Maybe I’m the last to know, but I just found out that the nominal outer diameter of a gauge-numbered machine screw is defined as the gauge number multiplied by .013″, plus .060″. The actual diameter is usually two or three thousandths or so under nominal. I know ’cause we tried it. And as you are all know doubt aware; once you reach a quarter inch, you’re going by fractional inch dimensions instead of gauge. Wood screws go by their own, as yet mysterious to me, system, probably developed by some guy and his partner making screws by hand 250 years ago.
Who cares? Well, we have run into problems with what we refer to as “stacking tolerances” in our production– a threading tap varies slightly (both initially and over time with wear) the anodizing depth varies slightly, and screw dimensions vary slightly even if you stick with one supplier. If these variations all go in the wrong direction at once, you end up with customers calling you saying the screws are so tight in the mount that some of them are breaking, even though you’ve been doing everything exactly the same for years and it’s always worked nicely. We started using +.001″ and +.002″ oversized form taps a few years ago, to make up for the thickness the anodizing adds to the threads, and then some, and the problem went away. Now at least we can measure screws and know exactly how they vary from “nominal” as opposed to making simply comparative measurements.
This new (to me) tidbit of information is just icing on the cake for you engineers out there, in the unlikely event that you were as ignorant of such things as I was a few minutes ago. What I still don’t understand is why we call a number eight screw a number eight screw instead of a .164″ screw. Too many digits? But then you’d not have to remember gauge x .013″ + .060″.
Some of these oddities come down from the past in “organic” ways. Firearm bullet and bore diameters are a good example. Who the hell came up with .223, .308 or .452, as opposed to, say .200, .250, .300 .350, etc? Some of these unlikely numbers, at least in part, come from the days of black powder, wrought iron barrels, soft lead bullets, and the manufacturing tolerances of yore. The realistic tolerances back then were nowhere near what’s possible now, and it resulted in some pretty weird numbers that became standards out of expediency and in response to backward compatibility issues. I use a .454 ball (that number’s still with us) in an 1850s .44 percussion revolver for example, because the oversized ball gets better purchase on the sides of the chamber and on the rifling. We would now refer to a .454 bullet as caliber 45, though you were shooting it from what was called a .44 caliber pistol back in the 1860s, and the modern 45 cal bullets are .451″ and .452″. Modern 44 caliber bullets are .429″. Huh? I definitely need to learn more about this stuff. In another .44 percussion revolver I have I use a .457″ ball– you want a ball that’s bigger than the cylinder, and a cylinder that’s bigger than the barrel groove diameter, so everything gets a sure, tight fit with the soft lead ball.
We still use grains as a unit of measurement, which came from some king somewhere telling us that the official definition of a pound was “seven thousand plump grains of wheat” (what poor saps had to count them, then recount them, and who verified their work?). Shotgunners use the dram, which converts to the tidy number of 27.34375 grains, or the “dram equivalent”, which is a charge of modern smokeless powder that generates about the same energy as that number of drams of black powder.
If we were to start all over and reinvent guns from the beginning today, we’d no doubt end up with simpler units and numbers, but the world doesn’t work that way. Each incremental development is built upon the previous one, and you don’t immediately re-tool everyone in the business, make all the old versions unusable, and change all the established experience and data, just for that little increment of improvement.
Still, I keep saying someone needs to reinvent the computer OS (or the very concept of the computer OS– maybe the very use of the term “OS” is thinking too much inside the box) from the beginning. There is of course no basis– no established school of thought or system of evaluation that would warrant such a claim.
Choose taps with a bigger “H” number if you’re anodizing tapped holes. Each .001″ of anodize thickness builds up app. .0005 on a surface (.001 split between penetration and buildup). But on a 60 degree thread the effect for .001 thick anodizing is .002″ pitch diameter reduction (or enlargement on an O.D. thread).
Use an H5 tap for the above example for a pitch diameter +.002 to +.0025″ oversize and let the anodize get you back to your preferred size.
If you made everything simple and rational, the publishers of Machinery’s Handbook would lose a significant amount of their income and engineering schools would have no way of confounding their students.
On the other hand, when was the last time you used (or even heard of) a “swinging choke” in a power supply?
I don’t know what you are making but have you tried looking for a precision screw manufacturer that is ISO-9000 and QA’s for tight tolerances? Also, have you thought about going metric instead of fractional?
There are Swiss companies that manufacture screws to extremely tight tolerances. Here’s one in the USA: http://www.americanswiss.com/
See what Lyle makes here.
ubu52; Our business philosophy is built around simplicity and convenience for the end user. On that note, we use the most common threads, available in virtually every hardware store in the country. If someone looses a screw (well, first we include one extra of each type) they can call for free replacements, or if they prefer they can walk into a local store in Anytown, USA and be assured of finding one exactly like the original. Thank you for the link, but sourcing an extremely consistent screw would be convenient for us, and less convenient for the user. What we’re doing now with the oversized taps has been working very well for several years, plus we now know what to look out for as we go along. I will check out Americanswiss to see about getting some better Allen wrenches, ’cause that has also been an issue, most especially in the smaller sizes. Thanks again.
Sennin; I’d never heard of a swinging choke, though I have built several power supplies. Some of them used capacitance alone and others required no filtering at all. I’m now in the process of reading about it. Interesting stuff. I’m still reading about the applications of this “swinging choke”.
This is what I like about you all- I show my ignorance and you come back with information. Thanks.
Ah, the memories. I can’t tell you how many tables of thread, pipe, and o-ring sizes I memorized over the years just to make my job simpler. Stack up is always an issue. You can make your tolerances tighter, but then the parts may fit together looser (than the worst case before) and still cost more (exactly counter to what someone would want).