iSixSigma

data collection confusion

Six Sigma – iSixSigma Forums Old Forums General data collection confusion

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
    Posts
  • #47289

    newbie
    Participant

    Forum,

    When gathering data for control charting, I assume the data is collected after a successful MSA, yes?
    Is that data then rolled into the overall data collection effort, or is a seperate effort?
    If one is to be careful to collect control chart data under similar technical conditions *same operator, same machine, etc* then how does this effect future collection efforts….ie arent we looking for different collection methods later on to highlight between group differences….isnt that what is wanted in ANALYSIS? 
    Thanks!  A real brief example would be great if anyone has the time….

    0
    #157568

    GrayR
    Participant

    1. You would like to know that you have a capable measurement system before you collect the data, and you would also like to know how much variation (or error) is related to the ms.  So yes, you should do a gage study — if you have collected data before the study and the ms is acceptable, then you can use that data as well.
    2.  We use the control charts to help understand the total process results, and measure ‘current state’ of the process, so the first control chart on the y-result should represent a good sample from the process (various operators, materials, etc.) — but we also don’t let it go very long before we start to identify trends, instability, etc. If possible, for variable measures, we try to get samples that represent within part variation, between part variation, and variation over time (different batches) — if we keep the sample sizes constant it is very easy to use variable control charts with subgroups (within, between, etc.) and use both x-bar and R-charts for analysis.

    0
    #157580

    newbie
    Participant

    Thanks! So we want to collect our subgroups in such a way as to minimize within group variation (eg small, sequenctial collections) , but the group of samples should span the range of the process, so as to capture between group variation (hence the need to collect samples across shifts, operators, batches, etc)…yes?

    0
    #157586

    GrayR
    Participant

    Without knowing what type of parts you are measuring — you wan to collect the samples to be representative of ‘within group variation’. And then span the range of the process as you mention.

    0
Viewing 4 posts - 1 through 4 (of 4 total)

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