iSixSigma

control chart issues

Six Sigma – iSixSigma Forums Old Forums General control chart issues

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
    Posts
  • #43304

    Ruddy
    Participant

    Hi, there
    My company has 3 shift, now we collected 3 weeks data about product( continuous data and collecting once hours), Now we want to draw control chart to see if process  is stable.
    The question is we can get several charts by different subgroup, such as by hour, by operator, by day, by shfit etc. And some charts appeared stable, some appeared unstable. Does anybody can tell me which chart I can use? thanks
    Michael

    0
    #137222

    Ricky.Wang
    Member

    Hi Michael,The way you group you data can influence your conclusion. I would suggest that you try all the methods, and make conclusions from each one of them. For example, based on subgroup by operators, you may find whether the operators are consistent. By hour or day, you can see the short-term and long-term variance.

    0
    #137643

    Jonathon Andell
    Participant

    Cosider using an anlytical method, such as ANOVA or GLM, to quantify the sources of variation. Also, think about a Multi-Vari chart to convert it into a visual display. Once you learn from them, the path may be a little more clear.

    0
    #137647

    Dr.MG
    Participant

    Dear,
    I recall yr statements  “….3 weeks data was collected….” and  “….whether the process is stable…”
    By this approach you are doing a postmortem. And the causes (for out-of-control) may not be valid today.
    My sincere submission is that past data may be used for practice or to train the operator on how to chart. The real benefit of control chart comes from charting the live data (real-time data) – especially to predict before the process actually goes out of statistical control from the trend – remember the out-of-control conditions and take corrective action.
    Best of luck.
    Dr.MG

    0
    #137651

    Learner
    Participant

    If I remeber well, you need to define what you want see and then plan a data collection that reflect your needs (rationale subgroup). As reported in other post, you can analize your data in every way you want, but if for ‘stable’ you mean as SPC, than you can use basic X-barRchart and maybe you’ll see that differencies seen between different subgroup, will be highlighted as out of control. Perhaps, you already have the answers to your out of control.

    0
    #137660

    Murthy
    Participant

    Hi Michael
    I would advice you not to look at all the control charts. It would only confuse and complicate matters. Plot the control chart with samples collected every hour. This will help you identify any early process deviations, thus preventing defects.

    0
Viewing 6 posts - 1 through 6 (of 6 total)

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