iSixSigma

Xbar charts with varying sample size.

Six Sigma – iSixSigma Forums Old Forums General Xbar charts with varying sample size.

Viewing 5 posts - 1 through 5 (of 5 total)
  • Author
    Posts
  • #39980

    jediblackbelt
    Participant

    I am interesting in looking at a process where I am checking the average of my set up times and see if they are in control.  The issue I am running into is that the setups come in on a daily basis and we may do 5 today, 7 tomorrow, and 2 the following day.  There is no set size of set ups.  What I am interesting in knowing is my average set up time is and if I am improving that metric.  However, I am also interested in reducing the variation of set up times.  What would be a good chart to use for SPC on this if my grouping is of varying sample size?  Or am I looking at this all wrong and should be investigating another method of tracking?  I have not entertained an individual chart due to also having to report what my average time per day is spent on setup.  I thought about a Xbar-S chart, but I know there are times that my grouping would be small and could potentially be only 1 setup.
    I am at a new company and right now playing the game of giving them what they want first before I can move to a different metric that may be of more value.
    Any help on this would be greatly appreciated. 

    0
    #122947

    Ken Feldman
    Participant

    You can do an Xbar/R chart with varying sample sizes which means the control limits will change for each sample.  Problem is interpretation and reaction, they will need to be tempered.  Small subgroup size is also a challenge.  I would not discount the I/MR chart.  You can still report out average setup time but keep in mind it might not be very meaningful if you are using such varying denominators.  Some sort of weighted average might be of value rather than just straight averages.

    0
    #122950

    jediblackbelt
    Participant

    Darth –
    What I am really thinking about doing is the IMR chart and then rolling a Multi-Vari chart together to come up with daily/shift differences in set ups.  What are your thoughts on that method?  Also, when you mention a weighted average chart how would I roll that out?  Could I still use a vaiable grouping for my dispersion or would I be breaking some rules?  I haven’t had the luxury of dealing too much with the weighted average charts.  I would need to do more research on their applications probably.
    Also, want to thank you for your help to me and this forum.  Your opinion along with several others (Carnell, Stan, just to name a couple)that answer a majority of the real issues is a great help. 

    0
    #122953

    BBMole
    Participant

    I would suggest that you be careful when interpreting the results. Often I have seen BB use Shift as a factor. This can only be assumed if the same personnel do the same job every shift. If the workforce is multi skilled you may observe effects that are due to different people doing different jobs during successive shifts. Net result is that each shift must be treated as different from the last one. Is how many combinations of 10 people can do 10 different jobs, and you begin to see the complications. To identify the cause and effect I have used a GLM which is crossed and nested for this. standard analysis may not show an effect while GLM should show the effect.
    Minitab is not good to set up for this as you have to determine the interactions and nesting for the GLM.
    Good Luck

    0
    #122956

    Ken Feldman
    Participant

    JediBB, I wasn’t refering to a weighted average control chart but the mere reporting of weekly average setup time.  This works when rolling up numbers.  Weekly might be too frequent.  You can roll up monthly averages by taking weekly averages but then you might want to weight the weekly averages when presenting the monthly average.  Bottom line is that taking averages of such small data sets is pretty tenuous.  Taking a s.d. for small data sets might be pretty meaningless as well.  A multi-vari chart is good at looking at snapshots but not over time.  How much action/reaction does management take with the weekly data?  If they are reacting to small changes in weekly averages, they are probably tampering.

    0
Viewing 5 posts - 1 through 5 (of 5 total)

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