we’re close, but oh! so far..
the figure of 37GW was to provide 2016 N thrust at effective velocity of 0.028 c,
(which was the thrust needed per tonne of mass of the ship!)
at 50% efficiency, that’s 37GW useful work, and an additional 37 GWt to reject.
the radiator is way too massive, so the only way to reject heat is to send it off as additional propellant,
which increases thrust per engine, but reduces the specific impulse.
