What diameter rod to start with, for a given sized thread die?

[Chewy]

What I find funny about this is the lack of specifications. MSC sells the 8-48 die for $28. Trout sells the 8-48 tap for $13-21. Brownells sells 8-40 tap for receivers for $29. All spec's in all my reference books from 1900 - 2018 stop at 8-36 NS. You can use formulas to figure out everything needed to cut and measure a thread. Funny how a commercially available product doesn't have published specifications.

 

[cazclocker]

To start with, depending on a "chart" is a PITA. My shop/charts are next door and I'm in a wheelchair. I do have a copy of Machinery's Handbook accessable from my desk, but I much prefer to have a good understanding of how any number is derived. That saves me a trip and I never need to depend on anything outside my mind.

In the case of machine screws, the basic concept is to start with 0.060 inch. That is size "0" or naught (zero). Each machine screw number adds 0.013 inch to the outside diameter. As in a #8 screw, a size I work with a great deal in electrical systems. The basic formula is: 0.060+(8 X .013), or .060+(.104), or .164 inch.

Machine screws go below 0.060 as well, down to the (metric) optical screws. Such as size 000, 3/0, a size that shows up on my brass models. 0.060 minus (013 per step) or 0.060-(2 X .013) or 0.060-(.026) or 0.034 inch. (3/0-120 TPI) Screw sizes get smaller, but I don't use them. 4/0 and 5/0 screws (and taps) are available from specialty suppliers. Metric screws are more common for the smaller sizes, a Nr4/0, at 0.021 inch is more commonly replaced today with a 0.5mm screw.

The depth of thread for 60 degree threads, from the metric system, is to subtract the thread "pitch" from the outer diameter. For a #8 screw, the "tap drill" size is the major diameter (0.164') minus the pitch, 1/32, 0.03125" or 0.13275". Lacking that exact size, using a 0.133" drill will suffice. That's where the "wierd" steps for number sized drills comes from. An 0-80 screw is (theoreticaly) 0.060 minus the thread pitch of 1/80, 0.0125", or 0.0475 (practicaly 0.048)" inch then becomes the tap drill for that size.

Thread pitch is a convenient source of calibration screws. Many lathes have 10TPI lead screws, giving 0.100 per revolution. A 20 pitch lead screw gives 0.050 per rev. A 40 TPI, such as a micrometer, 0.025 per rev. The next step down, 80 TPI, is seldom useful to most machinists. But does show up on some calirations. The same concept is often applied to fractional adjustments. Such as an 8-32 giving 1/32 inch per rev or a (#3)-64 TPI with a 1/64 per rev.

Some screw sizes are obsolete by today's standards. A #10-32 versus a 3/16-32 cannot be easily distingushed one from the other. #10 is 60+(10X.013) or 0.192. A 3/16 is 0.1875. If one goes to a hardware store for supplies, asking for one may well yield either. There are other curious numbers, a 1/4 inch screw is 0.250 inch. A #14 machine screw (now obsolete) is 0.242 inch. And metric falls so close it must be remarked on, a 6mm screw is 0.232 inch.

Bi1 Hudson, thanks for your comments. The info you provided is just the kind of background I like to learn. I never know whether I can trust a chart (at least the first few times I use one) so it's always best to get some background about how the data is derived. I'm just an old retired printing press operator who wishes he'd pursued machining for a living, but I never did. So here I am trying to learn stuff I know nothing about. I'm durned fortunate to find this forum with folks like yourself that have patience with folks like myself. Thanks very much.

I have very good results by the method of using nominal diameter less 10% of the thread pitch. Easy to calculate (or look up), and it covers everything you're likely to do. Standard threads, non standard threads, obsolete threads, and threads that you just made up on the spot. Obviously special threads (Acme, Buttress, etc) are different, but any conventional V thread that you can come up with, it works well. Essentially you "pre-cut" the peaks of the threads to finish size, before you run the die (or the single point cutter) along it. This is what they are doing on the Little Machine Shop link that ChazzC posted. You just have to remember (if you're single point cutting) that you're cutting the threads that are already "truncated", so you won't go fully to a sharp point, since it's already "trimmed". It also makes a die run a LOT smoother, and with a lot less cutting force when making threads that way.

Your comments answered a question about the Little Machine Shop calculator. I noticed that it told me to turn the rod down to an OD that is LESS than the major diameter of my target thread... I thought why would I do that? Now I see that the threads come out "pre-truncated" that way. I was doing some reading and seeing that thread standards don't have pointed threads - they're rounded at the tops and bottoms. I knew it was for fit purposes, but I didn't know math was involved. Thanks!

 

[Jake M]

I was doing some reading and seeing that thread standards don't have pointed threads - they're rounded at the tops and bottoms. I knew it was for fit purposes, but I didn't know math was involved. Thanks!

If there's a shape, there's math involved. It's really that simple. There's always math.

The standard thread profile for US threads isn't rounded, it's a flat crest and a flat root. BSW, or Whitworth threads have a round crest and a round root. Metric threads have a flat crest and a round root. In practice, it's very much allowable and in nearly any case preferable to round instead of flatten the roots, (as well as allow you to have a radius on your cutting tool instead of two hard corners). And crests cut flat work, fit, and function properly with no stress risers on that end. Rolled threads, which most of the threads that you see actually are- They take a LOT of liberties to get that process to work. If you accept that life isn't perfect, the crest flat/radius thing won't cause any stress concentrations in any way, and round roots prevent stress concentrations, you can kinda just make all your threads that way. So long as you get the dimensions right, and the overall thread engagement area stays the same (based on theoretical profiles), your stuff will work properly with everybody elses... Let me cite hardware grade nuts and bolts as an example. They are wrong in every possible way, but they work adequately, right? Key points right on, and you'll be good.

 

[Bi11Hudson]

What I find funny about this is the lack of specifications. MSC sells the 8-48 die for $28. Trout sells the 8-48 tap for $13-21. Brownells sells 8-40 tap for receivers for $29. All spec's in all my reference books from 1900 - 2018 stop at 8-36 NS. You can use formulas to figure out everything needed to cut and measure a thread. Funny how a commercially available product doesn't have published specifications.

8-48 is not a "common" thread, it is specific to a very few specific applications. By most of today's specifications, it is considered a "special" thread. I have run into this on occasion in the distant past when I was searching for 40 TPI taps of larger sizes. 4-40 is considered "normal", 5-40 and 6-40 are considered "fine" threads. Most anything larger is considered "special" threads and is difficult to easily find. 3/16-40 and 1/4-40 weren't too difficult, 3/8 up to 5/8 were. Some came only from China in recent years. I needed the larger sizes more to scratch a "personal itch" than for any real calibration builds. But I may one day, want to build a huge micrometer. . .

If there's a shape, there's math involved. It's really that simple. There's always math.

I tend to be a little more old school than my age (72) indicates. This is a "political" comment and may be somewhat out of line. If it is not tolerable, I won't hold a grudge if it is removed. I never attended high school, in my day there was no "middle school". I did attend the 8th grade, on occasion, but never had enough drive to attend full time. This was around 1964 when the Federal Govt got involved in state school systems, my parents supported me over the "official" school system. In my day, there was "Arithmetic" and then there was "Mathematics". In grammar school, basic Algebra (a^2+b^2) wasn't considered math. Trig and calculus was.

Today 2+2 is considered Math, when it is no more than arithmetic. Preposterous is all I can say there. Sort of like trying to call Electricity and Electronics the same subject. And yes, I am an EE, degreed by an offshore school. I did follow an electrical education as a child until the Feds started telling me what to study and what to ignore. For my own safety, if you will. From my own personal experience, I can only hold some government desk jockey in absolute contempt.

The standard thread profile for US threads isn't rounded, it's a flat crest and a flat root. BSW, or Whitworth threads have a round crest and a round root.

Most of the "flattened" or rounded thread forms came about when Whitworth, Metric, USS, and SAE were "unified" in form to allow better interchangability. The manufacturing standards were "loosened" to further allow for tool wear. "V-Sharp" threads are, even today, a valid thread form that is still used in some applications. But cutting a V-Sharp thread is a difficult and tool replacement nightmare. It is seldom run into outside of precision instruments. In most cases, Unified threads can be cut single point with a sharp cutter and then the outside diameter "flattened" a little with a file. That essentially makes the form a Unified Thread. The fastener can then be run into a Unified Threaded hole with no interference.

.