Earth-Sun L1 for VAST delay-tolerant computing

[Keith Lofstrom]

The following is kinda long, and discusses many related
extensions of our current ideas. It took a few months
to think about, and a few days to write - please read it
in little bits over a few days, and react after you've
had time to absorb it, and add your own well-thought
out and QUANTIFIED improvements.

We won't domesticate the solar system on a sitcom schedule.

----

One of many potential problems for low-mass-to-area space
solar power satellites in GEO is that light pressure
(absorbed, or worse reflected) transforms orbits from
circular to elliptical - see

http://server-sky.com/LightOrbit#Eccentric

This eccentricity effect occurs because the light pressure
vector rotates in relation to the Earth gravity (and
gravity gradient) vectors.

The eccentricity effect might be tolerable if ALL GEO
satellites had the same mass-to-area ratio. However, if
they differ radically (i.e. comsats vs. first-generation
and second generation SSPS), and there are MANY of them,
their velocities will differ and they WILL collide when
closely spaced near the GEO circle.

However, this effect does NOT occur near the Earth-Sun
Lagrange points. The "ESL4" and "ESL5" Lagrange points
are a 150 million kilometers away, but ESL1 and ESL2
are a "mere" 1.5 million kilometers from Earth, on the
Earth-Sun line. ESL1 closer to the Sun, ESL2 farther
from the Sun.

At ESL1 and ESL2, the sunlight and gravity effects are
approximately collinear, no large side-to-side
differential jostling.

----

The ESL1 point is actually the center of some metastable
not-quite-ellipses used by 5(?) active solar observatory
space probes. These are symmetrical ellipses centered
on the Sun ...

... NOT elliptical Kepler orbits with one "focus" on the
Earth-Sun line.

The plane of this ellipse is PERPENDICULAR to the Earth-Sun
line, and the space probes appear from Earth to trace a path
AROUND the Sun against the backdrop of stars in Earth's
daytime sky.

The paths may have a different north-south period than
east-west period, tracing out an aperiodic Lissajous
pattern over time.

https://en.wikipedia.org/wiki/Lissajous_curve

These odd orbits mean that the observatory probes are
never on the direct-Earth-Sun line, so that an Earth
transceiver dish can always communicate with them without
"looking at" the Sun's prodigious radio noise.

What's amusing about these traced paths is how very wide
they are. For example, The Deep Space Climate Observatory
Probe (DSCOVR) traces out a path that has a north-south
radius of 160,000 km and an east-west radius of 280,000 km.
Double both for the rectangle it eventually paints around
Earth-Sun L1.

Other smaller ellipses are possible, and here is the
IMPORTANT point - with proper synchronization and
light-sail maneuvering, it may be possible to "paint a
rectangle" in the sky around ESL1 with an area of AT LEAST
( 2*160000 km ) * ( 2*280000 km) or 1.8e17 square meters.

Insolation at nearer-to-Sun ESL1 is 2% higher than
Earth's 1361 W/m², or 1388 W/m², so the available
total insolation is 2.5e20 watts. Subtracting some
area to eliminate Earth/Moon-system shading and reduce
self-shading, a GYNORMOUS array of solar power satellites
could collect more that 1e20 watts - and radiate waste
heat to the 2.7K universe.

About 9 watts per square meter of extra infrared hitting
the Earth IF the heat is radiated isotropically. BUT:

A better design puts a low-thermal-emissivity surface on
the back of the sun-facing PV (inefficiency losses radiated
in other directions than Earth). The compute load (perhaps
distributed zetacomputation) can be on blades perpendicular
to the Earth and in the penumbra of the PV panel, so THAT
waste heat is emitted in a cardioid pattern with a null in
the direction of the Earth.

With many more clever minds considering design, materials,
radiated heat management, thermal conductivity, heat
tolerance, thermodynamics of computation (cold kTln2 bits
are "cheaper" than hot bits), there are vast opportunities
for improvement of my crude brainfart.

-----

Personal note: I'm a chip designer, and spent my career
shrinking circuit dimensions from micrometers to nanometers.
I'm way too immature to fully appreciate the vastly more
important but fractional percent improvements that combustion
engineers manage to squeeze out of nature. I am currently
working on a brainfart that could shrink CPU chips by a few
percent, and simplify the power supplies that drive them,
but thinking about solar-system-scale opportunities is more
fun.

-----

Earth insolation at ground level is ~1000W/m², or ~1.3e17W
for the entire planet, so an ESL1 array may collect 1000
Earths worth of power, 5 seconds away at the speed of light.

If the satellites are more gossamer, the light pressure-to-
gravity balance occurs a little closer to the Sun. Stable
orbits still occur with a power-to mass ratio of 20kW/kg,
especially if "PV useless" infrared is reflected sunwards
and provides more "light sail" thrust.

Note that a similar array COULD be built at ESL2, on the
"night" side of the Earth, but reflections would create a
VAST amount of night-sky light pollution. Some propose
"lunettas" that light up the night sky in lieu of
streetlights. This is INSANE, because much of nature
depends on a day and DARK night cycle to function, such
as garden plants that do poorly under street lights.
Sorry, Hermann Oberth and Kraft Ehricke, you should have
paid attention in biology class.

-----

So, what the heck would we do with >1e20 W, 10 round-trip
light seconds away? Not SSPS; focusing power beams would
be impractical at that distance. That much power delivered
to Earth would fry and radio-blind us.

Fortunately(!), two rapidly-growing electric power demands
are Bitcoin mining and artificial intelligence. Bitcoin
mining may be as much as 7 gigawatts worldwide, and AI
may soon "eclipse" (ahem) that.

Love it or hate it, Bitcoin occurs on a 600 second cycle.
With some lower-power Earthside support, most of the 7GW
(or more) can be powered at and radiated from an ESL1
computation constellation, with a seeming "time delay"
cost of 10 seconds round trip, 1.7% of the 600 second
mining cycle.

Possibly worse; any Bitcoin protocol experts (as opposed
to outsider knee-jerk opponents) on the list?

Over time, if "free space solar power" (which does NOT
need to be transmitted to Earth, just bits) is cheaper
than terrestrial generation, then ESL1 Bitcoin "wins"
will be cheaper than Earth wins, time delay disadvantages
will become advantages, and eventually replace Earth
Bitcoin mining systems. This would be "negawatt" Space
Solar Power; perhaps cheaper power than the raw power
delivery systems we have discussed for decades.

Note that Bitcoin can be surprisingly radiation bitflip-
error tolerant - 1 error per thousand hashes is a 0.1%
loss, and easy to detect, if a candidate "winning" hash
is re-tested a few times before transmission.

More importantly, after Bitcoin-mining-sat and deployment
systems are designed for mass production, economies of
scale will make small systems possible and large systems
cheaper per watt, enabling space-powered-computing "supply
chains" that can be used for other delay-tolerant (perhaps
delay ENHANCED) *valuable* applications.

----

For example:

Some of us worry about large-scale AI drawing ENORMOUS
amounts of electricity, and also out-competing humanity.

(perhaps not in the short term; AI-"assisted" Google
search behaves like artificial ignorance, and cannot be
forced to return the accurate search results that Google
returned a few years ago. Irritation yes, competition no)

However, if humans have a 10 second time advantage over
"AIL1", maybe we still have a chance, while AI can have
WAY MORE than 1e20 watts ... "50 AU AI Dyson shell" is
a whole 'nother blovating email.

Much depends on human wisdom conquering lazy impulsivity.
If humans refuse to GROW UP and CARE ABOUT EACH OTHER,
and we continue to damage our planet and poison polite
discourse, it may by better for AI to replace us.

Myself, I am still too immature to say.

Keith L.

--
Keith Lofstrom kei...@keithl.com

[Keith Henson]

Quick read-through,

"Subtracting some area to eliminate Earth/Moon-system shading "

L1 is way towards the Sunward of the E/M system.

The idea is similar to my thoughts about what might be causing the
deep shadows at Tabby's star. Only if you analyze the duration and
size of the dips, you get something 409 times the area of the Earth
and about 7.8 AU out from the star.

Why so far out? If this is a data center it might be to keep it cold
and reduce the error rate.

The habitable zone is close to the star where liquid water exists.
The computational zone might be out at the distance of Jupiter.

Best wishes,

KeithH

> --
> You received this message because you are subscribed to the Google Groups "Power Satellite Economics" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to power-satellite-ec...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/power-satellite-economics/20250425210754.GA6506%40gate.kl-ic.com.

[Keith Henson]

On Sun, Apr 27, 2025 at 2:22 PM Keith Lofstrom <kei...@kl-ic.com> wrote:

snip

> KHL>
>
> On Fri, Apr 25, 2025 at 08:58:37PM -0700, Keith Henson wrote:
> HKH> Quick read-through,
> HKH>
> HKH> "Subtracting some area to eliminate Earth/Moon-system shading "
> HKH>
> HKH> L1 is way towards the Sunward of the E/M system.
>
> KHL: the Moon is 380,000 km from Earth, Earth-Sun L1 is
> 1,500,000 km towards the Sun. ESL1 is a "point", but
> the vast region around it is a shallow metastable "gravity
> groove" that five solar probes (currently) are orbiting
> within.

Right, but L1 is not shaded by earth or moon.

KeithH

[Keith Lofstrom]

Regards the trajectory to Earth-Sun L1, here is a free
paper pdf describing in detail the mission to deliver the
570 kg DSCOVR deep space climate observatory probe to the
mission orbit around that point in space.

https://ntrs.nasa.gov/api/citations/20150019786/downloads/20150019786.pdf

An excellent and quite readable paper, assuming a basic
knowledge of rudimentary orbital mechanics and spacecraft
maneuvering. Pretty good introduction to this topic for
those who can read simple numbers. If you fear technical
detail, you may need to slow down to draw a picture of the
three dimensional XYZ plane of the Earth's orbit and the
"99% radius" L1 "orbit", but with that in mind the rest
of the paper isn't difficult, merely detailed.

The hero of this mission was a SpaceX Falcon launch vehicle
that delivered DISCOVR into a transfer trajectory with a
VERY small delta V error. That saved enough maneuvering
propellant that the spacecraft mission could be extended
from the planned 5 years to many more than 10 years,
overlapping more than an entire solar cycle.

The other heros were the team that made the decisions to
apportion the correction thrusts and related errors.
Imagine parking your car and coming to a perfect halt in
a narrow parking spot ... 1.5 million kilometers sunwards.

BTW, arrival delta V is around 200 m/s. The most amazing
(to me) aspect of this mission is that velocities were
measured using telemetry Doppler shift to Earth, accurate
to a millimeter per second.

----

BTW, I don't expect you to believe it, but the same launch
and maneuvering system could deliver 10 megawatts of space-
powered computation to L1, using a deep-space adaptation of
my http://server-sky.com brainfart. Note that delivery to
L1 is CHEAPER than delivery to GEO, which requires 1600 m/s
of apogee insertion delta V. Plus continuing large delta V
for station-keeping during the lifetime of the satellite.
Even more if the GEO satellite is "gossamer", with a high
"lightsail" ratio; see

http://server-sky.com/LightOrbit#Eccentric

Keith L.

---------------------

On Fri, Apr 25, 2025 at 02:07:54PM -0700, Keith Lofstrom wrote:
>
> The following is kinda long, and discusses many related
> extensions of our current ideas.

...

> The ESL1 point is actually the center of some metastable
> not-quite-ellipses used by 5(?) active solar observatory
> space probes. These are symmetrical ellipses centered
> on the Sun ...
>
> ... NOT elliptical Kepler orbits with one "focus" on the
> Earth-Sun line.

...

> What's amusing about these traced paths is how very wide
> they are. For example, The Deep Space Climate Observatory
> Probe (DSCOVR) traces out a path that has a north-south
> radius of 160,000 km and an east-west radius of 280,000 km.
> Double both for the rectangle it eventually paints around
> Earth-Sun L1.

...


--
Keith Lofstrom kei...@keithl.com