3 ms·

I am in complete agreement, with the exception of this statement: > there is virtually no way for the nuke to "fail" in a way that would be even remotely close

by Kadin 10y ago

I am in complete agreement, with the exception of this statement: > there is virtually no way for the nuke to "fail" in a way that would be even remotely close to a detonation sequence That is not true, at least not universally. Nuclear bombs are, by definition, designed to go off, and as a result there is a category of failure modes that could lead to inadvertent detonation. Basically, a failure that produces the same command that would cause the normal detonation sequence to begin, could cause an unintended detonation. The Mk. 39 bomb involved in the 1961 North Carolina B-52 crash [0] reportedly came close to detonating in this manner. Additionally, there are weapons designs that are not "one point safe", such that a failure which causes ignition of the high explosive could lead to a nuclear detonation. Weapons considered "one point safe" require multiple, carefully timed or simultaneous ignitions of the high explosive. The last non-one-point-safe weapons in the US arsenal were built in the 50s but not disassembled until recently; the W56 was reportedly not one point safe [1], or at least not provably so to the safety margins now recognized as reasonable. I don't know of any information in the unclassified literature about whether Soviet weapons are universally one-point-safe or not. It would not surprise me if there are a variety of weapon designs still floating around in the world that aren't one-point-safe. Single compression "gun type" bombs, like the Fat Man design, would require additional engineering to make one-point-safe beyond what is minimally required to get them to detonate, which makes me suspect that some marginal nuclear powers probably don't bother. [0]: https://en.wikipedia.org/wiki/1961_Goldsboro_B-52_crash https://en.wikipedia.org/wiki/1961_Goldsboro_B-52_crash [1]: https://web.archive.org/web/20070110205223/http://www.pogo.org/p/homeland/hl-061201-bodman.html https://web.archive.org/web/20070110205223/http://www.pogo.o...

mikeash 10y ago
Can a gun type weapon ever be made one point safe? You'd have to somehow make sure that the explosives could never accidentally drive the two pieces together, which seems impossible. Doing it for an implosion weapon is, relatively speaking, much more straightforward: implosion is inherently finicky anyway, so you design it to fail unless multiple detonations happen at precise times.
- Animats 10y ago
  The US hasn't had a gun-type nuclear weapon since 1963. The Hiroshima weapon was gun-type, and there was a serious risk of accidental detonation. Later gun-type bombs were used in the "atomic cannon" artillery shells. These apparently came in two parts intended to be assembled just before use. Other countries, who knows? The first Soviet atomic bomb was an implosion device.
- sandworm101 10y ago
  Yes, mechanically. You put something in the barrel that would shatter the 'bullet' should the explosives go off. Detonation then occurs only if the safety is moved prior to explosion. Then, as another measure, you can rig that mechanical safety mechanism to snap back in place in cases of external fire (ie some part is designed to melt and swing it back in place).
  - mikeash 10y ago
    Sounds like a good system! Does it really eliminate the possibility of any detonation, though? Seems like you could still potentially get enough stuff together to exceed critical mass. The explosion would probably be smaller than nominal, but more than the 4lbs TNT equivalent that the US military specifies as its criteria for one-point safety. But perhaps I overestimate the chances of this happening. I'm far from an expert here.
dfsegoat 10y ago
Having read the following, I am now convinced that Goldsboro was NOT anywhere close to a nuclear detonation. Author below has documents FOIA'd from NNSA and DOE to back his claims in a well written / detailed document: TLDR: These weapons require auxiliary aircraft power and aircrew intent to initiate the mechanisms and sequence (19 steps) required to even arm the bomb so that it can produce a nuclear detonation. http://nuclearweaponsaccidents.blogspot.com/2013/03/goldsboro-19-steps-away-from-detonation.html http://nuclearweaponsaccidents.blogspot.com/2013/03/goldsbor... "Bomb 2, the object of the Goldsboro controversy, was not "one step" away from detonation (nor was Bomb 1). The Mark 39 Mod 2 had two additional safety switches, the Trajectory Arm Switch and Rotary Safing Switch."
dogma1138 10y ago
The Goldsboro crash didn't came "close" to a detonation, the result of the crash were that all but 1 of the fail safes that would prevent it from going off were damaged, it doesn't mean that it would be able to detonate if the last one would have failed too from the crash. The detonation sequence of the nuke isn't just the conventional explosive part it's also the nuclear trigger itself (a fast neutron source) which has to fire at the exact time and power levels while the casing is still intact and can be used as a neutron reflector. Without that even if the core goes critical it would not detonate (might result in a fizzle if the core is really big) but it would go very hot and melt down. "one point safe" is a metric in which the US measures the nuclear explosion due to accidental detonation, anything above 4 lbs of TNT doesn't count as one point safe, this however doesn't mean that any none one point safe weapons can be detonated at full yield or anything close to that. So yes I would still stand by the fact that there is absolutly "zero" chance of an accidental damage to a weapon to bypass all the PAL's and other failsafes and initiate a nuclear triggering which would result in a substantial yield detonation. With modern weapons it's even less likely since technically modern designs can be considered "subcritical" (under normal pressures and conditions) and only undergo fission when they are configured properly via boosting/salting and neutron management, this is how you get "dial a yield weapons".

by Kadin 10y ago

mikeash 10y ago

Can a gun type weapon ever be made one point safe? You'd have to somehow make sure that the explosives could never accidentally drive the two pieces together, which seems impossible. Doing it for an implosion weapon is, relatively speaking, much more straightforward: implosion is inherently finicky anyway, so you design it to fail unless multiple detonations happen at precise times.

Animats 10y ago
The US hasn't had a gun-type nuclear weapon since 1963. The Hiroshima weapon was gun-type, and there was a serious risk of accidental detonation. Later gun-type bombs were used in the "atomic cannon" artillery shells. These apparently came in two parts intended to be assembled just before use. Other countries, who knows? The first Soviet atomic bomb was an implosion device.
sandworm101 10y ago
Yes, mechanically. You put something in the barrel that would shatter the 'bullet' should the explosives go off. Detonation then occurs only if the safety is moved prior to explosion. Then, as another measure, you can rig that mechanical safety mechanism to snap back in place in cases of external fire (ie some part is designed to melt and swing it back in place).
- mikeash 10y ago
  Sounds like a good system! Does it really eliminate the possibility of any detonation, though? Seems like you could still potentially get enough stuff together to exceed critical mass. The explosion would probably be smaller than nominal, but more than the 4lbs TNT equivalent that the US military specifies as its criteria for one-point safety. But perhaps I overestimate the chances of this happening. I'm far from an expert here.

dfsegoat 10y ago

Having read the following, I am now convinced that Goldsboro was NOT anywhere close to a nuclear detonation. Author below has documents FOIA'd from NNSA and DOE to back his claims in a well written / detailed document: TLDR: These weapons require auxiliary aircraft power and aircrew intent to initiate the mechanisms and sequence (19 steps) required to even arm the bomb so that it can produce a nuclear detonation. http://nuclearweaponsaccidents.blogspot.com/2013/03/goldsboro-19-steps-away-from-detonation.html http://nuclearweaponsaccidents.blogspot.com/2013/03/goldsbor... "Bomb 2, the object of the Goldsboro controversy, was not "one step" away from detonation (nor was Bomb 1). The Mark 39 Mod 2 had two additional safety switches, the Trajectory Arm Switch and Rotary Safing Switch."

dogma1138 10y ago

The Goldsboro crash didn't came "close" to a detonation, the result of the crash were that all but 1 of the fail safes that would prevent it from going off were damaged, it doesn't mean that it would be able to detonate if the last one would have failed too from the crash. The detonation sequence of the nuke isn't just the conventional explosive part it's also the nuclear trigger itself (a fast neutron source) which has to fire at the exact time and power levels while the casing is still intact and can be used as a neutron reflector. Without that even if the core goes critical it would not detonate (might result in a fizzle if the core is really big) but it would go very hot and melt down. "one point safe" is a metric in which the US measures the nuclear explosion due to accidental detonation, anything above 4 lbs of TNT doesn't count as one point safe, this however doesn't mean that any none one point safe weapons can be detonated at full yield or anything close to that. So yes I would still stand by the fact that there is absolutly "zero" chance of an accidental damage to a weapon to bypass all the PAL's and other failsafes and initiate a nuclear triggering which would result in a substantial yield detonation. With modern weapons it's even less likely since technically modern designs can be considered "subcritical" (under normal pressures and conditions) and only undergo fission when they are configured properly via boosting/salting and neutron management, this is how you get "dial a yield weapons".