Just read the papers. There are dozens on that archive covering the design, theory, and testing - all of those issues are covered. 20kg is actually a high numbe

Just read the papers. There are dozens on that archive covering the design, theory, and testing - all of those issues are covered. 20kg is actually a high number optimized for a certain plasma temperature that radiates in UV without losing criticality. All things being equal (enriched uranium hexafluoride), at higher temperatures and pressures the reactor needs far less fuel but it causes too much degradation to the reactor walls to be economical or useful except in cold boots. The terrestrial reactor and the rocket engine versions are different in how they transmit power from the reactor core to do useful work but not radically so - the core irrotational vortex compressing a high temperature plasma in a single crystal beryllium oxide chamber is the same in both designs and has been tested. The difference is in the cooling gas - hydrogen gas seeded with tungsten nanoparticles to absorb the black body radiation of the plasma for the rocket and water going through a steam turbine for terrestrial power generation.