How much can I cheat Gage RR

Six Sigma – iSixSigma Forums Old Forums General How much can I cheat Gage RR

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
  • #30901


    I have a measurement process where I load and unload a sensor and take a force reading off of it.  The process involves loading and unloading of a convertible top.  The ‘j-hook’ that latches the roof down mates with a sensor that gives me a force reading.  We are failing a gage R&R miserably doing it like this and was wondering if anyone knows of a rule of thumb that says how representative a process must be when doing a gage R&R.  For example, our convertible top is a very unstable and flimsy part relative to industry.  So one of the things we were trying in our gage R&R is not cycling the top up and down, only engaging and disengaging the ‘j-hook’.  Am I cheating?


    Carl H

    I dont think you are cheating – just using some engineering judgement!

    If you think the full top cycling does affect the sensors accuracy/stability, then you may want to do the full cycling to a least understand this.  Maybe run chart 20-30 full cycle measurements.
    If you are just interested in sensor repeatability, then take the short cut method and just load/unload the sensor without full cycle.   Maybe compare a run chart of these 20-30 cycles with those in step 1.
    Enjoy your turkey.



    There is an alternative to the standard gage R&R methods that is much more realistic.  Don Wheeler has been teaching it for years and it is now included in the AIAG MSA handbook as an approved method.  If you have access to the handbook it is listed as the Average and Range method starting on page 99 of the 2002 edition.  Dr. Wheeler also has a booklet titled ‘Evaluating the Measurement Process’ that covers this method and shows problems with the standard R&R guidelines. 
    I hope that this helps.


    Patrick Swift

    MSA = Measurement SYSTEM Analysis
    to get a meaningful R&R you need to conduct tials incorporating every part of the system. That’s the reason why we use several operators (i.e. the operators are part of the system). It sounds like removing the          j-hook removes part of the system, and will not give a true reflection of the variation in the System.
    Try and get homogeneous groups of parts covering the specification range concerned and then conduct the R&R. Having little or no variation between subgroups will show a very high R&R. It’s not cheating, just determining R&R over a particular spec range, afterall that’s what you want to know.

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

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