Hi, I am new to this Forum. I am also new to conducting GRRs. We mostly machine graphite with some steels and ceramics. Many of our parts call for tight tolerance (+/- 0.001 in or +/- 0.0005 in) from either our design engineer and/or our customers.

I have completed a few GRRs on tenth mics, mics, and calipers measuring the OD of a few different graphite parts and one stainless steel part. I am disappointed, not entirely shocked, by the results on the graphite parts, any where from 30%RR to 80%RR.

I adjusted my instructions in measuring the OD to specify a location where the parts where measure with no improvement. I then moved to testing mics on a metal part. My result where much better 1%RR. That at least told me that the gage (measuring tool and people) can perform.

Now I dont know what to do exactly. As far as I can tell the parts are mostly working for us and our customers. So, how could I justify buying gages. I say as far as I can tell becuase I have some hints of problems, but no real data on that yet.

Does anyone have experience with a similar material and gaging? Any suggestion on how to improve the gauging or types of gauges that maybe better?

Or, maybe it is more appropriate to prove that the parts need to be designed with a larger tolerance in mind. If so can anyone suggest how to calculate the tolerance range that is appropriate with the current measurement system? For this material and measuring tools.

Also, where is the best place to buy the MSA guide or any other good books on this subject? I need to review the nuts and bolt of this stuff.

Hello micsim:

Welcome and thanks for the interesting post.

With such a tight tolerance (comparatively), perhaps you could provide a bit more detail. Is there any other geometric feature, like roundness, cylindricity that has to be met too? What machining processes are you using?

In my experience, measurement of diameters to such close tolerances is susceptable to variation due to the characteristics imparted by the machining process. For example, grinding often causes lobes to form which can, as you know, cause different results.

It could be that you will have to use a comparitor and vee block type measurement system, between centres etc. to give you the accuracy you require. It also occurs to me that you should be aiming for a 4:1 ratio of the discrimination of the equipment, so if your tolerance is +/- 0.001 and +/- 0.0005, you should be using equipment which can resolve to 25% of these values (don't ask me to do the math it's too early!!). That's likely to take it out of the range of hand held micrometers/calipers.
:2cents:

micsim,
Since your results did not improve when you specified a location on the part to measure, I would think that lobing or other within part variation is not the issue. I suspect that differing gage contact force is causing differing elastic deformation of the part. Have you tried measuring with an indicator, or a "supermike" which uses a standardized gaging force?

Geoff Withnell

To AndyN,
These parts are mostly cylindrical and machined on a CNC lathes, some on 4 axis VCMs. The reason for the tight tolerances is to ensure concentricity or meet a TIR on a final assembly. Some situations, they are used for a press fit. In many cases these press fits will be sanded down by our operators to make assembly easier or possible (obviously this will bring them out of tolerance*rework*). Additionally, our machinist, in some of these situations, targets the lower or upper end of the spec to make assembly easier. The measuring instruments currently being used meet the 10:1 rule (tenths mics 0.0001 which is 10% of tolerance range +/-0.0005). Also, I am not sure if lobing is an issue, but I specified the location on the diameter to avoid this within part error on my study. I think what I am seeing is that we really cannot measure these parts and/or hit the tolerance on the prints. You cannot hit what you cannot see or atleast not all the time.

There are two choices find better gages and bring machining process into control which will reduce rework/scrap. Or live with the tolerance band we are/can machine to and make sanding(rework) part of the process and reduce some of the unnecissary inspection.

To Geoff,
I agree, the large variation in readings are probably due to a material surface related issue. I will try the gauges you suggested and look into some others. However, in the long run I will need to find a gauge that can be used quickly and in process.

What is a supermic? I cannot seem to find info on them. The mics we are using read to ten thousandths and have a rachet to improve the repeatablity of pressure, but does not seem to be working well on graphite.

macsim,
I'm sorry, the correct terminology is "indicating micrometer". these are hand held device that could be used on the shop floor. There is a disussion of them at

http://www.deterco.com/products/Mahr%20Federal/Gaging%20Tips%20Articles/O_HandTools.PDF Section M pg 6-7.

Mitutoyo among others has these gages. They really do improve operator repeatability.

Geoff Withnell

Geoff, thanks. Funny I was talking with a maintance guy here, who had experience at another machineshop where they used indicator mics(supermics). I will look into them.

I have seen over 100% of the tolerance used by micrometers in R & R studies but if 30 - 80% is the best you can get, then you will have to try to develop a variable checking fixture with a setting diameter.

I would love to be able to orient the part in some way but if that is not possible the a fixture with a V block holding the part with it butting up on one end. Then a digital indicator on a swing arm coming down in the same place all the time. One would need a setting diameter to zero off the indicator.

Micrometers are so subjective. Do we use the thimble at the end. If so, how many clicks. Most people feel it rather than use the thimble and where we take the reading must be the same each time.

Hope the thoughts help out.