I've been working through a thought experiment with a friend of mine, with some of my previous questions already being (very helpfully) answered on here :)

Here's the current issue I have though - Engines.

So my project is a fairly large 'ship', that doesn't need to get anywhere particularly fast, but can't be too ridiculously slow. As chemical rockets would require huge amounts of fuel, it seemed like the better solution was to go with an Ion drive instead.

Because of the mass of the ship though, this meant stacking up a few VASIMRs. Which, as far as Xenon (or whatever propellant) goes, it's pretty good. Constant (very small) acceleration will gradually pick things up. If you can generate 200MW of power.

And this is where I hit the next wall, and the 'false economy'. In order to generate that kind of power, you would need 150,000m3/h of hydrogen (in a hydrogen generator). I can't find information on the usage of nuclear fuel, but I suspect it's not insubstantial. It's also about 800 acres of solar panels, apparently.

So while the Ion engine is much more fuel efficient as far as propellant goes, it seems to trade that for another type of fuel. Which is fine for small-scale probes which can utilise solar energy, but not for the big boys.

So I have hit a wall, where even for relatively short range (such as mars or the asteroid belt) trips, we don't seem to have the capability to power a ship to our destination.

I've written this to mostly explain my process thus far, and so to ask the question -

What would be the way to power a ship with a mass of many hundreds (perhaps thousands) of tonnes, from LEO to LMO or the asteroid belt? Within the lifetimes of the people on board, preferably haha. Cos I'm stumped.

  • 1
    $\begingroup$ What is a hydrogen generator? $\endgroup$ Nov 6, 2019 at 12:29
  • $\begingroup$ Hydrogen fuel cell power generator. No idea how it works. $\endgroup$
    – nirurin
    Nov 6, 2019 at 16:39

3 Answers 3


Are you burning the hydrogen in a fuel cell to get the power for an ion drive? I don't think that's wise. You'd get more delta/v using the hydrogen and oxygen in a conventional rocket.

A nuclear reactor fuel mass I don't think would be prohibitive. According to this site, generating 200 MW for a year would require about 5 tons of fuel. You're bringing much more Xenon along than this for your reaction mass. At large enough scale, nuclear power is something of a no-brainer for an interplanetary ship.

Note: the 5 ton figure is for low-enriched fuel; for space applications, you'd probably want to use higher enrichment, which would lower the required fuel mass.

  • $\begingroup$ I was using hydrogen mostly as an example, but also as something that at least has a reasonable chance of being harvestable from other places in the solar system. Exotic fuel sources will run out and be much harder to replace (I think?) $\endgroup$
    – nirurin
    Nov 6, 2019 at 18:20
  • 2
    $\begingroup$ If you're "burning" the hydrogen in anything you'll need to be bringing some kind of oxidiser with you. The fuel cell is no exception. $\endgroup$ Nov 6, 2019 at 22:19
  • $\begingroup$ @niurin - I don't know why we wouldn't be able to find vast amounts of uranium in the asteroid belt-- smaller delta-v to get it out than any hydrogen sources I can think of. $\endgroup$
    – antlersoft
    Nov 7, 2019 at 17:15
  • 1
    $\begingroup$ There are no fuel cells working with hydrogen only, they need oxygen too. $\endgroup$
    – Uwe
    Nov 7, 2019 at 21:39
  • $\begingroup$ If you want to harvest fuel, a nuclear thermal rocket may be better. Or, once again, chemical. $\endgroup$
    – ikrase
    Nov 9, 2019 at 19:59

Yes, ion engines are somewhat of a false economy in the sense that they rely on electrical power on top of their mass and their propellant (e.g. Xenon).
So, compared to a chemical rocket engine where engine mass, thrust, specific impulse, plus propellant (and the mass of the tanks to store it) is about everything you need to plan an hypothetical spaceship, here you do have to add the mass of the thing you will get electric power from.

Some figures for nuclear:

  • TOPAZ space flown in the '80, 5 kw for 320 Kg = 15.6 kw/ton
  • KRUSTY in development, possible figure 43.3kw from 226Kg = 191 kw/ton

However, remember that nuclear usually produces a lot of heat on top of electric power, so you might need some big radiators. Another figure for KRUSTY in the same link gives 335Kg for 1 kw and 1120 kw for 10kw for the whole package (not just the reactor) including heat rejection components.
This is a lot worse than 191 kw/ton and more like 10kw/ton.
The encouraging part is that the 10kw reactor did not need 10x mass compared to the 1kw one. It may be that @ 200MW a nuclear reactor would have a much better power/weight ratio.
Note that the nuclear fuel mass is only a very small percentage of the weight in the above figures.

Some figures for solar:

  • You can get as high as 500kw/ton! See this question.

I think you can get 200MW, in 300kw increments, at 150kw/ton so 1333 tons total, with an area of $785000 m^2$ (this one scaled linearly from 20kw with $78.5m^2$), with somewhat "proven" technology (see UltraFlex, MegaFlex).

The THINS solar array technology referred in the above question might do even better (500 kw/ton) but I did not find the figure about $kw/m^2$, only $kw/m^3$ (which is when panels are stored).

I do not know how much assembling so many of them would be troublesome.
Indeed lifting 1333 tons is troublesome, but you are already assuming a thousands tons spaceship so... Also your figure of 200MW would mean having 1000 VASIMIRs (1 VASIMIR = 200kw)... so I guess they will weight quite a few tons already!

  • $\begingroup$ "Also your figure of 200MW would mean having 1000 VASIMIRs (1 VASIMIR = 200kw)" Yeh, I ended up with something like 5000N of thrust. Which only ends up (if my maths are correct) pushing 5 tonnes at 1 m/s2. So 1000tonnes (or two ISS's) would be ... 0.005m/s2? This still sounds extremely slow, but to be honest I haven't had a chance to work out how much actual acceleration that adds up to over an actual voyage. Seems that, for a 1000tonne ship (without fuel), it would take about 200 days to get to the asteroid belt. Or about 80 days to get to mars. Not sure if that's good or bad! $\endgroup$
    – nirurin
    Nov 6, 2019 at 17:00
  • 2
    $\begingroup$ 0.005m/s2 can add up to 6000m/s in about 14 days where 6000m/s is about the deltaV you need from LEO to Mars Low Orbit (I would imagine such a large starship would be built on LEO and then just go back and forth between LEO and Mars Low Orbit). 80 days to Mars is good! How did you get that? Note that 6000m/s deltaV is for "standard" orbit transfers between Earth and Mars using chemical rockets.. With a very low acceleration it is possible that you need MORE than 6000m/s. $\endgroup$
    – BlueCoder
    Nov 6, 2019 at 17:46
  • $\begingroup$ Sorry I did the foolish thing and did it by distance, rather than the required deltaV. I'm still not used to thinking in terms of deltaV. For a 1,000,000KG ship, getting 0.005m/s2, how long would that take to get from LEO to LMO? -- and yes, the ship would start and end in orbits, it would never be launched from earth, way too expensive! $\endgroup$
    – nirurin
    Nov 6, 2019 at 18:07
  • $\begingroup$ If you know you get 0.005m/s2, the mass is irrelevant. I do not know with such small acceleration what path you can follow - I have read somewhere else in this site that said ion-drive could take 50% longer than chemical for Mars trajectories. However, I have no source at the moment with a Mars trajectory planned for ion-drives. Starfish Prime answer gives one (detaching payload which might not be your idea). $\endgroup$
    – BlueCoder
    Nov 8, 2019 at 16:48
  • $\begingroup$ One thing I did not say in my answer (but present in Starfish Prime one) is that solar power is influenced by distance from the Sun (at Mars distance you get around half solar power i.e. irradiance). $\endgroup$
    – BlueCoder
    Nov 8, 2019 at 16:49

One mission profile thought up for the VASIMR involves initially travelling closer to the sun to take advantage of the stronger solar flux to run the engine at high power, releasing a payload on a long coasting trajectory and slowing the launching vehicle back down to return to Earth.

VASIMR Ejector Concept

The idea is still a little half-baked at present (the engine performance isn't there yet, the mass ratios are too high, it isn't practical to retrieve the launching vessel, no thought was given to returning from the destination) but it shows that it is possible to use an electrically-driven rocket without the need for nuclear reactors.

In this case, the flight time is 3 years, which should be confortably within the lifetimes of any crew, though those lifetimes may be reduced somewhat by spending 3 years in interplanetary space...


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.