Unless you’ve had to take care of an NMR facility, you might not have realized how many large chunks of ferromagnetic material might be moving around close to your building, and how much stray radio-frequency noise is banging around. Here’s a story on the University of Minnesota, where a research building sits right next to a light rail line, and I can easily believe that they’re having problems. A lot of folks in Cambridge and Boston can tell you stories about the trains (above and below ground) and their effects on NMR experiments.
It’s not just electromagnetic effects, of course. Good old vibration will hose things up, too, since a high-field NMR magnet needs very precise positioning of the sample and the probe for a sharp spectrum. That’s why the big magnets are always sitting on top of very expensive vibration-damping legs, and the bigger the magnet, the more impressive the technology that goes into canceling out the shakes. But radio noise is a real killer. The more machines you have, and the more nuclei you observe, the higher your chances of picking up police radios, having your observed frequencies wander into the commercial FM band (good luck there), and who knows what else.
Sometimes you can get around a specific problem by running your NMR at a bit less than its rated strength, which shifts the corresponding RF-observation windows. That seems like a shame at first (after all, you certainly paid for a 400 MHz magnet or what have you), but it’s a lot better to have a clean spectrum at 380 than it is to have unpredictable crap at 400. Good luck to the folks at Minnesota with their light-rail problems, though. At least they know where the noise is coming from!