Performing a Gage R and R with No Specification Range

D

duck5

Hi all,

I am used to doing G r and R but always I have been using a specification range and therefore getting the % error of the spec.

In this case I have no specification. Basically we are doing some work on liver tissue and we want to see how accurate our measurement technique is. We have the number of operators, samples and distinct categories but we have no spec. Essentially we don't need one, we just nee to make sure we can can measure the cut depth on a piece of tissue.

Any recommendations for this one? I am thinking of making up a spec range based on what depths we are getting but this will skew my data as the % error will be incorrect.
 

bobdoering

Stop X-bar/R Madness!!
Trusted Information Resource
You are not required to have a specification range to perform a gage R&R. %TV can be calculated based on the variation you present to the study (interpreted as the PV). My recommendation is to present as much variation as you expect to see (even if you have to fabricate come of the outer range) to get the most meaningful results.

Is the test destructive? Is there any impact on measuring the same spot each time? If not, be sure to do so. Identify a location and have each operator measure the same location.
 
Last edited:

Miner

Forum Moderator
Leader
Admin
Also, depending on your end goal, you may not even need a %TV metric as much as knowing the pure repeatability & reproducibility expressed as a standard deviation. You might also want to check my blog for the article on Wheeler's methods, which includes a metric called PME, which is the Probable Measurement Error of a single measurement.
 
D

duck5

Thanks all for your responses. It looks like I do not need a tolerance as mentioned. I shall be looking at my total % error
 
Top Bottom