iSixSigma

What is important LSL USL or LCL UCL

Six Sigma – iSixSigma Forums Old Forums General What is important LSL USL or LCL UCL

Viewing 11 posts - 1 through 11 (of 11 total)
  • Author
    Posts
  • #45818

    torres
    Participant

    I am working in an industry where accuracy of output required is 100%. It means my LSL & USL is set at zero, but if looking at my Control chart, I found out that my monthly mean is 2.51:My process is in control, because most of the point lies in UCL & LCL, but it is not capable.What should I do??

    0
    #150558

    Charles H.
    Participant

    If you are in control and stable (no out of control / special causes of variation are present), you must focus on the common causes of variation.  You have three choices in your improvement approach: shift the mean (usually the easiest to do), reduce the standard deviation, or do both at the same time (a tough one to pull off, but it can be done).
    You did state that most, not all, of the points are in control.  I would first look for assignable causes for those points and get rid of those special causes.  Then, I would reassess for stability and control.  Your control chart should be in control and stable before going after the common causes of variation.
    Best wishes.

    0
    #150562

    torres
    Participant

    It is only in one case that point is out of Control Limit, but everyday it is out of Specification Limits.I have a simple query, if the process is not in a specification limit, then what is the use even if it is in Control limits.Should we ask client to have a waiver in specification limit, its current expectation is 100% accuracyJulia
    New York

    0
    #150563

    The Force
    Member

    If you’re process is in control but not capable — focus on the common causes to ensure meeting your required ppm.
    Be sure that you consider different tests for your ucl/lcl to ensure that special causes are given due attention

    0
    #150564

    Charles H.
    Participant

    Julia,
    I would still look for an assignable cause for the out of control point.  It is telling you that, for that point, something from outside of the normal process (a source of non-random variation) has entered into the process).
    As for the use of control limits (control charts), they are useful for telling you “the voice of the process.”  When in control and stable, it is telling you the performance of the process, unless something changes.  It is a predictor of future performance of the process, based upon past performance.  Unless you change the process or some external (special) cause impacts it, this is how your process will naturally function through time.  If it is not capable, you must improve it by attacking the common causes of variation within the process. 
    Personally, I would be careful about approaching your customer and asking for a deviation, though that is one option.  First, the question I would ask yourself and your team is, can you get the process to meet the customer specs without going broke in the effort?  Are the specs realistic or just the wishful thinking that someone bought into to get the customer’s business?

    0
    #150567

    Charles H.
    Participant

    Just a thought.  if you would like me to take a look at your data, email it me at:  [email protected]

    0
    #150568

    torres
    Participant

    This is how my control chart look like. My Specification is 100% accuracy
    Mean:- 2.5
    UCL:- 3.91
    LCL:- 0.9
    Once the control point has reached 4.2, & once it has reached 0, It has reached 0 when there was no single error.Now, tell me is there any point in conrolling the process, which is not meeting the specification.Simple logic:- If tomorrow I say my client that my process is 100% in contol, but it is out of specification, will he enjoy that?Is 100% accuracy possible at first place??

    0
    #150571

    Charles H.
    Participant

    Julia,
    All your customer cares about is your ability to meet their spec. and the level of service that you deliver to them  The control chart is for your use in monitoring and improving the process. 
    How could I know if 100% accuracy is possible?  Theoretically, it is.  Obviously, without improving the process, it is will never be possible from what you have told us.  You need to improve it.  The control chart will tell you if you are indeed improving the process, standing still, or making it worse. 
    You seem to be looking for a way out of doing the control chart.  That’s your choice to make. I recommend that you become more familiar with control charting by studying the subject before jumping on that bandwagon.  Sounds like you have your work cut out for you.
    Without knowing more about your process and the data, this is about as far as we can go on this one.

    0
    #150573

    torres
    Participant

    Charles,Did u get my email. If in case you are not able to help me, then atleast please answer my this question.I have joined this organisation just a week back & they have given me this control chart,which I have emailed you.At least let me know, how did they calculated LCL & UCL

    0
    #150606

    Adam
    Participant

    The benefit of having a process that is in control means that it is consistent and predictable.  Being that it is like that, you just have to figure out how to configure your process to get the results your customer is looking for.
    However, you are couting the number of defects, therefore, you may not really have a process that is in control.  Are they the same defects?  Do you have continuous data output that you can monitor?  Or is it just pass/fail?  You need to find out the root cause of the failures and attack those.

    0
    #150608

    Ashman
    Member

    You sound as though you have been reading Bill Smith’s paper “Making War on Defects”, IEEE Spectrum, Sept 1993, that formed the basis of Six Sigma.  Bill Smith recommended that the way to improve quality was to change specification limits (and that this was equally important as reducing variation).
    One day people will wake up that Bill Smith’s ideas and six sigma is rubbish.

    0
Viewing 11 posts - 1 through 11 (of 11 total)

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