Auditor wants our Gage R&R Procedures to address Linearity, Accuracy and Stability


Greg Fote

Linearity, accuracy and stability

We have our R & R procedures and work instructions operating smoothly. Can anyone offer any assistance on what they are currently doing with linearity, accuracy, and stability as defined in the MSA manual?

Are you using specifically-written procedures and/or work instructions for these activities?

Are the auditors looking at these items closely during audits?

Thanks in advance for any advice or assistance and/or samples you can offer.


Fully vaccinated are you?
You might first want to read the following thread from the 'old' board:

(broken link removed) - from the old forums.

Mr. Adamek and I had a somewhat extended 'conversation' to which I admit I was pretty hard headed... I think he got mad at me...

In addition, you might want to check out:

Some Calibration - MSA related tidbits

The bottom line is auditors are beginning more and more to address MSA issues. They want to see evidence of understanding and utilization of MSA concepts. It would probably be wise about now to consider a procedure explaining how you address linearity and bias and such. It could be very simple and along the precepts of what is in the threads hot-linked in this message reply.

By the way - Did you know Gage R&Rs are not *mandatory*???

Hope this helps!
Last edited by a moderator:

Bill Smith

The UL auditor that just left our facility looked very closley at MSA and wrote us up for not measuring and trending stability. What he wanted to see was the past calibration results plotted for each instrument. I havent fully checked this out but it seems excessive. We were going the obligatory GR&R studies so we are still looking into the validity of this Action request.


Fully vaccinated are you?
Bill - My opinion is that is a crock of ----. Where does it say you have to trend on paper? If the calibration person has half a bean of intelligence s/he can easily look at the numbers and understand what is happening without putting together a graph. I just went thru this with LRQA in May. They didn't require a graph - but they did want the cal folks to prove that they considered this at each calibration and that the concept is understood.

UL is known to be asinine about MSA - But, other registrars are hitting it more now too. I will admit that many, many companies do not have people who really understand MSA, but the auditors are getting rediculous.

Christian Lupo

Bias and Linearity should be known before you do a GRR study. My company has a procedure for MSA, that includes GRR. In fact, GRR is one of the last, if not the last step in an MSA study. As far as stability is concerned, the procedure gives the QA Manager the responsibility to decide if a stability study is necessary. The QA manager's job description requires him (her) to have MSA and SPC training. Stability studies may not be necessary for computer controlled equipment or if the cost of the project outweigh the benefit.


Fully vaccinated are you?
I agree with you 100% Christian. You are looking at MSA as a high order system - which it is. R&R is done late in the game. It sounds like you have a well written procedure. I'd love to see it.

David McGan

Christian, when you're e-mailing a copy of your procedure to Marc, copy me in also. Or just post it here for all of us to read.


Fully vaccinated are you?
Comments, folks?


Date: Mon, 12 Jun 2000 10:49:57 -0400
From: Bruno Brazauskas
To: Greg Gogates

Per the previously recommended AIAG manual:

The data is the result of measurements taken with the gage being evaluated, utilizing the appropriate gaging instructions. The samples, which are measured, may be either production parts or masters. Collect the data in a manner which is similar to, but in terms of repetition not quite the same as a gage R&R study.

For Linearity, the samples should be taken from the expected operating range of the gage. The bias on each of these parts is determined to be the difference between the reference value and the observed average measurement. The reference value for each part is determined through a different means. These means may be plate or other inspection machine layout or through an independent or calibration lab.

For Stability, it is not necessary to have parts from the expected operating range of the gage but it is still a good idea. As the AIAG MSA states, sample size and the frequency of the measurements should be based on your knowledge of the measurement system. So you may use the same parts as you used for the Linearity Evaluation. For this study, the measurements are taken over time (i.e.. the same parts are measured again and again over a predetermined time).

Bruno Brazauskas
Inspection Technologies, Inc.

Greg Gogates wrote:

> Date: Fri, 09 Jun 2000 09:05:37 -0400
> Forgive me but I'm still confused on the data collection. Is this data from
> actual production part measurement, calibration data, Gage R&R data or how do I
> obtain the data? I guess I'm a little dumb but I'm not sure what data to use for
> these studies.
> ____________________Reply Separator____________________
> Author: Greg Gogates
> Date: 6/5/00 11:38 AM
> Date: Mon, 05 Jun 2000 08:44:18 -0400
> From: "Bruno J. Brazauskas"
> To: Greg Gogates
> These types of studies may be simple or more involved. The simple
> versions are as follows and reference GM's Statistical Process Control
> Manual:
> Basically, "Gage Stability" is the difference (range) in the average of
> a minimum of two sets of measurements taken with the same gage on the
> same samples at different times. For example, using gage "A" and sample
> "1" take 10 readings today and average them. Repeat this test tomorrow
> under the same conditions. The difference between today's and tomorrow's
> averages is the stability of gage "A" over one day.
> "Gage Linearity" is the range of values derived from at least groups of
> data. It is a little more complicated and requires that the actual
> values of samples to be tested are known. A known sample at the low end
> of the specification is measured with the gage and the difference
> between actual and measured readings is recorded. This is repeated with
> a sample from the high end of the specification. The range obtained
> from these deviations is the gage linearity.
> A more involved approach may be found within the AIAG's "Measurement
> Systems Analysis (MSA)." It includes more information in the form of
> "Guidelines for Determining Linearity" and "Guidelines for Determining
> Stability."
> Bruno Brazauskas
> Inspection Technologies, Inc.
> Greg Gogates wrote:
> >
> > Date: Wed, 31 May 2000 13:34:59 -0400
> >
> > I have been requested by our QS9000 registrar to begin performing Gage
> > Stablility and Linearity Studies. I have performed Gage R&R studies but never
> > linearity or stability studies. Can anyone give me some detailed information
> on
> > what type of data collection is needed to perform these studies and how to
> > perform these studies? I have software (Gagetrak 4.0) for analyzing the data
> but
> > it doesn't really specify exactly how and what type of data is needed.
> > Thank you.


Fully vaccinated are you?
Evolution of MSA Requirements

Just as an 'Oh My' fact, this thread sorta dates when registrars in QS started looking at stability and linearity - when MSA started reguiring more 'details'. Any comments? :thedeal:
Top Bottom