How to keep gage block sizes unknown for R&R factors of MU

[apestate]

Hello everyone

I know this question might be answered by study of my new AIAG MSA 3rd ed manual, but I thought I'd throw it out there for some plain talk discussion.

We have a problem with our internal calibration system. The problem seems to be that the amount of pressure to be applied to measurements is not defined well enough, because every time I pick up a micrometer it measures about +.0003" -- someone is measurement force biasing big time.

What would be the proper way to measure this variation? It's hard to do a GR&R on gage blocks, because the staff know exactly what size they are. It's hard not to use gage blocks, because that's exactly what we're measuring when we do calibrations.

Is the answer that we should use gage block stacks for R&R contribution to MU estimations?

Can gage blocks really be resurfaced so they'll wring, when wringability is bad?

What do we do about measurements over 6"? It's not practical to use an 81pc gage block set for anything over 5" in my experience...

This is a really discouraging amount of problems in calibrations. I would really appreciate any advice on how to fix these issues...

[True Position]

Hello Atetsade, long time!

My first question is why are you using gage blocks for GR&R? Isn't it the standard to use 10 different samples of a part so you have a random distribution throughout the tolerance, or at least throughout the variation typical to the part.

As to block resurfacing, it may be possible to lap them flat again, but you're talking about removing material, and hoping you don't remove too much to keep them in stock. If a chamois and the oil from your wrists aren't enough to get them to wring, it may be time to consider new blocks. One good quality set for calibration use only of a good brand, Mitutoyo makes a fine ceramic block I've been told; and if the operators are regularly using them, there's a number of very cheap chinese sets available, $100 or less for a full 81 pc set. You get what you pay for, but typically good enough for the tolerances you've told me you're working with in other threads.

For long measurements, they offer special long gage block sets, iirc my last supervisor kept those and he had a 6, 12 and 18 inch block set.

http://www.fvfowler.com/pdf/2304/2304_223.pdf

[apestate]

Thanks Hawat

That's probably the solution to the problem of micrometers over 5"

As for wringability, this 81 pc set has been in use for 7 years now. The main problem is with the corners on the ones that have been dinged, trying to wipe off the dent with a stone is bad practice in my opinion.

The 81pc set is up for calibration, I just realized. I can make the case for getting a new set and holding this one in spare.

The reason for repeatability and reproducability on gage blocks is because the people calibrating micrometers are missing the mark. The measurement is made on gage blocks, and I'd like to show that we're missing the mark and by how much.

If I go to the cabinet and look at a micrometer right now, it is likely to show +.0003" because someone is heavily biasing the measurement during calibration by cranking down too hard.

How this is going on I have no idea, but to address it, we should first measure the problem. That's why I was asking about how to keep gage block sizes unknown for an R&R study, which should be possible by incorporating various 1/10,000 sized blocks into the stack when we run the test.

[True Position]

Actually, removing dings/knicks with a stone is common:

'Minor burrs can be removed by applying a fine-grain dressing stone, made of some natural rock, such as "Arkansas" stone or black granite' - Handbook of Dimensional Measurement, page 71.

My recommendation is to remove the old set from the system, make it reference only or lock it up in non-conforming, and just use it to replace broken blocks by sending the individuals out for calibration as you need them. Otherwise you're risking a huge NC to an auditor.

The issue of incorrect force isn't an issue of GR&R, the gage is repeating fine(I'd assume), it's just set wrong. Your problem is in your calibration system, not your GRR. How many people do you have calibrating micrometers? Would some training be impossible, perhaps friction thimble micrometers?

[Claes Gefvenberg]

In Reply to Parent Post by atetsade

We have a problem with our internal calibration system. The problem seems to be that the amount of pressure to be applied to measurements is not defined well enough, because every time I pick up a micrometer it measures about +.0003" -- someone is measurement force biasing big time.

The sad truth is that a depressing percentage of micrometer users do not know how to properly use the instrument (As an example, many will simply turn the screw too fast, thereby achieving too high a measurement force). Even worse: most of them will keep telling you, in no uncertain terms, that they do know... until you force the issue by testing them, that is. Have a look at this old post.

In Reply to Parent Post by atetsade

What would be the proper way to measure this variation? It's hard to do a GR&R on gage blocks, because the staff know exactly what size they are. It's hard not to use gage blocks, because that's exactly what we're measuring when we do calibrations.

But you can use the blocks... Put a combination together: That will hide the figures on the blocks, and a digital micrometer, connected to a printer and with the display taped over, will complicate matters even further.

/Claes

[Ederie - 2007]

I agree with Claes,
Don't let them know the size of the blocks or let them see their readings.
That should get them to learn to go by feel.
Ed
Throw in some .1003's

[Dave Dunn]

In Reply to Parent Post by atetsade

Thanks Hawat

The reason for repeatability and reproducability on gage blocks is because the people calibrating micrometers are missing the mark. The measurement is made on gage blocks, and I'd like to show that we're missing the mark and by how much.

If I go to the cabinet and look at a micrometer right now, it is likely to show +.0003" because someone is heavily biasing the measurement during calibration by cranking down too hard.

How this is going on I have no idea, but to address it, we should first measure the problem. That's why I was asking about how to keep gage block sizes unknown for an R&R study, which should be possible by incorporating various 1/10,000 sized blocks into the stack when we run the test.

This seems more a training issue. Why not set up a training session with all the personnel in question to go over proper measurement technique? Give each person in the group a gage block of any size, and a properly zeroed micrometer. Tell them to measure the block without rezeroing and note the reading. After that, go into the explanation of how measurments should be taken with a mic.

[Hershal]

The situation is common. Claes is correct, this is at least in part a training issue.....and only one of the reasons I suggest having metrology professionals do the calibration, just as a machinist would be employed for a CNC machine.....

However, your situation requires some immediate work.

Train your folks in proper use techniques as Claes suggests, calibrate the gage blocks (don't forget to specify the accreditation logo and specific uncertainties in your P.O.), and work the MU.

When you wring blocks, remember that requires an additional entry of 0.00005" divided by square root three (1.732) for EACH wringing and each is its own Type B entry, plus the other Type B contributions, and the Type A also.....oh and gloves only slow down the thermal transfer to the gage blocks, they do not prevent it.....

No need to prevent the folks from seeing the numbers.

For longer measurements, say over 4", acquire a 6" block. They are not cheap but will give you what you need without wringing.

Hope this helps.

Hershal