As of Saturday morning, the situation in Geneva is looking a touch grim: according to the CERN press release repairs to the LHC after Friday’s accident will require two months or so of downtime, which given that the machine is expected to shut down in December for a winter break anyway means that proton-proton collisions may be off the menu until 2009.

It seems that as the magnetic fields were being increased in the machine as part of the commissioning (there was no beam in the machine at the time) there was a massive quench in sector 3-4 of the machine. A chart of the spectacular temperature rise can be seen here. According to the Resonaances blog

LHC-progress addicts report that pretty scaring entries were appearing in the LHC logbook this morning (fire alarm, power failure, helium leaking into the tunnel), though all the record seems to be deleted now.

When a magnet quenches it stops being superconducting and is reduced to being merely conducting — at which point it starts to resist the current flowing through it, and heating up in a hurry. As Cosmic Variance points out, this is something you’d expect to happen to the magnets now and then, and they are designed to withstand it. But a basic part of withstanding it would be not rupturing the cooling system and leaking helium out into the tunnel, which is what seems to have happened, with a loss of about a tonne of helium, according to physicsworld.com (which is apparently home to some of the LHC-progress addicts of which Resonaances speaks: respect). According to CERN, “a faulty electrical connection between two magnets … probably melted at high current leading to mechanical failure”.

The question now, I assume, is faulty component, faulty installation — or faulty design? Doubtless we’ll know a bit more soon.