Part One of Two Parts: The Study
This question is prompted by another round-robin type study involving the calibration of thread plug gages, the results of which were predictable because similar studies over the past twenty years or so have produced similar results.
To keep the record straight, you should know that I have been involved in such studies that were conducted by the American Measuring Tool Manufacturers Association. I participated in the preliminary discussions for this one involving the Boeing Company, the Industrial Fasteners Institute, and the Chair of the ASME B.1 committee that writes/reviews the related standards such gages are made to. I also confess to being a member of some of these committees.
In the beginning, the study organizers had hoped to complete it in less than a year but I offered an opinion that with about fifty participants involved it would be more like two years. I was wrong, of course. It took nearly four years with the results coming available in late 2014.
Unlike other such studies, this one looked at more than just simple pitch diameter and major diameter. It included linear pitch or lead as well as thread angle. The organizations participating included military, aerospace, gage manufacturers and independent calibration companies. All in all, a group of people who take their metrology seriously many of whom have above average equipment and expertise.
Like similar studies, participants were given detailed instructions respecting where on the gages the measurements were to be taken so that gage geometry problems such as taper didn’t mess up the data. NIST participated at the beginning and end of the study to ensure the gages had not changed during the process and to also provide some references for evaluating the results.
The final report recognized that a similar study based on the equipment types used would be a worthwhile study. It could explain some of the variations noted, but might skew the numbers another way. Since I am not a mathematician or statistician, I won’t wade into those mysterious waters.
The study allowed participants the option to use the hardware of their choice, so several general equipment types were represented, which is a good thing. Why? Because it gives those who use outside calibration sources an idea of what to expect as far as results are concerned. It can also be a good guide if you happen to be involved in a dispute over measurements involving such gages.
Participants in the study were assigned a number so embarrassing results wouldn’t haunt them forever. Some would argue that this protects the guilty but even I wouldn’t say that. It does ensure more companies take part and follow the rules, which makes it worthwhile. Comparing their results with such a large number of participants gives each one a good idea where they may need to make changes to their procedures to improve their overall performance.
The results from this study with all the data is over ninety pages in length. I found the results interesting even though the data was massaged to eliminate ‘outliers.’ ‘Outliers’ are generally values that are so far out in space they lack any credibility compared to the majority. Treating them this way is common in such studies and usually means the summary of the data looks much better without them.
Folks who send their gages out for calibration don’t have the luxury of being able to compare their source to so many others and may send their work to a space cadet without knowing it. This is why I like to look at the raw data from reputable studies such as this one that provides it even if the participants are not identified.
If you look at the numbers from Lab #XXA and see their numbers are all quite a bit off, you may conclude they are beyond ‘outliers’ status and are probably into ‘outlaw’ territory. If only one of their numbers is off, it may be a one-time glitch. Analyzing the “why’s” becomes an interesting exercise.
In my next column I’ll show some of the numbers from the study which may cause some folks to review their thinking on this specialized area of metrology or call their pharmacist for relief—or both.
To be continued...