On Saturday I investigated SEP systems for spacecraft going from LEO to the Moon and having a relatively small LEO mass of 20 tonnes. In hindsight, some of my numbers were a bit dodgy; more realistic values are for 15 tonnes of non-propulsion equipment.
But what happens when there's a possible initial mass of 100 tonnes, as could be launched to LEO by a Saturn V?
At this scale, SEP doesn't make sense. Sticking to my maximum transfer time of 2 years, the math works out as follows: mf / m0 = 0.843, so 84.3 tonnes of payload and systems and 15.7 tonnes of fuel. Using just one ion thruster as before, a Δv of 6.33x103 ms-1 would take 100 years. That means that to transfer in 2 years, I'd need 50 thrusters, corresponding to 250 kW of power consumption.
That's an awful lot of power, and a solar panel array that big is just ridiculous, without even taking radiation degradation into account. So, nuclear power could be used instead. Up to 1993 as part of the Space Defence Initiative NASA developed a reactor designated the SP-100 . This was capable of delivering power levels up to about a MW at between 10-50 kgkW-1 (better specific mass at higher power). By the way: for the environmentalists, the SP100's nuclear fuel container is designed to reach the ground intact in the event of the reactor unexpectedly breaking up in the atmosphere. Additionally, I'm assuming a 1000 km starting orbit, and an orbital height of 700 km is considered by NASA to be "nuclear safe".
Pulling figures out of thin air, let's assume that a 250 kW reactor can be built at 30 kg/kW. That means that 7.5 tonnes would be power source. Additionally, I'd need 250 kW of ion thrusters, at about (yep, more made-up numbers) 13 kg/kW, another 3.3 tonnes.
Adding all that up, for 100 tonnes at LEO, I'd get 15.7 tonnes of fuel, 7.5 tonnes of nuclear reactor and 3.3 tonnes of thrusters, leaving 73.5 tonnes for payload. Very nice.
The real problem with what I'm doing here is the fact that no-one's done any research on really big spacecraft in the last decade or so, because there's been no requirement for such research. Everything's been kept small, cheap and efficient as far as possible (apart from the ISS, which is big, expensive and inefficient). I want to see things that are big, not too expensive and extremely efficient.
 'SP100 Power Source', NASA SpaceLink