Cleaning mounds of swarf out of the chip pan of my Clausing 5914 lathe
was a problem.  Fingers are not a good idea with steel chips, as one
always needs to pull steel splinters out of the skin afterwards.

I got an idea from a food tool called a "cake breaker" (US Patent
1,858,790). Cake breakers are for dividing cakes without squishing them,
and look like a big comb, with about twenty 4" long steel tines.  I have
also seen bagel splitters that consist of two combs hinged together with
alternating tines: one pushes the double comb through the bagel, and
then opens the two combs, pulling the bagel apart.  Anyway, this kind of
design seemed likely to be effective at getting under tangles of swarf,
allowing mounds of swarf to be shoveled into the wastecan.

The cake breaker tines are perpendicular to the handle, which is awkward
for swarf removal duty, where one wants the tines more or less parallel
to the handle.  And cutting oil does not improve cakes.  So I made a
dedicated tool for shoveling swarf.

The handle is a piece of 0.500" diameter stressproof (1144) steel rod,
and the crossbar is a piece of 3/8" by 3/4" mild (1018) steel rectangle
4" long; these were used because I had them.  The times are made of
0.080" diameter music wire, bought for the purpose.

The cross bar is attached to the end of the handle rod with a 1/4-28
flat head screw with hex socket.  At first I was just going to machine
the rod end off flat, and depend on the screw to keep the bar tight to
the handle.  But my experience is that this usually loosens.  What to
do?  I recalled lots of Starrett tools where a locking thumb screw
prevents rotation.  The general design is a male cone is forced into a
female conical seat by the thumbscrew, everything being made of greasy
hardened steel.  It's the wedging action of the cones that makes the
difference, and provides the mechanical advantage so a thumbscrew can
generate sufficient clamping force.  So, I decided to implement a
conical seat in the bar, with handle rod tip machined to fit.

First the conical seat in the bar.  This is easy to do on a lathe
faceplate.  The bar was firmly clamped to the faceplate, with center of
rotation somewhat offset from the bar centerline, to accommodate the row
of holes that will accept the tines.  Drill through with 1/4", rough out
the seat recess with a 3/8" drill, and then make the conical sides and
flat bottom.  This was done with a boring bar held on the tool post
slide, with the compound set to 60 degrees.  Crank slide in and out,
back compound out a bit (increasing the cut), crank slide in and out,
until the outside diameter matches that of the handle rod.  Lubricated
with lots of black sulfur oil.  This yields a more or less flat-bottomed
hole with conical walls.  The depth of the recess is about 3/16, the
outer diameter is almost exactly 0.500".

Then the handle rod.  Change to a 5C collet chuck, center drill and tap
for 1/4-28, using lots of black sulfur oil.

Now the trick.  We want the cone angle of the rod end to exactly match
the cone angle of the conical recess, but angle settings on a lathe
compound are not all that precise.  So, without unclamping the compound,
run the lathe in reverse and machine the cone on the back side of the
rod, running the slide in and out as before.  One can use the same
boring tool for this, or (what I did) a regular tool bit.  I used a
insert tool with triangular insert to make the rough cut by moving a
long edge against rod tip, and then did a finish pass with the tip
moving in and out.

The machining was a bit rough (from impatience - no power feed on the
toolpost slide), so I ground things together with the rod in the mill,
the bar on the table, and some valve grinding compound between.  After
all this, the bar centers and palpably nests on the rod.  This will not
wiggle when clamped by the 1/4-28 screw.

The other side of the bar then got a countersink, and a row of 19 blind
holes to accept the tines.  The spacing between tines is 0.200", and the
holes are drilled ~0.250" deep with a #44 (0.086") stub drill. A pass
with a spotting drill was made so the #44 bit would not be tempted to
wander.  The 0.086" drill was used to ensure enough clearance that
solder could easily wick into the space between tine and blind hole.

[If one has a taper attachment, one can machine a tapered female thread
in the bar and the matching male thread in the rod, again by running in
reverse for the bar.  An alternate non-lathe approach would be to use a
pipe thread tap and die to make tapered threads. In both cases, the
1/4-28 screw is not required.]

Also made was an aluminum fixture, a 1/4 by 1/2" by 4.5" with a row of
#45 (0.082") holes on 0.200" centers.

Then made the tines.  The tines are 6" long when cut, ignoring the curve
of the wire as removed from the ~12" diameter coil.  Each tine is
mechanically deburred and one end sharpened on a vertical belt sander,
and the blunt end is abraded clean by hand for at least 1/2" using a wet
3M polishing pad.  When pointing the tines, it's useful to use the
aluminum fixture to hold the tine while the other hand rotates it as
grinding proceeds.  There are 19 tines to make.

The cross-bar was also cleaned and polished where there will be solder,
and the 1/4-28 screw liberally coated with anti-sieze compound.

The tines are then soft-soldered to the cross bar, using the aluminum
fixture and some black iron wire to keep the tines aligned during
soldering.  Soft-soldering does not seem to soften the music wire.  The
solder I used appears to be 60/40 radio solder, but this isn't all that
critical.

For soldering, the bar was assembled to the rod and the rod was clamped
vertically in a vise, arranged so the top of the bar is horizontal and
the tines are vertical, resting in the blind holes.  The tine-bar joint
area was liberally doused with acid flux (tinners' flux), and the bar
was heated from below using a MAPP torch, with care not to heat the
tines directly (which would anneal them).

Flowed the solder till everything had nice fillets, turned the torch
off, and went off to read newsgroups, without touching anything.  This
step is critical, as the assembly will cool far slower than impatience
demands, and the joints will be ruined if anything is moved too soon.

Wash it all off with hot water and a brush, replace the now softish
1/4-28 screw with a new one, and we are done.  The anti-seize compound
prevented the screw from being soldered to the bar.  And will prevent
the screw rusting to the handle.

A number of photographs were taken.  Look for Chip_Fork_Photo-1.jpg and
Chip_Fork_Photo-2.jpg.  The overall length is 15".

Initial tests show that the rake works quite well.

Joe Gwinn


Last updated on 23 January 2012.