In article ,
"DoN. Nichols" wrote:

On 2008-02-26, Joseph Gwinn wrote:
In article ,
"DoN. Nichols" wrote:

[ ... ]

As a hobbist, the only time I would buy a full-form thread
insert would be when I needed to do a production run for something where
the thread must have the rounded bottom to minimize stress risers. And
most of those I would cut with the Geometric die head anyway -- if it
did not need to be so long a threaded section that it would hit the center
pivot rod for
the turret.

I don't really know what a "geometric die head" is. Can you suggest a
URL that leads to a drawing or photo?

O.K. I've found a URL which gives some images:

http://reno.craigslist.org/tls/563369567.html

but it is short on description, since it is simply someone selling a
used one, and it is presumed that the purchaser already knows what it is
and how to use it.

The picture helped a lot.


So -- I'll do the description here.

1) Picture something which looks like a 4-jaw chuck with strange
levers coming out radially.

2) Into the ways in the chuck fit four "chasers" -- chuck jaws
whose inner end is one quarter of a threading die. A full set is
sequentially numbered for which slot each one goes into to make
sure that the threads are continuations of each others. You
need a bin of chaser sets to cut different sized threads.

3) It is mounted to the lathe bed turret by the cylindrical
shank. This one has a 1-1/2" shank, while the largest which my
lathe turret will hold is 1".

4) You use the lever with the red ball handle to close the chasers
down to the proper diameter. This rotates part of the body to
close the chasers. Some Geometric die heads, like this one,
have a smaller lever which selects between roughing (a bit
oversized) and finish (to precise size). A pair of setscrews
(not visible in these photos) tune the final size of the thread
produced, so if you need a different fit class of thread, you
can achieve it.

5) With the workpiece spinning, you advance the turret to bring
the chasers into contact with the workpiece (which should start
out just a little oversized). Very soon, the chasers will bite,
and you simply need to follow the head as it is drawn onto the
workpiece. (I typically cut 5/8-27 threads in brass at 800 RPM
as an example, and with coolant in steel could cut just as fast.

6) Eventually, the turret ram reaches a preset stop, so the shank
of the die head stops, but the body and the chasers are drawn on
for a short bit more travel.

7) When it draws the right distance, it releases a lock pin and
releases the red-ball-handled lever and the collar to which it
is mounted to rotate, withdrawing the chasers radially so they
totally clear the OD of the workpiece.

8) At this point, you can withdraw the ram of the turret with the
workpiece still spinning at 800 RPM. When the turret ram
reaches its full back travel, it will rotate to bring a new
station and tool into position. However, if you are using the
rough/finish lever, you flip the lever from rough to finish and
use the red ball handle to close it again for another stroke.
(Without the rough/finish lever, you complete the whole thread
in a single stroke. And because the form of the chasers is
correct, you can simply check the setting by measuring the OD of
the finished thread, instead of having to use a thread pitch
micrometer, or a set of thread measuring wires.

This sounds very useful in a production environment.


The extra lever opposite the one with the red ball handle is
used by cam surfaces in an automatic screw machine so there is no need
for an operator to operate the handle. This lever, or the ball-handled
one, can be unscrewed if it is in the way.

Some models have a separate pin beside the body which can be set
to release the chasers by bumping into a surface beside the workpiece,
normally used for automatic screw machines as well.

Smaller die heads (I have down to 5/16", with 1" being my
largest) take less pull, so you want them for smaller threads so the
pull is not likely to strip the threads. You don't want to use a 1" die
head for an 0-80 thread for example.

the third photo shows the end of the die head with no chasers
installed. You can (barely) see the steep thread in the face inside
which moves the chasers in and out. The knurled knob at 2:00 (just
above the alternate lever for the Automatic Screw Machine) can be pulled
out while the lever is in the released position to allow the lever
to move just a bit more, bring an interruption in the threads into line,
so you can remove the chasers and replace them with a different set.

All in all, it is a very nice way to make a lot of threaded
objects in a batch.

Yes. Think I'll hold off until I have a batch to do.


There is also an inside-out version, called a "Collapsing Tap
Head" which can be used to make fairly large ID fine threads, such as
those used for making a set of extension tubes for a camera lens. I
have a couple of these just because -- not because I have a need for
them -- and I don't know what a set of chasers for that would cost. I
know that the normal chasers for the 5/16" die head are now about
$100.00 new. I've got four sizes of Geometric die heads:

5/16"
1/2" ***
3/4"
1"

*** I have never found chasers for this head -- I suspect that it is
totally obsolete.

As I said, this would be a real asset in production.


Even on the Clausing, I would use scissors style knurlers --
less wear on the cross-feed leadscrew if nothing else. I normally use
the BXA size knurling tool with two arms on a common vertical dovetail
and a left-and-right handed leadscrew to move the arms together and
apart in a balanced manner, so one height setting will work for all
sizes of knurls.

Of course, on the turret, I would use the T-style knurlers,
which also have two knurls pressing form opposite sides of the
workpiece.

I don't think I will add much to the wear already experienced by the
Clausing in production use versus HSM use.

Well ... the cross-slide leadscrew (which had been mostly used
under power feed for parting off) on my Clausing looked like this in
the middle:

__/\__/\__/\__/\__/\__/\__/\

instead of like this:
_ _ _ _ _ _ _
_/ \_/ \_/ \_/ \_/ \_/ \_/ \

thanks to long service. And that was a leadscrew which was just being
used for parting off. The lathe did not even have the threading dial
mounted, because it was being used with Geometric die heads in the
turret for threading.

I assume that this crossfeed screw has been replaced by now. Mine
didn't look that bad, although the corresponding nut did seem a bit
loose. What kind of backlash is reasonable?


Hmm ... that sounds interesting. But do you mean that you will
cut them as a partial torus around the hold-down bolt? I'm afraid that
won't produce the compound curve you need. Instead, turn a ball on the
end of a section of rod of perhaps 1-1/2" diameter or so (double the
measured radius of the ring), and then cut radial sections out of that.
On the horizontal mill, I would use a slitting saw and an index head
with a 3-jaw chuck to cut those. You could even set up two slitting
saws on an arbor with a spacer defining the width of the sector. I
might consider leaving the boring of the center hole until after the
sphere is turned and the slots cut.

I don't quite visualize your proposed approach.

1) Make a ball on the end of a sufficient diameter rod to match the
radius of the ring around the lantern style toolpost.
Obviously, the ball can't be complete, but rather must have a
neck holding it to the chucked part of the workpiece.

2) Cut a series of slits into the ball from the OD towards
the center, but not reaching it. Ideally -- use two slitting
saws on a single arbor on the horizontal milling machine, so one
(if continued through the center) would be half the needed width
to one side of the center, and the other would be the same
distance to the other side, resulting in parts just the right
width.

3) Drill the center, and then bore out until you separate all of
the crescents from the main stock. Duck as they fly past your
head -- or catch them in a steel wire basket. :-)

Now I understand. This will yield a section of a sphere, not a torus.

That said, it would make mechanical sense if the surface were in fact a
sphere, not a torus. I will remeasure the two curvatures. It's hard to
measure the radius across the thickness, and I may have it wrong.


Let me re-describe my approach. Take an octagonal or roughly circular
piece of steel plate 5/8" thick, and mount it firmly on the lathe such
that the rotation axis is perpendicular to the plane of the plate, and
more or less in the center of the piece.

This mounting may be accomplished in a number of ways. For instance,
one can bore a 2" diameter hole in the center, and use the 3-jaw chuck
with the jaws pressing firmly outward on the inside of this hole.

O.K. Your mention of mounting it to the faceplate did not show
the edge being totally clear of the faceplate, so I misinterpreted where
the contour was being cut. I was picturing a flat piece with one face
firmly against the faceplate, and the curve being cut on the other
something like this:
___
\
|
/
|
|
---+
-- Bolt here to center of faceplate -- perhaps a drawbar
through the faceplate and spindle
---+
|
|
\
|
____/

Ahh. No. The plate edge would be out in space, well away from the
chuck.


Machine the outer edge of this plate into a section of a torus. The
larger radius is about 3", and the smaller (thickness) radius is about
0.75", both measured by eyeball.

The reason to use a torus is to allow toolbars that are not perfectly
rectangular to nonetheless be clamped quite firmly.

Oh -- you want to relieve the center of the top of the crescent
so it contacts at the ends. I thought that you were trying to match the
compound curvature of the support ring around the toolpost. And that
would be difficult to do with anything other than a section of a sphere.

Right. I'm wondering if it's a torus, or a sphere. A sphere would make
mechanical sense, and is easier to machine as well. Maybe I'll make a
disk out of thin aluminum, for fit testing.


Note that the rockers which I have seen were forged, and have a
diamond pattern grip surface on the top.

I have one of those. It almost fits, but is a cylinder, not a sphere or
torus. It is also hardened, making trimming arduous.


Hmm ... is a Woodruff key made large enough? :-)

That would be quite the woodruff key. Seriously, the rocker is far
shallower in proportion than a woodruff key.

You cut off part of the Woodruff key. It just gets you closer
for a starting point.

Except that the woodruff key is a cylinder, just like the forged rocker.


I'm still using a lantern style toolpost -- but it is on my 7"
shaper, not on a lathe. And it does not have a rocker, but rather a
flat-topped support ring.

No height adjustment needed.


Also -- there is another trick for putting a spherical surface
on the end of a rod. Mount a boring head in the mill with the cutter
facing inwards instead of outwards. Then set up an index head at
something like a 45 degree angle, put the end under the center of the
boring head and start cranking the index head as you slowly bring down
the spindle with the boring head. Stop as the boring head is cutting a
circle around the center of the intended sphere. You can adjust the
size of the neck left by tuning the angle. I've seen this done to make
the ball-handle on a rifle bolt.

That is a classic sphere generator. I'm not so sure it can generate a
torus, as the intersection of a torus and a plane is not necessarily a
circle. I think it's an ellipse, as the in-plate and cross-plate radii
differ. If these radii were equal, the torus would be a sphere.

O.K. But I was trying to make a ball to cut into slices for a
proper match to the support ring curvature.

Ok. It all depends on the true surface, sphere or torus. I will
revisit this.


Joe Gwinn