The MT3 collet chuck modification (I guess that's what you'd call it) is done. Her's my report:
This
Avey drill press wants to be a mill. It has double angular
contact bearings in the spindle, less than .0005" runout on the inside of
the MT3 taper, and less than .00025 sideplay 6" below the spindle so it
would seem adequate. However, among other things it needs a way to retain
collets. The usual method of drilling a center hole down the splined shaft
won't work: the thing is 3' long. Fortunately the slot for driving out
tanged tools with a drift is easily accessible: we'll work through that.
This is the approach. A steel bar, tapped on the ends, will slide up and
down in the slot. Attached to its center will be a bit of shaft tapped
3/8-16 to fit MT3 collets. An internally-tapped "nut" (actually a 2"
merchant coupling) will screw onto the bar. A collar will support the nut
from below so that when it is turned clockwise it will draw collets screwed
onto the shaft up tight. Spacers will be needed to hold the nut concentric
with the 2" spindle and thrust bearings will be needed at each end (the top
of the nut will press against the bottom of the quill when driving collets
out).
Calculations and research on the Web indicate that the collets will require
something between 85 and 650 pounds of force. The 11 1/2 TPI threads on
the nut and bar will handle that with a modest amount of torque,
In order to make assembly possible the 3/8-16 threaded shaft must be
coupled to the bar via a 1/4-20 threaded shaft. This is the weakest point
in the mechanism. Using a piece of grade 5 bolt for the 3/8-16 section and
grade 8 for the 1/4-20 makes pullout from the bar the failure point on
overload (just as well as that means if it breaks I can get the thing
apart). At about 1500lb this is good enough as long as I don't use a pipe
wrench.
Materials:
The bar will be cut out of the lump of mystery metal. The PVC will become
spacers.
Finished parts:
There were change orders. PVC did not work out as a bushing so I acquired
a thrust bearing.
Wrecked fixture:
I pressed the bar into a piece of 2" pipe and had the mechanic at the coop
threaded it on his machine. Almost worked. Fortunately the dies weren't
damaged and the bar got cut enough that I was able to finish the job with a
file.
Completed unit:
Tests:
I assembled the test fixture shown above and measured torque on the nut and
force on the bolt at various points. Gnuplot tells me that F =57T where F
is upward force in pounds and T is torque in foot-pounds. Matches
reasonably well with theory.
I made a couple of wrenches (the nut and collar are drilled for them)
And then I made some chips. Works! Turns out not to be hard to get the
collet to release, which is something I worried about.
To do:
Still need a bushing and a better spacer for the top of the nut.
Need a shield for the thrust bearing.
Problems:
The nut is hard to get started properly onto the threads on the bar. The
nut binds a little, probably due to the hand-filed threads in the bar and
the lack of concentricity of the PVC spacer inside the nut.
The way I should have done it (if I had the tools and materials): The nut
should be a running fit on the spindle, eliminating the spacers. The
threads should be double-start Acme. The bar and nut should be harder
material: the mild steel is going to wear. The collar should be a split
collar rather than set screw. The collar should have gear teeth cut in it
and the nut drilled so that a Jacobs chuck key could be used.
