I was working to my thesis in nuclear physics, and I was developing a new analytical method based on beam-induced radiation spectrometry. In one of the beam lines we had an alpha particle detector (basically a large Si crystal with a gold-plated surface) that was driving me crazy. When we used the detector for measurements it worked perfectly, but when we turned the beam off and I entered the radiation facility to do some measurements, it was insanely noisy and could not be properly calibrated. I was staring at the screen of the scope looking at the huge noise floor and scratching my head, when at once the noise disappeared, then came back again. This repeated for a few times and suddendly I realized what was happening. The vacuum chamber where the detector was installed had a glass window hidden to my sight that was sometimes used for remote inspection with a TV camera. A technician was doing some maintenance work, walking back and forth in front of the window. When he was close to it the noise disappeared, and when the window glass was clear of obstacles the photoelectric effect in the detector caused the noise! During irradiation tests nobody was inside the facility, so we turned off the lights, and this explained why the detector worked without problems. A piece of duct tape on the glass window fixed the problem for ever, as always...
