Lunar fission power transmission: cables win the mass trade

What the work claims

For the 40 kWe Fission Surface Power (FSP) system, with loads that may sit up to 3 km from the reactor lander, a NASA Glenn trade study evaluated power beaming, superconducting cables, carbon nanotube (CNT) ropes, and conventional metallic conductors as bulk transmission options.¹ The authors' preliminary conclusion is that metallic conductors operated at elevated voltage are the most efficient and feasible solution, and that proposed DC-DC conversion architectures need scrutiny because accessible flight-rated switches cap out at 175 V, forcing a complex stacked design for kilovolt transmission. They state explicitly that the work is intended to inform requirements and does not represent NASA consensus or preference.

How it works

The study's central physics is that the fission source dominates system mass, estimated at 0.27 kg/W of generation, so every watt lost in transmission must be paid for roughly 0.27 kg at a time in extra reactor, shielding, and radiator mass.¹ That conversion factor turns transmission efficiency into the deciding variable. On that basis the four options separate quickly. A wired system carries an assembled end-to-end efficiency of 77% (DC-DC conversion at 90% each end, aluminum line at 95%), needing 12 kW of additional generation and about 3,200 kg of extra FSP mass. Laser beaming assembles to 14% to 18% end-to-end from literature links: DC-DC 90%, laser DC-to-light 50% to 60%, diffraction and optics 90%, propagation 98%, collection 95%, and light-to-DC conversion at 40% to 45%. Each of those links is an assumption taken from references, not a measurement by the authors; the 98% propagation link is an Earth-atmosphere figure carried into what would be a vacuum application. The assembled result requires 210 kW of additional generation and roughly 57,000 kg of added FSP mass, before counting any beaming hardware. Micrometer-wave and millimeter-wave options land at 26% to 27% and 13% to 18% end-to-end respectively, with correspondingly large mass penalties.

The one measured anchor in the chain is the Navy's PTROL demonstration of 2019: 20% efficiency over 0.33 km, with 2 kWe sent by laser and 400 We received, on equipment whose laser and receiver together weighed more than 2,500 kg and was, in the authors' words, not minimized.¹ The authors conclude that beaming carries over ten times the mass penalty of a cabled system once equipment mass and loss-compensating generation are both counted.

Superconducting cable scores a respectable 42% end-to-end efficiency but loses on its support system: three or more cryocoolers at about 5,500 kg, roughly 600 kg of nitrogen coolant, about 1,500 kg of cryogenic hose, 5,500 W of heat removal requiring about 2,300 kg of radiator, and roughly 55 kW of cryocooler power that itself demands dedicated generation and backup. Total superconducting system mass comes to about 25,000 kg against 3,200 kg for the wired case. CNT rope, produced in prototype form by DexMat and BORONITE (a company claim cited by the authors from vendor interviews), offers four times the tensile strength of copper and far better fatigue and corrosion behavior, but its volumetric conductivity is six to ten times lower than copper. At the high transmission efficiency the FSP application demands, an equivalent-conductivity CNT cable masses 34 to 57 kg/km against 47 kg/km for 10 AWG copper: the mass benefit is minimal or nonexistent.

The metallic-conductor baseline is sized for 43 kW over 3 km at 95% efficiency, with cable loss limited to 0.7 W/m. Cable mass falls steeply with voltage and is minimized at 3 to 6 kV. Here the switch ceiling bites: accessible flight-rated switching devices are limited to 175 V, so a 3 kV DC link needs an 18-level stack of 72 series devices, with high-voltage capacitors and a part count the authors flag as a reliability problem. Their proposed alternative is modest-frequency AC transmission at 1 to 5 kHz and roughly 2 to 5 kV, which permits voltage step-up and step-down with transformers, eliminates high-voltage capacitors, and uses existing flight-rated switches without stacking. The heritage precedent is a 2008 demonstration at Glenn's Lunar Power System Facility: a 1 km line model, 14 AWG conductors on 1/8 inch spacing, designed for 50 kWe at a 3-phase 1750 Hz fundamental, with the line impedance represented by 500 microhenry inductors and 4 ohm resistors.

Steelman: why the result matters

This is the rare study that prices exotic transmission honestly. Space power discourse routinely treats beaming and superconductors as defaults and wires as legacy; this analysis shows that when the generator is the massive element, a 77% efficient aluminum line beats a 14% to 18% efficient beam by roughly a factor of eighteen in added generation mass before any beaming hardware is counted. The methodology is transparent: specific mass stated, efficiency chains itemized, hardware mass included where others neglect it. The AC option deserves particular attention because it sidesteps the single hardest gap in space power electronics today, the lack of radiation-hardened high-voltage switches, using 175 V devices that already exist and magnetic components that do not care about DC ratings.

Where a skeptic should push

First, the load-bearing number is the 0.27 kg/W FSP specific mass, an estimate; every mass conclusion scales with it. Second, the beaming chain is assembled from literature links, and one of them, the 98% propagation figure, is an atmosphere link applied to the Moon, where there is no atmosphere. The authors' own assembled end-to-end numbers inherit that sloppiness, so treat the laser penalty as an upper bound rather than a measurement. Third, the 2008 AC "1 km demonstration" was a bench line model: lumped inductors and resistors emulated the impedance of a kilometer of cable, with only about 100 feet of physical cable in the loop. It validates the converter and transformer concept, not a deployed lunar transmission line. Fourth, the CNT comparison rests partly on vendor interviews and prototype products, which are company claims rather than qualified data. Finally, the authors themselves label the results preliminary and non-consensus; the AC-versus-DC mass trade depends on rad-hard switch availability, which is exactly the technology that does not yet exist at 400 to 1,200 V.

What it means for power wiring on the Moon

Implication: the binding constraint on lunar power grids is not cable conductivity but switch voltage. Copper and aluminum are good enough that the mass optimum sits at 3 to 6 kV, and the architecture cannot reach it with 175 V flight-rated parts without stacking 72 devices into a conversion chain whose part count is itself the reliability risk. The non-obvious corollary: a kilohertz-class AC architecture, last demonstrated in 2008 and largely absent from modern lunar power planning, is the shortest credible path to kilovolt transmission with existing parts.

Threat: if the field defaults to DC microgrid thinking imported from terrestrial practice, every landed power system carries either a heavy copper penalty or a stacked-converter complexity penalty, and the radiation-hardened high-voltage switch gap becomes a single point of failure for the entire surface architecture. The study's own caution applies: proposed DC-DC systems have unproven maturity, and high part count reduces reliability in a system that must run for years unattended.

Opportunity: the leverage points are named precisely. Radiation-hardened switching at 400 to 1,200 V, and component-level development of modest-frequency AC converters, are the two investments that move the whole trade. A credible AC transmission demonstration, extending the 2008 line model toward hardware with flight-like insulation and connectors, would retire more architectural risk for a 40 kW lunar fission plant than an equivalent spend on almost any other subsystem.

The bottom line

Within its stated assumptions, the conclusion is robust: for moving tens of kilowatts a few kilometers on the Moon, wires win, superconductors and nanotubes are support-mass and conductivity dead ends respectively, and beaming is a mass trap for bulk power regardless of how elegant the physics looks. Confidence is high on the wired-beats-beamed ranking and moderate on the exact margins, since the efficiency chains are assembled from literature and the FSP specific mass is an estimate. What would confirm it: a component-level AC transmission demonstration at kilovolt class, and measured efficiency of a complete beaming chain rather than multiplied link assumptions. What would break it: a measured FSP specific mass far below 0.27 kg/W, or a beaming end-to-end efficiency demonstrated near the micrometer-wave assembled figure of 26% to 27% with hardware mass included.