Thread milling vs single point threading

[Conrad_R_Hoffman]

There is a fundamental difference between the various thread generating methods. When you single point with the top of the tool level, you get an Archimedes thread. In cross section the thread flanks are straight lines, just like the sides of the cutting tool. If you tilt the cutting tool to match the helix angle, you get something slightly different. If you cut the thread with a V-disk tool (round cutter or grinding wheel), you get an involute surface (I think), with no straight lines no matter how you cross section it. This is discussed in Vogel's book, The Exact Overwire Measurement of Screws, Gears, Splines and Worms. I've never seen it mentioned anywhere else, but it would have to affect the contact area between a screw and nut, depending on how both were generated.

The only advantage I can see for thread milling on the lathe is that it's easier to make a clean stop without an undercut, or if you want to do a Higbee cut.

[Harold_V]

If you cut the thread with a V-disk tool (round cutter or grinding wheel), you get an involute surface (I think), with no straight lines no matter how you cross section it.

Not true, although I have NO experience on thread grinders. I am open to correction if what I say is wrong.

The wheel head is tilted to the helix angle, and is dressed @ the angle, so the wheel generates a true flat surface on the thread flanks. Thread grinding is the method by which precision thread gauges are produced.

H

[Conrad_R_Hoffman]

I've been wrong before... just ask my wife! But, I'm home now and can copy from page 34 of Vogel's book on V-disk tools like grinding wheels-

"Screws generated by such tools have in common the fact that, in general, their surfaces do not contain any straight lines or any curves which could be characterized geometrically or analytically in a simple manner (see refs. 23 and 26). For all practical purposes, including their over-wire measurement, we may consider them as screws of unknown character."

23 seems to be a gear hobbing book and 26 a German math book, neither of which I own. The Vogel book is quite remarkable. It constitutes a quarter century of research by Vogel, done for the Van Keuren company. I believe it was also instigated by increasingly precise needs of the military. There's a cutaway illustration in both the Van Keuren catalog, and Vogel's book, of a big V-disk generated screw, where the cutaway shows the curved surfaces in several different planes. If you have a Van Keuren book #36, one of the more common ones, it's on page 199, Appendix D, where they summarize Dr. Vogel's work and give not quite enough information to really put it to use!

FWIW, I think the reason for the curved surfaces has to do with length of the cutting surface. Imagine an infinite diameter grinding wheel, essentially a big V, removing whatever it touches as the threading proceeds. More complex than a simple 2-line contact and sort of a relative of gear shaping, at least math-wise.

I have a 3-wire utility on my website using the Vogel formula that one can use to check other methods- http://www.conradhoffman.com/vogel_3_wire.zip

[Harold_V]

I think I understand where you're going with this. The larger the wheel, the greater the error.
I suspect that the degree of error is miniscule, likely in the realm of a few millionths. I'm not smart enough to run the figures.
This is akin to the idea of generating a helical gear with a milling cutter without the benefit of a universal mill.

H

[Conrad_R_Hoffman]

AFAIK, for the smallish single start threads we normally use, it's not of concern, save for the fussiest of fits in instrumentation applications. Where the error becomes significant are multi-start screws with significant thread depth compared to the diameter. The usual 3-wire formulas can give large enough errors to cause trouble there. I just find the whole topic fascinating and am amazed that nobody ever discusses it. Turns out some threads don't look like the ISO standard! Much of the supporting math in Vogel's book is beyond my abilities, but I can understand enough to make use of it. The external thread is well enough understood, but I don't think we know enough about nuts. I don't know what kind of thread a tap gives and suspect it doesn't reproduce the method used to make it. Ah well, trivia for the day!

[pete]

A very interesting detail and that's an effect I'd not read about anywhere else Conrad, it now makes sense the way you explained it, many thanks. If any company knows what there talking about then Van Keuren or anyone good enough to be doing research for them sure would. For my lathe, it's own built in lead/lag errors in it's lead screw are most probably multiple times more inaccurate than whatever method I have available to generate any thread. I'd sure like to know how they compensate for it when grinding high accuracy thread gauges.

If you haven't read it yet? There's supposed to be an online PDF out there somewhere of the Moore Tools book Foundations of Mechanical Accuracy. There's some extremely interesting details of the lengths Moore Tools went to in an effort to produce error free feed screws and nuts for there jig borers and grinders prior to the invention of high accuracy dro's. I just tried a search and it was easy enough to find, try this link. https://archive.org/details/Foundations ... alAccuracy And not to throw your post off topic but fwiw, Wayne Moore's diagrams explaining the possible inaccuracy's that may or may not be present in any machine tool slide was in my opinion a large step up in my understanding of machine tool accuracy. Even more so if your trying to improve what you already have.

[Conrad_R_Hoffman]

I'm familiar with the Moore books. I've got another good one but have to check at home for the exact title. Might be the Slocum Precision Machine Design book.

Yes, my leadscrew and everything else, being make in 1947, is probably a larger source of error than anything else. Buried deep underground near here, in Rochester, is the Richardson Grating Lab. The ruling engines for gratings are pretty amazing and the work that goes into the leadscrews is probably the fussiest in the world. Some interesting stuff here- http://loke.as.arizona.edu/~ckulesa/bin ... ndbook.pdf but I know there are many more books about the engines.

[Wolfgang]

Hmmm, interesting topic!

To summarize, and please correct me if I am wrong, a standard thread turning tool with straight sides and appropriate clearance angles and no top rake, will produce a true straight sided thread. Utilized in the usual manner: The work piece turns and the tool proceeds from the right to the left, producing a right-hand thread.

The same tool used as a fly-cutter in a motorized spindle on the lathe, proceeding as described above but the tool turning on an axis parallel to the work, would this not produce a straight-sided thread?

The fly cutter becomes a multi-tooth milling cutter, with each tooth identical in shape as the fly-cutter or normal thread tool and turned on an axis parallel to the work and proceeding as described above, would this not produce a straight-sided thread?

Since the last paragraph above in essence describes the thread-milling operation, does thread milling produce a straight-sided thread?

Just curious, and thanks for your response. w

[Harold_V]

The complication revolves around the contact of the circular cutting tool as it relates to the helix angle of the work piece. The larger the cutting tool (diameter), the greater will be the contact beyond the centerline of the work piece.

Conrad's states

Where the error becomes significant are multi-start screws with significant thread depth compared to the diameter.

Fact is, multi-start screws suffer from helix angle (lead), not depth of thread. The reduced depth actually works in one's favor.

As I stated, above, the greater the helix angle, the greater the error in form, due to contact beyond center of the cutting wheel of choice. The smaller the diameter, the lower the error. That's true even when the cutter is set at the helix angle.

H

[Conrad_R_Hoffman]

I don't know much about thread grinders but, correct me if I'm wrong, isn't the wheel size usually pretty large, like 10" or so? Or do they vary all over? So, if you thread mill on the lathe (or mill), the cutter is probably much smaller?

[Harold_V]

I don't know much about thread grinders but, correct me if I'm wrong, isn't the wheel size usually pretty large, like 10" or so? Or do they vary all over? So, if you thread mill on the lathe (or mill), the cutter is probably much smaller?

I have no thread grinding experience, but your assessment of the wheels being larger than the typical milling cutter is correct. I suspect you are spot on with the concept of 10" diameter wheels. If memory serves, they use resinoid bonded wheels in lieu of vitrified wheels, so they operate at much greater surface speeds and tend to hold form much better.

Needless to say, the wheel size is ever changing, as the wheel gets dressed to guarantee thread form and maintain a free cutting wheel. I do NOT know if the wheel is capable of incremental speed changes, which is quite desirable if one hopes to maintain a constant surface speed. You likely recall that surface speed is related to how hard a wheel performs. Slow it down, it behaves softer, and that's the result of reducing wheel diameter by dressing.

H

[pete]

Fantastic link Conrad. I've saved that, but even with a quick scroll through a whole lot of it is well above my IQ and education level. Interesting they now use a servo to compensate and correct for the measurable mechanical inaccuracy to replicate a theoretical perfect screw pitch. And the age of when there 3 ruling engines were built with improvements made as the technology got better. Information like that always makes me wonder what's really possible when it takes a security clearance to even find out it can be done.