X-Message-Number: 1089
Date: Sat, 1 Aug 92 13:14 PST
From:  (Keith Lofstrom)
Subject: Re: cryonics: #1083 - #1086

>... you can take 20 years if you like to do the
>repair provided you can do it at a low enough temperature. At worst,
>you can use the patient's own circulatory system as a cooling duct
>system, once you've made some minor temporary adjustments to it.
> ...
>Its almost impossible to find anything that operates effectively at
>liquid helium temperatures; designing functioning nanomachines that
>operate near the glass transition temperature at -130C is going to be
>hard enough. As for "colder" than LHe, its very hard to get anything
>colder than liquid helium. There are no gasses at that temperature to
>use as working fluids for conventional cooling techniques. Maybe you
>could use some sort of strange Peltier effect device, but if you could
>I haven't heard of it -- hell, I'm not even sure that things like the
>Peltier effect still operate as advertised down that cool, (everything
>starts acting strange that close to absolute zero -- liquid helium is
>a superfluid, many if not most metals start superconducting, anyone
>know if the Peltier effect or other similar effects still function?)
>and certainly building equipment at those temperatures is a bitch.
>Thermal motion is not really a problem -- your body works fine,
>doesn't it, and its full of nanomachines. Drexler's arguments about
>this in EoC are pretty solid. Another problem is that getting that
>cold causes even more thermal stresses on large inhomogeneous objects
>like humans -- thermal stress cracking is bad enough just at LN2

Regards 20 years - people do NOTHING that takes 20 years.  Sometimes they
initiate a process that takes longer, and look in on it from time to time, 
but where options are possible they try to do things as fast as possible.
1 33MHz 386 is worth a lot more than 10 6MHz PCs, for example.  Given what
I expect will happen with the time value of money in a nanotechnological
age, decision makers would be silly to initiate a process that completes
1000 money doubling times from now, if there is any way to complete it

Regards temperature:
Do I pull rank here?  Well, I'm a microcircuit designer, who studied
superconductivity in school.  While it's true that the structures I work
with are hundreds of nanometers in size, I have had a chance to work with
these large structures for a while.  I am working in an area in the
transition region between bulk handling and individual handling of
electrons ( the charge that makes it from a dynamic RAM cell to the
sense amp can be measured in the thousands of electrons).  Thermal
problems - where the electrons jiggle from where you want them to
where you don't - is a hell of a problem.

The biggest problem with nanotechnology is observation.  Bucko, when a
wavelength of light is 1000 times the size of what you are looking at, AND
there is enough energy in that light quanta to tear a hole in what you are
looking at, you use something else to do the looking.  And believe me, you've
GOT to look (I'll eat my words if you can find a machinist that will work
in the dark, with no sound, and no sense of touch) because you don't know
what you are going to find.  I bet a cookie that the processes to look at
things will be "acoustic", that is, you yank on something and see how it
moves in response, then infer the structure.  If it is moving thermally,
it will mask out the tiny echos you are looking for.  Phonons - acoustic
quanta - are the only "particles" I know of that have short wavelengths,
low energy, and don't stick to things, but do interact with them.

Solids? Glass?  You are going to be working in vacuum, in tunnels you have
constructed through the frozen cells.  There won't be any liquid water around,
so wet chemistry is not going to happen.  I am assuming you are doing 
machine work at this scale, not bulk chemistry.  Pick up a radical from the
toolbox, and bond it into place.  Who needs viscosity, or Brownian motion?
You want things to move when you want, and stay put when you don't.   
Macroscopic heating and cooling stresses are absurdly simple to take care
of - you just put the right things in the right places at the small scale,
and you don't have any macroscopic stresses.

Regards cooling - they MAKE liquid helium somehow, don't they?  That isn't
the coldest you can get.  4.2K is the temperature of liquid helium at standard
pressure;  it gets a good bit colder if you boil it off in a vacuum chamber.
When we went over to the physics building to get LHe for our experiments,
I could watch the multi-stage compressor working to make more - there were
parts of the machine much colder than 4.2K.  There are labs that are working at
0.001K right now.  If a patient isn't spending long at that temperature, it
doesn't cost that much to dump them in LHe for a while.  Actually, I would
be using some sort of solid state cooler - not Peltier - probably little
nanotechnological phonon pumps of some sort.

A lot of nanotechnology I expect will be happening at room temperature, or
hotter - the processes will be simpler, the end product simpler, and the
process can be done "blind", without reference to the previous structure.
We obviously have demonstration proofs that people can be constructed at
room temperature.  This is NOT a demonstration that people can be
RE-constructed at room temperature, in some fashion arbitrarily close to
a design ideal.  This is a much more complicated process - certainly doable,
but involving much more measurement and control.


Keith Lofstrom                Voice (503)-520-1993
KLIC --- Keith Lofstrom Integrated Circuits --- "Your Ideas in Silicon"
Design Contracting in Bipolar and CMOS - Analog, Digital, and Power ICs

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=1089