[Server-sky] A kind of thinsat to help deorbiting small debris

[Keith Lofstrom]

On Mon, Nov 18, 2013 at 09:59:50PM +0900, Michael Turner wrote:
> The ideal solution might be a neutralization scheme that feeds on the
> debris itself. For example, in the above comment I propose that the
> LODR mirrors be distributed in orbit. (This admittedly gets tricky in
> terms of scheduling if you're dependent on the intensity of deposited
> energy for net thrust vectors -- speed-of-light delays become even
> more significant.) What if those mirrorsats could be (partly) made of
> debris themselves? 

Orbiting mirrors under high intensity illumination for laser ablation
are tricky but /perhaps/ not impossible.  The targets are shiny aluminum -
or they are after the first pulses take off the paint.  The mirrors will
be pitted, after microdebris and meteoric sanding.  The mirrors will
accumulate a lot more energy flux than any single target will.

Optical grade surfaces are made with sophisticated tooling;  there are
only a handful of places on earth that can make large mirrors.  So the
mirrors themselves will be shipped up.  Space mass will be useful as
ballast.  Those mirrors will be light sails, and subject to tidal as
well as optical torques;  extra mass added in orbit could add a lever
arm that balances the torques in all the rotation axes, so the momentum
wheels don't have to work as hard.

If the illumination of the mirror is off center, the mirror will torque.
If aiming at a sub-meter-sized spot 10,000 kilometers away,  then the
pointing accuracy must be less than 100 nanoradians, which translates 
to a fraction of an optical wavelength for a meter-scale mirror.  The
shape of the mirror surface will need to stay figured to an even smaller
fraction of a wavelength for uniform energy deposition on the distant
target; we can assume piezo actuators against a non-resonant bulk mass
backing the mirror, and actuators to damp out the vibrations in the
bulk mass, and ...  

Again, I won't say it is impossible, but this will be very challenging.
We encounter these kinds of optical problems designing lens stacks for
integrated circuit photolithograpy;  the light must be focused to tiny
fractions of a wavelength on a wafer target.  Fortunately, that target
is only a few wavelengths away from the business end of the lens stack,
which is as big as a couple of beer kegs, weighs tons, and contains a
couple of dozen lenses figured to nanometer precision.  Oh, the things
we do to bring you musical birthday cards.

Those lens stacks and optical systems are shipped from Japan and
the Netherlands (Canon and ASML, mostly) in specially modified 747s
with air-bearing actively-damped cargo holds, by pilots who land
those planes so gently you don't know that it happened.  For Intel
Hillsboro, they use a lengthened runway at KHIL airport, after flying
flightpaths that avoid all turbulence.  Very different than typical
high-vibration rocket launches!

"Figuring" server sky thinsat arrays is tricky, too.  If we blindly 
launched 60GHz energy from all the antennas, the result would be an
incoherent mess.  The lowest frequency vibration mode for a thinsat
is seconds; thinsats will slowly bend and flex, very slowly dissipating
low frequency flexural energy left over from thrust changes.  One of
the low intensity measurement and computing tasks for a thinsat is
determining the shape of the surface at any given instant, then
adjusting radio phasing to compensate for it.  Fortunately, the
mechanical phasing calculations are ten orders of magnitude slower
than the packet aiming calculations, though they are much faster than
ground receiver tracking calculations.  Tenth-wavelength adjustment of
5mm wavelength microwaves is a heck of a lot easier than tenth-wavelength
adjustments of 1μm infrared.  That's one of many reasons why we aren't 
yet ready to for optical crosslinks and downlinks for server sky.

So yeah, mirrors are tricky.  At nanometer scale, everything is jello.

Keith

-- 
Keith Lofstrom          keithl at keithl.com