iSixSigma

XBar/R basic questions

Six Sigma – iSixSigma Forums Old Forums General XBar/R basic questions

This topic contains 8 replies, has 6 voices, and was last updated by  lin 12 years, 7 months ago.

Viewing 9 posts - 1 through 9 (of 9 total)
  • Author
    Posts
  • #46343

    ilovesausage
    Participant

    We are just scratching the surface with control charts and i was hoping to get some advice.
    Is it correct in thinking that a control chart could be used to help establish specification limits?  I know control and spec limits are not directly related but for example; We have a spec limit that is very wide and when I look at out capability we use very little of the available spec.  We do occasionally get a very high “in spec” reading that would show up as “out of control on a control chart.  Typically the high reading is well over 3 sigma from the center line.
    I believe (and am trying to prove) that even though the units pass they have undesirable characterisitics.  The problem is that it will take a few months to actually harvest enough samples to do a proper analysis.  I’m not a big fan of easy way out type short cuts but i also dont want to miss an opportunity if these control charts are telling us more than we understand.
    Also, subgroup size in control charts (minitab specific) seems to be critical.  The control chart can change significantly by adjusting the subgroup size.  Is there an optimum subgroup size that we should be shooting for to maximize effectiveness of the control analysis?
     
    thanks for the help.
     

    0
    #152924

    PV
    Participant

    If you are well within spec but not in control that is a good case for a DMAIC project to find out what your special causes are that are driving your process out of control. I would assume that your spec limits are set by customer and cannot be changed. Anyway why would you want to establish new spec limits?
    There is nothing called an optimum subgroup size. You need to plan a rational subgrouping strategy to collect samples, calculate sample size and depending on the process decide the number of samples within each subgroup. The control chart to use will depend on subgroup size

    0
    #152939

    ilovesausage
    Participant

    The spec limits we have are a standard spec that our division engineering group developed 4+ years ago.  We use it for every customer regardless of model. 
    We have a new customer that believes the spec is too wide and they are requesting that it be re-established.  Granted, they are requesting this with no data or supporting evidence which is another issue.  I’m trying to make sure we approach this correctly.  I am almost certain that the spec is out dated and should be investigated.  I believe that the high end of the spec allows too much risk to our customer and will be trying to prove that but it will take considerable time due to the low frequency of samples in the ranges that we need to test.
    I am by no means trying to change just because the customer is requesting it.  I just want to make sure we are where we really need to be to protect ourslelves and the customer.
    I’m a little confused about the subgroup explanaton you gave.  We normally use the standard minitab Xbar/R chart found in the Control Chart tools menu.  Are you saying that different subgroup sizes may require a different control chart selection?  The standard chart lets me select a subgroup size.  I normally try to keep that at a subgroup size of 5.
     
    thanks for the help.

    0
    #152948

    Darth
    Participant

    Plug your spec into Mini and do a process capability.  Assuming that your process is stable you can see whether the new spec will cause you any problem or not.

    0
    #152950

    Chris Seider
    Participant

    Control charts have no bearing in the discussions of long term process capability especially if you are using XbarR charts.  It is considered by many, including myself, to be a cardinal sin to have specification limits on a control chart.  The main reason is you can have points within the spec limits but one has to remember the Xbar is an average and one of those items measured was out of spec.
    Keep in mind that customers set specifications except in those rare cases where the customer has no alternatives.  Specs are the voice of the customer and the control limits are the voice of the process.
    A practical example of what to do in your case would be to process capability analysis on your long term data based on the proposed specs and hope you have at least 1.0 and preferably a 1.5 for Ppk.  Hopefully the customer is not asking for a different target or center between the specs or else you will have to consider controlling your process to differing targets depending on the customer–leading to probable increase in process variability.  If you don’t have a good process capability on the proposed specs, then you will want to instantly launch a project to find the sources of variation or consider the option of not taking the new business if you can’t prove good sustainable processes.

    0
    #152954

    ilovesausage
    Participant

    I agree that specification has no place on the control chart.  I’m not getting the two confused but moreover trying to use the two together to understand my options.
    The customer has not provided a specification.  In my industry they provide us with SOME spec but mostly the performance/testing spec.  It is up to our engineering group to determine the correct spec (in most cases) to meet the customers needs. 
    The issue that I have here is that this spec has not been looked at for a long time by engineering. Primarily because none of our customers have questioned it.  It just so happens that at the same time we began having some warranty issues a potential customer challenged the spec.  Two seperate issues but its forcing us to dig real deep.
    I have no target value.  I have a spec but there is no target that we are trying to achieve to optimize the assembly. 
    The thing is, when i look at a Xbar/r chart AND a capability analysis i can see pretty quickly that without the rare high end of the spec part we are actually pretty capable.  Pp 3.3 and Ppk 1.9.  To me this is evidence that when we have a part at the high end of the spec i.e. out of control; on an Xbar, that although i have an “acceptable” part there is something special about it.
    We are in the process of collecting samples that fall into that category BUT since it happens so rarely it will take a long time to get enough samples to do anything worth while.
     
    thanks for the input so far.  you guys are great.
     

    0
    #152956

    Chris Seider
    Participant

    Great approach to let data lead your efforts…..
    Nice real, serious questions….
    I don’t post this often unless I’m overseas and fewer distractions than when I’m back home…..

    0
    #152960

    Jim Shelor
    Participant

    If there really is a risk to the user of your product while it still meets your current spec, I would think an FMEA would get somebody’s attention, especially when you show that with no failures there is still a risk to the user.
    I would also think a root cause analysis on the OOC points you have to date would give you something to go on rather than waiting for more points.

    0
    #152962

    lin
    Participant

    Sorry, not a stats guy here (don’t really deal in minitab speak), but if your customer told you they think your spec is too wide and you have observed out of control measurements that are within some spec that was set by engineering 4 years ago, your spec is probably wrong and you need to go back to the DM part of the process to determine proper VOC.  Call me crazy, but if you work for a company that considers profits important (most do), your company really should value your customer’s input with regard to their requirements. 
    That being said, if you observe an out of control measurement that is over the favorable line of an upper or lower spec limit, you would be foolish not to try to determine the cause of that particular variation.
    FYI, I’m not GB, BB or any other B.  I just find this forum strangely interesting and informative….sometimes.  So take what I say with a large grain of salt.
     
     
     

    0
Viewing 9 posts - 1 through 9 (of 9 total)

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