iSixSigma

Batch Frequencies for SPC

Six Sigma – iSixSigma Forums General Forums General Batch Frequencies for SPC

This topic contains 5 replies, has 4 voices, and was last updated by  Seamus Maguire 1 year, 8 months ago.

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
    Posts
  • #55925

    Timma
    Participant

    Hello,

    I was wondering if the powers that be might have any insight into situations where it does not make sense to perform SPC. In the pharmaceutical industry, for example, it can be common to have a minimal amount of batches produced per year for certain products (1-2). Does it even make sense to trend?

    Reading up on some of the other postings here, I can confirm that trying to improve a process (other than retraining operators, fixing instruments, etc) can take an act of Congress since we are governed by regulatory filings which tend not to change very much.

    The other problem is the dictate from management that “thou shall trend” even though getting enough data to do so might entail a data pull well into the stone age.

    Any thoughts and advice are welcomed in advance.

    Regards,

    Tim

    0
    #202176

    Chris Seider
    Participant

    Don’t forget control of X’s is just as important if you were thinking of the Y’s.

    Things such as center lining, poka yokes, controlled recipes, etc. can help with zero defect environment.

    0
    #202180

    Brunehilde Carniel-Perrin
    Participant

    I would say it really depends on the batch size. If you produce 10 products, twice a year, then I agree that trending does not make sense. However if your batch sizes are big then it would make sense.

    As long as the inspection is done in process (instead of when the batch is finished), and as long as your measurements provide a quantitative result (not just a pass or fail), then you can plot 1 in every X parts, and observe the trend: if there is a trend (values slowly increasing for example), or a point outside the statistical limits (even if it is within your customer specs) then there is something impacting your process. By plotting you can see in advance that there is a problem creeping up and solve it before your parts get over the customer limit and get rejected.

    So long answer short: if you have 1) large quantities 2) quantitative measurements 3) in process inspection then it is definitely worth doing.

    Otherwise if you have small batches I have seen an adaptation of the SPC concept where instead of using statistical limits, the customer limits were used and a “danger zone” was defined. If the parts were in the danger zone, the process was tweaked until the parts were back near the middle limit. Not ideal and not SPC, but better than nothing.
    If any of that is unclear or if you would like more information I am happy to help, I realise I am being very succinct here.

    0
    #202181

    Timma
    Participant

    Thanks for the reply @cseider and @brunehilde. Unfortunately, and this is where the rub is, we receive the samples after production. Our batch size is large enough and they are quantitative measurements but to try and fix the process is not in the cards since the product is manufactured offsite. Perhaps I could appease them with some sort of SPC light where one doesn’t have to respond to out of control signals since there is nothing really one can do about it until the next lot which might be the next year.

    I’ve been trying to convince management that trending just doesn’t work in some cases but it is falling on deaf ears.

    0
    #202182

    Brunehilde Carniel-Perrin
    Participant

    @timma, when you say off site do you mean a subcontractor or your own company just a different site?

    If they ask for trending without understanding the background you might actually be dealing with people who don’t understand what they REALLY want… Maybe the first step if to understand what THEY mean by trending and what is the outcome they want out of it. You might realise that what they want is just proof that available data is being analysed, and for a lack of another word call it trending.

    Maybe as a trade off you could make a Pareto of the failure types (if there are any) and ask the supplier/other site to focus on reducing the top reason? It’s not “trending” as such, but it might satisfy them.

    In any case good luck! :)

    0
    #202194

    Seamus Maguire
    Participant

    Hi,

    I work in the Medical Device industry and agree to some extent. Making process changes can be near impossible once they are validated and have undergone regulatory filing.

    But improvement is not!

    Focusing on waste between steps (no process changes) is where you will get a huge improvement. Be unrelenting in reducing the 7 wastes and applying TOC. You can yield amazing results in even the most restrictive of industries.

    Best regards,

    Séamus

    0
Viewing 6 posts - 1 through 6 (of 6 total)

You must be logged in to reply to this topic.