![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
So, there's a couple thoughts about realistic starships. One is that we can't do them and they're centuries of tech away. Another is what we could do them, or probably good with a decade or two of engineering research, but they'd be really expensive. Most people with a clue tend to think the first. But! The second might be more accurate, at least as far as the propulsion goes. The classic one is
http://en.wikipedia.org/wiki/Project_Orion_(nuclear_propulsion) and variant http://en.wikipedia.org/wiki/Nuclear_pulse_propulsion#Medusa
where you blow up nukes against a pusher plate. Advantage is that it utilizes the one kind of energy productive fusion we can actually do, fusion bombs. Disadvantage is people's nervousness about huge quantities of nuclear bombs, plus since a bomb has a minimum size, the vehicle has to be large. Which can be good if you're really out to send something big, but if you just want a probe, is problematic.
Recently I learned of http://en.wikipedia.org/wiki/Project_Longshot which seems like it might have a chance of working. Unlike http://en.wikipedia.org/wiki/Project_Daedalus which needs a fusion reactor cum drive, using D-He3 which is harder than the D-T we can't do, Longshot uses a fission reactor for basic power; For propulsion it would use lasers to induce D-He3 fusion pulses, the results of which would blast out as thrust.
AFAICT, in the worst case this is a fission powered ion drive with mild fusion boost. If 100 J of laser come in, and 25 J of energy is produced by fusion, it's not "positive", but so what? You've still got 125 J of heat -- there's nowhere for the input energy to go -- vs. the 100 J you started with, and the fusion hopefully boosts exhaust velocity to the fractions of lightspeed you want. More optimistically, this works where fusion reactors don't precisely because it's a rocket, not a reactor, and hot plasma leaking out all over is a feature, not a bug -- it's a rocket, squirting plasma is exactly what you want. And in this case Longshot would be a fission-triggered fusion rocket.
Design envisioned 4.5% c cruising speed, and stopping at Alpha Centauri to orbit stuff, vs. Daedalus's boosting up to 12% c and zipping by as a fly-by. Of course those mission designs are independent of the rocket type, merely reflecting the goals of different teams. 4.5%c would mean travel time of 96 years.
So it's still best for starship *probes*, not crewed missions, and assuming the engine works, that might well be the easy part -- you also need automation that'll work for 100 years (including a fission reactor with lots of fuel), and that can make decisions of how to explore the system and such. OTOH those don't feel to me like problems as hard as, say, fusion reactors, which we've spend 60 years failing at. Robustness and redundancy for stuff lasting, and the exploration doesn't seem AI-hard, especially since the final probe could have lots and lots of delta-vee. (Mission delta-vee of 9% c is 27,000 km/s; having the payload have 1000 km/s to play with doesn't feel like a big deal, but would give lots of room for even dumb mission calculations to orbit various objects.)
Oh, and you also need He3 -- it's nice because most of the energy is in charged particles, so you can use a magnetic rocket nozzle. Though I wonder if D-T fusion might not be that bad; 80% of the energy is in neutrons but hey, they have momentum too. Anyway, He3 can be transmuted from lithium via tritium if you wait a couple decades. Or, more dubiously, sucked from Saturn. Or even more dubiously, taken out of Lunar regolith -- though since we're collecting special space fuel, the fact that it allegedly takes more energy to extract the He3 than it would provide isn't a fatal drawback.
Oh, and D-T wouldn't work, at least for deceleration; most of the T would have decayed to He3 by the time you got there, so you'd need a D-He3 capable fusor anyway. But there's D-D.
Cost? The Wikipage and original paper don't say, but I can make up some numbers. Total probe mass is 400 tons, or 4e5 kg. Launch costs for LEO are $3-10K/kg, we're going past LEO, say $10K/kg, for a total of $4 billion, suspiciously low but I can't see an error. Launch costs tend to be only 20% of the cost of a satellite, so that suggests $20 billion, and we're talking about cutting edge has-to-last-a-century stuff, so let's say 100x launch cost is more reasonable for development, for $400 billion.
$400 billion for an interstellar probe that could get there in a century? That's a lot of money, but it's not a LOT of money. We have wars for more than that. The ISS is like 1/4 that cost and a lot less useful for science. Apollo was 1/3 of that. Humans to Mars is probably comparable. Also, unlike satellites which I assume consist of a supermajority of space-rated miniaturized electronics, the bulk of the probe is simple fuel and tanks, which shouldn't be that much more expensive than launching them. (Though having your hydrogen not leak out of the tank over a century is a challenge too.) Of course, I haven't included the cost of the He3.
One website claims He-3 is currently about $4 billion a ton, which would mean $800 billion in a simple but dubious linear extrapolation. So, more cost. But again, US GDP over a decade is $140,000 billion; world GDP is over 4x that. We're talking a total mission cost that's 1% of world GDP for a decade; politically unlikely, especially at the moment, but far from *impossible*.
Why is an interstellar probe comparable in cost to a manned mission to Mars or the ISS? Probably because of "probe" vs. "manned" (or of course because I'm ignorantly overlooking some massive source of costs.) Final payload would be 30 tons, and the probe can basically 'sleep' most of the way; keeping people alive and healthy -- and breeding -- for a century would take a lot lot more. And then there's the cost of a return capability. *That* probably *would* bankrupt the planet, at least under my "100:1" development cost assumption.
And there's another nuclear rocket design, http://en.wikipedia.org/wiki/Fission-fragment_rocket where you don't even both with fusion, just let the fission fragments zip off at their natural 3% lightspeed.
Need lots of uranium for that. U308 is currently $52/pound, let me call that $200/kg for space purposes, 400 tons, $80 million. U3O8 probably isn't purified enough, but even another 100x is just $8 billion, trivial compared to the other costs.
I guess the takeaway lesson is that you need nuclear energies to have even kind of crappy interstellar ships, but we *do* have nuclear energies, and throwing large but reasonable amounts of fission, and maybe explosive fusion, at the problem will suffice for automated probes, such that propulsion may well be the easy bit.
http://en.wikipedia.org/wiki/Project_Orion_(nuclear_propulsion) and variant http://en.wikipedia.org/wiki/Nuclear_pulse_propulsion#Medusa
where you blow up nukes against a pusher plate. Advantage is that it utilizes the one kind of energy productive fusion we can actually do, fusion bombs. Disadvantage is people's nervousness about huge quantities of nuclear bombs, plus since a bomb has a minimum size, the vehicle has to be large. Which can be good if you're really out to send something big, but if you just want a probe, is problematic.
Recently I learned of http://en.wikipedia.org/wiki/Project_Longshot which seems like it might have a chance of working. Unlike http://en.wikipedia.org/wiki/Project_Daedalus which needs a fusion reactor cum drive, using D-He3 which is harder than the D-T we can't do, Longshot uses a fission reactor for basic power; For propulsion it would use lasers to induce D-He3 fusion pulses, the results of which would blast out as thrust.
AFAICT, in the worst case this is a fission powered ion drive with mild fusion boost. If 100 J of laser come in, and 25 J of energy is produced by fusion, it's not "positive", but so what? You've still got 125 J of heat -- there's nowhere for the input energy to go -- vs. the 100 J you started with, and the fusion hopefully boosts exhaust velocity to the fractions of lightspeed you want. More optimistically, this works where fusion reactors don't precisely because it's a rocket, not a reactor, and hot plasma leaking out all over is a feature, not a bug -- it's a rocket, squirting plasma is exactly what you want. And in this case Longshot would be a fission-triggered fusion rocket.
Design envisioned 4.5% c cruising speed, and stopping at Alpha Centauri to orbit stuff, vs. Daedalus's boosting up to 12% c and zipping by as a fly-by. Of course those mission designs are independent of the rocket type, merely reflecting the goals of different teams. 4.5%c would mean travel time of 96 years.
So it's still best for starship *probes*, not crewed missions, and assuming the engine works, that might well be the easy part -- you also need automation that'll work for 100 years (including a fission reactor with lots of fuel), and that can make decisions of how to explore the system and such. OTOH those don't feel to me like problems as hard as, say, fusion reactors, which we've spend 60 years failing at. Robustness and redundancy for stuff lasting, and the exploration doesn't seem AI-hard, especially since the final probe could have lots and lots of delta-vee. (Mission delta-vee of 9% c is 27,000 km/s; having the payload have 1000 km/s to play with doesn't feel like a big deal, but would give lots of room for even dumb mission calculations to orbit various objects.)
Oh, and you also need He3 -- it's nice because most of the energy is in charged particles, so you can use a magnetic rocket nozzle. Though I wonder if D-T fusion might not be that bad; 80% of the energy is in neutrons but hey, they have momentum too. Anyway, He3 can be transmuted from lithium via tritium if you wait a couple decades. Or, more dubiously, sucked from Saturn. Or even more dubiously, taken out of Lunar regolith -- though since we're collecting special space fuel, the fact that it allegedly takes more energy to extract the He3 than it would provide isn't a fatal drawback.
Oh, and D-T wouldn't work, at least for deceleration; most of the T would have decayed to He3 by the time you got there, so you'd need a D-He3 capable fusor anyway. But there's D-D.
Cost? The Wikipage and original paper don't say, but I can make up some numbers. Total probe mass is 400 tons, or 4e5 kg. Launch costs for LEO are $3-10K/kg, we're going past LEO, say $10K/kg, for a total of $4 billion, suspiciously low but I can't see an error. Launch costs tend to be only 20% of the cost of a satellite, so that suggests $20 billion, and we're talking about cutting edge has-to-last-a-century stuff, so let's say 100x launch cost is more reasonable for development, for $400 billion.
$400 billion for an interstellar probe that could get there in a century? That's a lot of money, but it's not a LOT of money. We have wars for more than that. The ISS is like 1/4 that cost and a lot less useful for science. Apollo was 1/3 of that. Humans to Mars is probably comparable. Also, unlike satellites which I assume consist of a supermajority of space-rated miniaturized electronics, the bulk of the probe is simple fuel and tanks, which shouldn't be that much more expensive than launching them. (Though having your hydrogen not leak out of the tank over a century is a challenge too.) Of course, I haven't included the cost of the He3.
One website claims He-3 is currently about $4 billion a ton, which would mean $800 billion in a simple but dubious linear extrapolation. So, more cost. But again, US GDP over a decade is $140,000 billion; world GDP is over 4x that. We're talking a total mission cost that's 1% of world GDP for a decade; politically unlikely, especially at the moment, but far from *impossible*.
Why is an interstellar probe comparable in cost to a manned mission to Mars or the ISS? Probably because of "probe" vs. "manned" (or of course because I'm ignorantly overlooking some massive source of costs.) Final payload would be 30 tons, and the probe can basically 'sleep' most of the way; keeping people alive and healthy -- and breeding -- for a century would take a lot lot more. And then there's the cost of a return capability. *That* probably *would* bankrupt the planet, at least under my "100:1" development cost assumption.
And there's another nuclear rocket design, http://en.wikipedia.org/wiki/Fission-fragment_rocket where you don't even both with fusion, just let the fission fragments zip off at their natural 3% lightspeed.
Need lots of uranium for that. U308 is currently $52/pound, let me call that $200/kg for space purposes, 400 tons, $80 million. U3O8 probably isn't purified enough, but even another 100x is just $8 billion, trivial compared to the other costs.
I guess the takeaway lesson is that you need nuclear energies to have even kind of crappy interstellar ships, but we *do* have nuclear energies, and throwing large but reasonable amounts of fission, and maybe explosive fusion, at the problem will suffice for automated probes, such that propulsion may well be the easy bit.