Launch Loop Frequently Asked Questions
What has Keith done lately with Launch Loop?
Not much. See http://server-sky.com for my main focus. 5 gram satellites mean that even at $10,000/kg, people can have their own $100 satellites.
I lost interest in launchers when I realized that launch is limited by demand, not supply (which is subsidized). The world launches perhaps 300 metric tonnes of useful payload into orbit per year; That is 40 milligrams per person on earth. Even at $10,000 per kilogram, that is 40 cents per person spent on launch. Those 300 tonnes are the focus of almost 100 billion dollars worth of space-related or space-enabled activity - $14 per person. Each tonne earns more than 30 times its launch cost. If we find applications that can earn a thousand times their launch cost, they will grow quickly, and there will be incentives to build a lot more launch.
Some space advocates want space travel to be cheap enough so they can do so out of their entertainment budget, as a consumption activity. Perhaps if they cut their entertainment budget to zero, and spent the time and money learning and earning $40M, they can join the other space tourists who have spent that for a brief trip to orbit. If they want a brief experience of zero gee, find 35 friends and spend $5K each on a charter parabolic flight. If they want to float cheaper than that, go scuba diving, or indoor skydiving.
I want launch loop to happen someday - it will be cheaper and cleaner than rockets. But I spent my career helping make integrated circuits a billion times cheaper, and that happened because of customer demand, and was funded by customer money. No big leaps, just decades of relentless improvement by hundreds of thousands of people competing with each other for what is now a $350B/year market. Economics and focus on the needs of others (not thrillseeking, selfishness, and fairy tales) will open the universe.
When will Launch Loop get built?
When there is a market, and after the technology is developed for other applications that can pay for the research, development, tooling, and manufacturing.
See PowerLoop - we can develop the most important aspects of loop storage technology by storing energy. A small loop can launch aircraft (quiet airports!), large loops can load-level intermittent alternative power sources, giant loops can load level the Northern hemisphere for a year.
Before that happens, decisionmakers need to learn the difference between energy and power, and that electricity doesn't store. Demand is fickle, and not driven by price-regulated supply. Batteries are expensive, and dangerous in huge quantities; pumped storage is the only practical alternative, and most of the good sites are already in use. Loops can store a LOT of energy.
8 tons of iron (perhaps $5K worth) moving at 14 kilometers per second stores 800 gigajoules, and occupies a cubic meter of volume. That is as much as a cubic kilometer of water falling 80 meters, or $100M worth of lithium ion batteries.
This is silly!
Yes. I design magic rocks that sing, others that grow fingerprints. Launch loop is almost as silly.
Does Launch Loop need nanotechnology, carbon nanotubes, superconductivity, quantum computing, ...?
The first launch loop designs presumed technologies available in 1975 - 5 micrometer gatelength CMOS integrated circuits, Kevlar 49 aramid fiber (derated 2x), transformer steel. Since then, fiber optics and diamond-coated tool surfaces have become common, and simplify some aspects of the launch loop. All the "magic technologies" listed above might become well developed someday, and make launch loop implementation simpler, but for now they are unwanted and unnecessary complexities, distractions from the task at hand.
Supercondictivity in particular. My masters thesis at UC Berkeley was about superconducting tunneling logic, working under Dr. Ted Van Duzer, arguably the most knowledgable superconductivity engineer in the world at the time. His team worked very hard to produce fragile, finicky superconducting devices; those never "broke out" into common use, and that wasn't due to lack of engineering talent. My fellow grad students run businesses employing thousands of people, making ultra-high technology devices like film bulk acoustic resonators (FBAR, there are dozens in your smart phone, wifi access point, etc.). We failed to do superconductivity, but some of us made billions of dollars doing the slightly easier stuff.
Superconductivity is still very difficult. The Large Hadron Collider failed in 2008, probably because of a bad weld causing a quench, an explosive loss-of-coolant accident, which cost 14 months to repair. The proton beam enery stored is high, about the kinetic energy of a speeding freight train (est. 1e10 J), and the energy stored in the magnetic fields is higher (hence the explosive quench). The energy stored in a launch loop is 1.5e15 J, 150,000 freight trains.
Superconductors must remain extremely cold - "room temperature superconductivity" is a myth propagated by a few dabblers who do not know how to measure superconductors (Clue: measure Meissner effect and quantum coherence, not two-point ohmmeter connections). The rotor of a launch loop gets up to 800K, dull red hot. Not a good environment for a superconductor.
And what would superconductivity provide? Less electrical loss in the ambit ("D") deflection magnets at the ends. The current plan is to use ordinary copper windings and transformer iron - that will consume perhaps 100 MW of electrical power, perhaps $2000/hour of "waste". However, copper and transformer iron can operate at high temperatures, and are relatively easy to cool with seawater to keep them under 100C, even when they are radiantly heated from the rotor.
Good technology is really boring "under the hood". Cleverness consists of using cheap inputs to make valuable outputs. As a favorite boss used to say, "too many oughta works is a not-oughta work". Add in a few "it might possibly works" and the chances of success are practically zero.
Why 80 km altitude? Isn't there a lot of residual drag force below the "edge of space"?
The air density at 80 km is 1.85e-5 kg/m3, about 1.5e-5 of the density at the surface. The drag force on a 5 m2 frontal cross section at 10 km/s, with a drag coefficient of 0.2, is about 1000 Newtons, less than a percent of the acceleration force. The drag would be much less at 120 km altitude (the original proposal), but reducing 1% to 0.001% loss is trivial compared to other losses. The higher altitude creates three difficulties:
- 1) the drag is also a lot less on re-entering space debris. 1/1000 drag means more slowly decaying orbits - we can expect 1000 times higher debris density at the higher altitude. At 80 km, debris is falling fast, not lingering at loop altitude.
- 2) higher altitudes means longer, heavier stabilization cables. The cables are tapered, fatter at the top, so more top means more fat. This greatly increases the mass of the structure, thus scaling it up further.
- 3) top-heavy stabilization cables "reflect" transmitted forces, making them less effective stabilizers. Also, the delay from the ground (where most of the stabilization actuators reside) is longer.
So, the big question is actually why as high as 80 km? Why not 70 km, where the air is only 4 times as dense? No good answer for that; the engineers who actually design the first launch loop will make that decision. From here in their past, I am sure they will be smart people (they are working on a launch loop) and will make wise choices.
BTW, it may make more sense to put west station at a lower altitude, perhaps 50 kilometers, and follow a curved track up to 80 kilometers over the first 20% of the acceleration path. That will lower downward gee forces a little, helping with lateral stability. It will also help with magnetrail testing, launch aborts, and allow some airfoil stabilization, too. This will put west station in warmer, 50 times denser air; even more meteoroid and debris protection. Deflection magnet cooling will probably be done with seawater haulled from the surface, and that gets easier, too.
Isn't 14 kilometers per second very fast?
I am an electronics designer - 14 kilometers per second is 14 millimeters per microsecond, and my $100 nVidia graphics card can do 500,000 floating point multiplies in a microsecond. You can't move that fast, and your multiplication rate is even slower, so we will control the launch loop electronically, with millions of cheap integrated circuit controllerss, managed with thousands of microprocessors. We will learn how with power loops. If you don't like computers, get outta my century (and don't use planes, cars, phones, and the internet).
And if velocity bothers you, stay out of space. The earth hurdles around the sun at 30 kilometers per second, and the galaxy is plunging in the direction of Hydra at 600 km/s.
Is this dangerous?
Yes, but so is strapping several tons of explosives to your back and blasting into space. It is important to prototype, test, and develop the system so that you find all the ways it can fail (and how to prevent them) before going full scale. It is also important to locate the Launch Loop somewhere devoid of people like the southern equatorial Pacific. And buy lots of insurance.
How does Magnetic Levitation Work?
There are two kinds - attractive and repulsive.
Attractive levitation works because magnetic fields store energy in volumes of space, and decreasing a field gap decreases the volume and the energy - and typically increases the field That is why magnets slap together and you should never EVER put your hand between loosely anchored magnets; it can get smashed to jelly. Since the field tends to increase with shorter distances, the pull increases, too. That makes the system unstable. However, if you measure the distance, and reduce the magnetic field if the distance is shrinking, you can stabilize the system. "Can" meaning that your control circuit needs to keep track of the space, and the velocity, and the acceleration, and model everything with a clever analog circuit (I've done this) or a speciallized low-latency computer (I do not have one) and keep everything in balance. Magnets snap together quicker than an eyeblink - but an eyeblink lasts millions of computer clock cycles.
Repulsive levitation works because a moving magnetic field generates rings of voltage, which generate rings of current in a conductor. Those generate an opposing field, pushing out the field and pushing away the magnet that creates them. This is stable - the closer the magnet gets, the higher the repulsion force - but the currents create heat in the conductor. This is how magnetic levitation trains work - the power losses are significant, but not nearly as big as airdrag on a train. Electromagnets also consume power and make heat. The number one question you should ask of any maglev system is "where does the heat go?"
The launch loop rotor and track are held together with attractive levitation - with lots of control electronics to keep it stable. The vehicle rides on the rotor with a combination of both, attractive and repulsive levitation, mostly counteracting each other. This creates a LOT of drag, which provides the force that pushes the vehicle, and slows down the rotor to compensate. The rotor carries off a LOT of heat - enough to heat it dull red hot. This is wasteful, but systems to recapture that energy are expensive and failure-prone.
The rotor is made of magnetic, at least somewhat conductive metal which passes repeatedly through a very large magnetic field which lifts it and curves it, and it does this at extremely high speeds (14 km/s). This can be expected to generate at least some eddy currents. Would the energy loss and rotor heating due to this effect be prohibitive to a launch loop? How hot does it get?
Eddy currents occur when there are changes in the magnetic field that pass through a conductor. If the lift magnetic field is sufficiently uniform parallel to the motion of the rotor and changes slowly over distance, the currents will be small. However, vehicle acceleration puts very high eddy currents into the rotor - the 150,000 Newton force from a 5 ton payload accelerating at 3 gees heats the rotor by 85K as it passes. If 80 such payloads are launched per hour (averaging one every 45 seconds), the rotor heats up to 900 Kelvin, about 620 Celsius or 1150 degrees Fahrenheit. It cools by black body radiation to the sheath and from there to the surroundings.
Eddy currents entering and leaving the large deflection magnets are minimized by spreading the field change out over meters rather than centimeters. The induced voltage is proportional to the change in field with time - which in this case is the velocity (very high) times the change of field with distance (very low). The induced current is proportional to the induced voltage, and the power loss is that current through the "skin effect" resistance of the rotor - the skin resistance is proportional to the inverse sqaure root of the field change. Overall, the losses are inversely proportional to the length to the 3/2 power.
So, if we accelerate a 5 tonne vehicle at 3 gees, we need 150 KN against 14 kilometers per second, 2.1 GW. We do that with a 10 meter magnet rail, with field cycles of 10 centimeters. 200 transitions 4 centimeters long, each 4 centimeter transition adding a 10.5 megawatt thermal boost to the iron moving past it. A 5 meter transition to deflection magnets, scaled by the inverse 3/2 power, will add 8 kilowatts to the rotor moving past, in perhaps a dozen places in the whole system, perhaps 100 kW total. We will use megawatts of power to feed current to the copper and iron deflection magnets.
If the rotor is constructed so that eddy currents can't occur, how can acceleration of the vehicle be achieved?
Eddy currents are minimized when field changes are small and spread out over a very large distance. The turnarounds are almost 50 kilometers in radius, for example. The vehicle magnets generate rapidly varying magnetic fields at the 10 centimeter scale, which induces much larger eddy currents and lots of drag. The vehicle magnets have an attractive "DC" component working against the repulsive "AC" component, increasing the repulsive AC forces and drag. The drag may be further increased by making a more complex rotor cross section that resonates electrically in synchronization with the passage of vehicle magnet sections - this may be used to control acceleration versus speed.
Do the catenary-like tensile cables give sufficient lateral stability to the structure?
The tensile cables give a static pull on the structure sufficient to curve it down to the horizontal, as well as help keep it pointed in the right direction. The cables allow lateral forces to be transmitted from spooling motors on the surface. These may take many seconds to arrive, but they can be used to correct long-wavelength low-frequency motions that exceed the capabilities of moving counterweights near the track.
Is there an issue with the vehicle balancing on top of the rotor like that?
The centre of mass of the vehicle is maintained near the cable, and the vehicle attitude can be controlled with attitude jets.
The magnet rail can also be stabilized with long skinny poles hanging down and to the sides; in near-vacuum, the drag is low. The poles would act as torque multipliers for thrusters. The poles can be jettisoned or folded back as the vehicle approaches 8 km/s, and centrifugal forces dominate.
If the rotor was to escape containment, can the structure survive the fall from 80 km?
Some portions of the structure will be destroyed. Most of it can be lowered on parachutes. There must be re-entry vehicles for station crew.
How much energy is stored in the rotor at full speed? How big an explosion would containment loss be?
The rotor is designed to not be lost. The magnetic levitation is massively redundant, sections can fail completely without any problems as neighbouring levitation sections just work slightly harder to compensate.
The total energy in the rotor is huge, equivalent to a several hundred kiloton bomb. However, the round trip circuit for the rotor is nearly 7 minutes; it will not be released all at once. Perhaps the best way to release the energy is to use the turnarounds to fan the rotor outwards and downwards into the ocean, vaporizing the rotor and a lot of sea water.
The rotor that escape horizontally are moving faster than Earth escape velocity - they will go into long orbits around the sun. The Earth is a pretty small target relative to the radius of its orbit - the fragments will eventually impact the Earth, but it may be hundreds of millions of years before they get close enough to do so. They will be radar trackable, so there is plenty of time to find and collect them.
However, the same complex structure that amplifies vehicle drag may make the rotor fragment into small pieces when it encounters high gee forces and lateral stresses, which will be much higher during a reentry event than during normal operation. The small pieces will have a high surface-to-mass ratio, and will either burn up or fall relatively gently.
During startup in particular, as well as shutdown, will the cable go through resonances that may potentially damage it?
The rotor is actively stabilised, so will not go through resonances, as it is damped at all times. The "damping" is the removal of perturbations from the intended alignment of track and rotor, and the forces available are between local segments of rotor and of track - which, as the rotor propagates, means that the segment pairs change rapidly. More slowly, and perturbations are not as easily transported and shared (though measurements and correction signals are still propagated electronically at near-lightspeed).
Startup and shutdown occurs with the track and rotor near the surface - running at full speed, but not yet deflected into the sky. So the transition from stopped to full speed occurs with the acceleration track near the ocean surface, where it is easy to add lateral forces from temporary external actuators. Modelling errors may lead to unpredicted resonances and damage.
This is another reason why "power loop first" is so important. Almost all of these problems will be found and fixed during the development, deployment, and operation of power loops; and power loops will have plenty of paying customers in regions with inadequate pumped storage. Those regions are everywhere, when huge amounts of intermittent energy harvesting (AKA solar/wind/tide) feeds the grid.
Why is there a loading dock at 80 km altitude, rather than at the ground?
From the 1985 paper: "The long elevators to the stations are supported by pulleys from the anchor cables. The vehicles are brought up these cables rather than up the west incline to simplify the spacing controllers on the incline. Other benefits of this approach are minimized incline weight, shorter upward transit times, and less likelihood of sheath damage."
What's the difference between the $10 billion version and the $30 billion version? Could there be an upgrade?
They're mostly the same. The cheaper version has a quicker payback on investment, so the launch costs are set at $300/kg, and a much smaller power generation capacity, which limits the launch rate. (After the first year, the costs should go down significantly, once the hardware is paid off.) It also needs less powerful linear motors due to the lack of available power for them. The $10 billion version is mostly upgradeable, but for safety and obsolescence reasons the loop may need to be "undeployed" before the modifications are made.
There is room in the equatorial ocean to build hundreds of launch loops. It may be easier to build larger launch loops elsewhere while continuing to use the original.
Would the rotor fry due to runaway erosion due to the extremely high temperatures reached in the near vacuum of the sheath?
The rotor would be coated with a coating that prevents significant erosion, probably carbon nanotube or graphene. With an atomic weight of 12, a carbon atom moving at 14km per second has an energy of 14 electron volts. A glancing blow from this atom on the sheath will deposit some energy, perhaps kicking loose two carbon atoms, which could kick loose 4 when they hit the rotor, etc. This can lead to a "hypervelocity spalling cascade".
If the coating on the rotor and sheath is atomically smooth, the energy lost per collision will be small. If the coating is strong, it will absorb the energy without spalling off more atoms. And if the sheath has "diverter traps" along the bottom, loose atoms driven by the rotor velocity will be driven downwards and outwards away from the sheath, perhaps to be absorbed or pumped out by turbomolecular pumps below the sheath. However, the threat of spalling cascades does limit the maximum speed of the rotor; at some point, the energy per particle is too high.
Is the rotor suspension mechanism extraordinarily unstable?
Electromagnet suspension requires electronic systems to stabilize it. With an electronic system active, provided the system doesn't break, it is completely stable. In case of failures the rotor suspension mechanism has at least 10X redundancy; it needs at least 10 consecutive units to fail for the rotor to crash. Barring common mode failures (which have to be designed out of any system), such a failure is essentially impossible. Most normal systems have a redundancy ratio of 2 or 3. If the system did fail there would be a big explosion, but most of the expensive parts of the loop would be expected to survive; as would anybody launching at the time.
The rotor goes faster than orbital velocity, does this give a net antigravity effect and lift the structure?
No. The loop is pulled downwards by gravity in more or less the same direction along its whole length, and this net force must be resisted by the magnetic bearings. The east end and west end turnarounds are absorbing thousands of tons of force, and are spaced by many degrees of longitude so they point upwards relative to each other. The vector sum of these very large forces is upwards, and supports the rest of the launch loop in between them.
Does it take a lot of force to bend the rotor against its own stiffness?
The rotor is about the same weight and diameter as U.S. Schedule 10 "1.5 inch" (actually 1.93 inch, 4.9cm) pipe. If it was as stiff, then the bending force is about 100,000 Newton-meters2 divided by the square of the diameter of the turn in meters. To bend the rotor into a 1 meter turn would require 100,000 Newtons, about a 10 ton weight (and it would break). To bend the rotor into a 100 meter turn would require 10 Newtons, about the weight of 1 kilogram. The launch loop uses 28 kilometer diameter turns, so the bending forces are the same as a 13 milligram weight - far too small to notice, compared to the 42000 Newton/meter force needed to deflect the rotor into a 28 kilometer diameter turn.
For small scale, lower speed experimental launch loops, the bending forces will be significant. If a "back yard" 260m/s circular loop has a diameter of 10 meters, the bending forces will be 1000 Newtons, a noticable fraction of the deflection force, and perhaps enough to fracture welds.
How big can a launch loop get?
Launch loops are scaled by wind loading, maximum acceleration, and heat dissipation. If the resources and electric power were available, a launch loop could be built around an acceleration track perhaps 5000 kilometers long, with a rotor two meters in diameter, allowing 200 ton vehicles to be launched at 10 second spacings with accelerations of 2 gees to interplanetary velocities. Hundreds of gigawatts of power would be needed for this, but if previous launch loops put arrays of solar power satellites into orbit, the power would be available. Again, many of these jumbo launch loops could be built along the equator.
A more modest scaling would be to a 2 gee 30 ton launch loop, which can launch unmodified and prepacked intermodal ship/rail/truck containers. This would simplify logistics, although inspecting prepacked containers for explosives will be a challenge.
Scale-wise, we should stop thinking "space port" and start thinking "airport" and eventually "container seaport". An interplanetary society will require interplanetary-scale trading and transportation logistics.
Can launch loops be built in orbit? On the moon?
Probably not in orbit. The launch loop uses the earth as an enormous "tensile beam" between east and west stations, as well as a momemtum source. Constructing a long, stiff tensile beam in orbit would be enormously expensive. And it would be tidally and mechanically unstable - it wants to align vertically with the gravitational field, and even in that position it wants to buckle and vibrate. There are better ways to accelerate vehicles in orbit - rotating tethers, for example.
The moon is an excellent place for launch loops. Launch loops can be operated only a few meters above the maria. Power can be provided by solar collectors on the surface, and supplemented by power from SPS during the long lunar nights. A short lunar launch loop with modest rotor speeds can reach lunar escape velocity. A longer and faster loop will be able to launch to Mars and beyond. A sheath is still needed - the lunar dust acts like a very nasty kind of abrasive "atmosphere" and must be shielded from the rotor.
A lunar loop with alignment rings at the east end could capture and re-accelerate vehicles in Hohmann transfer orbits from Earth, circularizing orbits or perhaps pushing vehicles up to interplanetary velocities. Keep in mind that it must be kept in vacuum sheath - lunar dust is nasty, extremely sharp tiny fragments, just the thing to cause hypervelocity spalling cascades. Lunar dust is thrown into the sky by UV from the sun and the occasional cosmic ray; the moon does indeed have an atmosphere, a few hundred meters of tiny rock particles in ballistic trajectories. Very very low density, but enough to cause havoc in machinery.
==== How are the rotor segments joined? ===
The rotor is segmented so that it can stretch without axial tension. As it speeds up or slows down, by motors or payload passage or the drop from 80 kilometers, the mass density per meter changes - Bernoulli effect written in iron. Iron is too stiff, so we divide it into segments that can slide against each other. We do want lateral stiffness, resistance to bending - that helps the control system maintain stability. And we want magnetic continuity between the segments. We can achieve this with sliding joints. We can also include some dampened springs to help maintain spacing. For startup, some kind of magnetically activated slide lock would be helpful, to distribute the force of the startup motors down the long length of the rotor. Lots of mad handwaving for the last.
All these behaviors will be developed for power storage loops. A team of mechanical engineers with budgets and schedules will decide how to do this, I can only speculate what they will do.
How is energy added to the loop?
Very much like maglev trains - a linear motor, which is like a circular motor unwrapped and straightened out. These create a horizontal version of repulsive levitation - the currents induced in the rotor are pushed by a magnetic field. A cool (literally) thing happens with such motors - the power loss is proportional to the force, while the acceleration power is proportional to the force times the velocity. The faster you go, the smaller the relative fraction of loss. 14 kilometer per second linear motors can be > 99% efficient; which is literally cool!
These motors will NOT operate at 60 Hz line frequency; they go far too fast. They will be powered by 100 KHz switched power, and the motor "stator" (stationary magnets and coils) will be ferrites, like the cores of the transformers in the wall wart for your cell phone. At scale, these "switchers" can be very efficient, too, which is why they have replaced all the old, heavy, inefficient iron core wall warts. Not because wall wart manufacturers are saints, or because they are forced by regulators, but because shaped ferrite is cheaper than bulky iron, shipping costs are lower, electronic wall warts are safer and less legal liability, and electronics are cheaper than extra cooling.
What about war and enemy attack?
Launch loops are incredibly vulnerable. Do not operate one in a threatening way; at least not to anyone with missiles or submarines. Like hypothetical space elevators, and ordinary rocket launch pads, launch loops are strategic vulnerabilities, not assets. The best way to protect a launch loop is to share it, abundantly, with all potential adversaries. And inspect the hell out of everything that is launched, to make sure it is not a threat to any adversary (or to the loop itself).
So - rockets launched from airplanes (based at many different airfields) are the only strategically stable way to launch military satellites. Cheapass fools who launch military threats from any fixed site are asking for a world war to stop them. Launching mutually protective systems, for example spysats whose results are shared with the world, are inconvenient for some but protective for most. The "most" can band together to protect their assets and world peace against maniacs. Hopefully, space projects can teach cooperation.
Tsunamis and Earthquakes
Most of the near-surface portions of the launch loop will operate below water, anchored with active attachments to the deep sea floor. Tsunamis are millimeters high in deep ocean, they get huge as they approach shallow land, and are magnified enormously by convergent riverine bays and estuaries.
Earthquakes are more of a problem - we must adjust attachments to accomodate ground moving at up to one gee in any direction. For those keeping score electronically, one gee is 9.8 picometers per microsecond squared. We will need at least enough compliance in the attach cables to deal with meter-scale seabed displacements.
We will not do this without telemetry and computer models. Earthquake S waves, the nasty kind, move at hundreds of meters per second. The computers will have plenty of time to measure, adapt, and model a response. Again, we will learn how to do this with power loops.
Launch path length
This will of course depend on what you are launching, and what orbit you are launching it to. The canonical launch loop design is 2000 kilometers launch path, capable of pushing a payload at 3 gees to 11 kilometers per second. It can push a smaller payload to nearly 14 kilometers per second, sending it to Mars or Venus, or (more practically) a payload with a couple of small rocket stages to planetary or outer solar system destinations. But it won't do everything - no military stuff, no interstellar near-lightspeed craft, etc.
Why 3 gees? Because we can launch cheap stuff at 3 gees. Some things can tolerate 10,000 gees (frozen fuel or food), others cannot tolerate acceleration changes of 0.1 gees (semiconductor fabrication equipment). Everything launched will need to be packaged for the gee forces involved. I chose the sizes and masses commensurate with typical satellites in the 1970s - 3 gees, 5 tons is a typical comsat.
Launch Loop for Server Sky
The first launch loops will probably be small, fast, and short. Stacked server sky thinsats can tolerate hundreds of gees, and practical array packages are 100 kilograms; if we launch 300 packages (30 tonnes) per hour (5 million second generation thinsats, 40 MW, or 350 GW of server capacity per year), a much smaller 1 GW, 12 km/s, 300 kilometer long loop would suffice for that. We will still need stations and stabilizing cables, and the wind forces on stabilizing cables scale as the square root of cross section area, so there is a lower limit on how spindly and light weight we can make a launch loop. The cables also need enough conductor cross section to survive a lightning strike, and that puts a minimum scale on things.
Let's assume that server sky has caught on, and is rapidly enhancing the world with gigawatts of computing. Thinsats might cost $50 each in quantity 10 million. Assuming a 75% learning curve, 80 billion of them would cost $1.19. Assume weight-reduced thinsats massing 3 grams, up to 5 grams with deployment and boost rockets from LEO. 10 million thinsats is 50 tons (5 Falcon 9 launches) to LEO, about $260 million dollars worth, or $76 total per thinsat at the 10M level.
80 billion thinsats is a launch mass of 400,000 tons, 40,000 launches of SpaceX Falcon 9s. Assuming those cost $54 million per launch at quantity 40, with a 75% learning curve, 40,000 launches would cost $3M per launch. About 120 billion dollars worth of rocket launch, adding $1.50 to the cost of each deployed thinsat, total around $2.70 per thinsat at the 80G level (about 40W of computing per person).
Assuming a 5 billion dollar "mini-loop" (WAG), two year payback, and $20/MHhr fuel costs, the cost per kilogram is $ (5E9/(2*8766) + 20*1000) / ( 30,000 ) = $10 per kilogram, a penny per gram, or 5 cents per launched thinsat. The total cost to put 1 watt of thinsat capability in orbit will be 30 cents, $0.30. Higher volume higher velocity launch loops will be perhaps twice as expensive and launch 10 times as much. Amortized over 5 years and launching to the Lagrange positions, that might result in $0.20 deployed thinsats and $0.04 per watt of generated capacity. 10 kW of computation per person would cost $400 per person.
In the Real World, Jevon's Law applies - more efficient production lowers price, and lower price increases resource demand faster than efficiency lowers it. So we won't stop at 10 kW of computation per person - we will probably demand 10 MW, then more and more and more. That, and not the demand for thrill rides in space, will drive high launch demand, first from Earth, then materials from the moon (aluminum for subtrates and ??? for radiation-resistant PV). When we have a hundreds tonnes per person in orbit, we will have enough "industrial ecosystem" to sustainably support physical human residence, and build chip fabs in orbit.
You did not read about this in your science fiction books, but AFAIK, there are no chip-designing science fiction authors.
Elevators - Incline? Three Stage? Friction Elevators?
There are many ways to get payloads to West Station for launch.
- John Knapman prefers climbing up the inclines. This is problematic because there may be a lot of wind stress on the bulky payloads, and that adds to the complexity of the control problem through the atmosphere. Not sure of the best way to power such a long escalator, sending up power via an electrified trackway and rollers makes the most sense, a long linear motor would be hugely expensive and heavy.
- A straight pulley elevator means starting and stopping an 80 km tall elevator cable loop - extra stress and extra power.
- The three stage elevator has the 80 km tall loop going at a continuous rate, with two "small" 8 km loops accelerating and decellerating payloads, then transferring attachments. Robert Williscroft uses this in his novel "Slingshot", see below.
A straight pulley elevator with a friction wheel and incandescent filament dissipators is my current favorite.. This is somewhat energy inefficient, like the launch loop itself, but it minimizes complexity and maximizes control.
Fictional Launch Loops
Heechee Rendezvous I met Fred Pohl at an Orycon. He included a launch loop in his novel. A dramatic and flawed version of launch loop that vehicles land on - with the inevitable consequences, a miss and collision and destruction. A great illustration of why vehicles should return with aerobraking, not a very precise rendezvous. Besides, if a vehicle must launch during abort, or after one-orbit-and-return, it will not enter atmosphere anywhere near the loop that launched it. Fred and I were sporadic penpals until his death - more on that below.
Starquake Robert Forward was a more frequent friend. Bob read my AIAA paper, and coined the neologism Dynamic Structures to describe structures shaped by the deflection forces on streams of mass. Bob included launch loops in Starquake, made from neutronium by the hypothetical inhabitants of a neutron star, living and evolving at relativistic speeds. This launch loop broke in the devastating neutron star "earthquake" that wiped out the neutron star civilization.
The Last Theorem Sir Arthur C. Clarke wrote about space elevators. The only way space elevators will happen is half a dozen material advances and problem solutions, and the non-possibility of launch loops. So, the novel he co-authored with Fred Pohl dismisses the "Lofstrom Loop" with one sentence and the word "friction". No explanation, but if the loop was not completely contained in a vacuum sheath pumped down to 1e-3 Pa or so, gas drag would dissipate way too much heat. Fred and I wrote each other about that, and decided that writing a novel with Sir Arthur was worth stretching the truth. And I have the privilege of being the victim of an egregious violation of Clarke's First Law, committed by none other than the Law's author himself. Fred's experience was rare, mine is perhaps unique!
Slingshot. I met Robert G. Williscroft when he was still in the Navy, after his return from a research expedition to Antarctica to measure the "ozone hole" A.K.A. the "ozone not generated during long polar winters without solar UV" hole (you can read his opinions about that in his book of essays "Chicken Little"). In any case, we hit it off, two intelligent guys with differing opinions, but more interested in learning from each other than in arguing. Three decades ago, Argee decided to write a science fiction novel about the building of the first Launch Loop, and will debut it at the August 2015 Space Elevator Conference. He would be entering the lion's den, if the space elevator community weren't such kind and adorable pussycats!