h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

There are several other types of reactors other than BWR and PWR; in fact most UK reactors are gas cooled. (Gas is pumped through the reactor core, and then through a heat exchanger that heats water to drive the turbine). The other popular coolant is pressurised liquid sodium; this is used in fast breeder reactors (experimental things that don't work properly, but would be wonderful if they did. Alledgedly) and in nuclear subs (which work in the same way as a power station only there's a propeller where the generator would normaly be). Using liquid sodium in nuclear subs always worries me a bit; if it touches water, it will explode.

Yeah, I wrote this from the perspective of American nuclear industry.

There's a new kind of reactor in South Africa that's gas cooled, and rather than fuel rods, they just have a pressurized drum full of uranium spheres, about the size of a baseball.

And of course, there's the Soviet RBMK (Chernobyl) style reactor, which uses graphite as a moderator instead of water. The disadvantage is that while water doesn't burn, graphite does, which we learned in the Chernobyl disaster.

Uran baseballs? That doesn´t sound very good to me. From what I know baseball-size is excactly when the uranium reaches critical mass. So the whole thing should be able to explode spontanously.

The British gas cooled reactors are also graphite moderated and I'm sad to say we Brits demonstated the ability of graphite in reactors to burn at Winscale in the 1950's when one of the first british (military)reactors went up in flames, well before the much worse disaster at Chernobyl.
http://www.british-energy.co.uk/media/mn_factfiles_windscale.html

The explosion at Cherobyl was caused not by the graphite but by hydrogen released when the reactor overheated and the steam produced reacted with the fuel rods.

See this link for a better written description and the whole sorry saga which led to it.
http://www.british-energy.co.uk/media/mn_factfiles_chernobyl.html

Well, a nuclear reaction reaching criticality does not necessarily result in an explosion. That's what I would call a super-critical event. Nuclear reactors operate AT critical mass - otherwise they'd "go out" like a candle with a very short wick. Don't confuse a nuclear reactor with a nuclear weapon - the two are very different in terms of how material is reacted, how much and for what reason (obviously).

I know the difference between a nuclear plant and a nuclear bomb. But most ppl don´t, that´s why they´re so radiophobic. But the old style nuclear bombs were made by having two half-baseballs of uran. Then the explosion was started by putting them together, and Nature took care of the rest. That´s why I´m worried.

I'd guess it's actualy a ball containing a mixture of Uranium and a moderator (graphite at a guess), so there'll be a lot less than a whole baseball full. That way, the only difference between it and a normal gas-cooled reactor is that the 'fuel rods' are 'fuel baseballs'. YOu need to get some moderator in there somewhere.

Without going into too much detail, there is quite a difference between letting "baseballs" of uranium sit around each other in a reactor and how "baseballs" of uranium are brought together in the old atomic bomb. Not only that, but the isotopic content, or grade, of the uranium is different in reactors than in weapons. All I'm trying to say is that I think those South African reactors are nothing to worry about, at least from a nuclear detonation standpoint. I don't mean to sound condescending when I say this, but the engineers typically design things like reactors so they won't explode. That would just ruin our donuts, you see.

Old bombs didn't really have "two half-baseballs" of U; one design had a hollow sphere of U into which a "pit" of U was fired by a special cannon. The bringing together of these caused the U to go critical and detonate.

Bomb material needs to be at least 70% fissionable material in order to explode. That means in the heart of a nuclear weapon, a majority of it will be U-235 or Pl-239.

In commercial fuel, the portion of fissionable material is around 5% or so. Depending on the reactor design, it might be more or less. That concentration is picked so that no matter how close you bring the fuel rods (or spheres), they won't go so supercritical as to explode. They might get hot and melt, as in the case of Three Mile Island, but a nuclear chain-reaction explosion is impossible.

In the South African reactor, they mathematically showed that if the same design were scaled down to generate 100 megawatts or less, the fuel would never be able to get hot enough to melt, even if the operators turned off every active form of cooling. Unfortunately, the South African reactor isn't that small; I think it's around 600 MW or so.

As a former US Navy nuclear professional and a current US Nuclear Industry professional, I want to clarify a few things:

1) ALL reactors operate at critical levels. Super-Criticality is when power is rising, while sub-criticality is when power is lowering. When they scream in the movies that the "Reactor has gone CRITICAL!!!!!" it's the same as saying "The reactor has gone STABLE!!!". Movies are sooo full of it these days....
2) Nuclear submarines don't use Liquid Sodium reactors. In fact, I don't know of ANY non-research US reactors that do anymore.
3) Criticality applies to reactors, and critical mass to atomic weapons, they are NOT interchangeable.
4) Both the Cherynobyl and Three-Mile Island accidents were caused by Operator Error and exascerbated by equipment design and procedures. Nuclear Power, if done correctly, is perhaps the most safe and boring thing in the world.

Hope this all helped....