Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!usc!zaphod.mps.ohio-state.edu!caen!rphroy!cfctech!rel!edsews!uunet!psitech!jim
From: jim@psitech.com (Jim Rothrock)
Newsgroups: comp.robotics
Subject: Re: Nitinol message
Message-ID: <437@psitech.com>
Date: 16 Nov 90 16:45:23 GMT
References: <1990Nov11.205704.24949@watmath.waterloo.edu>
Distribution: comp
Organization: PsiTech, Fountain Valley, CA
Lines: 26

In article <1990Nov11.205704.24949@watmath.waterloo.edu> mwtilden@watmath.waterloo.edu (M.W.Tilden, Hardware) writes:
>
>...All the kids toys you might 
>have seen using nitenol use ice as the flex medium.  Sub 
>room-temperature mediums help nitenol keep it's 'memory' but 
>are very difficult to implement in a robot limb (unless you pump 
>freon through it, but to what point.  If you're going to do all
>that plumbing, why not just go hydralic?).

How about using Peltier-effect heat pumps to cool/heat the wire?
There is a company that markets these as "Frigichips."  Run current
through them one way, and one side of the chip gets hot and the other
gets cold.  Reverse the current, and the hot and cold sides exchange.
Thus, you would get quite a bit of control over the wire's temperature.

The main problem with this technique is that the frigichips are flat and
square, from about 1/4"x1/4" to 1"x1".  The wires, of course, tend to
extend along an entire length of robot arm.  The solution I've been
thinking of is to have a "clump" of very short Nitinol wires attached to
one side of the chip with some sort of heat-conductive glue, and have
the other end of the "clump" attached to a reverse block-and-tackle.
The block-and-tackle cable would then attach to the next segment of the
robot arm.

Jim Rothrock    | Disclaimer: My opinions do *not* represent the opinions
jim@psitech.com | of PsiTech, Inc.