[funsec] Fun with (In)Security
drsollyp at drsolly.com
Mon Jan 30 09:36:18 CST 2006
On Mon, 30 Jan 2006 Valdis.Kletnieks at vt.edu wrote:
> On Mon, 30 Jan 2006 11:19:28 GMT, Drsolly said:
> > You get a gun (that's a real gun, not the popguns that it's legal to buy,
> > something in 88mm or so) and fire one piece at the other; that's how
> > Little Boy worked. I don't think it's rocket science.
> Actually, it's trickier than it looks - the "obvious" solution of simply
> slamming two sub-critical hemispheres together will almost certainly fizzle.
> The problem is that as it gets close, the surfaces start to melt down and
> vaporize, and the resulting high-pressure vapor tends to push the pieces apart
> before you get a really good chain reaction going.
> And if you want to get good efficiency, the projectile part is riding the
> pressure wave of some really high energy explosives to compress the fissionable
> material. Yes, it's metal, but a really good shaped charge can compress metal
> by a factor of 3 to 5 or more, which makes the reaction go *lots* faster by
> reducing the average neutron travel distance (and hey, that explosive only has
> to keep it compressed for a few ten-thousandths of a second). Oh, and you want
> to try really hard to make the compression symmetric, as otherwise the more
> dense side will start to fission first, and blow itself apart before the other
> side gets going, resulting in a fizzle or a low yield....
> Of course, there's not much real difference between a dirty bomb and a fizzle,
> so if you're a terrorist you can just bang two hemispheres together... :)
It's like programming. Version 1 tells you what you need to do to make a
good version 2.
More information about the funsec