iSixSigma

RR studies with one operator

Six Sigma – iSixSigma Forums Old Forums General RR studies with one operator

Viewing 14 posts - 1 through 14 (of 14 total)
  • Author
    Posts
  • #33146

    mortez
    Participant

    Hi everybody,
    Consider the actual situation (such as in a small company), when there is always ONE inspector performing some specific measurement with some specific instrument everyday. In other words, usually, there is no one else to perform that measurement.
    Is it still necessary to perform a (hypothetical) 2 or 3 apprasier R&R study, or a Repeatability study would be sufficient?
    Regards & Thanks on beforehand.
    Mortez
     

    0
    #89175

    tsiriud
    Member

    My company perform only repeatability of the tester using only one operator

    0
    #89188

    mortez
    Participant

    tsiriud
    Could it be that operator cannot affect your test result (i.e. no appraisr to appraiser variation).
    The question I had put, concerns the situation where there CAN be appraiser to appraiser variation but, practically there is only one appraiser available to do the measurement.
    Regards
    Mortez

    0
    #89194

    Mike Clayton
    Participant

    http://www.itl.nist.gov/div898/handbook/mpc/  is a link to the NIST-Sematech online textbook, this part on measurement systems analysis.
    The over-simplified R&R studies often miss the point that measurement systems have many sources of variation that need study and control.   Bias, linearity, stability over time, etc.   So even with one operator, many other things can affect the system, including room temperature changes, drift of calibration, etc.  
    The simple spreadsheet approach is a START only.  If the gage looks marginal, then do a broader study.   

    0
    #89210

    David Oakley
    Participant

    Mortez
     
    I always have more than one operator. In some cases, I have been one of the operators. In that way, you get data that will tell you if the operator will make a difference.
     
    Those times when I was told that there was only one operator, that was always wrong. Operators take vacation, get sick, act as leads, give or recieve training and on and on. I have seen some large variation when the measurements weren’t done by the “only” operator.
     
    David

    0
    #89567

    Larry MBB
    Participant

    I am asked this question frequently.  In the case of a GB project, where data is being gathered primarily for the purposes of calculating process capability, performing statistical tests, and improving a specific process over a specific period of time, one operator is sufficient.  Why introduce another operator when the 2nd operator is not going to collect data for your project?  It seems like a “fill in the box” exercise, and will likely turn people off to the important stuff, like using Six Sigma to improve a process.

    0
    #89572

    David Oakley
    Participant

    Larry
     
    I have never seen the case where there was truly only one operator. In some cases, I didn’t find out about the other operators until after I had done the R&R. I then had to go back and redo the study. Now that will really turn off people.
     
    If you have bad data, your data analysis part of SS will lead to bad conclussions. In that case, there is a good chance that you will not have a successfull project. That will also turn people off.
     
    David

    0
    #89576

    Kimberly BB
    Participant

    Larry, 
    I agree totally with your perspective.  Too many times we try to check boxes and lose sight of what we’re really trying to accomplish.  I recently closed a BB project in which two of my MSAs were done with  only one operator.  The solution?   Use the P/T ratio to support your conclusions on a given test method.  Hopefully your test variability is significantly lower than your spec range.

    0
    #89578

    Sam Orez
    Member

    Occassionaly, we will perform a 1 operator, 10 part, 2 trial repeatability study. Just regress the first trial on the second trial. We look for R Squared to be greater than 80%.
    Sam Orez

    0
    #89619

    marklamfu
    Participant

    It is a interested case!
    One operator to perform R&R stduy is not adequate, as you said, only one R (Repeatability) is done.
    You known, the aim of perform R&R study is to control the measurement as a suitable condition to meet customer requirement. Lack of another R, the status of this measurment can not be confirmed. 
    In fact, the persons involved  the process should be more than one , the measuring data may be verified by supervisor or other person, the outgoing product may be inspected by customer,etc. So, I propose you to invite related persons (even indirect) to do another R with this operator together.
     
     

    0
    #89635

    Ram
    Participant

    Hello Mortez,
    To do GRR, you need another person for comparison to determine its repeatability and reproducibility as well. And, note that the other person has to know and uses the equipment as well otherwise you defeat the purpose. The production output before it reaches to Inspector their is always setup or in-process inspection. Try to utilize these people and do the GRR.
    I hope this help.
    Good Luck,
    RAM
     
     

    0
    #89636

    Robert Butler
    Participant

    Situations with one operator and one machine are fairly common. As was mentioned in one of the other posts to this thread there are a lot of things besides operator and machine that impact R&R.  If you have the situation where one operator and/or one machine are the norm then you should test it that way. 
      If the situation exists where “fill in” work is done by someone else when the regular is on vacation you can include the substitute but you need to have an identifier to allow you to tell who did what and you should analyze the results separately.  In every instance that I can recall where it was one machine and one operator with the option of either an occasional fill in operator or substitute machine the differences in means and/or variance was such that pooling of the regular and substitute results gave a false estimate of the “typical” target and tolerance that could be expected from the process.
      In those cases I reported out “typical” to be the measurements associated with the “standard” operator/machine and I made certain that everyone understood how things could change when substitutions were made.  In several instances management took note of my findings and either decided to guarantee the existence of another qualified operator and/or substitute machine or they made a policy decision that when the “standard” wasn’t available they would choose to backlog needed measurements rather than have them run by a substitute.
      In the case of a single machine/operator you need to make sure the operator is getting blind samples and that the samples are introduced into the system over the usual maintenance cycle of the single machine.  You also need to make sure that everyone is aware of the one operator one machine situation and you need to have an agreed upon plan in place to revisit R&R the minute either of these situations change.

    0
    #89678

    mortez
    Participant

    Dear Marklamfu,
    What you and some of the guys are saying about the need to include another appraiser in the study, has of course some logic behind it.
    But suppose I included another person who usually does NOT do such measurement, just for the sake of evaluating the Reproduceability.
    We all can guess what happens: Reproduceability is bad and as a result the total R&R is unacceptable, because the new appraiser is less skilled in that kind of measurement.
    But what kind of corrective action can I take to improve the measurement system? Should I train that appraiser to do a measurement which he has nothing to do with?
    I hope I could explain my point.
    Regards and thanks for everybody’s input,
    Mortez
     

    0
    #96761

    howe
    Participant

    I have the same situation with a Green Belt I’m coaching. Only one person performs this function, if she is out it doesn’t get done. Therefore, I instructed the GB do conduct a repeatability test only. There’s no sense in determining the variability from how another operator measures the item if no other operator will measure it…

    0
Viewing 14 posts - 1 through 14 (of 14 total)

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