I certainly have no detailed knowledge of how nuclear weapons work or worked, but I have read that nuclear weapons into at least the late 50s (ie early fusion designs), required significant maintenance. A nuclear weapon left without maintenance (at least replacing the tritium) would become unusable in six months to a year. While in stories some hand-waving science fiction would fix such problems, it was a limitation that the real world had to deal with. A nuke in space would have to be replaced regularly, and bringing back to old one would be a bit dicy for any real military to plan for. 
Do recall that he and most other authors of the time was operating under the assumption that space would be colonized. So sustaining people in orbit would be no problem as there’d be space stations, moon bases and so forth to do so.
I’m starting to think that the premise was based on nuclear powered “rocket ships” that could haul warheads into space en masse cheaper than putting them on their own individual launchers.
Yeah, pretty much. They had single stage to orbit ships that we can’t match even now, with what companies like Space-X are doing.
This is not really true. Until the 'Eightes, most nuclear weapon designs in the US nuclear arsenal were not “one point safe”; that is, that “the probability of achieving a nuclear yield greater than four pounds of TNT must not exceed one in a million for any event involving the initiation of the warhead’s high explosive at a single point on its periphery.” (and such ‘unsafe’ weapons remained on deployment and in the Active Stockpile until the late 'Nineties). Most Russian nuclear weapons are not “one point safe”, and the same is probably true for India, Pakistani, and almost certainly North Korean weapons. (I don’t know about Chinese and Israeli weapons but there is enough cycles of development that I expect that modern designs meet the intent if not the specific requirements.) Nuclear weapons that malfunction will not likely achieve anything close to their full yield with a less-than-optimal detonation field, but even a ‘fizzle’ is still a fraction of a kiloton of TNT equivalent energy release, as well as the radioactivity from the bomb materials and any other fallout produced by interaction with the local environment.
The fissile material of the weapons will be stable for decades (although eventually discomposition caused by neutron flux can cause unexpected asymmetries in the fission energy profile) but as you note the tritium that is used as a booster charge in boosted fission and most fusion weapons has a half-life of 12.32 years, and the daughter product of decay of 3He is a neutron absorber (referred to in the field as a ‘neutron poison’) that in sufficient abundance will result in a fizzle (a weapon that blows itself apart before achieving the design yield). So any weapon in orbit would have to be designed to not use tritium, which is certainly possible but is a constraint.
There is another security and reliability consideration for space-based nuclear weapons: that of cybersecurity. Ground based intercontinental ballistic missiles (ICBMs) are sent launch codes via hardline (the Hardened Intersite Cable System) from the launch control center (which has various methods of communication with USSTRATCOM) and submarine launched ballistic missiles (SLBMs) are launched locally by crews sent launch orders via a virtually unjammable extremely low frequency (ELF) (and potentially other methods) so there is always a man-in-loop failsafe but weapons in space would have to be controlled from the ground, and sent the actual launch commands by encrypted radio or laser. Radio signals could be jammed, and a high altitude electromagnetic pulse (HEMP) could fry that entire system, but even worse an opponent which broke the encryption scheme and acquired launch codes could take over such a system with no man-in-loop intervention.
The destabilizing effects of putting nuclear weapons in orbit, thus eliminating the potential for discovering a launch alert is manifestly destabilizing as noted in discussion above. The issue of ‘basing’ such weapons where they can be as responsive as ICBMs and SLBMs is also a problem, because while their is a known period of time between launch and impact on fixed points or a patrol area in broad ocean area, the time that an orbital weapon can be on target depends upon the phasing of the orbit. If a weapon can’t be in position to strike a ground target for ~90 minutes then it is essentially useless as a prompt counterforce weapon (although, frankly, the value of ‘counterforce’ in an apocalyptic nuclear exchange is mostly theoretical, anyway, but does factor into deterrence strategies). In addition, there is really no hiding in space, and weapon platforms would be pretty obvious by their orbital elements and characteristics. Such platforms would be vulnerable to pre-emptive kinetic attacks of various kinds by an opponent, and of course if those attacks are detected before they are successful the reflexive response is to launch the weapons while they are still usable, spastically instigating nuclear exchange.
So, basing nuclear weapons in Earth orbit is a really terrible idea for numerous reasons in addition to the fact that it is just an enormous pain in the ass and not able to be readily recovered at some future time when a nation might elect to deproliferate or just retire obsolete weapons.
Stranger
Which does bring things back to the erroneous assumption among sci-fi writers of the time that space would be heavily colonized. And probably also an assumption that manual control would be necessary. They were probably mostly thinking in terms of a crewed military space station with nuclear weapons on board; not an automated launch system. The idea that there wouldn’t be a “man in the loop” probably wasn’t even a consideration.
Also, the same goes for maintenance. They’d have assumed that any needed maintenance of the missiles would be done by technicians on the space station from onboard supplies, likely shipped in from all over the Solar System on nuclear-powered rockets. Not that they’d have to be maintained by shipping the supplies and technicians to orbit on a chemical rocket from Earth, every single time.
AIUI, the main problem with a nuke going off high in the atmosphere would not be so much the fallout as the EMP wave it would release that would knock out power grids and devices for a huge area.
Another issue would be a launch failure that did not cause a nuclear explosion, but rather distributed highly radioactive material all over the downrange area. To case-harden the bomb against this would require serious armour which implies much more weight, the opposite of what you want in a payload. Also, in a bomb the radioactive material is surrounded by explosive, which adds to the risks.
Which rocket scientist was it that said a weapon in orbit could be defeated by a bucket of nails?
That would certainly be a concern, but nuclear fallout mostly comes from two things – using certain elements/materials in the fission stages of bombs (fusion bombs are far cleaner per yield than fission bombs, though they do contain small fission bombs as their initiator) and the materials they’re detonated next to. Ground bursts create much more fallout than air bursts both because the soil/rock becomes more irradiated and it kicks up debris which ends up spreading around. A nuclear burst high in the atmosphere with a relatively clean fallout design would not create too much fallout.
But more importantly, the spacecraft blowing up wouldn’t set off the nuclear detonation – it would end up being just a regular old dirty bomb. A concern, but the fact that the debris would be scattered so widely would probably mean that nowhere was hit particularly hard. It would be a huge diplomatic incident because the public is terrified of anything radioactive, but it probably wouldn’t have much of an effect of worldwide background radiation.
Well, the main difference, of course, is that these launches would occur in peacetime. If an ICBM blew up during liftoff no one would care – the world is about to get blown up and having it “just” be a dirty bomb is going to do less damage compared to it reaching its target and detonating anyway.
From memory and googling, it was Arthur C. Clarke. Found a quote on it here:
This bucket of nails isn’t related to a USSR satellite, but was theorised by Arthur C. Clarke.
>2 For the origins of Clarke’s “bucket of nails” argument; Clarke gave the speech “Where is Mankind Headed?” on May 25 1983 for the Reader’s Digest Worldwide Editorial Conference. In the conference’s summary “Memories of Monaco,” Clarke’s speech was described as follows: “After dinner, Arthur Clarke, the noted British science-fiction writer, spoke on ‘The Militarization of Space.’ To put nuclear war into perspective, he quoted a chilling statement from Carl Sagan: ‘A full-scale thermonuclear exchange would be the equivalent of World War II, once a second, for the length of a lazy summer afternoon.’ Clarke went on to demonstrate that there’s no security in technology, concluding that the only defense is to prevent weapons from being used. Thus, the problem is political, not military. At one point, he described how a sophisticated, multi-billion-dollar laser weapon system could be destroyed by a bucket of nails,” see “Memories of Monaco from Reader’s Digest Worldwide Editorial Conference,” May 21-28, 1983, Folder 4, Box 144, Arthur C. Clarke Collection (Acc. 2015-0010), Air and Space Archives, National Air and Space Museum, Smithsonian Institution, Washington, DC, 14-15.
Reentry vehicles (RVs) are already designed to withstand to survive the intense shock and thermal conditions of flying back into the atmosphere at hypersonic speeds, and the explosive lenses which focus the blast inward to compress the primary, producing the supercritical configuration that initiates the prompt nuclear reaction, are made from insensitive high explosive (IHE) intended to survive catastrophic destruction of the boost vehicle.
Stranger
It sounds like you, and/or @snowthx, are asking about LV reliability, which may not be 100% but I do not see why it would vary radically as a function of the trajectory. Note also that during your Global Thermonuclear War you are not going to launch only one ICBM.
As for what happens if the rocket does happen to explode, it is not a secret that the modern gen of nuclear weapons are designed to “one-point safe” specifications, as has been pointed out, so you are not going to nuke your own silo.
Which is amazing considering that the Germans thought of it almost immediately after the development of the V2. And developed a typically Teutonic design for a towed submarine missile launcher to attack the US with.
Me, too! The secret to keeping the peace with an arms race is relative equality. If one side gets an advantage so large that it makes the other side feel doomed, look out.
That was pretty much the function of the “Davy Crockett” tactical nuclear weapon, which was actually intermediate between a true tactical nuke and a dirty bomb. The Davy Crocket was simply too small to achieve a full nuclear explosion, but it could produce a substantial pulse of prompt radiation like a weaponized “Demon Core”. This would kill any unprotected enemy forces within half a mile, and leave the site (such as a narrow pass invaders would need to use) too radioactive to safely pass through for 48 hours.
Although there’s a concept called the Liberty Ship, which would be a gas-core “lightbulb” fission rocket of Heinleinesque performance. Engineers vary on whether this would be merely enormously challenging or flatly impossible. Still, it’s close to what the science fiction writers of the 1950s envisioned for the future of space travel. Writers of the era had a tendency to use the word “atomic” as basically a stand-in for magic; or even if they considered the hard science technical challenges simply presumed that atomic energy was in still in its infancy and more sophisticated technology than a mere “pile” would eventually be developed.
Here is a discussion from the Center for Strategic and International Studies (CSIS) on space-based nuclear weapons and specifically focused on high altitude nuclear explosions (HANEs):
Stranger
How much fuel would be needed to decelerate a space-based warhead in (presumable) LEO to hit an arbitrary point on earth? It’s definitely not like dropping rocks off of a highway overpass…
You would need a near-polar orbit and a reentry vehicle with the largest practical cross range; otherwise you’d be limited to what sites happened to be in the ground track of the warhead’s orbit. You might want one constellation of near-polar orbit warheads for higher latitude targets and another for equatorial/tropical targets. A comparatively modest amount of fuel is sufficient to drop the orbit to where atmospheric drag does the rest.
That’s LEO. An interesting question is how much deorbiting burn is needed to intersect the atmosphere from different starting altitudes. LEO can be nearly grazing the atmosphere and so even a modest change of delta-v drops you into reentry. A very high (e.g. L-4 or L-5) orbit at lunar distance has a low transverse velocity and so a modest burn drops you like a rock. In-between might be worse than either extreme; needing a lot of delta-v before your trajectory intersects the atmosphere. Further away makes target selection easier but lengthens how long it takes; depends on if prompt response is necessary or if an eventual Sword of Damocles will suffice.
That depends on the orbital parameters relative to the target and the mass of the RV and descent vehicle. The Space Transportation System Orbiter Vehicle (“Space Shuttle”) Orbital Maneuvering System (OMS) had a Δv of about 0.3 km/s for a relatively gentle descent profile but for prompt delivery of an RV you’d probably need something like 3-4 km/s even assuming the target is on the ground track over the weapon platform. That is a sizable ‘kick motor’.
The amount of crossrange momentum transfer available to make azimuth changes sufficient to cover a ±45° range would be enormous, and would require turning the RV into an impractically giant glider. A ‘practical’ system for being able to target anywhere in Eurasia (we’ll leave Australia alone and assume Africa isn’t worth nuking for the moment) within 30 minutes or so would require satellites at multiple orbital azimuths and phased to be in range. In fact, it would look very much like the proposed ‘Brilliant Pebbles’ spaced-based anti-ballistic missile system (see Figure 1-1 in the linked GAO report) and would be subject to many of the same issues addressed therein. So it really isn’t any kind of ‘improvement’ over the current system save that there isn’t the evidence of launch flares from satellite early warning systems, which makes its only utility a first strike weapon.
Stranger