SSPS "too many email addresses" messages

2024 Feb 15 Thursday, 1530 PST

"The fault, dear Brutus, is not in the stars, but in ourselves ..." ... Shakespeare's Julius Caesar, Act 1, Scene 2

Could it be ... that broadcasting terawatts of ISM-band microwaves to Earth is a Really BAD Idea, and there are cheaper, safer, and more profitable ways to exploit many EXAWATTS of space solar energy at the collectors, IN SITU ?

Perhaps NASA and their contractors know this, and also know that arousing the people who killed the Moon Treaty could also kill NASA. Don't wake the fire-breathing dragon, even if that dragon really should wake up and learn to be neighborly.

Real radios are electronically and geometrically imprecise, Real radios make harmonics - DC doesn't want to be microwaves (exception - Josephson effect); Even with heavy (in both senses) filtering, the best transmitters might emit 100 ppm of unwanted harmonic energy, splattered in different directions than a power beam. So, 10 TW of SBSP might radiate a gigawatt of broad spectrum radio interference.

I've read technical journal papers about SBSP emitter elements that bragged about high efficiency, with "only" 2% (2000 ppm) third harmonic. 200 GW of communication-obliterating microwave-band noise for 10 TW of power beam. WHAT THE HELL were those authors SMOKING?

Yes, we've plodded in the same blind circles for almost 50 years, ignoring new opportunities all around us. One definition of insanity is doing the same thing over and over, and expecting a different result next time.

What better opportunities are available in 2024?

In my own field of integrated circuit design, in 1980, typical chips had 30,000 transistors and clocked at 5 megahertz. Today, I can purchase 8 terabit, 560 megabyte per second solid state drives for $75. The mass-to-function ratio will never be zero, but the mass and cost trend has followed Moore's exponential Law for FAR longer than any sober person would DARE to predict. Moore's Law is driven by mind power; it also amplifies mind power, and connects minds everywhere, from pole to pole. It will soon enable vastly more mind power beyond Earth.

In 2010, I made (with my partners at Hitachi) retail RFID chips 40 by 40 by 10 micrometers on a side, storing 256 bits of ID information and costing a small fraction of a penny. My Hitachi colleagues called the chips "smart dust".

The 2011 Tōhoku earthquake and tsunami diverted Hitachi from retail RFID and towards Japanese economic recovery. A GOOD THING! RFID-supercharged storefront retail, out-competing online commerce, would have exposed far more of us to COVID. But I digress, except to emphasize that different economic paths come with different benefits and RISKS.

The risks of SBSP are NOT trivial; I'll go into those after telling you more of the "silicon miracle story".

Silicon density is 2.7 grams per cubic centimeter. SiO₂ glass and aluminum (for insulation and wiring) are similar. That is 2.7 PICOgrams per cubic micrometer. Hence a smart dust chip massed less than 50 nanograms.

In 2024, after 7 more generations of Moore's Law scaling (with rapid layer growth in the "vertical" direction), we can cram vastly more transistors into a "dust-sized" chip.

Each centimeter-square, 200-micrometer-thick memory chip in my Samsung drive contains more than 2 trillion transistors, including many real-time self-testable hot-swap-able spare pages of memory. That is 25 femtograms per transistor. It will be less than 1 femtogram per transistor-equivalent a decade from now.

Note that the photons that image these transistors today are 100 electron-volt extreme UV, verging on soft X-rays, vastly denser energy flux than above-atmosphere extreme solar UV. Future transistors will be practically invulnerable to space radiation, AND redundant, AND reconfigurable. How? Don't know, not my specialty. Thousands of Intel engineers are solving these problems a few miles away from me. If they don't, Asia will eat their lunch (and also yours and mine).

Good for us, good for Asia, and EXCELLENT for 21st century space tech. NOT to make gigawatt microwave arrays in space. Instead, gigawatt COMPUTING and ARTIFICIAL INTELLIGENCE arrays in space, transmitting information beams, not power beams, to Earth. Peta-transistor per kilogram systems. Note that a human brain is about 0.6 peta-synapses, and consumes 12 watts. Electronics will do better, soon.

NEXT: Rephrase Gerard O'Neill's excellent question: "Is the surface of a planet really the right place for an expanding ARTIFICIAL INTELLIGENCE civilization?"

I've explored this line of reasoning for more than a decade, starting with in near-Earth MEO orbits. I'm currently considering vast arrays of gossamer "thinsats" in Lissajous orbits sunwards of the Earth-Sun L1 point. A vast ring-like region in space, where AI can collect MUCH more than "200 EARTHS" of solar energy. Intercepting a small fraction of the light currently reaching Earth, tunable between 0% and 5%.

AI arrays mostly made of lunar-abundant elements, with one huge and pesky exception - the recyclable carbon anodes needed to process lunar regolith into silicon and aluminum and glass.

Where-oh-where will AI find 500 gigatonnes of unwanted carbon? :-)

What riches will AI trade for that carbon? Perhaps teaching our children to NOT be ignorant superstitious paranoid buttheads like their parents and ancestors? We can dream; giga-intelligence can teach us how.

Much depends on the path dependencies that we bake into the AI systems. We can do better than the hominid path dependencies that baked Hiroshima.

I can write a LOT more, but I write MUCH BETTER with collaborators. A few of you are mentally agile enough to pivot, catch up with me, and blow right past me. I will cheer as you zoom by.

Ad astra per AI!

Keith L.

2024 Feb 15 Thursday, 2100 PST

- - - administrivia, content follows - - -

I'll start this by saying my outbound mail software complains about a zillion recipients. So the people at the end of the Cc list won't get this email, or the prior one.

A Mailing list server is the Right Way to do this. I can set that up on my server, but not without explicit permission of everyone on that specific list. Others may not be as "ethically encumbered".

I will also post my last missive and this one at


That server will be migrated from MoinMoin to mediawiki, https arranged, etc., REAL SOON NOW. Sadly, I am better(?) at grinding numbers than arranging secure websites. Help?

- - - - - content below - - - - -

On Thu, Feb 15, 2024 at 2:29 PM Keith Lofstrom <> wrote:

> Could it be ... that broadcasting terawatts of ISM-band microwaves > to Earth is a Really BAD Idea, and there are cheaper, safer, and > more profitable ways to exploit many EXAWATTS of space solar > energy at the collectors, IN SITU ?

On Thu, Feb 15, 2024 at 05:20:40PM -0800, Keith Henson wrote:

> Cheaper is what the studies are about. I don't know that anyone has > proposed a power satellite beam that has any safety concerns, not even > for birds.

You are missing the point. We already use the spectrum from DC to daylight to move information, vastly more data every year. Some life-saving information - routing emergency vehicles, airport radar, detecting and intercepting incoming warheads. Half the planet uses cell phones. Most of the rest would like to.

Learn about radio receiver design. Filters are noisy; modern radios put a low noise amplifier (LNA) in front of the first filter stage, to boost signals and minimize the relative effect of filter-generated thermal noise.

Sadly, low noise amplifiers are slightly nonlinear, and that nonlinearity mixes unwanted signals into the desired band, CUBE LAW relative to the strength of the unwanted interfering signal.

Actually, EVERY circuit is nonlinear; including the devices (Power FETs? Magnetrons?) that convert generator or photovoltaic power into microwaves for communication and radar (or hypothetical powersats).

Increase the power to megawatts for radar, or gigawatts for SSPS, and inevitably increase the unwanted signal radiation in other bands.

For most uses, we design the converters to meet strict limits on out-of-band radiation. Either by generating "class A" sine waves between DC rails, or by generating chopped "class C" square waves then filtering the hell out of them. Class C is more power efficient, but the filters can be lossy and heavy. Doing a Really Good filtering job is neither cheap nor energy efficient.

And when you are moving a million times as much power as a radar, you need a million times better filtering. Otherwise, you will destroy important uses of the spectrum.

Such as radar, which you will need to defend your power satellite from those who get REALLY ANGRY about your out-of-band emissions in their band, or your high power main signal blinding their receivers with intermodulation. If we develop cheap enough launch technology to loft power satellites, that same launch technology can launch much cheaper "devices" to convert power satellites into confetti.

"But my power isn't pointed at you," some will yammer. "I can focus my power with 150 decibel suppression of sidelobes." Nope, only with PERFECT zero-error components, which cost infinite dollars. By the same "reasoning", I can give you a Really Close Shave with my cheap Harbor Freight chainsaw.

"Well then", asks a properly skeptical critic of my AI-at-L1" brainfart, "how do you propose to deliver a gazillion bits per second over millions of kilometers to a cloudy Earth?"

"Er, uh," I stammer, "laser communication relays".

Which, frankly, we haven't done much of yet. What we have done is LIGO, which controls mirrors to "fraction of a proton diameter" accuracy. Clouds and heat diffraction will interrupt space-to-surface optical links often, but not over the entire Earth, all the time. Someday, a net of millions of laser space-ground links, with a global net of fiber optics connecting functioning links to customers anywhere on the Earth.

Perhaps, given with that global capability, we can screw up microwave-band communications and get away with it. Until then, be ready to clean up a mess before you make it. I don't use my power saw before I find the broom.

Keith L.

-- Keith Lofstrom

SSPS (last edited 2024-02-16 05:10:48 by KeithLofstrom)