Small Variation in Thickness Measurements in Gage R&R Study

D

dbulak

When we are doing our GR&R studies, it ususally involves taking measurements on the thickness of the parts. The problem is that there is very little if any variation in the thickness. For example, .032 +/-.001 on the thickness, we might get from .031950 to .032005 on all the readings. The micrometer reads to 5 decimal places. When we do the study, we do not get good results. Looking for comments on how to deal with this situation because we are getting frustrated.
 
A

alspread

Re: Thickness measurements

It may be due to the distribution of yoour results. On the face of it, it may seem that a micrometer mAy be adequate for a =/-.001 thickness that only varies .0005. But if the process/results are showing a wide distribution of the data, then the GRR may fail because, statistically (i.e. probability), it needs to compensate.

Is there a relatively wide variation in either one or more evaluators (operators). Or is there a wide (relatively) variation between operators?
 

bobdoering

Stop X-bar/R Madness!!
Trusted Information Resource
Re: Thickness measurements

The best suggestion is to post the GR&R data so that we can review the issues.
 

Miner

Forum Moderator
Leader
Admin
I agree with Bob. In addition, please specify the purpose of the gauge (i.e., inspection or process control). Also what type of process forms the thickness?
 
D

dbulak

The micrometer is used in production. The thickness is a steel product. I have attached a pdf file of the results. Any comments would be appreciated.
 

Attachments

  • Document (13).pdf
    51.2 KB · Views: 170

bobdoering

Stop X-bar/R Madness!!
Trusted Information Resource
The best way to interpret the data is to note that EV is "within variation" and AV is "between variation". That is a far better way to look at it than the more casual "appraiser" or "equipment"! So, your biggest issue is "within variation" at over 30%. If you have the AIAG MSA book, a good list of things to look at are on page 52-53 - potential causes of "within variation". Within-appraiser or within-instrument are quite common with micrometers, because of their lack of gage pressure control - as in appraiser "feel". There may be some awkwardness in trying to measure the specimen, which can also increase the error. That is why when doing a Gage R&R - especially one with issues - the gage professional should observe the entire study, looking for potential causes of error. Without knowing what the error is, you will not be able to eliminate it - either with another gage or a better measuring process with the same gage.
 
D

dbulak

We will be doing another study soon. When we do it should we be looking to get all the mesurements as close to the thickness we will be cutting by the operators. Would this give us better results?
 

bobdoering

Stop X-bar/R Madness!!
Trusted Information Resource
We will be doing another study soon. When we do it should we be looking to get all the measurements as close to the thickness we will be cutting by the operators. Would this give us better results?

Could you clarify what you are going to be doing to "get all the measurements as close to the thickness we will be cutting"? How is that different from your last study?
 
D

dbulak

First, we are measuring strips of steel. Second, what I meant by getting closer to the actual gauge was to make sure that range of readings was not too big.
 
Top Bottom