iSixSigma

High Cpk and Control Charts

Six Sigma – iSixSigma Forums Old Forums General High Cpk and Control Charts

Viewing 13 posts - 1 through 13 (of 13 total)
  • Author
    Posts
  • #67865

    Neil Polhemus
    Participant

    I don’t think you’d ever want to stop collecting data and plotting it on a routine basis. You can always learn from looking at your data. You’d also like the ability to detect changes before they cause a problem, and that’s what control charts are designed for.You might consider modifying the control limits, however, to allow for the wide specifications. Doug Montgomery’s book has a good chapter on charts such as Acceptance Control Charts, which relax the control limits for high Cpk processes.

    0
    #27605

    MSAFAI
    Participant

    Hi everybody,
    I processes, where we have high Cpk’s (like 2 or more), is it justified to stop using control charts (like X-bar and R) and instead, calculate Cpk every month or so ?
    Thanks on beforehand

    0
    #67876

    MSAFAI
    Participant

    Neil,
    Thanks for the comment. You are absolutely right for saying control charts help in early identification of trouble.
    Modified control limits is  also a good idea.
    Thanks
    MSAFAI
     

    0
    #67881

    Ken Myers
    Participant

    Acceptance Control charts allow the user to make decisions on both the process and product using the same data. Process decisions are made using the control limits, and product disposition decisions can be made using the acceptance control limits.The acceptance control limits are determined by selecting “k” standard deviations in from the upper and lower spec limit. The value of “k” can be determined from the desired AQL and LTPD along with established subgroup sample size.

    0
    #67884

    Hartshorn
    Participant

    Having that continuous data history would also come in handy if the customer calls with a concern.  Several times we have been able to prove that it is not our product that is causing an issue at the customer using this data.

    0
    #67885

    Savage
    Participant

    One additional thought, to those already posted, is that for Cpk to be a “valid” statistic, you must be able to show the process is in control. If you do not continue to use a control chart, how will you know if it is in control?

    0
    #67890

    picassobox
    Participant

    While it may be safe to assume your process will not flirt with it control limits any time soon, a control chart can still detect special causes of variation beyond the errant data point beyond the control limit.  Control charts help study variation between material lots and work shifts.  They allow us to monitor process improvements, and ensure upstream changes haven’t impacted the operation. 
    Running monthly process capabilities, for the sake of ensuring last month’s operations were within control is a dangerous trap, and one should be wise not to end measurement simply because it appears as though one form of special cause variation will never be detected

    0
    #67900

    MSAFAI
    Participant

    Very good points raised by all the contributors. I thank everybody on my part.
    Just may I explain on the point raised about “using past data to calculate Cpk,  …  and  making sure the process is under control before Cpk calculation …”
    What I meant was: each month or so you collect data (25 subgroups or so), establish control using control chart, only THEN you calculate Cpk.
    Thanks again
    MSAFAI

    0
    #67907

    “Ken”
    Participant

    Guy,Good point. Having time series inspection data organized on a control chart goes a long way at convincing a customer that your process was working fine when making product for them.I might also suggest that if one is performing line inspection using attribute sampling plans that a plot of the p or np chart could also point out whether the process was producing on average to the AQL level in a stable fashion.Ken

    0
    #67908

    kamlesh shah
    Participant

    I would ask customer and show them cpk,  they might reduce inspection on products.

    0
    #67914

    MSAFAI
    Participant

    Kamlesh,
    Thanks for the hint.
    Regards
    MSAFAI

    0
    #67918

    Eoin Barry
    Participant

    Yes – eliminate SPC – it is a useful tool but ultimately is associated with your cost of poor quality – Cpk of 2 is six sigma – i.e. is no longer a poor quality process. If your process model fits well e.g. is strongly normal and you have long term data on which to base your CpK calculation (e.g. 400+ points) then yes, by all means eliminate SPC on your output. This is one of the sweet “fruits” of six sigma.
    It is a leap of faith, faith in your measurement system and process controls and data collection, I would urge you consider it.
     SPC is still a valuable control tool – however at this point the real benefit is in controlling your key input variables rather than the output variable. Generally it is easier to do this e.g. automated control devices such as temperature sensors which alarm etc., Poke Yoke source inspection techniques are very valuable at this stage.
    Well done on a Cpk of 2 and best wishes,
    Eoin

    0
    #67920

    Sorour
    Participant

    I would just like to say that I share your view completely.
    I would now be looking to move what are sometimes limited resources for monitoring SPC to part of the process that is not yet six sigma.

    0
Viewing 13 posts - 1 through 13 (of 13 total)

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