Gage R, not Gage R&R? Gage R&R study of my calibration equipment


Torque Guy

I am doing an R&R study of my calibration equipment. However I am the only person in the lab. So I can calculate my gage repeatability but I have no way of doing my gage reproducibility. Since I am the only person doing calibrations would this mean that reproducibility is not an issue?


Hi TG welcome to the Cove

Question has to be: what if you go on vacation or get hit by a truck? Who does the calibration then? Surely you have a back-up or someone else who is trained. All good labs are supposed to have cross-trained employees :vfunny: so it wouldn't be a bad idea to find a warm body someplace.

For the R&R could you bring in some of the people who use the equipment you are calibrating?

I'm one of the lucky ones there I guess - I don't have that problem.

Once again, welcome.


Torque Guy

I thought someone would post that answer. I should have added this.

I am the owner of the company. I wear all hats and do all jobs.

I don't think bringing in two of my customers is an option.

I'm all by myself on this one.
Last edited by a moderator:

Atul Khandekar

Torque Guy,

I think, as you said, in this case you will not have any variation due to appraiser (AV=0). R&R will be = equipment variation (EV).

I am not quite sure what are the implications of doing only one 'R' on 'Calibration Equipment'. May be one of the calibration experts on board would be able to help.

Dave's point is well taken. Are your customers asking for both the R's? Not that you should have employees for the purpose of R&R alone, but (assuming you are an external calibration lab for your customers) don't you think your customers will be rather nervous if they find there's no back up?

Ryan Wilde

Okay, I'll admit, I know nothing about GR&R. I do uncertainty analysis, because that is what cal labs do. That said, there are many types of calibration in which variation due to operator is zero (such as a scale - level is level, center of pan is center of pan, etc.). The trick is that I still include the variation of zero in my uncertainty budgets, basically for standardization's sake. Therefore, I would surmise that the same would apply to a GR&R, do the calculation, and include the assessor variation of zero. It's still a G R&R, but that second R is really nice. :vfunny:

As for "backup", it seems to me that a single person operation's backup is called "subcontracting", which probably falls outside of the scope of a GR&R.


Torque Guy

I am doing this gage R&R in order to add the result to my uncertainty analysis. I believe that I am performing a gage capability study rather than a GRR. I think that I will use an OV of zero though for standardizations sake.

My next question would be how to complete the R&R with more than 3 trials? I have a total of 11 trials so far, should i combine them all or should I do a GRR for every three trials and chart the results? If I was to combine the trials what would my K1 factor be if GV = (R bar) X ( k1) ?

As for my customers being nervous. I am a small lab and am working with small automotive plants. I give a lot of attention to customers that other vendors find to be too small. My customers are loyal to me because I provide them better than average service. I don't really worry about it much, however I do get frustrated with the industry belief that bigger is better. I find that these large companies do not care about their customers only with the bottom line.

Atul Khandekar

Factor K1, it is just the inverse of factor d2star (d2*). You can look up the value of d2* from the table provided with AIAG MSA Manual.

11 trials seem to be a bit too much. Also I don't understand what is meant by 11 trials "so far". How is the study being conducted? Are you taking the readings in quick succession or are you following fixed periodicity for taking readings?

If you are taking readings under two (or more) different environmental conditions, these conditions can be treated as two (or more) appraisers. The AV in this case would give you the reproducibility under different conditions.

Torque Guy

By 11 trials I mean that I do a trial every month. I have had the system in place for 11 months. This is part of my on going verification of my equipment in order to avoid a catastrophe.

I am trying to fulfil the interlaboratory comparisons element of 17025 also.

Right now I am just trying to get a number to use in my uncertainty calculation.

I believe what I am going to do is complete a gage capability study based on 3 trials one from each month of the year. I will then plot the results in order to detect any trends that may indicate my equipment is drifting out of spec. I will take an average of the studies in order to get a number for my uncertainty.

Sound good?
Top Bottom