Rotor Dip Windings
The velocity transformer track couples windings inserted into a magnet sled on the vehicle with windings inserted about 10 millimeters into the rotor. The windings are stationary; the rotor moves at 14 km/s [1]. However, inserting windings along the entire length of the track and rotor is unnecessary and risky; windings are only used in the immediate vicinity of a vehicle sled, which briefly pass by for a few milliseconds every 45 seconds or so.
Dip windings only insert into the rotor from the track as the vehicle approaches, and withdraw just after it passes. Assume the windings move one centimeter, and accelerate and decelerate at 100 m/s² (10 gees) during this insertion. Peak velocity occurs at 5 millimeters insertion; from D = 0.5 × a × t², t = 2 × 0.005 / 100 = 10 milliseconds, and the total movement time is 20 milliseconds. If we launch the insertion 30 milliseconds before the arrival of a 10 km/s vehicle, it will begin as a wave 300 meters in front of the vehicle, be halfway inserted 200 meters before the vehicle sled, and fully inserted 100 meters before the vehicle sled. The total insertion length will be on the order of 500 meters, while 45 second spaced vehicles accelerating at 30 m/s² will be 420 km apart.
Assuming the first km/s is provided by 20 km of alternately-powered special track, there will be only six or seven vehicles accelerated by dip windings, and perhaps 2 km of inserted windings spread over 2000 km of rotor and track. Drag and spalling cascade opportunities are reduced 1000x compared to a continuous stator winding.
Assume that the insertions are discrete and independent; perhaps 20 meter sections insert when needed. We can do test insertions and observe what happens while we do; an anomalous insertion can be terminated quickly, and that section disabled and not used. A 20 m interruption in acceleration at 10 km/s is a 2 millisecond interruption; at 30 m/s² acceleration, that is a 0.06 m/s velocity jerk. That can be smoothed out somewhat by neighboring windings. Presuming an MTBF per section of 2.4 million hours, that is one section failing per day, and we can expect two neighboring sections to fail in about a year. After three years, there may be a dozen paired failures, and a few days of repair downtime may replace perhaps 1000 failed and 10,000 marginal sections.
Most failures will be "soft"; each of the sections will have memory and "intelligence", and will make algorithmic estimates of quality. Launch loops will "feel pain" and learn to avoid it, detecting pre-failure sections and cautiously disabling them, triggering manufacturing quality control reviews and more robust sections.
The energy used to insert the rotors will pass through power FETs and thermal fuses. The FET gate drive will be switched by a tiny "smart chip" and a capacitor; the capacitor will be charged with series PV from energy tapped from a thin optical fiber. Control information will be delivered by the same fiber in a different wavelength band. Much handwaving; a very thin low-loss graded-index fiber, precisely welded and properly terminated, should be able to carry perhaps ten watts (WAG) in a -70C ambient, with power tapped off by some form of collinear coupler (W-WAG). Power taps along the track will pull power from the rotor, mostly to power the spacing and levitation electromagnets, and will also feed the power lasers that feed the (redundant) sensor and power FET gate drivers.
Gate drive power, WAG
The FET gates will fire every 45 seconds. If the FET gate is 10 nanofarads, and the gate voltage is 10 volts, that is 100 nanocoulombs and 1 microjoule per gate. 200,000 sections is 200 millijoules; perhaps 50 milliwatts total average gate power. Trivial; the power switched for the mechanical dip will be larger, and the power transferred to the vehicles will be vastly larger.
5 tonnes every 45 seconds is 111 kilograms per second; 10,000 m/s is 5e7 MJ/kg. At 95% efficiency, that is about 6 GW of system power, with most of the 300 MW of "waste" power feeding perhaps 720 degrees of deflection magnets.
Note [1]: . I considered a 12.5 km/s rotor through most of 2018, joping to reduce errant particle energies and the risk of a hypervelocity particle cascade for a full length 14 km/s winding. Dip windings greatly reduce the interaction with errant particles, and remove the winding sections with the highest risk. Nothing is perfect, but imagine a hard drive with many spare heads; the chance of a head failure disabling read/write operations will be greatly reduced, and there would be no need to keep a near-failure head in service.