iSixSigma

Stability in MSA

Six Sigma – iSixSigma Forums Old Forums General Stability in MSA

Viewing 13 posts - 1 through 13 (of 13 total)
  • Author
    Posts
  • #32605

    ROSS
    Member

    Hello:
        I am working in MSA, and as we all know, the gage accuracy has three component which is linearity/ acuuracy and stability. And we can use Gage linearity study in Minitab to study the lineariy and accuracy( using the linearity/ %linearity and %bias), but what can we do for the stability? Some material tell me we can use the control chart to study the stability, could you tell the procedure to do it, such as: what control chart? the same part? how to determine the control limited? and how to judge the stability?….
         Thanks for your help!

    0
    #87266

    Puppy Love
    Participant

    Tony
    you can monitor stability from R-chart … If the R-chart is going out of control ,that means repeatability is going bad (your equipment is not stable for measuring) so you have to prepare to recalibrate your equipment.
    Puppy Love

    0
    #87270

    faceman888
    Participant
    #87271

    Gabriel
    Participant

    You mean the process R-chart? If you find an OOC in the R-chart that means that the process is going out of control. The measurement is just one part of the process, but not the only thing that can go wrong to put the R-chart out of control.

    0
    #87273

    Puppy Love
    Participant

    Dear Gabriel
    If you want to monitor stability of equipment so you get some data from that equipment by repeating from 1 operator so you can make r-chart , right ? and if your chart is going out of control … what does it mean ??? it means the capability of repeating is not good so what is the cause of that ??? that is equipment is going to error , right ? and why can’t you use it to apply for testing stability ??? I think it’s make sense to do it…
    Let you discuss about that…
    thanks,
    Puppy Love

    0
    #87275

    Mike Carnell
    Participant

    PL,
    That is a huge leap from an OOC on a range chart to a single cause particularly after truncating the input to the chart to one operator. That means there is only one variable that controls the range. It might happen but I would be inclined to say not very frequently.
    Good luck.

    0
    #87288

    Gabriel
    Participant

    Oh! You meant the R-chart of the R and R study using the Xbar-R method. I missunderstood.
    Well, that chart is not intended to check stability of the measurement system either. Only to check that the “between operators” variation is stable. You wil not detect any kind of drift over the time with it.
    The typical way to conduct a stability study is to take one part and measure it several consecutive times (let´s say 7) every days. Each group of consecutive measurement is a subgroup, and you plot Xbar and R for each (1 point per day). After several days (let´s say 20) you calculate the control limits for both Xbar and R and look for OOC signals. If it is free of them, then you failed to reject that the measurement system is stable.

    0
    #87292

    Mike Carnell
    Participant

    Gabriel,
    I’m glad you got the methodology because I still don’t see it.
    We use 25 groups of 5 for the study. Probably mre of a legacy than anything else.
    Regards,
    Mike

    0
    #87297

    Puppy Love
    Participant

    Dear,
    I think my idea as same as the way you said in previous post…. I think it’s o.k….
    Puppy Love

    0
    #88183

    Tim
    Member

    Tony, the most recent MSA Manual from the Automotive Industry Action Group (3rd Ed, March 2002) was rewritten to make it more user friendly and is a great source of information on MSA, even for those not involved in the automative. The manual can be obtained from the AIAG website, http://www.aiag.org.   The manual recommends finding a master for the low-end, mid-range, and upper-end of the expected measurement range (I recommend gauge blocks of various sizes).  You need to establish an X-bar/R or X-bar/s chart for each master, take recurring measurements (sample size & frequency based on knowledge of the system), establish control limits, and then monitor for out-of-control or unstable conditions using standard control chart analysis.  Stability should be confirmed as a prerequisite to bias, Is the system proves to be stable, that same data can be used to determine bias.  If the measurement system is unstable, more sophisticated problem solving methods such as DOE or ANOVA may be necessary to determine the cause of instability.
    Hope this helps.
    Tim
     

    0
    #107260

    Sean Steele
    Member

    My only Question is, How long do you have to record reading in order to evaluate stability?

    0
    #107509

    Tim
    Member

    AIAG MSA leaves it to the discretion of the evaluator to determine sample size and frequency based on their knowledge of the measurement system.  If you are asking how many sub-groups must be collected before control limits can established and decisions made on the stabilty of the process based on those control limits, once again no specific criteria is given (the example provided in the MSA Manual uses 20 sub-groups).

    0
    #107514

    Sean Steele
    Member

    Thank You Tim. That is the way I understood it as well. I was just trying to get another point of view to confirm.

    0
Viewing 13 posts - 1 through 13 (of 13 total)

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