Clutch Mechanism

IBM typewriters (and probably others) used to use spring clutches.

Take a continuously revolving shaft, wrap a spring round the shaft. Connect
one end to something to be driven.

Now, if you move the other end of the spring in the opposite direction, the
spring tightens on the drive shaft and is driven round. Kind of difficult to
describe, but a very simple idea.

This way, a small solenoid can control quite a lot of power. These gadgets
are also pretty quick, inertia is low.

Drive is only one direction of course, but there are plenty of solutions to
that ...

YMMV

Dave

I never new anything like that was used in typewriters !  Sounds like
a very interesting mechanism.

I have had a similar idea in mind for years based on the principle of
the capstan - the rotating, indented shafts that are used on ships to
hall them in to a harbor wall.  Anyway the capstan is a device which
amplifies the force pulling on a rope.  It means that you could use a
bunch of small servo or stepper motors to control much larger pulling
forces.  The key to doing it successfully would probably be accurate
measurement of the large forces and incorporation of these
measurements in a feedback control loop.  My guess is the big
disadvantage would be wear and tear on the cables involved.  It is
also difficult to effect correct tensioning of the low force side.  If
the tension is not well controlled then 'sticking' can make the system
unstable and difficult to control.  I have always thought that
materials science should come to the rescue and give us good materials
for the capstans and cables that would reduce these problems. Maybe by
using springs IBM has done just that.

The advantages of a multiple capstan system are -
 -one big motor can make all of its force available on any single
cable .
 -the motor's power can be shared between multiple cables
simultaneously.
 -much lighter than using lots of big motors if all cables are not in
operation simultaneously.
 -output is linear instead of rotary, very handy in robotics.

There is also the double-capstan which can pull in both directions.

If your Google-fu is strong you can find the maths of capstans and a
few control projects on the web.

 > There is also the double-capstan which can pull in both directions.

     That's a well known form of mechanical amplifier.  It was used
in early Sperry autopilots and in the Bush Differential Analyzer.

     It belongs to the list of "clunky things people had
to do before amplifiers worked", along with Ward-Leonard drives,
thyatrons, and amplidynes.

                    John Nagle

I know about capstans. I have a patent on this topic.
And the wear and tear issue can be addressed with suitable
materials.

But I was not aware they were used in IBM typewriters. Can you
tell me where these are found?

Also, I am not aware of a capstan that pulls in both directions.
The double capstans, I am aware of, use two drums to avoid the
cables from tangling.

Brad Smallridge
AiVision

     Electromechanical clutches are widely available.  See

http://www.ogura-clutch.com/products/industrial/howtheywork/electromagnetic-clutch.html

     If you like purely mechanical solutions, see Camco,

http://www.camcoindex.com/products.htm

     Their stuff is boring, reliable, and rock-solid.

     What's the scale you're talking about?  Are you feeding thread through
a loom, or driving winches for docking a ship?

                    John Nagle

http://www.ogura-clutch.com/products/industrial/howtheywork/electromagnetic-clutch.html

Thank you. These are sites I haven't seen.


I am looking for strokes of a few inches and forces ranging from 3 to 100
lbs.
If you look at my old hand design on my website, and envision the rest of
the
body, you can get a good idea of where I want to go with this.

Brad Smallridge
AiVision