5 ms·

Really, really, really good. Making a nuclear weapon actually detonate in a nuclear manner is one of the hard problems of making a nuclear weapon. Basically y

Really, really, really good. Making a nuclear weapon actually detonate in a nuclear manner is one of the hard problems of making a nuclear weapon. Basically you have to set the explosives off just right. An accidental detonation of the conventional parts of the bomb can't do that (short of like space re-entry speeds or some other thing that could provide the force needed even without an explosion). The fear has always been around the detonation systems since those are the only thing realistically capable of setting off a nuclear explosion. Since the 1950s as detonation systems have gotten more reliable they've added more and more redundancy so those have all sorts of interlocks to prevent them from going off if not armed and fired. It's not like an airbag where if you hook up 12v to the right pins it goes bang. They actually need to be armed and then fired which requires a whole bunch of systems doing their thing in a specific order. The systems that arm weapons (like the aircraft and rocket side hardware) are relatively sophisticated and robust and a hell of a lot of man hours have been put into them over the years because the last thing you need is a risk of ordinance going off when you're already trying to fight a fire.

rtkwe 6y ago
Though a conventional detonation would turn it into a dirty bomb which while almost infinitely better than a full blown nuclear detonation would still ruin a good number of lives.
monocasa 6y ago
Adding to this, I've heard on the grapevine that the parameters of the detonation sequence are generated from a hash of the arming code number. The arming code number isn't like a traditional code lock where there's a chip that's doing "strcmp(entered_code, CORRECT_CODE)", but instead the arming code number is a fundamental piece of how the bomb works. Without that you need the kind of state nuclear apparatus you'd need to build a bomb in the first place to reverse engineer (think a state run nuclear research lab), or at best you can remanufacture it into a much crappier bomb that probably duds to just being a dirty bomb.
- megous 6y ago
  Seems more likely that the code is just a [part of] decryption key for the parameters, rather than some weird hashing scheme.
  - monocasa 6y ago
    I think the design goal was to never have a decryptable set of the parameters stored with the device.
    - megous 6y ago
      https://en.wikipedia.org/wiki/Permissive_Action_Link https://en.wikipedia.org/wiki/Permissive_Action_Link Dunno, wikipedia says the parameters are encrypted. Anyway it would probably be good enough to just rip all the safety/firing electronics out, and replace it with your own best effort attempt at controlling the fireset capacitors charging and firing. What's there to vary anyway? Charge voltages, and discharge timing? Odds would be you could probably make it go boom in a more controlled manner than what happened in past accidents when the bombs exploded after hitting the ground. The trouble would be getting past the tamper-proof barrier, because the modern bombs are probably made to self-destruct, from what I've read, so instead of a nice ready-made bomb internals waiting for a simpler arduino control circuit for a fireset, you'd be dealing with a hot mess, after trying to get in. Just some uninformed thoughts. I also found some images of fat man firset online: https://lanlmuseum.pastperfectonline.com/webobject/3D33BB7F-ABB3-4D52-A7C2-678145610359 https://lanlmuseum.pastperfectonline.com/webobject/3D33BB7F-... Interesting device. :)
      - monocasa 6y ago
        > Dunno, wikipedia says the parameters are encrypted. Laymen refer to any use of crypto as "encrypted". > Anyway it would probably be good enough to just rip all the safety/firing electronics out, and replace it with your own best effort attempt at controlling the fireset capacitors charging and firing. It's really not though, it's very very specific to the design of the warhead, and the electronics are embedded in it in a way that you can't just take it apart and put it back together. It's all nasty stuff like FOGBANK that if you could remanufacture, you wouldn't need to steal a nuke in the first place. > What's there to vary anyway? Charge voltages, and discharge timing? Odds would be you could probably make it go boom in a more controlled manner than what happened in past accidents when the bombs exploded after hitting the ground. Yep, nanosecond precision detonation on the shaped charges, and not just "all at the same time", but a very specific sequence that's dependent on the exact geometries, isotopes and these days, how they've decayed over decades. That's what Oak Ridge uses those fancy super computers for. > The trouble would be getting past the tamper-proof barrier, because the modern bombs are probably made to self-destruct, from what I've read, so instead of a nice ready-made bomb internals waiting for a simpler arduino control circuit for a fireset, you'd be dealing with a hot mess, after trying to get in. They are very much not designed to self destruct. Really the opposite. And yes, the threat model is that without the codes, you leave someone with components that they essentially have to remanufacture into a new bomb, which takes a well funded state actor. Or they just use the bits as a dirty bomb, which is why that gets so much play in the age of terrorist cells. But you don't get a full on usable nuke by just sticking a microcontroller on a stolen bomb.
_jal 6y ago
> Really, really, really good. They made up for it elsewhere, though, with the sort of behavior everyone expects: https://www.warhistoryonline.com/war-articles/us-nuclear-weapons-launch-code-terrifyingly-00000000.html https://www.warhistoryonline.com/war-articles/us-nuclear-wea...
lostlogin 6y ago
> Really, really, really good. Are they though? This accident had 5 of it’s 6 safety switches arm during the accident. https://en.m.wikipedia.org/wiki/1961_Goldsboro_B-52_crash https://en.m.wikipedia.org/wiki/1961_Goldsboro_B-52_crash
- dmitriid 6y ago
  "A single switch out of four (not six) was unarmed". So, the failsafes worked as intended. And yes, they should be even more failsafe.
  - lostlogin 6y ago
    Hmm, the wiki says “ Former military analyst Daniel Ellsberg has claimed to have seen highly classified documents indicating that its safe/arm switch was the only one of the six arming devices on the bomb that prevented detonation.” The linked story says “ Robert McNamara, who’d been Secretary of Defense at the time of the incident, told reporters in 1983, "The bomb’s arming mechanism had six or seven steps to go through to detonate, and it went through all but one.”” Where is your quote from? I can’t find it in the Natgeo article.
    - dmitriid 6y ago
      That very same wiki. The very next sentence after Daniel Ellsberg quote: "In 2013, information released as a result of a Freedom of Information Act request confirmed that a single switch out of four (not six) prevented detonation."
      - lostlogin 6y ago
        Thanks - sorry I was looking in the National Geographic article.
- Teever 6y ago
  That is a nuclear weapon that was Wikipedia says was in service from 1957-1966. Wiki also says that the first nuclear explosion occurred on July 16, 1945. This means that the development of the weapon in question occurred sometime in the first 12 years after the first nuclear weapons were tested. I think that it is safe to assume that this very old weapon design has only been approved on since the 1950s.

rtkwe 6y ago

Though a conventional detonation would turn it into a dirty bomb which while almost infinitely better than a full blown nuclear detonation would still ruin a good number of lives.

monocasa 6y ago

Adding to this, I've heard on the grapevine that the parameters of the detonation sequence are generated from a hash of the arming code number. The arming code number isn't like a traditional code lock where there's a chip that's doing "strcmp(entered_code, CORRECT_CODE)", but instead the arming code number is a fundamental piece of how the bomb works. Without that you need the kind of state nuclear apparatus you'd need to build a bomb in the first place to reverse engineer (think a state run nuclear research lab), or at best you can remanufacture it into a much crappier bomb that probably duds to just being a dirty bomb.

megous 6y ago
Seems more likely that the code is just a [part of] decryption key for the parameters, rather than some weird hashing scheme.
- monocasa 6y ago
  I think the design goal was to never have a decryptable set of the parameters stored with the device.
  - megous 6y ago
    https://en.wikipedia.org/wiki/Permissive_Action_Link https://en.wikipedia.org/wiki/Permissive_Action_Link Dunno, wikipedia says the parameters are encrypted. Anyway it would probably be good enough to just rip all the safety/firing electronics out, and replace it with your own best effort attempt at controlling the fireset capacitors charging and firing. What's there to vary anyway? Charge voltages, and discharge timing? Odds would be you could probably make it go boom in a more controlled manner than what happened in past accidents when the bombs exploded after hitting the ground. The trouble would be getting past the tamper-proof barrier, because the modern bombs are probably made to self-destruct, from what I've read, so instead of a nice ready-made bomb internals waiting for a simpler arduino control circuit for a fireset, you'd be dealing with a hot mess, after trying to get in. Just some uninformed thoughts. I also found some images of fat man firset online: https://lanlmuseum.pastperfectonline.com/webobject/3D33BB7F-ABB3-4D52-A7C2-678145610359 https://lanlmuseum.pastperfectonline.com/webobject/3D33BB7F-... Interesting device. :)
    - monocasa 6y ago
      > Dunno, wikipedia says the parameters are encrypted. Laymen refer to any use of crypto as "encrypted". > Anyway it would probably be good enough to just rip all the safety/firing electronics out, and replace it with your own best effort attempt at controlling the fireset capacitors charging and firing. It's really not though, it's very very specific to the design of the warhead, and the electronics are embedded in it in a way that you can't just take it apart and put it back together. It's all nasty stuff like FOGBANK that if you could remanufacture, you wouldn't need to steal a nuke in the first place. > What's there to vary anyway? Charge voltages, and discharge timing? Odds would be you could probably make it go boom in a more controlled manner than what happened in past accidents when the bombs exploded after hitting the ground. Yep, nanosecond precision detonation on the shaped charges, and not just "all at the same time", but a very specific sequence that's dependent on the exact geometries, isotopes and these days, how they've decayed over decades. That's what Oak Ridge uses those fancy super computers for. > The trouble would be getting past the tamper-proof barrier, because the modern bombs are probably made to self-destruct, from what I've read, so instead of a nice ready-made bomb internals waiting for a simpler arduino control circuit for a fireset, you'd be dealing with a hot mess, after trying to get in. They are very much not designed to self destruct. Really the opposite. And yes, the threat model is that without the codes, you leave someone with components that they essentially have to remanufacture into a new bomb, which takes a well funded state actor. Or they just use the bits as a dirty bomb, which is why that gets so much play in the age of terrorist cells. But you don't get a full on usable nuke by just sticking a microcontroller on a stolen bomb.

_jal 6y ago

> Really, really, really good. They made up for it elsewhere, though, with the sort of behavior everyone expects: https://www.warhistoryonline.com/war-articles/us-nuclear-weapons-launch-code-terrifyingly-00000000.html https://www.warhistoryonline.com/war-articles/us-nuclear-wea...

lostlogin 6y ago

> Really, really, really good. Are they though? This accident had 5 of it’s 6 safety switches arm during the accident. https://en.m.wikipedia.org/wiki/1961_Goldsboro_B-52_crash https://en.m.wikipedia.org/wiki/1961_Goldsboro_B-52_crash

dmitriid 6y ago
"A single switch out of four (not six) was unarmed". So, the failsafes worked as intended. And yes, they should be even more failsafe.
- lostlogin 6y ago
  Hmm, the wiki says “ Former military analyst Daniel Ellsberg has claimed to have seen highly classified documents indicating that its safe/arm switch was the only one of the six arming devices on the bomb that prevented detonation.” The linked story says “ Robert McNamara, who’d been Secretary of Defense at the time of the incident, told reporters in 1983, "The bomb’s arming mechanism had six or seven steps to go through to detonate, and it went through all but one.”” Where is your quote from? I can’t find it in the Natgeo article.
  - dmitriid 6y ago
    That very same wiki. The very next sentence after Daniel Ellsberg quote: "In 2013, information released as a result of a Freedom of Information Act request confirmed that a single switch out of four (not six) prevented detonation."
    - lostlogin 6y ago
      Thanks - sorry I was looking in the National Geographic article.
Teever 6y ago
That is a nuclear weapon that was Wikipedia says was in service from 1957-1966. Wiki also says that the first nuclear explosion occurred on July 16, 1945. This means that the development of the weapon in question occurred sometime in the first 12 years after the first nuclear weapons were tested. I think that it is safe to assume that this very old weapon design has only been approved on since the 1950s.