smaller RTY independent of how well the line is actually performing.
    In addition, it doesn’t seem like a very good way for benchmarking.
     

    0
    #122564

    Mikel
    Member

    JP,
    This is simple –
    A

    0
    #122565

    Mikel
    Member

    JP,
    This is simple –
    A defect is any use of resource greater than the minimm necessary to create a product or service correctly.
    An opportunity is a value added operation necessary to create a product or service correctly.
    Tests and inspections are economic decisions. Do them to minimize costs.

    0
    #122566

    Tierradentro
    Participant

    DP,
    “In addition, it doesn’t seem like a very good way for benchmarking”
     is what you referred to RTY. I suggest you go & acquaint yourself to RTY & to then try it. It is very similar to DPMO, yet you still seem to be quite vague in how you are carrying out your DPMO exercise? are you looking at defects per process step, or just defects that pass through the one testing station? (that is why you may be confused about the definition)
    john

    0
    #122569

    JP
    Participant

    Again, I’m trying to determine the quality of our product rather than minimizing the costs.
    “An opportunity is a value added operation necessary to create a product or service correctly.”
    Define ‘value added operation’. Eg.when installing a module with 4 screws and spreading thermal paste under it are there
    a) 4 (screws) + 1 (module) + 1 (thermal paste) b) 1 (module) + 1 (thermal paste)c) 1 (module)or evend) 4 + 1000 (from each value added step in the module) + 1value added operations?

    0
    #122571

    Mikel
    Member

    a)

    0
    #122573

    JP
    Participant

    I’ve understood that only screws in electric connections would count as an opportunity. At least none of the case examples I’ve seen mentioned anything about non-electric screws, although they were very likely present.

    0
    #122574

    Ken Feldman
    Participant

    Stan, you forgot; fivvvvvvve golden rings, four calling birds, three French hens, two turtle doves and a partridge in a pear tree. 

    0
    #122576

    Mikel
    Member

    You have understood incorrectly.

    0
    #122578

    Mikel
    Member

    My, my – aren’t we in a feisty mood this morning?
    Too much cheap tequila last night?

    0
    #122581

    JP
    Participant

    Based entirely on the info here: https://www.isixsigma.com/dictionary/Rolled_Throughput_Yield_-_RTY-168.htm I’ve gotten the picture, that RTY would give better results for less complex processes, and so wouldn’t be good basis of comparision in neither between our two assembly lines nor between other manufacturers with less or more complex products. Is this right?
    “are you looking at defects per process step, or just defects that pass through the one testing station? (that is why you may be confused about the definition)”
    Generally defects for the whole process. The problem is should I use faults in testing, faults in end user products or both of them. The goal is to calculate a sigma level for the process that would be comparable with other processes and the six sigma limit.
     

    0
    #122582

    Tierradentro
    Participant

    JP,
    Ok, an even more simple approach, use both the proportional data of failure internally (P chart or NP if the amount produced is the same everyday) & also a P chart for ‘proportion of warranty claims’ group these together & put them into a sigma calculator (there is one on this site i believe) that will give you a Sigma level for both these Y’s. To understand DPMO you need to know defects per step (rework, scrap etc etc) of the process, so only using the one testing station (i presume end of line) will only tell you what the defect rate is when the assembly ‘hits’ the final point in the process (not all the hidden reworks alog the way until the assembly gets to the testing/end of line)
     
    John

    0
    #122583

    JP
    Participant

    Do you actually know or are you just throwing quesses?

    0
    #122585

    Tierradentro
    Participant

    JP,
    “Generally defects for the whole process. The problem is should I use faults in testing, faults in end user products or both of them. The goal is to calculate a sigma level for the process that would be comparable with other processes and the six sigma limit”
    “Generally”? are you looking at DPMO or total proportion defects, come on. The reason that i am guessing (not quessing) is that you are being vague. if you want to look at DPMO (or RTY) you need to look at each process step defects/assembled to get a sigma value, orfor the whole process use P (or NP) it is hardly rocket science (or is it?)
    I would still look back at your first posting & use (A)
    John

    0
    #122588

    Mikel
    Member

    Yes, I actually know

    0
    #122590

    Tierradentro
    Participant

    hnnn?
    agreeing – you?
     
     
    John

    0
    #122599

    Alvarez
    Participant

    You have no idea sensei stan
    ha, ha, ha.
     

    0
    #122603

    UFO
    Member

    A defect opportunity must have a “chance” to be defective.
    If a defect opportunity has no chance of being defective, it
    should not be factored into the DPMO metric. This means
    that a CTQ must have a defect probability greater than
    zero if it is to be “counted” in the denominator of the
    DPMO metric. If the defect probability is zero it is NOT a
    “defect” opportunity. It may be some other kind of
    opportunity, but it is certainly not a “defect opportunity.”

    0
    #122816

    Ken Feldman
    Participant

    Yes Stan, you know me too well.

    0
    #122817

    Dr. Scott
    Participant

    Call me!

    0
    #122818

    Ken Feldman
    Participant

    Scotty, what do you want me to call you??? Stan and Dufo seem to have captured all the really good verbal slams the past few days.  Too bad I was out of pocket and missed all the fun realtime.

    0
    #122820

    Dr. Scott
    Participant

    Scotty, Dr. Scott, or Scooterbug (as my boat friends call me) will work just fine.
    Dr. Scott

    0
Viewing 28 posts - 1 through 28 (of 28 total)

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