iSixSigma

SPC charts and non-capable processes

Six Sigma – iSixSigma Forums Old Forums General SPC charts and non-capable processes

Viewing 12 posts - 1 through 12 (of 12 total)
  • Author
    Posts
  • #46076

    amhender
    Participant

    With a process not capable of meeting six sigma, why don’t you develop SPC charts?

    0
    #151804

    accrington
    Participant

    Indeed!

    0
    #151811

    Ola L
    Participant

    I agree. SPC is a good (and nescecary) tool to use both to control both robust and non bobust.
    One question that I have is how to determine sample intervalls. Anyone got any idea?
    (We use SPC to control a “halv-robust” process, but the question was regarding how frequent sample intervals we needed to ensure quality. To measure a lot of products to determine sample size is not something we would like to do).
     
     

    0
    #151815

    Furneaux
    Member

    “halv-robust” ?? !! 
    What school of six sigma idiocy did you go to ?

    0
    #151818

    Jim Shelor
    Participant

    Dear Amhender,
    Process that you wish to monitor to confirm the process is producing a consistent product should be monitored using SPC regardless of the process capability.  The purpose of SPC is to tell you whether or not your process is in control.  If your process is in control, you will be producing a consistent product.
    The data gathered from the SPC is then used to determine your process capability.  This tells you whether your process is capable of meeting the customer specifications and the out of tolerance rate you can expect from the process.
    Both of these techniques should be used to monitor processes that produce product for a customer (internal or external).  If you are not capable of meeting the customer’s requirements, you are not going to be in business very long.
    Best regards,
    Jim Shelor

    0
    #151819

    Jim Shelor
    Participant

    Dear Ola L,
    The frequency of sampling and the amount of each sample depend on what you want to know about your process and how long you are willing to wait for that answer.
    The process for determining the sample frequency and sample size is called “Rational Subgrouping” you can find references to rational subgrouping in any good Six Sigma text.
    Let’s talk about an example.
    Assume you have a process that
    1.  operates 24/7
    2.  has two input material streams
         a.  one that can greatly effect the quality of your product
         b. one that has little effect on the quality of your product
    3.  Is shutdown at the end of each shift for lubrication and the setup prior to startup can effect the quality of the product.
    Rational subgrouping would give you the following options to think about.
    1.  Size of sample –
         a.  if you are going to use an X-bar, R control chart, the normally accepted sample size is between 3 and 7 (most processes use 5)
         b.  if you are going to use an I, mR control chart, the sample size will be 1.
    2.  When to sample –
         a.  consider sampling at the start of each shift following startup of the process to check for consistency of setup.  If you know your operators are consistent and proficient in the setup process, you may want to forgo this sample.
         b.  If you do not sample at the start of each shift, consider how long you are willing to wait before you get indication that your process may not be in control.  For most processes of this type, samples are taken no less than once a day.
         c.  You have an input material delivered by batch that can greatly effect the quality of your product.  You should consider sampling after the initiation of the process using a new batch of this material.
    In this example, I would chose:
    1.  use an X-bar, R chart with a sample size of 5.  The reason I would chose this is that when I do the capability analysis, R-bar/d2 for an X-bar, R chart provides a better estimate of sigma(common cause) than does R-bar/d2 for an I, mR chart.
    2.  Sample at the start of each shift to check for possible variations in the setup process.  If after sampling for some time such that I am confident that the setup process is not vausing special cause variation in the process, reduct the sampling rate to once a day.
    3.  Sample each time a new batch of the input material that can greatly effect the quality of my output product to check for input material variation.  Further, if it is possible to sample the input material directly and compare it against known specification that produce a quality product, I would do that to keep fom making bad quality product before I find out the input material was bad.
    In the end, there are no hard, fast rules for determining how often or how much to sample.  The sampling method is determined by you and your management team based on risk of producing bad product, how long your are willing to take the risk before finding out, and how much sampling costs in both direct cost and lost production time (if any).
    I hope this helps.
    I recommend you look up “Rational Subgrouping” in a good Six Sigma text for the understanding of how this is done.
    Best regards,
    Jim Shelor
     

    0
    #151837

    Ola L
    Participant

    Thank You very much, Jim!
     

    0
    #151842

    Furneaux
    Member

    You obviously have no idea what rational subgrouping is.Forget crappy SS texts.  Read Wheeler, “Advanced Topics in SPC” for the facts.

    0
    #151847

    Jim Shelor
    Participant

    Dear Will,
    Now that you have gotten that off your chest, do you have anything constructive to add to this conversation?
    Best regards
    Jim Shelor

    0
    #151849

    Jim Shelor
    Participant

    Dear Ola L,
    Will is correct in that I should not have referred to the method I was describing as “Rational Subgrouping”.
    Rational Subgrouping refers to subgrouping samples in a manner that provides for the least opportunity of having assignable variation (special cause variation) within the sample subgroup (i.e., within the 5 consecutive samples you take, if you are using that method).
    I should have referred to the method I was describing as “Rational Sampling” wherein you decide which parameters to sample and at what frequency to provide the best analysis for the items of concern (i.e., quality effecting items).
    My apologies for the misinformation.
    Sincere regards,
    Jim Shelor

    0
    #151859

    mcintosh
    Participant

    If rational subgrouping is an advanced topic of SPC  heaven help us!!!!

    0
    #151872

    Jim Shelor
    Participant

    Dear Tom and Will,
    As I said in the post following this, I was incorrect in the use of the term Rational Subgrouping.  I should have used the term Rational Sampling.
    Failure to employ Rational Subgrouping (subgrouping that contains minimal, and hopefully no special cause variation) CAN LEAD TO control charts that are less sensitive, provide incorrect answers and cause incorrect decisions to be made.
    I was in error in the use of the term Rational Subgrouping when I should have said Rational Sampling.
    My apologies for the misinformation.
    Regards,
    Jim Shelor

    0
Viewing 12 posts - 1 through 12 (of 12 total)

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