iSixSigma

How to use the variable limits in U charts ?

Six Sigma – iSixSigma › Forums › Old Forums › General › How to use the variable limits in U charts ?

Viewing 8 posts - 1 through 8 (of 8 total)
  • Author
    Posts
  • June 4, 2001 at 4:00 am #27359

    Vishy
    ★ 20 Years ★
    Member

    Hi all :

    I’m applying u charts for deriving proces capability limits in a software environment using defect per line of code as the variable. Since the UCL and LCL are variable , what is my process capability limits at any point in time ? Is it the last UCL/LCL that the chart shows ?

    need some inputs on this…

    0
    June 5, 2001 at 4:00 am #66844

    Niraj Goyal
    ★ 10 Years ★
    Participant

    I am interested in the replies to this question too please.

    Vishy, I have applied control charts(not u) to IT enabled services – errors in punched data very successfully.

    Perhaps we can gain from sharing experiences.

    Niraj

    0
    June 5, 2001 at 4:00 am #66847

    Neil Polhemus
    ★ 20 Years ★
    Participant

    It is important not to confuse the concept of capability with that of control. The control limits do not provide your estimate of process capability. That comes directly from the mean DPU. The control limits simply allow you to determine whether the process has been stable over the sampling period.

    0
    June 5, 2001 at 4:00 am #66865

    Jim Parnella
    ★ 20 Years ★
    Participant

    Vishy,
    I think you misspoke about Process Capability limits and really meant CONTROL LIMITS. Since you are doing errors PER LINE OF CODE, you won’t have VARIABLE LIMITS on your control chart (because per line of code gives you a constant sample size). As far as process capability from an attribute chart like this, I’ll defer to Neil’s explanation. I’ve never seen an adequate explantion of process capability for attribute data so I’ll defer to him.
    Jim

    0
    June 7, 2001 at 4:00 am #66887

    Vishy
    ★ 20 Years ★
    Member

    Though mentioning the word ” capability ” when doing a control chart may sound wrong , in actual case , the control limits are derived based on Natural Process Limits and they show the “capability” of the process without any references. They reflect what the process can be expected to do as long as it continues to operate as it has .

    I see it this way (thank to Anita Carlton’s book ) – Control limits are the “Voice of the process “
    Specification limits are the “Voice of the customer”.

    If we have both, we can do a process capability analysis and calculate capability indices. If I don’t have the specification limits , I can atleast look at what my process is “capable” of , right ?

    Any other views ?

    0
    June 7, 2001 at 4:00 am #66891

    Dr Phil Whateley
    ★ 20 Years ★
    Participant

    Suggest you take a look at ch 13 of Bissel “Statistical Methods for SPC and TQM” (London : Chapman & Hall) 1994, especially sec 13.3. In general, for attributes data, capability is best expressed in terms of the average (u-bar) ! The control limits only serve to indicate when the process goes out of control.

    0
    June 7, 2001 at 4:00 am #66906

    Jim Parnella
    ★ 20 Years ★
    Participant

    Dr. Phil,
    I believe you are correct when you said,
    “In general, for attributes data, capability is best expressed in terms of the average (u-bar)!” I believe that is the way most Kodak practitioners handled it when I was with them up until 1988. I’ve never seen anything really convincing on this one way or the other. I’m not familiar with the reference you mentioned.

    Anyone else have a different, and reliable, reference on this topic?

    I’m still interested to find out if the original post misspoke and should have said control limits instead of process capability limits. Perhaps he will respond with an answer to that.
    Jim

    0
    June 11, 2001 at 4:00 am #66981

    Paul E
    ★ 20 Years ★
    Participant

    Vishy, I will try to explain my point of view about this GOOD QUESTION, Control limits at any chart is only to identify those assignable causes versus common causes, there is not relationship with the capability except by the posibility to calculate the sigma value, now in regard to attributes data, ans expect result or voice os customer should be Zero defects, in other words a capability index would be when no defects are found in a defined period, think about this, probably somebody else would like to share other point of views,

    I hope helps or start another dicsussion.

    0
  • Author
    Posts
Viewing 8 posts - 1 through 8 (of 8 total)

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