iSixSigma

Time Data MSE

Six Sigma – iSixSigma Forums Old Forums General Time Data MSE

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
    Posts
  • #29304

    Baker
    Participant

    Quick question, with a long answer I hope.  Does anyone out there have experience with performing an MSE on time data?  The problem I have is there is no simple “observation vs computer” data available.  Since a certain time can only be performed once (destructive), I am having trouble figuring a way to show repeatability.
    The procedure I was thinking:  workers enter in a time manually on a log, someone times the process simultaneously, hypothesis testing.  Would this work?  There is still no repeatability, but is that acceptable? 
    Jason

    0
    #74698

    Joy Cowling
    Participant

    Jason,
    I’m not completely sure I understand the question…but I’m going to stick my neck out and give you an example that a client used.
    The process had a time element and the operator needed to be able to measure the event accurately.  They performed a study with multiple operators by video taping the process close up.  Each operator viewed each video segment (same event, multiple measurements) to get repeatability and reproducibility.
    I’m not sure if that helps or not.  If I’m on the wrong track, provide more info and I’ll try to help.
    Joy
     

    0
    #74704

    Baker
    Participant

    Joy,
    You are right on track.  I have heard that argument.  If this case is done, what is the only discrepancy?  Start time and end time?  Why not develop a standard operating procedure to handle operator differences?  These are questions that I challenged BB and MBB with and have yet to hear a definitive answer.  Is this method industry standard?  Would love to entertain more ideas.
    thanks
     

    0
    #74706

    Joy Cowling
    Participant

    Jason,
    I’d have to see your process to give you help with setting up the actual test (what the criteria were, etc).  I’m a visual person and sometimes need to SEE what I am giving advice about.  In the case I mentioned, the measurement was the time from the start of the test to the event being measured.  Obviously start time would be a factor in determining that.
    However, the standard operating procedures should ABSOLUTELY and ALWAYS be addressed.  I am surprised that any MBB wouldn’t give you that same advice.  You would want to set up any kind of video test to as closely mimic the real process as you can.  The distance, clarity, etc of the camera should match what would be seen in real life. (Don’t put the video camera 6 inches away if the operator has to stand 12 inches away.  Don’t zoom in if the operators have no ability to do that in real life.)
    Hope this helps.
    Joy

    0
    #74735

    Baker
    Participant

    Joy,
    Think of the process as taking an entity from one location to another, and I want to know the time it takes.  If I clearly define a start time and an end time, what is the need for an MSE?  Also, considering the long time of the process, the time can be rounded to minutes.  What is the need for a resolution calculation?
    thanks
    Jason 

    0
    #74740

    Joy Cowling
    Participant

    Jason,
    What are you using the data for? 
    If you define the start and end time, that is a good start.  How will you verify that your defined procedure is the one being used to take the measurement?
    For something with a time that is in minutes, I’m not sure that the video tape method will be that effective, but that may depend on the resolution you need.  I can’t answer that for you.  Again, what are you using the data for?  What kinds of differences are you trying to detect?
    I know I am asking more questions than I am answering.  I think I put in an earlier post that I need to ask at least 3 questions for every one that I answer (especially when I can’t see the process).
    Joy

    0
    #74747

    Hemanth
    Participant

    Hi,
    I would say if you are confident of your measurement system then there is no need to go for an elaborate study. Just make sure that resolution of your measurement system is atleast one fifth of the difference in present and desired level you are looking from your project. Secondly, in cases where time is a factor (and looking at your question, I think your span of time is in atleast hours) then a digital clock would further reduce the chances of error. Just clearly define the start and end point and that should do. In case if you still feel then you can use cameras and do a full fledged MSA. Hope this helps.
     

    0
Viewing 7 posts - 1 through 7 (of 7 total)

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