Good morning!

The quality staff at my new job is using a recent GR&R from one part number as proof that the gage is repeatable and reproducable for other part numbers. The thought is that once you prove the gage repeatable/reproducable, it will be so for all parts using that gage. I need proof whether this is true or not.

I was always taught to perform the GR&R on the part that the customer is requesting the study for? I'm getting a bit of resistance from the quality techs. They want to know where in the MSA manual does it says a GR&R has to be done on the part in question. I can't answer that one.


:thanx:

Paul

It depends on how similar the part are.

- are the characteristics being checked the same or very similar? if so, might be ok, depending on next question
- is the tolerance the same? if tolerance on GR&R is looser than on others, then it wouldn't qualify

The dimensions are similar and the tolerance is the part studied is much tighter than on the part not measured. Does the MSA Manual spell this information out?

I am not saying this is correct, But I could not find anything to prove me different and it hasn't been questioned on a audit.

I do a GR&R on every SCC item for all new parts. (CMM runs these parts when they are in PPAP process) I group gages as families and choose the tightest tolerance to complete the GR&R. This allows me to do one GR&R on one set of 6" calipers say for a bushing step depth. Our internal auditor at one time tried to get us to do a GR&R on every part, every hand, every characteristic, and every gage being used, YEARLY! Of course this was shot down because I would have had to hire about 4 extra people just to do GR&R's full time. We have close to 27 different parts and each part has a left and right hand. Some of these have up to 20 characteristics that are tracked per hand. As you can see that alone would be a great undertaking not including I have 2600 active gages in the system.

So that is how I handle the GR&R's and it is working well. Someone else may be able to give you the exact wording on what the MSA manually states and how it is interpreted.

Good Luck
Dale

MSA = measurement System Analysis

the system includes part feature, any fixturing and the operators. If it si measuerd using a piece of equipment (CMM) as opposed to a handheld (calipers) the system drift over time must be included. It also includes the tolerances as stated above.

IF the part features are very similar and the operators are the same I perform a generic R&R. if not I wouldn't.

Bev,
I would like some more information on how you decide which parts can be measured on the CMM. I would love to be able to have my GR&R done by the CMM operators and leave the Quality techs to do their job of watching the floor. I just can't justify having the GR&R not being performed by the people that use it and in the environment that it is normally used in. You probably have other circumstances that make it possible to have the GR&R done in the lab. Any information you would be willing to share would be great.

Thanks
Sean

Bev,
I would like some more information on how you decide which parts can be measured on the CMM. I would love to be able to have my GR&R done by the CMM operators and leave the Quality techs to do their job of watching the floor. I just can't justify having the GR&R not being performed by the people that use it and in the environment that it is normally used in. You probably have other circumstances that make it possible to have the GR&R done in the lab. Any information you would be willing to share would be great.

Thanks
Sean

Sean - I only have GR&R done on a CMM when it is the gage of record. There are dimensions that can only be measured properly by a CMM.

If a hand gage is your gage of record for product acceptance then it will require a GR&R if your custoemr requires it.

If the part to measure by gage is much similar to the part which has done gr&r, I think, it's ok to do as that.

IF the part features are very similar and the operators are the same I perform a generic R&R. if not I wouldn't.

The GR&R should use a representative skill level of any operator that may use the device, so that the results do not need to be repeated if there is a specific change in operators. The point is correct that the part tolerance should be equal or tighter in the GR&R specimens and general geometry should be similar to be used on other parts.

The GR&R should use a representative skill level of any operator that may use the device, so that the results do not need to be repeated if there is a specific change in operators.

I think that in some cases this might be a form of circular reasoning; part of the purpose of MSA (of which GR&R is one form) is to verify the skills and continuity of operators. How can you objectively determine a "representative skill level" without actual tests?

I think that in some cases this might be a form of circular reasoning; part of the purpose of MSA (of which GR&R is one form) is to verify the skills and continuity of operators. How can you objectively determine a "representative skill level" without actual tests?

GR&R is a powerful tool - it can collect a vast number of variables. Sometimes too many, if used without an understanding of its power. It can also be used ad infinitum by testing every gage and every individual on every part until you are doing more gage testing than part measuring. The typical point of its use is to verify that the gage is the correct gage for the job no matter who is expected to use it. For that, I typically use a newer operator, a seasoned operator and a qc lab tech as the individuals to provide a varied enough skill set to make the point. If, for example, it is a specialized gage that only lab techs use, then I only use lab techs.

If you are using gages that are so operator sensitive that each operator needs to be tested -and retested over time for continuity - then first I would look for a more robust gaging system. But if not an option, GR&R will be the tool to verify their skill (compared to others).

The typical point of its use is to verify that the gage is the correct gage for the job no matter who is expected to use it. For that, I typically use a newer operator, ad seasoned operator and a qc lab tech as the individuals to provide a varied enough skill set to make the point. If, for example, it is a specialized gage that only lab techs use, then I only use lab techs.

I agree. :agree: I think I might have misinterpreted you originally.

.... I just can't justify having the GR&R not being performed by the people that use it and in the environment that it is normally used in. You probably have other circumstances that make it possible to have the GR&R done in the lab.


No, the GR&R must be performed by the people who actually use the gage system, or it is not a valid test. And as Bobdoering points out, better to test the weaker users, to determine the maximum variation. In fact, I would select two weaker users, and only one expert tech as the control.