(I've posted a little bit about this before.)

Oil well boreholes are helical, because the drill bit drills in a curve;
so, to go straight, you twist the curve into a helix.  The curve allows
for horizontal drilling, or indeed drilling in any shape at all, subject
to the limitations of a minimum radius on any curve and a smaller
minimum radius on the helix induced by limits on twisting.  You could
use this for fabrication, too.

Protein synthesis, for example, uses exactly this system for fabricating
arbitrary shapes.

You could use it to fabricate small items from wire, too; force a spool
of wire through a greased, curved nozzle with a curve a few times larger
than the wire's radius, which bends it into a circle by default; by
rotating the nozzle, you form some shape.

The weight of the article will eventually make it hard to twist the
nozzle, and will add its own bend to the wire, which will tend to
accumulate as shape error.  I don't know what to do about that, other
than float the wire in mercury.

Other possibilities exist.  You could extrude molten polypropylene into
water, for instance; you'd have to get it to harden in the nozzle
without sticking to the sides, perhaps with a lubricant coating.  I
think many thermoplastics or thermosets could work this way with salt
water of the right density, not even getting into exotic salts like
ferric chloride.  Or you could extrude something that sticks to itself
and/or expanding, like spray foam insulation or shaving cream.

All of these probably need some way to measure and compensate for
rotations of the resulting shape, which subtract from the nozle
rotation.  A cheap digital camera should suffice.

All of the above variants share the advantage of producing shapes with
six degrees of freedom with only one axis of control.

As a variant, you could imagine an extrusion machine whose nozzle emits
one round strand of liquid which solidifies quickly, but with varying
curvatures, by controlling the extrusion speed of the three thirds of
the circular cross-section separately.  You'd actually only have to
control two of the thirds, if the speed of the third was known; so you
get 6DOF for the price of 2.

Reply via email to