Alt-BEAM Archive
Message #11836
To: beam@sgiblab.sgi.com
From: "Jim Taylor" ace27@icdc.com
Date: Sun, 12 Mar 2000 19:52:38 -0500
Subject: [alt-beam] Re: driving muscle wire
ugh muscle wire......how I loved thee......I too tried to biuld a super
small walker using nitinol, some SMT parts, some really small lithium
batterys, and some really thin music wire. I tried a 4 legger, a 6 legger,
and a tilden worm. The 4 legger showed the most promise, and still does
looking back at my work. However Nitinol is very tricky stuff. You mind
as well be directly shorting the batterys. Not only do they draw alot of
current, but it doesn't transfer very well into torque. It seems (even with
BORIS, and STIQUITO) no matter how small your driver, and control board is,
you need alot of battery storage. Unless you like 4 minute strolls :)
Don't want to discourge you though.....My mini 4 motor walker walks fine on
a tether. However I don't think he will do any walking autonomously
>
>
> > To drive a muscle wire you do not even need half of an h-bridge. The
> > second transisotr is used to provide a path for current flow. All you
> > would need is an NPN transistor with the muscle wire connected between
Vcc
> > and the collector, the emitter connected to ground and the base
connected
> > to the input through a resistor. You will want the base resistor small
> > enough to drive the transistor into saturation and will probably need a
> > resistor in series with the muscle wire to control current flow.
> > I do not remember off the top of my head, but I think the 50 micron
wire
> > nominally needs a 100mA current. I do not remember the resistance (ohms
> > per unit length) off the top of my head. There are a couple sites out
> > there with properties on muscle wires (www.dynalloy.com does I am sure).
I
> > would definitely reccomend checking them out because they give other
> > helpful hints on using the wires, especially about attaching them.
> > I did a concept project a while back using muscle wires for an
> > autonomous robot design class. From my experience, which is anything but
> > all inclusive, I found that the technology is fun and promising but
still
> > needs some work. You only get about a 5% deformation and the power
> > required is still on the high side. I finally settled on a design that
> > used four muscle wires for movement of legs down and back (2 frames of 3
> > legs) and the reverse stroke was handled by a spring system to save on
> > power. I was using microprocessor control at the time, never got around
to
> > trying it with BEAM control. If I find time one of these days, I may go
> > back and finish that... I would be interested to hear what your
experience
> > with the wire is like. Best of luck.
> >
> > John Gonser
> >
> >
> > On Tue, 14 Mar 2000, David Perry wrote:
> >
> > > Okay - haven't asked a question for a while....
> > > Okay i'm making a robot, quite advanced, very, very tiny, purely
because
> i
> > > had the idea for it and couldn't get it out of my head all day and
it'll
> > > drive me mad if i don't build it.
> > > The question is - what is the best way to drive muscle wire? You would
> only
> > > use half a hbridge, (no need for switching wires) but i'm thinking
thats
> > > a little too large. Perhaps a 7hc245 would suffice. I need to drive 6
> > > pieces of 050 LT, any advice would be appreciated.Oh and why am i
using
> > > muscle wire? - well i can't find 6x 4mm dia. gearmotors for under $31
> > > :-)
> > >
> > > thanks,
> > >
> > > David
> >
>
>
James Taylor
"Edgar The Hate Bug"
URL: http://fly.to/springmeadows
ICQ 14888587
Home