As building alarm systems get more pervasive and sophisticated, all best efforts can still not prevent unanticipated operation. Here’s one incident that I don’t mind not being a part of…
…On August 7, 1997, a training department representative at the Haddam Neck nuclear generating plant was taking pictures of the fire system indicators and controls to develop training aids to upgrade the Halon system training manual. The training representative was using a Canon PS1001, Power Shot 600 digital camera to take pictures of the ANSUL Autopulse 2000 Halon control panel mounted on the wall in the control room. When the training representative took a flash photograph of the alarm reset/silence push buttons inside the Fire Detection System (FDS) control panel, the first flash caused an annunciator inside the panel to sound. The cabinet door on the panel was closed, and an examination of the front panel showed no lock-in alarm indications. The cabinet door on the panel was reopened, and a second flash photograph was taken within two minutes of the first picture. The second flash caused a second alarm with a different tone, indicating that system actuation was imminent. Within three to five seconds of the second flash, Halon discharged from the overhead nozzles.
Subsequent testing confirmed that the light from the camera flash affected an EPROM microprocessor located inside the Halon FDS control panel. The camera flash caused the normal one-minute delay to be bypassed and resulted in an almost immediate actuation of the Halon system. Discussions with the manufacturer indicated that a strong light source could cause an unpredictable perturbation within the EPROM, depending on the light intensity and the angle of incidence to the circuitry through the EPROM's window…[Excerpted from:
http://www.hanford.gov/lessons/sitell/ll98/199805.htm After indoor Halon discharge, it is routine to conduct hearing tests on the occupants. It is supposed to be a very noisy event.]
