Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Do Nuclear Weapons Need to be Tested?

  1. Oct 5, 2006 #1
    There has been a lot of speculation over whether North Korea has nuclear weapons and now they are announcing that they plan on testing. I'm a software engineer and I know that no matter how meticulously you plan a piece of software, you do not know if it will work correctly until you start it up and more often than not, it does not work exactly the way you planned. I imagine any nuclear engineering involves a great deal of complexity and many places where a miscalculation could be made. It seems to me that having a nuclear weapon that does not work would be more than useless if actually launched, considering the retaliation would be devastating. My question is, is it possible for a country to build a nuclear weapon and have any confidence that it will actually work without testing it?
  2. jcsd
  3. Oct 6, 2006 #2
    why not? if they send the nuke to wheverever they are probably pretty desparate and are asking to be nuked by the major nuclear countries anyways if it works it works if it dosnt their screwed and still no body would have the right to nuke that country since no damage was done
  4. Oct 6, 2006 #3
    I'm an EE so I'm not an expert on the topic but I guess you can use simulation's rather than testing, because to my knowledge Israel haven't tested any of it's nuclear weapons.
  5. Oct 6, 2006 #4


    User Avatar

    Staff: Mentor

    There are two major systems to be tested - 1) the critical mass, and 2) everything else - more or less.

    The critical mass (of some fissile material, e.g. U-233, U-235 or Pu-239) is the principal component. Everything else in the package is simply designed to transform the CM from subcritical (safe) to prompt supercritical as quickly as possible - on order of microseconds.

    The CM can be tested up to a point - just short of being supercritical. The detonation is the actual test.

    The rest can be tested in 'separate effects' tests which test the electronics and explosives, without the fissile material.

    Ultimately though to see how effective the device is, one usually wants to test a prototype to determine yield.

    Simulations can be performed, but one needs actual experimental results with which to compare the results of the simulations. Separate effects experiments can be used to test modules of the simulator, but the integrated simulation still needs to tested against an integrated system.

    One has to wonder if NK has fabricated more than one, such that if the US attacks NK following a test, then NK uses the remaining set.
  6. Oct 6, 2006 #5


    User Avatar
    Science Advisor


    Your analogy above to software is a good one - as are your instincts here.

    Let me separate the question into two questions. Is it necessary for a nascent
    proliferant - i.e. a country that is producing its first nuclear weapons, to test?
    The other question is whether it is necessary for an experienced nuclear power to test.

    As in the case of your software, and as Astronuc points out; you can test the
    components and subsystems of the device without producing a mushroom cloud
    You can thus assure yourself that the components work to specification.

    You do that with software too. [ I'm a computational physicist; I write programs to
    simulate physics.] You test the various modules of a program you are writing
    individually. You also assure yourself that the various modules conform to the
    various interfaces between modules that you define.

    After doing that level of testing; the program SHOULD work!!! The operational
    word is SHOULD. As your instincts tell you, there may be something you overlooked.
    It's hard to plan and address the "unknown unknowns"

    So a nascent nuclear power could be fairly confident in their new weapon; but they
    don't know FOR SURE until they test it.

    Now for experienced nuclear weapons powers like the USA; the situation is
    different. The USA is looking to replace some of its old warheads with new ones.
    Since the USA has decided not to test; these new warheads would have to be
    fielded without nuclear testing.

    So why put out new untested weapons when you have old tested weapons?

    Because weapons "age". Consider the following analogy. Your small town is
    fortunate enough not to have had any structural fires in the last 40 years. The
    town's fire engine has been run, i.e. "tested" - but that was 40 years ago. The
    old fire engine has just been sitting there for 40 years.

    Somebody tows into the firestation, a brand new, never run, fire engine fresh off
    the assembly line from a company that has been making working fire engines for years.
    Question: "Which fire engine are you more confident will start and run if the
    call comes in to go to a fire?"

    Dr. Gregory Greenman
    Last edited: Oct 6, 2006
  7. Oct 6, 2006 #6


    User Avatar
    Science Advisor



    It's an open question.

    Dr. Gregory Greenman
  8. Oct 6, 2006 #7


    User Avatar

    Staff: Mentor

    How about a third question: Was it necessary to test the very first nuclear wepons?

    When the theories governing nuclear reactions themselves were new (even still under development), people weren't entirely certain what would happen when the little red button was pushed for the first time. I'm sure by the time they got there, they were relatively confident, but I bet they were still sweating a little.

    I even heard that Heisenberg incorrectly calculated the critical mass requirements, which was a key reason why Germany never made a nuclear weapon. I'm not sure of the specifics of that, though.
  9. Oct 6, 2006 #8


    User Avatar

    Staff: Mentor

    The very first nuclear weapons test was Trinity - Alamogordo, NM, on July 16, 1945. http://en.wikipedia.org/wiki/Trinity_site

    This was an implosion device (Pu-239) and was essentially proof of concept, which was different than the "Little Boy" gun-type device, which used U-235. http://en.wikipedia.org/wiki/Little_Boy So the test didn't necessarily demonstrate that the gun-type system would work.

    The design calculations did not demonstrate how the devices performed, but rather that they would work. And even with all the calculations and the 'crème de le crème' of scientists, they still wanted it tested to 'know' that it worked. If it had been a dud, the bombings might have been delayed.
  10. Oct 6, 2006 #9


    User Avatar
    Science Advisor


    Just to add to what Astronuc said above; the first weapon tested at Trinity was a
    plutonium-fueled implosion or "Fat Man" type bomb the twin of which was dropped
    on Nagasaki. The other type was the uranium-fuel gun-assembled "Little Boy" bomb.

    Even though they had tested the "Fat Man" device at Trinity; the first bomb dropped
    which was on Hiroshima, was the "Little Boy" device. The reason was that, even
    though "Fat Man" had been tested; they still had more confidence in the untested
    "Little Boy" than the already tested "Fat Man".

    "Little Boy" could not be tested prior to use, because it took Oak Ridge YEARS to
    enrich the required amount of uranium for use in "Little Boy". On the other hand,
    the reactors at Hanford began operation in Sept. 1944; and in 9 months had produced
    enough plutonium for the Trinity device, the Nagasaki bomb; and one more "Fat Man"
    that wasn't used which was eventually tested in the "Crossroads Able" test on
    Bikini Atoll.

    The implosion bomb is more complicated than the gun-assembled type. So even
    though "Fat Man" was tested at Trinity; they had more confidence in "Little Boy"
    About 2 weeks before Trinity, a test was made of the implosion system without the
    plutonium core. Measurements indicated that had the plutonium been present; the
    bomb would have failed to work. This test however was made with rejected parts
    for Trinity that chief explosives chemist George Kistiakowsky had "repaired". So
    they weren't the best parts; as the test showed.

    Dr. Gregory Greenman
  11. Oct 6, 2006 #10
    The US is currently working on testing its next series of warheads via computer. Thanks to inertial confinement fusion experiments and improvements in computers a few research labs are pretty close to being able to simulate the entire detonation.

    I saw an article the other day about scientists at los alomos simulating the first second of a hydrogen bomb detonation.
  12. Oct 6, 2006 #11


    User Avatar
    Science Advisor


    Do you know where you saw that article?

    I'd like to see the specifics.

    Lawrence Livermore can also do weapons simulation,
    Dept. of Energy / NNSA press release of May 7, 2002:


    Note the difference in CPU times; Los Alamos required 122.5 days while using 1,920 processors
    while Lawrence Livermore accomplished the same thing in 39 days while using 1,024 processors.


    Dr. Gregory Greenman
    Last edited: Oct 6, 2006
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?

Similar Discussions: Do Nuclear Weapons Need to be Tested?