Latest Space Based Solar Power Concepts and Experiments at Caltech

60 thoughts on “Latest Space Based Solar Power Concepts and Experiments at Caltech”

1. They did tests long ago. It was called “Broadcast Power”. The relevant factors are wavelength, distance, size of transmitting antenna and size of receiving antenna. The receiving antenna will measure in sq miles. There are formula to calculate the Aperture on the web. The satellite will be geostationary.

2. IR laser implies a beam director, and you aren’t likely to use a single monster one due to modular layouts and a desire to reduce cross-connect power links, so the limitation is how many reasonable beam directors per sat, versus the how many different spots can be multiplexed/serviced with microwave, at a full transmitter level, and at a portion of the array if you are using sparse array cheats (think 1 out of 100 modules in a line are members of a unique sparse array). A 500kW-1MW standard spot for delivery/ordering purposes would allow for some calculation for how many spots can be serviced.

   For a given microwave array, how would one go about estimating the limit off a pulse dwell time to determine the multiplex spot maximum number? Is it wavelet generator ability to keep up with the duty cycle, along with the chosen frequency setting the single wave timespan working frame?

   I guess which hard limit for spots would come up first? the max serviceable spots for a given microwave array at some frequency, versus the minimum diameter/mass of the IR laser beam director at 12000km limiting the practical number of beam directors.

   Then there are now noises regarding optical phased arrays (chip size level), but if they could be made into thin film rolls and phase locked…

   https://www.nature.com/articles/nature11727

   https://www.nature.com/articles/nphoton.2016.104

3. You know, I see that argument. (“from Luna”). Thing is, at perhaps on the order of 350 to 1,000 tons per km², it isn’t all that much once the likes of SpaceX pull out in front an actually deliver $1,000,000 a ton (all in) pricing for launch services. 

   Basically ⅓ to 1 billion a km². In the big picture, considering each km² is going to produce (and beam, and have received at Dirt) no less than 330 MW/km², the budget seems to fit. 

   From a simplicity point of view, I am seeing merit in and embracing the idea of conversion-to-coherent IR laser wavelengths on a per-module basis. Fiber optics can carry hundreds of kilowatts in single fibers these days, with essentially 1% losses for under–10-kilometer hanks. About right for a 5 GW (transmitter power) space-based setup. 50,000 fibers, 100 kW each (IR). 

   As a designer, I think there’d be some hard tradeoffs between using infrared and microwave: certainly the size of each transmitter “antenna” whether synthetic aperture or just a parabolic, is substantially different. But the MW allows thousand-time-a-second (or some duty cycle) beam tracking and switching. IR doesn’t. 

   GoatGuy

4. Test systems should come from earth. Later, PV, and structural components could be made with Lunar materials, with the balance of plant launched from earth. Eventually, everything could be manufactured in orbit from Lunar, and asteroidal materials. Imagine how easy vacuum deposition, additive manufacturing, and smelting would be at L5. Unlimited vacuum, unlimited high temperature heat at the cost of a mirror, the backside of which is a waste heat radiator, zero g crucibles of sound, plus all the products of smelting would be useful, even the oxygen.

5. I’d say microwave transmission is best suited for huge rectenna farms at latitudes of 35 degrees or less, and very large powersats with huge phased array antennas. Between the large aperture of the transmitter, and the large receiver, a reasonable transmission efficiency can be reached.
   For the US, I envision a rectenna on the gulf coast, near the Texas-Louisiana border with a capacity of many gigaWatts. The petrochemical refineries of the area could be powered by it, so that less of the petroleum would be burned to power the installation. The area is also well situated to help power the Texas, and eastern US grids.
   Later the existing petrochemical refineries could be repurposed to creating synthetic fuels, and ammonia using the beamed energy. The pipelines, right of ways, deep water port, and riverways are already in place to distribute the fuels/ammonia. Need more energy? Easy, Peasy, add a square kilometer of rectenna on the ground, and PV at GEO. With a large enough rectenna, the beam could be steered around thunderstorms. The single huge rectenna farm would also cause fewer problems, with aviation, since it would be well known.
   Smaller loads, like electric airplanes in flight, or beamed power launch vehicles would be better served by laser transmission. It’s possible to stop transmission quickly if the beam goes off target. For launch vehicles, a chemical booster, or ground stations could hande lift off, with orbital power finishing the job.

6. Yes, yes and yes.

   RADAR now uses phased array non-moving (mostly) systems to synthetically ‘point’ the outbound (and high-gain receiver) beam thousands of times per second. To track hundreds of targets, yet still scan the area for targets not yet acquired.

   Electronically ‘pointing’ a beam with high accuracy is an exercise in subtly shifting the phase of each tiny dipole antenna in a large rectenna with respect to its adjacent neighbors. The phase shifts are VERY small, in fact smaller directly in relation to the accuracy with which the outbound beam is to be sent. 

   On curiously attractive approach however is for each intended ground receiver to beam its own “pilot beam” at the SBSS rising in its Molniya orbit, establishing the exact phase coherence and delays necessary for the satellite to emit a return beam, at high power. A “laser pointer” beam out, and a 100-to–5000 MW pulse back. Such a system could fairly easily support hundreds to even thousands of ground stations. 

   Geometrically however is the time-of-flight and how far the return-beam misses the target. I’m pretty certain however that the needed ‘fudge factors’ could either be done mechanically (i.e. the “laser pointer” transmitted ahead of where the delayed return pulse is calculated to hit), or even just mathematically. 

   In any case, it also FEELS like there are definite size-vs-safety-vs-usefulness issues involved. 

   Just saying,
   GoatGuy ✓

7. Ah, didn’t see the discussion below.

   So if I am reading that right, 660m transmitter diameter for 60 GHz to a 50m spot from 4400km, 1800m transmitter diameter for 60GHz to a 50m spot from 12000km. So a roughly 2km transmitter for a Molnyia orbit, which is not entirely insane (compared to the 45km transmitter diameter at 2.4GHz).

   But, how far can you cheat with a sparse array rather than fully filled in diameter for the transmitter?

   If at ¼ kW/m² you are only getting half a megawatt on a 50-150m parabolic array receiver that sounds like neighborhood level power. How would that compare to a wire mesh rectenna though? Is the 500kW dedicated only receive power level or available even with reasonable multiplexing using that pulse delivery distribution trick?

8. Um, what frequencies were you thinking of with a 150m diameter effective receiver? Is this for a GEO SPS? I thought the nominal microwave frequencies basically forced you to multi-km receivers due to the distance from GEO? If you had a tight beam to hit 150m, using a huge array of 5cm dishes is not insane (would look like a field of bubblewrap), but aren’t most SPS designes baselined with a rectenna covering a 5km circle?

9. Yes and no, collection/concentration can be separated out, but a lot of SPS designs are now going to modular integrated conversion and transmission concepts, notably the sandwich panel shown here and the Z panel. The aim is to minimize the distance between conversion and transmission systems, and highly modularize them so the power level is low per module, such that low power level electronics and lower amperage/voltage wires can be used. This is a reflection of the realization that older SPS designs had a single monster PV panel feeding power to the 1GW transmitter through a single rotary joint, and it was this joint design that was troublesome. A lot of the weird SPS designs go through various contortions to avoid a single high power rotary joint. The NASA butterfly design would use sandwich panel modules for instance, using an unpowered rotary joint to deal with the main collector mirror array. Mankins SPS ALPHA design uses an earth facing Z panel pagoda so they can reject heat away from earth.

10. My just-purchased 128 GB thumb drive has 36 chip layers, each lapped from almost 800 µm down to 25 µm without even 1 in 1,000 chips being accidentally broken in the process. The raw wafer making and cutting (and lapping) folks are doing wonders in this area. 

    Takes a lot of black diamond dust, but thankfully it is 99.99% recyclable. Almost indefinitely. 

    The only reason I’m not hopping on your metabolized Mylar bandwagon is because of the long term heat dissipation and flexure-pad-bonding (delimitation) issues that Mylar has, like so many plastics. Don’t get me wrong: I like the stuff, and it makes great modest-durability space structures. One might even consider carbon-fiber reinforced spools of it as a flat stock, at the 10 µm-per-layer level as “roll stock” to be fabricated into struts, beams, other hollow-but-strong structural spider-fab-maker wares. 

    I think the elevated temperature and single-crystal stability of the lapped silicon substrate though is markedly superior over decades of in-service operation. And we make it copiously, cheaply, today. Just have to lap it wholesale to 25-to–40 µm.

    Just saying,
    GoatGuy ✓

11. psssh, that’s just engineering… /sarc

12. so to keep it shorter this time:
    Use 60Ghz in the oxygen-absorption band. Then those GW of power just dissipate into the atmosphere. Of course, you can’t get your power down to the ground!

    Additional cost needed for stratospheric balloons/airships to convert from 60Ghz to 2.4Ghz for terrestrial transmission. It’s expensive, but the only way I see that would provide physical guarantee against the ‘death-ray’ scenario.

13. are you freaking kidding me? I just lost a nice post in response to this. grrrr 🙁

14. When I did the math (a long time ago), I just couldn’t get to a reasonable mass balance with silicon of any thickness. It has to be true thin-film – metallized layers on polymer sheets (aluminized mylar) with perovskite sputtered/printed on that.

15. Good point. While the idea of modular puzzle pieces is really tempting, I think in the end you’re going to have to segregate responsibilities for capture and transmission.

16. Panels = 5um thin film (e.g., aluminized mylar); perovskite is looking very good (so far) and hitting terrestrial targets seems doable.
    Panel support = inflatable; with spider trusses where needed – I probably didn’t budget enough mass and associated launch costs for this
    Satellite infrastructure = primary trusses, comms, station-keeping
    Ground station = a fraction of solar, since (theoretically) the rectennas are simpler, along with higher power densities and higher areal efficiency
    Power electronics = $0.25/W terrestrial solar; account for heat dissipation, longer life, tons of redundancy, vacuum-proofing; I think $1/W is mildly optimistic.
    Launch costs = Musk once claimed $200/kg but $1000/kg seems reasonable with extensive reuse. Assumes LEO/SSO only.
    Design = a cool billion for R&D, design, project management, engineering, etc? Yeah, easily. Good news is that this should effectively be a fixed cost.

    Ugh, i did forget microwave transmitter. That’s additional power electronics, heat dissipation, antenna mass, and additional support.

    So I think maybe $10/W is more likely for the first few GW. 🙁

    I think it’s interesting that it is feasible for SBSS to hit costs similar to nuclear; assuming a few billion in R&D… 🙂

17. The real “biggest problem” with 60 GHz is coming up with an efficient way to generate it, at power levels that are meaningful.  

    If (in my number examples herein) a 5 GW transmitter is endowed with a 1,800 m diameter synthetic aperture phased array transmitter, at 60 GHz (5 millimeter), then it takes over 90,000 dipoles per circular m² or a total of  230 billion of ’em at 22 milliwatts each, or, something rather more complicated (but still functionally equivalent) of 2,000,000,000 of them, at the focal point of well aligned parabolic reflectors… running about 3 watts each. A bit less… 2.5 watts.  

    And the problem then becomes EFFICIENTLY synthesizing the 60 GHz with sub-picosecond phase jitter, programmable over 3 orders of magnitude range. To be able to focus the oblate ellipse in real time, tracking a rotating Earth, a rapidly changing Molniya orbit azimuth, and who knows what accumulating orbital aperiodicities due to the precession of the apogee and perigee over time. Quite a workload. 

    60 GHz also doesn’t efficiently “synthesize well” even without the picosecond phase jitter issues. Ideally, one would want to “digitally synthesize” it by gating picosecond duration pulses at the peaks and troughs of the wave. Can get to high efficiency that way. But at 60 GHz, the rise/fall time is too long for that.  

    Just saying,
    GoatGuy ✓

18. ⊕1 … but:

    Buck-a-watt for panels is posible, but seems — given the special treatment (alluded to elsewhere) of them — to be closer to $2/watt. Moreover, the frames that hold the square KILOmeters of lapped PV together are going to be costly.  

    1 kg per kW … if lapped to 40 µm and with a mean density of 4,500 kg/m³ (including perovskite, silicon, copper, possibly titanium), comes out to 180 tons per km². For the PV alone. Not the frames. That’s only 180 g/m² for the PV, leaving a 820 g/m² budget for the frames, tensioning, power conduits, etc. Doable? Mmmm… feels like a maybe-kinda-sorta thing. 335 MW (microwave out) per km² too. 

    Buck-a-watt ground station? Possible. But if its using (as it should) modest sized parabolic 5 cm diameter reflectors and liquid-nitrogen cooled 5 mm dipole receivers (receiving 4,000 W pulsed, from the 5 GW station transmitters), and covering 50 m diameter (realistically more like 150 m, with slop-over skirt)… dunno. Doesn’t seem right, especially since — again, realistically — one is going to want dozens-to-hundreds of similar stations time-division-sharing the output of the Big Furnace in the Sky.

    Buck-a-watt for power electronics? 5 billion for a 5 GW transmitter and all the associated stuff? Heady! I want the contract. 

    Buck a watt for design? 
    Ooo … ooo… that’s the contract WE want, you and I.  

    Anyway, just wondering quantitatively out loud. 
    Just saying,
    GoatGuy ✓

19. Also, there’s another really significant problem. With SBSS of any size that comes close to replacing terrestrial conventional (including nuclear) power: potential for civilization-destroying mayhem. 

    A 5,000 MW (just did the numbers) SBSS … 5 GW transmitter, with a 1.8 km phase-array, 60 GHz transmitter, would be able to image a 50 m spot from 12,000 km; the mean power is over 2,000,000 W/m² (!!!) dirt-side; it wouldn’t take very much military conviction to melt down whole battalions (at the least) of field troops. 

    Poach ’em like balut¹ in their shells. 

    But if aimed at cities… anything above-ground is destroyable. Telephone exchanges, power substations, pumping plants, metropolitan civic governance buildings, police stations, freeway interchanges, crossovers, bridges, flyways. Airport runways, fuel depots, refineries, stuff-of-living bivouacs (produce, Walmart crâhp, all that). 

    At 2 MW/m², the destructive potential is … profound. Probably wouldn’t take more than a few weeks to kill off a city the size of New York. 

    Imagining hundreds of these 5 gigawatt birds flying in their Molniya orbits is, well, sobering to the point of warranting against the idea completely. Yet, if SBSS is going to replace everything including conventional nuclear, … we’ll need hundreds-to-thousands of the birds flying, all the time. 

    Sigh… 
    Problems.

    Just saying,
    GoatGuy ✓
    ________________________________________
    ¹ boiled fertilized eggs
    ht

20. That last bit is key: even at an astounding 40% conversion (I wrote, and lost, a post estimating 33% for variously defensible reasons), dissipation of heat is key: as shown, the relatively thick panels, flexible-or-not, tend to encumber the outward radiation of heat (infrared) so necessary to keep the conversion efficiency up, and the lifetime (thermal degradation rate) of the cells conserved. 

    It is one of the chief reasons I see high value in lapped silicon over other substrates. The perovskites can (should, must) be sputtered over a conventional silicon PV base layer. In manufacturing, 500-to–1000 µm (½ to 1 mm) thick silicon is perfectly normal. However, once diced up and max-stress tested, the triple or quadruple layer cells should be ground and lapped down to 25 µm or lower thickness.

    This has been perfected in the last 15 years, and is commonly (and invisibly) deployed in our ubiquitous “thumb drive” SSD devices. There are over 36 layers in a single unit storing 64 GB, it turns out. Each of them only 25 µm thick. Except the bottom one. 1,000 µm all told. 

    Grind-and-lap to 25 µm makes the cells far more flexible (harmlessly). Also, shaves away the mass (crucial for SBSS). And being flat, the radiative waste-heat problem is well contained. Both top-and-reverse surfaces dissipate. Especially if reverse is given high-compliance super-mossy black etching just before deployment. 

    Just saying,
    GoatGuy ✓

21. Without getting TOO gushy, thank you for reading what I wrote. I think you and I actually agree more than disagree on the various technological advantages and shortcomings of space-based solar power. 

    You are right on the money, viz a vis hundreds-to-thousands of ground station targets time shared in short pulses, rapidly switched. Too high of a microwave illumination intensity, and the receivers tend to become lossy and complex to handle the blasts-of-gigawatts delivered. But, still, doable. 

    Likewise, if we free our minds of always latching “production must be beamed now” thinking, (i.e. allowing some short term (seconds-to-minutes) storage aboard the SBSS), then even the power-PEAK-of-each-pulse becomes tunable. For larger ground spots, even “just parts of the transmitter array” need be employed to deliver the power. (From the math, above). 

    Which in turn really frees up the scheduling of deliveries. But puts substantial overbuild-capacity requirements onto the SBSS synthetic aperture phased array power wavelet generators. 

    Again, good discussion. 

    Just saying,
    GoatGuy ✓

22. “Enough battery to get through a windless night is really expensive.” Really? Compared with $2/kWh…I highly doubt it. And how much are you going to use at night? LED lights don’t use much. If you are not opening and closing the fridge all the time, it does not use much. There are efficient televisions. I wouldn’t suggest using electricity heat your house, run your hair dryer, run your dishwasher, run your water heater, or run your clothes dryer, when no electricity is being generated. And some of this can be done other ways entirely. There are more direct solar means to heat houses, water, and dry clothes. And washing clothes by hand is not that much more involved. Or there is gas.

23. New underwater circuits are generally HVDC. Dual-conductor systems are balanced, with minimal magnetic and electrical fields compared to HVAC.

24. thanks for doing the math. You can quickly see why 60Ghz frequencies are so tempting. I really doubt microwave power will ever ‘scale down’ given the laws of physics. Even distributed aperture or active phase array transmitters don’t help reduce the spot size. Size matters in this case.

    There’s no direct relationship to concentrated solar space, but it does help shortcut the conversations about reusing existing solar power plants by reflecting sunlight from space mirrors.

    Interesting idea regarding pulsed power. While phased antenna do allow multiple targets, pulsed power would provide support for a higher number of concurrent ground receivers, increase efficiency, and justify larger transmitters. Rather than 10 100MW SBSS satellites, each with a duplicated large transmitter, you can now have a single 1GW SBSS satellite (assuming coverage isn’t a problem) with a very large transmitter; reducing the ground station spot size. When power density is the prime bottleneck, not your collimation or spot size limits, the ‘safety’ of pulsed microwave allows higher densities. Scaling is easier too, all a ground station needs is more power management and a request for additional cycles. Phased antenna should be able support rapid switching and beamforming necessary to support this model.

25. “has anyone actually transmitted big power over those sorts of distances yet?”

    Nope. Not even terrestrial “playpen distances”. (i.e. unnamed tall hill to another unnamed tall hill, 1,000 km off, but still “on the visible horizon”). Not yet. Not gigawatts, not megawatts. Not even dozens of kilowatts. I think some group or another has transmitted a hundred watts so far or so. 

    Moreover, the notion of it being 90% efficient is baseless.  

    At the transmission side, converting 60% of incoming DC to single-resonant frequency microwave power is considered excellent, in practice. Laboratories can do more, approaching 80%, if the wavelength is long enough.  

    The atmosphere column would itself absorb no less than 10%, from nothing more than aerosols, dusts, water vapor, organics and all the rest.  

    Then, there is the “problem of the receiver”. Ideally, one imagines having a kind of super-duper microwave diode that has no forward-voltage bandgap.  That lets the tiniest forward power thru, and blocks the reverse… so a pair can gather both + and — pulses to nice sweet DC power with some filtering. 

    Reality bûggers that. Even superior FETs have body resistance. And getting them to synchronously rectify is no cakewalk.  

    Anyway. I’d be surprised at more than 50%. Really surprised.  
    Just saying,
    GoatGuy ✓

26. Please see what I wrote, above, in response to this very old Science Fiction idea. The problem remains the same: trigonometry and optical physics. No way around its limitations without frequency conversion. While one might imagine sunlight-optical-pumping of long laser rods, it turns out that the complexity of that system is without exaggeration, hideous. 

    Better to use PV, make DC, use that to excite high-efficiency microwave transmission semiconductors, generating microwave at very high efficiency, and potentially picosecond jitter (think phased=array beam steering) thousands of times per second across the whole thing.  

    Again, see the write-up to David Freeman

    GoatGuy

27. Part 3

    The TMD ranges from 13.2 meters (60 GHz, 2.5 km spot, 4,400 km orbital distance) to 36 m, same GHz, spot … but 12,000 km apogee. However, with other parameters, it isn’t so sweet.  

    330 / 900 m for 2.4 GHz, same orbit. 
    660 / 1800 m for 60 GHz, but for a 50 m spot size, same orbit
    16,500 / 45,000 m for a 2.4 GHz, 50 m spot, … same orbit. 

    Your mileage may vary, as the saying goes. Thing is, that the size of that transmitter array (AKA mirror) is independent of power. It just depends on distances, wavelengths and borrowed-but-accurate optical physics. And trig. 

    And if we’re making smaller spots, the problem then becomes not desiring to have TOO high a unit-area (1 m²) power density down here on Dirt.  

    50 m spot with ¼ kW/m² has 500 kW of microwave energy hitting the bull’s eye. That’s not very much of a space satellite. Even if the transmissions are heavily duty-cycle modulated (i.e. intrinsically much more powerful, but only delivered in short pulses, perhaps 1 millisecond wide, of 100 megawatts, 5 times a second.)  

    At least that way, variable-delivery power would be possible. Not very good to walk through (it’d feel like insects crawling on one’s skin). But manageable. 

    Just saying,
    GoatGuy ✓

28. … and so, it then becomes obvious that none of the orbits does a good job at delivering reflected sunlight itself. What is needed is to get s smaller spot size, ideally since it’d take frequency conversion, to a source that can be highly collimated, spectrally simple (not a bunch of wavelengths), and offering high conversion efficiencies both in turning space-based power into waves, focussing them much better, and on Dirt, synchronously (super-high efficiency) rectifying them to DC power, which thru modern DC→AC conversion, becomes grid-friendly directly.  

    What spot size?

    Well… that depends a lot on what the application is. If merely delivering a bunch of reliable power to an interstate or inter-national grid, then a big spot is OK… 2 or 3 km across is fine.  If however, one is hoping to deliver power to a critically vital micro-customer, such as a battalion or company (50 to 500 soldiers), well … I’m sure that 50 meters is pretty big from their point of view. Setting up a power antenna array half the size of a football pitch is asking a lot.  

    But those then define the collimation problem. Those, and choosing a wavelength for the radiation (power) to be transmitted. David espouses 2.4 GHz and 60 GHz bands … they in turn:

    transmit mirror diameter = 1.5λ/receiver-angle
    tmd = 1.5λdistance/spotsize
    tmd = 1.5 × ((0.3 / (2.4, 60)) • ( 4,400,000 to 12,000,000 ) / ( 2500, 50 m)
    tmd = … 

    MORE…

29. Recall … from PHYSICS (specifically, optics, and actually in turn, trigonometry…) we have a sad, but still mathematically marvelous finding.

    Ideally, the 9th grade space-science (and 1950s SciFi authorship) hits on the idea, “lets just beam a reflection of sunlight down on a patch of land, concentrated with a lens, and use that to make power Earthside.”

    The problem is “the focal plane” and the image size of the Sun. In the shortest representation, the angular size of Sol is approximately:

    Asol = arcsin( DIAsol / DISTsol );
    Asol = arcsin( 1,391,000 km ÷ 149,500,000 km );
    Asol = 0.0093045 radians
    Asol = 0.53311°

    And with that, one can also say with good authority that the minimum size of the circle of the image of the Sun, projected down here on Planet Dirt would be:

    ImageDiameter = Asol • lens-distance;

    Which for almost any MEO (Molniya style) orbit of 12,000 km apogee (furthest) and a ½ sidereal day period and a focal horizon somewhere between 35° to 145° of sky-angle, is a distance of 4,400 → 12,000 → 4,400 km so:

    ImageDiameter = 0.0093045 × [4,400 to 12,000]
    ImageDiameter = 40 km to 111 km, respectively.

    So, these solar images, being rather large, aren’t terribly helpful for generating electricity earthside, although the simplicity of a great slightly curved solar mirror has its attraction.

    MORE FOLLOWING

30. There were microwave beaming demos down in Hawaii (I vaguely remember Goldstone also being mentioned) by NASA/JPL, which was in the roughly 15 mile range. While not huge amounts of power since besides radar nobody has transmitters that powerful, it did confirm the basic physics with reasonable amounts of power (I think they used a DSN dish?) so most of the researchers felt it was a solved problem.

31. Take a look at HESPeruS and CASSIOPeiA SPS concepts for Molnyia style orbit designs

    https://www.researchgate.net/publication/309671368_Hesperus_-_Highly_elliptical_solar_power_satellite

    https://www.researchgate.net/publication/331231886_CASSIOPeiA_SPS_Part_1_-_Introduction_and_Background

    Stratosolar was previously working on high altitude tethered aerostats as an intermediate receiver as an endrun around atmosphere absorption.

32. At the moment, I agree, but if SpaceX pulls off the super cheap launch costs that Musk talks about for the BFR, that’ll change. Enough battery to get through a windless night is really expensive.

    The beam doesn’t blast down to your location, it blasts down to a collector well outside of town. Except it doesn’t exactly blast; birds can fly through the beam and just warm up a little bit.

33. As for costs:

    • $1/W panels (perovskite thin films)
    • 1kW/KG = $1/W launch costs (assuming reusable rockets bring costs down to reasonable levels; now imaginable)
    • $1/W ground station
    • $1/W power electronics
    • $1/W satellite
    • $1/W design

    So we’re looking at $6+/W, which is pretty high compared to turbines (~$1/W), but not that much higher than modern ‘clean’ coal, and certainly cheaper than nuclear.

    I have not considered the time value of money, costs of maintenance and replacement, or fuel costs for comparison. However, as a ‘back of the napkin’ exercise, using mass-production costs that are not unimaginable, we can see that SBSS has the potential be a competitive solution with modern terrestrial energy sources.

34. In the end, SBSS will replace almost every other source of terrestrial energy, with the exception of solar during the day to power our air conditioners.

    Goodbye nuclear :'(

35. Oh, another option. Use SBSS in SSO orbit, providing power in a shifting band that is optimized for the ‘duck’ demand curve. As the sun sets and people get home, solar is no longer meeting demand for about 4 hours demand/supply is lop-sided. If SBSS could provide additional power during that gap from a relatively low SSO orbit, that would mean capacity would be present when most needed.

    That opens additional options. By not attempting to meet baseline needs, and just filling-in for solar, then we can overlay microwave receivers on terrestrial solar plants. Since we’re looking for a 1:1 replacement, effective power density only needs to be 200W/m^2 (or lower, since microwave would have greater areal efficiency). With such low densities and relatively large area available to a relatively ‘low’ SBSS satellite, 2.4Ghz could be used in favor of 60Ghz.

    Downside is that crossing the Pacific, arctic, and southern oceans is a complete waste. The advantage is that it’s possible to generate revenue quickly at lower cost.

    I still think using SBSS in reverse to power electric drives to GEO and beyond is great, cost-insenstive niche to start with. Then figure out how to deliver terrestrial power to places that are also cost-insensitive (offshore oil platforms, military bases). Here, though, other considerations may play a role (e.g., land needed for collectors).

36. I read an article somewhere, which I can no longer find, proposing a very different, and much more realistic, alternative to 5GW GEO monstrosities operating at 2.4Ghz.

    1. Launch into Molniya orbits. You need 3 satellites to provide full coverage, but 2/3 of the time you’re on a useful track to provide power to US/EU/RU/CN/JP.
    2. Use 60Ghz bands, avoiding the oxygen-absorbing bands. This significantly reduces dispersion and sideband (or side-lobe – I’m definitely not a microwave engineer) losses; at the cost of greater atmospheric losses.
    3. Avoid the mess of satellite-to-satellite transfers in an attempt to work around limitations of GEO/SSO/LEO and 2.4Ghz.

    This is probably the most straightforward approach to bootstrapping for terrestrial purposes. This model allows low-power satellites (~100MW) begin transmitting useful amounts of power almost immediately, without the need for large ground installations, large transmission antennas, and massive multi-GW installations.

    My only tweak on this approach would be to embrace the oxygen-absorbing bands to make it ‘safer’, and then figure out how to create a intermediate transmitter working at 2.4Ghz (or a variant of 60Ghz). Past proposals have considered lots of LEO power transmission satellites for that purpose; but I think stratospheric balloons or airships could provide a viable alternative.

37. In the end, it’s all about cost. Does 60Ghz reduce mass (and therefore cost) of SPSS? Does it reduce complexity? Improve efficiency and thereby reducing heat dissipation requirements?

    Does the cost of a stratospheric balloon and antennae assembly, including costs of keeping it on-station, come in at the same or less of a large ground station?

    Are there benefits to allowing GEO or Molniya orbits that 60Ghz allows, and is it worth the additional cost of GEO/Molniya?

    Are there niches where 60Ghz could actually bootstrap better than 2.4Ghz?

    I would argue that while we usually consider terrestrial applications (e.g., rural power, long-rand power transmission, military bases), there is another customer that would bear high costs initially in return for unique capabilities.

    Launch low-cost LEO or SSO SBSS. Direct the power, not at earth, but at space. Imagine space tugs that are able to receive, consistently, multiple MW of power to feed VASIMR or ion drives. If it’s cheaper than schlepping propellant to deep space, it’s a good deal, no matter the actual cost. In addition to carrying cargo to GEO, it’s possible that with big enough drives we could gain enough velocity to ferry cargo to the moon or beyond.

38. Yes. This would allow the use of ‘safe’ 60Ghz frequency. At 15-18k meters, oxygen absorption is almost a non-factor.

    The challenge is making the balloons cover enough area to capture a decent % of incoming power. You would probably end up throwing away a lot of sideband power in order to minimize the size of the balloons.

    This can be mitigated. The high safety, low dispersion, and higher power density of 60Ghz would probably allow 10+kW/m^2 effective power, instead of the 200W/m^2 of terrestrial solar.

    A single circular array ~120m diameter (perhaps an antennae array suspended between multiple balloons?) would collect ~100MW of power. The downside is that then needs to be converted to 2.4Ghz and beamed down to a ground station. Even there, though, the advantages are that with being so close we can use much smaller ground stations, low dispersion or sideband loss. Ground station is determined only by ‘safe’ microwave densities, and with other considerations handled those could be higher than pure SPSS. Safety-wise, the balloons are effectively fixed position, we know where they are, they have limited ‘range’, and we can shoot them down. In such a case, it’s likely we could reach 2kW/m^2 ground density (10x solar). 100MW transmitted by the ballons would require <500m diameter antennae (including extra for sidebands).

39. I think space solar, is just not viable at the moment.

    And any solar system on the ground with batteries would be cheaper (unless you were in the Arctic or Antarctic in their winters. And then you can probably use wind instead).

    I’d be pretty nervous too with that beam blasting down on my location to power my dishwasher.

40. for crying out loud can we please increase the message posting size?

    Most 2.4Ghz proposals try to match power output of solar, no more, due to safety concerns. Using ‘safe’ 60Ghz microwave, it’s likely that the perception of safety, combine with lower dispersion, and inherent higher power density would allow power densities an order of magnitude greater than solar.

41. Around 2.4/2.5Gh is pretty good, though there is still some absorption by water (rain is not good); but even on a cloudy, rainy day you’d get most of the power through. Unfortunately, the relatively ‘low’ frequency means high levels of dispersion, and in order to maximize ROI of ground stations, most operators would probably ‘let it go’. High dispersion means GEO is pretty much impossible until you reach multi-GW installations, and of course the land and antennae required for the ground stations gets expensive.

    Interestingly enough, due to fears of “death-ray weapon” and “cooking our cities”, taking the +opposite+ approach could be handy. Use 60Ghz with low dispersion means smaller GEO installations are feasible, long-distance transmission is more efficient, and ground stations can be much smaller. The real advantage is that 60Ghz is in the oxygen-absorbtion band, so should some evil hackers try to microwave NYC, all of that energy is immediately absorbed by the atmosphere.

    Downside is that ground stations are useless…

    Upside is that satellite-to-satellite power beaming could actually be feasible.

    Another upside is that we look at it in reverse – how to get power from LEO to GEO instead of GEO to LEO/surface – then power-beaming from cheap LEO satellites to GEO ‘tugs’ is much easier, and the target satellite requires a much smaller and lighter receiver.

42. The military should use little nukes for powering temporary bases. Bury them maybe 100 feet underground. That can power all their air conditioners, lights, stoves, electronics and whatnot. Don’t make it too heavy to fly in and out. Moving lots of Diesel around the globe gets very expensive. Could even extract water from the air using the power.

43. You don’t want that UHV stuff underwater. On high towers it is fine.

44. Grid-generation efficiency is closer to 55% for modern CCGT; 20% if calculating from solar insolation then <20%.

    I don’t think I would calculate terrestrial (or space) solar from insolation, but rather from peak power produced by the panels, since sunlight is ‘free’; it’s the panels that aren’t.

    I also think you’re very optimistic about microwave efficiency. Microwave transmit/receive is optimistic, but not drastically so. However, losses to the atmosphere, sideband dispersal, probably limit the end-to-end efficiency (from antennae to antennae) to <80%.

45. Modified Molniya orbits may be optimal for SPSS.

46. GEO is a fool’s errand. Too expensive, too far away. I think Molniya orbits would be better.

    In the end, I suspect thin-film perovskite cells will win the mass/W race. Complex collectors, concentrators, and cell structures are simply too expensive to assemble.

    What I do find interesting and valuable is the panelization work. Back-to-back solar–>microwave with integrated solid-state power electronics makes it easier to assemble and deploy.

    Finally, how to manage heat dissipation? Even at 40+% efficiency, that’s still >700W of heat per sq m that somehow needs to be dispersed.

47. Orbital Solar to distribution on the ground is comparable in efficiency with overall generator-to-wheel for an electric car.

    Solar-Convertor: Microwave-Transmit: Microwave-Receive: Line-Inverter/Driver: AC-Line

    .4x.95x.95x.97x.96=.32

    Grid-Generation: AC-Line: DC-Charger: Battery-IN: Battery-Out: Control Module: Traction-Motor

    .4x.96x.98x.90x.90x.97x.95=.28

48. Sure, radars do. In normal practice the beam is rapidly scanned, but hold it still, and conductive things, and molecules with the right natural frequency get hot quickly. Military radars can have a peak power in the MW range.

49. The earthbound Global Energy Interconnection sends renewable energy around the world on existing UHV cables.

    Given that it has 2,000 mile cable runs working already, it seems more promising and vastly cheaper than space-based generation.

50. It’s not quite 90% but it is close. I advocate tethered high altitude balloons, that have a wire mesh impregnated in the skin and a conducive tether. It might be possible to get a grouping of them 30km up (above the clouds) where the atmosphere is thinner and use the heat energy from the microwaves to maintain the balloons for extended periods. It could decrease transmission losses by getting the receiver to an altitude where atmosphere molecular density is much lower and will absorb less energy.

51. I know the theory is there, but has anyone actually transmitted big power over those sorts of distances yet?

    I think that you’d assume 15 km at ground level to be about the same amount of atmosphere as going from ground to space. But that’s clearly much less distance so I don’t know what a valid test distance would be.

    As with the first applications, your first actual ground-orbit test would probably start on the ground and send the power into space, just because the setup cost would be so much less.

52. Some microwave wavelengths have very little atmospheric absorption.

53. The US Army has started looking at small mobile nuclear power plants for forward operating bases in places like Afghanistan.

    Perhaps rather than beaming power to space and then back down to a combat outpost, you could just build a radio tower style structure at the forward operating base and combat outpost and use microwaves to beam power between them, or perhaps teathered steerable blimp if that is cheaper?

54. Note these tiles by themselves can’t maintain solar pointing and fire their beam at the earth 24 hours a day due to the geometry. They could be earth pointing and use relay/reflector mirrors to maintain sun angles, or the core techniques be applied to a Z plate system rather than a true sandwich (more like Mankin’s SPS designs which utilize Z panels in a central receiver pagoda structure).

55. Why? An unnecessary and complicated system to provide super expensive electricity. New generation nuclear power plant designs will provide all the cheap electricity we could possibly use for millennia. Even Old King Coal is better when you consider the damage high price electricity does to the poor. And poor people have no regard for the environment. Poor people burn whatever is available to stay warm, cook food or for light at night.

56. Using microwave beaming for long distance power transmission can be a feasible idea. The biggest cost will be the cost of launching the large satellite need to capture, convert, and transmit the beam. A set of satellites at a much lower orbit might be more feasible.

57. Using microwave you can get conversion efficiency over 90%.

58. I remember an old concept of orbiting mirrors that could focus light on ground based solar panel 24hrs a day. I think the concept was called SOLARES. If they used the Space-X launcher it would be economical.

59. I’m pretty impressed with $1-2/kWh. That’s way ahead of where I thought anyone would be.

    Still. I think that if it IS feasible to send bulk MW between orbit and ground via microwaves, then the first option that becomes viable is to have your power generated on Earth, and send it to orbit, to power space things. A lightweight wire microwave antenna will be much lighter and more compact when folded than solar cells. If you can get dozens of MW to a lightweight orbital ship you can play with a bunch of seriously interesting ideas. A high thrust high ISP Vasimir type plasma drive for example.

    The second option is to have big, cheap power sources at location A (solar in a desert, hydroelectric on huge remote rivers, geothermal in a volcanic field, nukes in a country that chooses to have them) and then send it up to a relay and down to a highly populated area B.

60. Converting sunlight to electricity to some form of microwave or laser, having it pass through the atmosphere and then converting it to electricity entails impossible losses. I will wait till we have very light nano carbon cables to transmit electricity from Geo directly to earth.

Comments are closed.