Daniel Holz of Cosmic Variance passes along this tale from Lyn Evans, project leader of the Large Hadron Collider:*
They use superfluid helium to cool the superconducting magnets. One of the many weird properties of this stuff is that it has zero viscosity. Which means that, if there’s any sort of hairline fracture anywhere in the 27 kilometer long tunnel, the stuff comes spewing out, and very, very bad things happen. Every component, every joint, every one of the tens of thousands of tiny connections has to be perfect. It is this sort of failure which brought the machine to its knees shortly after commissioning, over a year ago.
The magnets are kept very, very cold; the superfluid helium is at 1.9 Kelvin (-271 Celsius), or a couple of degrees above absolute zero. We’re not talking a little vial in a laboratory being kept at this temperature. We’re talking many thousands of tonnes of magnets, kept just above absolute zero (using 96 tonnes of liquid helium). As things cool down, they naturally contract. The decks on bridges do the same thing, hence those serrated grills at the ends of bridges to absorb the expansion and contraction due to weather (if you’ve ever motorcycled across a bridge, you know exactly what I’m talking about). There are equivalent serrated joints in the LHC beam pipe to ensure that it doesn’t contract and rip open upon cooling (which, needless to say, would be bad). But upon reheating a section of the LHC, it turned out some of these devices left little fibers in the beam tube. Not good. How to find them, without ripping open the entire collider (costing millions of dollars and setting the project back precious months)? They ended up blowing a ping pong ball (with electronics embedded) down the tube, and tracking where it would get stuck. A simple, elegant, cheap solution to fix a multi-billion dollar enterprise.
* Initializations may be preferable in this case.