Sundial List,

There are many ways to make longitude gears - I've been exposed to some
ideas since I do gear design as part of my job.  The ideas suggested so
far will work well - I just offer a few more - brainstorming if you
like.

First - you don't need gears.  I've seen some clever things done at the
local vocational school using chain sprockets and pins.  The pins are
spaced as if they were part of a chain. This is an inexpensive way to
get semicircular motion.  Usually the pins are fixed and the sprocket
moves.  A related trick is to stretch (not to the point of distortion)
the chain around the convex portion of a semi-circle (outside bottom of
a capital "U"), and then use a sprocket to drive it - allowing for
sprocket tooth clearance.

THK makes linear bearings (huge Japanese company - should have offices
everywhere).   One unique type they make is curved - this could be used
to get very smooth circular motion.  The bearings come in segments - I
think 45 degrees.  One way to drive these is with a belt or a chain on
the outside of the bearing (convex side - or the outside bottom of the
capital "U")  They typically use three rollers with a flat belt looped
through.  Picture a capital "Omega" character - and place the three
rollers in the obvious spots, the belt loops through in the "Omega"
path, with the bottom of the "Omega" in contact with the convex part. 
In this set up, the belt is fixed to the ends of the semi-circle, and
doesn't slide over it, but remains in contact with it.  This is used to
control banking in high speed rail cars (I'm told).  The roller in the
inside top of the Omega is usually the drive roller.

As far as gear solutions, there are many - bevel gears, spiral bevels,
internal gears, etc.  Two ideas that come to mind are:

1) contact a manufacturer of planetary transmissions - and get an
internal ring gear from them (or a tranny from a scrap yard?) - they
often broach these in quantity, so they may be affordable.  (Shaping is
typically done for custom work - therefore more expensive.)  Then just
cut a section out of the ring gear as needed. Buy a standard spur gear
to fit.   There are always commercial sources as well - not cheap -
contact me if interested.

2) This works on a conventional knee mill:   Another idea is to use a
ring gear with straight sided internal teeth.  Its been largely
forgotten technology - most gears use involute shaped teeth because of
standardized tooling and insensitivity to errors in center distance
spacing (except for increased backlash of course).  However - other gear
tooth shapes can provide perfectly smooth (called "conjugate action")
motion with no backlash (in practice a small amount of backlash is
always designed in to avoid interference).  Then you could make the
individual ring gear teeth on a conventional knee mill, and use a shaped
cutter to make the spur gear teeth - again all possible on a
conventional knee mill.  The math for this is in a book "Analytical
Mechanics of Gears" by Earle Buckingham, Dover reprint (in print!) ISBN
0-486-65712-4 (15$ paperback), original copyright 1949, reprint 1963. 
Look at page 40 - the "Williams internal gear".  Also of interest is a
pin drive gear on page 38-39.

Final note - you can use your mill as a shaper if your patient - it
works really well.  Just keep the spindle turned off, and manually run
it up and down with a hand sharpened HSS tool bit in a collet.  Mount
the part on a rotory table if needed, and shave .005 to .010 of material
off on each pass (nice to have digital read outs here).   I do vertical
key ways this way - it works fine.  You could do an involute internal
gear this way too!  (let me know how it works). By the way - with a
rotary table - I'd use a rack shaped cutter, and rotate the table a
small amount on each pass - to actually generate the involute on the
workpiece.  I suggest this because it is easier to hand make a straight
sided cutter than an involute shaped cutter.

Best Wishes,
 - Ben Hoffmann  

Content-Type: text/x-vcard; charset=us-ascii; name="vcard.vcf"
Content-Description: Card for Ben Hoffmann
Content-Disposition: attachment; filename="vcard.vcf"
Content-Transfer-Encoding: 7bit

Attachment converted: MAC Hard Disk:vcard.vcf (TEXT/ttxt) (00009977)

Reply via email to