A question for which I have been unable to find an answer in the archive (though
I limited myself to 3 hours...):
What is the nominal thread for attachment of the pad? I have a couple examples
that all are in the ballpark of 0.810"-12, or 0.810"-11-1/2, or.... The less
than three complete threads do not give a good measure, and the originals that I
have are rough enough that all measure between 11-1/2 and 12 pitch using
traveling microscope and parallel wires.

My objective is to make a proper tap, since I have two Stanley era  with barf-
smelling plastic pads in need of replacement, and I have a threading demo to do
with some students. But, as the replacement pads will be in wood, being off will
lead to eventual failure, even with only 2-1/2 threads engaged.
 *** This message was sent from a convenience email service, and the reply
address(es) may not match the originating address

Erik Levin asked:

> A question ...):
> What is the nominal thread for attachment of the pad? I have a couple
> examples that all are in the ballpark of 0.810"-12, or 0.810"-11-1/2,
> or.... The less than three complete threads do not give a good measure,
> and the originals that I have are rough enough that all measure between
> 11-1/2 and 12 pitch using traveling microscope and parallel wires.

> My objective is to make ... replacement pads will be in wood, being off
> will lead to eventual failure, even with only 2-1/2 threads engaged.

There is one ideal way of doing this, and that is to make a plaster
cast of the male threads on the remaining (?) quill (s) and then to find
a Wade Bench Profiling Machine:

>> http://www.lathes.co.uk/wade-bench-profiler/img0.jpg

I have the all-apart original made by the American Watch Tool Company
which I got from Phil Lathrop many years ago, and for which I now have
an incentive to get it up and running. There is one severely bent lever
which I can replicate, rather than try to straighten, which is doomed
to failure unless I first make a replacement. Not any time soon, though.

The bench profiler works with a stylus or follower that is kept against
a model part by said lever, which controls the travels of a cutter in
the main spindle with a one-to-one lever arrangement. Alas, it does not
do mirror images ... hence the need for a plaster cast. The cutter and
follower each need a vee shape just like the thread profile, which ought
to be easy to find or make.

The bench profiler was mainly used to make a great many copies of the
frames of pocket watches, and one could shape the basic profile as well
as spot all the holes to be later drilled & tapped, for screws or 
jewels.

One might cobble together a wooden model of such a machine with a high
speed grinder as the prime mover ... or build an auxiliary spindle for
a lathe driven one-to-one by equal size pulleys and a Vee belt, and then
use a follower coupled stiffly to said high speed grinder.

The follower will be pressed against the front side of the quill's male
thread, while the cutter will bear against the rear side of the female
thread-in-waiting, using the cross-feed slide (without any feed screw)
to serve as the rigid coupler. The lathe must turn very slowly while
the operator's left hand traverses the carriage and the right hand keeps
the follower in contact with the male thread.

For a right hand thread with the spindles turning counterclockwise, both
threads travel right-to-left, but it's the front side of the male thread
and the rear side of the female thread. However, the opposite sides of
both move the same way ... so the threads will have the same hand. Whew.

George Langford, cutting & running in SE PA

On Fri, May 26, 2017 at 2:08 PM,  wrote:

>
> There is one ideal way of doing this,


Well heck, there goes my one great idea.  I was going to suggest visiting
Georgesbasement.com, as I was certain he would have had the answer to this
question.

To the Lumberyard (or Google).... (Young Frankenstein joke)

Do the threads really have to be "that" precise?   Seems to me that most of
the braces I have taken apart have had smaller screws holding the main
screw in place (stop the turning of the pad).  11-1/2 vs 12 should be a
trivial difference.

How about filing a sharp edge on the leading edge of your screw, and just
work it back and forth into an appropriately sized hole in the new pad.  It
is wood, so you can be a little less precise and more aggressive than metal
working or (clock making).
-- 
Kirk Eppler in Half Moon Bay, intrigued by the idea of cutting threads

George replied:
>There is one ideal way of doing this, and that is to make a plaster
>cast of the male threads on the remaining (?) quill (s) and then to find
>a Wade Bench Profiling Machine:

As much as I would love another machine, I do not see a profile machine on my
horizon.

If all else fails, I may use this as an opportunity to make a thread tracer for
the small lathe. Part of the difficulty is that the flank angle isn't the same
to either side, being about 50 degrees on one side and about 55 on the other,
for an included angle of about 75 degrees (again, not consistent from point to
point. Come on, Stanley. It is steel, not chewing gum). Given that the design is
post turn of the 20th century, I would guess that it is some reasonable nominal
pitch, even if it isn't a standard machine thread. It just comes in at less than
12TPI and greater than 11-1/2TPI on both braces, any position I measure, any
method (thread gauge, traveling microscope, over parallel wires, direct scale).
I even pulled out the oddball scales for comparison (true point and pica rule
among them. No clear match.

I no longer have access to a CMM (really? who besides me would think of using a
CMM for this?) though that would be the cats meow for this. Oh well.

Given the short timeframe I have if I want to make my demo useful, I may just go
with 12TPI and see how the pads hold up, but I would really like to make the
tool so it is useful to others, as I am likely to have a lifetime need of two
brace pads.
 *** This message was sent from a convenience email service, and the reply
address(es) may not match the originating address

Kirk opined:
>Do the threads really have to be "that" precise?   Seems to me that most of
>the braces I have taken apart have had smaller screws holding the main
>screw in place (stop the turning of the pad).  11-1/2 vs 12 should be a
>trivial difference. I tend to get fixated on details like this. Probably the
threads don't need to be that precise (wood moves, after all) but an accurate
nominal minimizes the chance of misfit or failure given the variation in the
parts I have in hand and other parts that may be seen in the future.
>How about filing a sharp edge on the leading edge of your screw, and just
>work it back and forth into an appropriately sized hole in the new pad. That
may be the final endpoint. I am still looking at this project for use in my
threading demo.
As to the lumberyard, I have the wood on hand. A really nice chunk of curley
maple from a tree i helped a friend take down several years ago. Truely lovely.

*** This message was sent from a convenience email service, and the reply
address(es) may not match the originating address

Well, after a bit of digging, thinking, measuring, and comparing, I came to the
conclusion that the thread on this brace is just plain odd. I tried fitting a
hole tapped with a 13/16-12 tap (60 deg thread, 0.812 dia) and no go, and it
appeared to be the pitch rather than the thread included angle, but  I am not
sure. Tried another Stanley pad with the same two screw setup and the threads
were close but no match. Still seem to be between 11-1/2 and 12TPI. So, I made
the new pad, bored the center to 0.795-ish, let the thread smoosh itself in very
lightly, and the two wood screws do the primary holding. If I get the incentive,
I may burn a few electrons on the CNC machine and make up test pieces at 11-1/2,
11-3/4, and 12TPI to see what the best fit is, but first, I need to make (or,
perish the thought, buy) a thread milling tool suitable for the thread pitch...

Looks and feels great. I love maple. *** This message was sent from a
convenience email service, and the reply address(es) may not match the
originating address