Are Thread Gage Tolerances Practical?
Part 2: The Numbers.
Editor’s Note: This is part two of a two-part series.
My last column gave the background to an extensive study of results on the calibration of some thread plug gages. As noted last month, the study results were similar to those I’ve encountered before with similar tests. Where measurement uncertainties are quoted in the following notes, remember they are all plus/minus values. As you read the numbers please remember that I am only reporting them—I didn’t create them—so send your emails of outrage or disbelief to someone else.
This should be a no-brainer; after all, there are six ways to Sunday this simple outside diameter can be measured. However, the study reports the uncertainty for it at 57 micro-inches with the outliers removed. The range of the reported data including the outliers was .000290” for a diameter with a .00005” tolerance. One often overlooked reason for this may be due to some labs measuring it using the same measuring forces required for pitch diameter measurements while others did not.
The report shows the uncertainty for this measurement at 97 micro-inches or .000097” again with the outliers removed. This level of uncertainty comes up in test after test under similar conditions. The range of measurements on one of the gages for this was .00119” or more than three times the tolerance of .0003” for this element. At 97 micro-inches, the uncertainty is about 33% of a class ‘X’ tolerance, quite high if you’re disputing measured values close to the gage size limit, worse if it’s supposed to be a class ‘W’ tolerance which would be about .00015. In either case, don’t bet the mortgage money if your readings suggest a rejection. There are a number of reasons for whichever number you’re looking at—something I’ll comment on another day.
Some of the labs reported 20 micro-inches as their uncertainty for this measurement which indicates they need to re-think how they arrived at a number so much better than NIST. Four participants reported no uncertainty in defiance of all the laws of physics and metrology. To be fair, I don’t know whether they didn’t know what it was or didn’t believe they had any.
This is a much more difficult measurement to take and the study results show this with the uncertainty at 199 micro-inches or about .0002” over five pitches of thread. I’m surprised it is that good given the equipment that is often used. Since the tolerance for this feature is .00015”, you’d have a problem in a dispute. The range of readings was .0011” approximately, which doesn’t instill much confidence in the process either. Interestingly, eleven labs did not provide readings for this element even though they are calibrating thread plug gages in their normal activities.
Another tricky measurement with a reported uncertainty of 20 minutes with the outliers removed. The tolerance for this is 8 minutes or .133°. The range of readings: 1.220°. Once again, check your weapons if you are going into battle over this measurement on a gage. Eight labs did not report this feature I assume because they can’t or won’t measure it. This, along with a similar situation arising over pitch/lead measurements makes me wonder how they can claim to be calibrating thread gages. Four of those who did the measurement omitted a value for their uncertainty.
I believe this brief look at the study results answers the question that leads this column. The uncertainties noted for the various elements that make up a thread gage are quite high in relation to the tolerances they are supposed to comply with. This means that arguments that can’t be won will continue to plague the industry.
The next question has to be: What do we have to fix? Considering the fact that the uncertainties from this study are similar to those done in Europe, it would seem the metrology is unlikely to be dramatically improved any time soon. That leaves the tolerances for such gages to be considered. The European tolerances for such gages are more in keeping with calibration capabilities. It would seem our tolerances would benefit from a similar study.
The folks who put this study together and saw it through to completion have provided a great service to everyone who calibrates thread gages. It’s up to everyone to keep these realities in mind when the millionths or microns start flying.