The old design used passive inductive drag on a moving rotor. For a launch loop payload with an exit velocity of 11 km/sec, the first 30% of the velocity gained generates 44% of the heat in the rotor, because the slip is high and almost all the 14 km/sec rotor drive power is being turned into heat. What if that could be reduced, by increasing the efficiency of the energy coupling from rotor to payload? This would increase payload efficiency and reduce energy cost.
30% of exit velocity occurs 9% of the way down the track. Enhancing this much of the track with extra hardware and weight is possible - see LowerWestIncline. A mass driver would be too expensive, but similar results might be achievable by magnetically coupling the payload to the track with velocity reduction coils, loops of ordinary wire with perhaps a 1x to 3x horizontal ratio over the first 9% of the track.
This shows the relative pitch of the windings; the actual connection will be single loop to properly phased single loop, with some approximation of twisted pair connections between them. The first few poles of the launch sled's Halbach array will be slightly phase-offset compared to the "thrust" poles farther east on the sled. The first-encountered "exciter" poles will build the magnetic field in the rotor.
The alternating magnetic field moves 3 times as fast in the lower section - this might be the rotor moving at 14 km/s. The field moves slower in the upper section, in this case 4.67 km/s . So if the payload is moving at 3km/s, and slipping behind the field at 1.67 km/s, it will appear to be moving at 9 km/s and slipping at 5 km/s to the rotor.
The rotor should be optimized for accelerating the payload close to exit velocity; with long field coils, it will probably drive less optimally. However, significant efficiency gains are possible - that means more payload launched per megajoule, less heat into the rotor, and significantly higher throughput.
MORE LATER - needs analysis.
Motor thrust and power loss is proportional to "slip", the vehicle drive power is the thrust times the velocity. The slip and power loss can be high because both rotor and stator are only heated for milliseconds per segment, while the velocity (and total power) is enormous. I presume launch sleds will be in the range of 10 to 100 meters long, mass limited by the weight of the magnets in the Halbach arrays on the sled. Much depends on circumventing parasitic inductance in the motors and arrays; I can't afford the weight of large capacitors on the track, and they would heat up far too much on the sled.
This can be applied in the opposite direction, for accelerating vehicles moving faster than rotor speed. This may have applications for some versions of the space cable, using 3 km/sec bolt streams to launch 9 km/sec vehicles. However, if the rotor and vehicle are moving at the same speed, the relative speed is zero and it will be difficult to transfer power from the relatively stationary rotor to the vehicle.
A newer drawing
Note: the design may evolve to a 4 or 5 lobe design, with tapered slots and stator windings, for better stiffness and heat conduction out of the center. The stator windings will hopefully be centered by diamagnetic repulsion (maglev style) from the rotor magnetics.
For 99.9% of the 2000 km of the acceleration track, there will be no launch sleds on the track and very low excitation; some power will be drawn off for electronics and actively-controlled stabilization magnets.
The track/rotor system is mathematically unstable. Two "natural" right-hand poles in the linearized control equation. Recent work by Dr. John Knapman demonstrates that this can be stabilized, but the stabilization process requires global spacing and positioning information to compute local force corrections; otherwise, the track and rotor oscillate and meander. The math suggests that a skein of wires to an "inverted suspension bridge" cable, connected to stabilization cables to heavy anchors on the surface, will also be required. The control system will be computationally intensive and complex; on the other hand, the rotor moves at a mere 12.5 micrometers per nanosecond, plenty of time for computation and signalling, and the correction of millisecond-scale instabilities.
Deflection Track Cross section
The bolts are "exploded" from the rotor and very rapidly (tens of milliseconds) reconfigured to pass through the deflection magnets. My calculations show deflection radii less than 8 kilometers ... I don't believe it, but it will be less than the original 14 kilometers.
The triangular Si/Fe laminations of the rotor are pulled towards the DC magnet pole at the bottom. The flux moves vertically through the laminations to the poles at the side. The flux through each side pole is half of the bottom flux; half the flux density means 0.25 of the force. The vertical component of the two diagonal forces is about 1/4th of the bottom force. This isn't 100% efficient, and leads to a larger radius deflector, but it is more magnetically efficient than attempting to pull the bolt towards a split field gap (like earlier designs ).
I do NOT expect to see such tight and optimized deflectors (or a launch loop using them) for many years. This technology will be developed, optimized, and custom-integrated-circuit infused for a series of increasingly larger and faster power storage loops, a scalable and very profitable near term application of loop technology. Hopefully that technology will mature, accumulate clever fail-safes, establish a perfect safety record, and spread around the world to optimize the use of terrestrial wind and solar. Space power will happen when we are rich enough to fret about the ecological damage that surface energy scavenging incurs.
New bolt design and cross section
The bolts might be 10 meters long and mass 4.32 kilograms. There are 6 rows of bolts in the rotor. The complete rotor h may be 6000 kilometers in circumference; 36,000 kilometers of bolts.
Most of the bolt mass will be thin 60 degree triangular laminations of high-flux transformer steel, suitable for kilohertz magnetic field rates. These will be stacked perpendicular to the bolt's long axis, perhaps 3,000 laminations per meter and 30,000 laminations per bolt. Two faces of the diamond shape will be the "motor" side of the bolt, one will be the deflector sides.
The laminations may be cut out of lamination strip, the same material used to wind audio-frequency toroidal coils and transformers, but in ten thousand tonne lots (!!!). Hundreds of cutters will cut diagonally across the strip to make the laminations. During automated mass production, these thin laminations will be maglev-transported past many inspection and finishing stations at high speed, perhaps 100 per second (WAG) on each production line. One production line at 50% utilization might turn out 1.5 billion laminations for 300 kilometers of bolts per year. 500 production lines can turn out one launch loop worth of bolts every six months.
The motor side of the bolt will have slots for aluminum rotor windings, running down the sides and the top edge, and periodically across the two top faces of the bolt. These aluminum windings may be cast in place, but stiffened, embedded wires with lubricious and insulating coatings may be more durable and electrically optimal.
Aluminum or copper? TBD. Aluminum won't melt below the Curie temperature of the Fe/Si laminations, and the atomic weight of aluminum is smaller, making loose atoms less damaging in the vacuum plenum, and easier to deflect outwards to pumps. Detailed analysis may uncover advantages for copper; perhaps stiffness and electromigration when the windings are shock heated by millisecond vehicle sled passage.
The deflection magnet side of the bolt will be the coated edges of the laminations, without windings. The field of the deflection magnet will pass through these faces, and hopefully that field will penetrate rapidly so deflection magnet entry and exit lengths will be short.
Kevlar rotor bolt stringers:Kelvar is a dielectric, as is epoxy matrix. Kevlar is somewhat cheaper than carbon fiber, which is resistive and lossy as a motor facing. The faces of the rotor bolt will be strung longitudinally with Kevlar fibers, with some manner of strong attachment (hands waving madly) to the edge of each lamination and aluminum wire. The attachments must survive the high speed shock of passing into the deflector magnet field, and the somewhat smaller shock of a payload passage.
This longitudinal array of Kevlar fibers may be encapsulated and smoothed with a very thin coating of epoxy resin, but an insulating vacuum-compatible matrix material may be more suitable for the vacuum environment that the bolt will be deployed in. At this and subsequent stages, the bolt will be processed in vacuum, and should be heated so it will outgas before encapsulation.
The smooth matrix will be coated with a series of bonding layers to minimize strain to the outer diamond coating. One of the layers will be embossed with a bar code running longitudinally down the motor side of the bolt. This bar code will probably not be apparent in visible light, perhaps "read" with millimeter waves that penetrate the outer coatings.
The top coating will be DLC, "Diamond-Like-Coating". Acree Technologies coats tools with DLC; the bolts and magnet facings for one launch loop may require 10 square kilometers (!!!) of DLC.
At the ends of the three core bolts will be voltage generators, simple coil-resonator-and-diode assemblies that rectify MHz-range power transmissions from the track to generate small longitudinal bolt-to-bolt voltages, plus-or-minus. Responding to controllers in the track, the generated field will be rectified to produce longitudinal voltage fields in the rotor, matching voltage fields induced in the incline outer sheath by lightning strokes.
Like so many other aspects of launch loop manufacturing, bolt production capability must be expanded exponentially through other profitable applications, such as Power Storage Loops. A daunting task, but conquerable with automation manufactured by other automation. This does not require nanometer lithography or hard-disk-platter smoothness, "merely" large numbers of robotically-manufactured, robotically-operated assembly lines, using a clean-but-lower-precision version of the wafer-handling and electronic-assembly equipment we manufacture now. A company like Applied Materials could make the equipment; my guess is A.M. will scoff at this market, a startup vendor will tackle it and grow exponentially, while feeding an exponentially growing power storage market. They will overshoot A.M.'s mature market, and use their "automation-of-automation" capability to outcompete A.M. in their primary market. Innovation is disruptive, incumbents beware!