Articles about CMM vs. Hard gaging?

[Sl33stak]

Ok, here we go again. My shop is now wanting go back to "master" gages to control locations and diameters. I can understand diameters, but it would seem that locations (TP) would be better measured/tracked with the CMM. Does anyone know of any articles duscussing the case FOR CMM use?

[#2#]

Well, a hard gage will only give you one of two results. Good or Bad. A cmm will tell you if it is good or bad and how close to being bad it is.

I believe in statistics it is refered to as attribute data (good/bad) or variable data (numbers). It is hard to analyze attribute data other than we have this many good and this many bad. With variable data you can do a lot more.

HTH

Sorry, no articles that I know of but they are bound to exist.

[Jan d.]

I am not going to be very helpful, but we do have lots of experience here with this issue. Position is something a CMM is VERY good at measuring. However, it MUST be programmed right.... and there is where many times the problem lies.

The biggest problem is how EXACTLY you have programmed the features and how EXACTLY you evaluate. If you have a cylinder, do you use that cylinder to evaluate TP or just 1 circle? How well are your datums established?

Examples of problem parts: if you have a part that has datums that are real close together, and then you want to evaluate TP of a hole that is far away from that datum structure, you can expect lots of corralation problems. Small variations in establishing the datums can have HUGE effects on the TP of the to-be-evaluated feature. Even hard gauges are going to mess up.

Other problems: the datum plane is measured with only 4 points. Make sure you cover the COMPLETE surface, with as many points as time permits. You'll see you data improve.


Jan.

[kbotta]

The big "plus" when using a (correct) hard gage, is the establishment of the datum reference frame (DRF). With the correct components used on a hard gage you would have true geometric counterparts (TGC's) that actually exist. These "Datum feature simulators" pick up the Actual Mating Envelopes (AME's) and provide a fundamental approach to inspection. Conversly, one would need to "create" these "TGC's" via a CMM (to get the AME's), and that can be very tricky, if not almost impossible in a lot of cases.

[Start Soap box:::::]
One of the biggest pitfalls of the CMM (and programmers too!) Is the lack of understanding of the DRF, and how to correctly simulate one with a CMM.
[end soap box::::]

ASME Y14,5M-1994 in para. 5.3.2.1 has the explanation of positional tolerance at MMC.
Note the "note" below section 'B'. It says that you can think of positional tolerances in two ways. Interms of the axis (enter CMM), or interms of the surface (hello hard gage).
If there is a descrepency, THE SURFACE INTERPRATATION ALWAYS WINS. Interesting note I'd say...

As far as books, I'm sure you could google (verb) your quarry.

HTH...
Kev

[MrComment]

I agree with Jan, and Goodluck, but I'll also add that each gauge has its strengths and weaknesses, and when used correctly and appropriately both can be helpful.

Hard gauges are useful for spot checking parts quickly and occasionally but only when good or bad is all you need to know. Also they can become a rather costly to build if your only checking a few parts.

Of course if you need to move something or see how parts deviate over time you'll need more resolution then just Good or Bad, but again not always is a CMM the best machine to use.

A CMM does seem IMO to adequately fit most general applications Ive seen and at my shop we have both hard gauges on the plant floor and programs for the CMM in the lab as well. We are also making the same wiggits everyday so economically speaking this was decided to be the best way to go. Obviously this approach my not going be practical for everyone.

Anyway thats my 2.5cents (inflation).

[cmmguy]

There are times when a fixed limit TP gage can do a better job than a CMM... Because bestfitting routines dont always represent the part perfectly.

But as pointed out, they are good for product acceptance only, not process control.

------
dang, I pause for a second and two posts are ahead of me. WKS

[Sl33stak]

I guess it all comes down to how good the hard gaing is too right? I mean, should the gage tolerances match the part print tolerances? all our gages are made "in-house" and interestingly enough, half the gages have non-matching tolerances. I guess I am spooked about us going back to the old gages because I wonder what I will spend my time doing. We only do so much new product a year...unless I can look forward to s4it tons of capability studies...
yay!!!(sarcasm)