Virtual Capstan

[Richard Brekne]

Ron Nossaman wrote:

> >>But the main problem has always been finding something
> >>that didnt get in the way of regulating the standard capstan. This
> >>solves that problem.
> >
> >Again, congratulations.
>
> Has this been tried in actual hardware, or just as a vaporware conceptual
> exercise? There's a considerable gap between magnets in the drawing, and
> the inverse square law of magnetic field strength would indicate to me that
> you'd need magnets of considerable size, and mass, to have much effect with
> this sort of gap. I think I'd like a performance report of the prototype
> before any discussion.
>

I have a working model at the shop. Tho it's a single key action model I use
for lots of dinking around. And the magnets I've used up to now are far from
optimal. Still the principle seems to work well enough. The drawing is meant to
show the general concept, not a final engineers draft. The space between them
need only be sufficient to allow for regulation of the standard capstan as
well.

One concern is the mass of the magnets vs their affect. Course you know that a
magnet is not just a magnet, and that some of these are pretty darn powerfull
compared to their weight. A link to some European suppliers is below for anyone
interested in browseing websites to see whats really available.

Remember tho, all thats needed here is enough force to lower balance weight a
bit. I dont envision this as a way of replacing 20 grams of key leads or
anything. Rather a way of giving say 5 - 7 grams latitude in BW while allowing
for plenty leeway relative to standard capstan placement.

Jon Page suggested intergrating this into the existing capstan / heel
arrangment... might be very doable as well.

This is an Open Source project, meant to help myself and anyone else who's
interested learn a bit and hopefully come up with a very decent method for any
action rebuilder to put to use some of these new ideas flourishing about
touchweight. So by all means feel free to develop the idea.

I'll have some initial performance data over the weekend for a single key
model.

Cheers
RicB

--
Richard Brekne
RPT, N.P.T.F.
UiB, Bergen, Norway
mailto:rbrekne@broadpark.no
http://home.broadpark.no/~rbrekne/ricmain.html
http://www.hf.uib.no/grieg/personer/cv_RB.html