LAUSR

Abstract The volume of smoke alarm sound in rooms (other than room of sound origin) in real houses and smoke alarm activation time in rooms in full-scale model houses using ionization, photoelectric and dual detector smoke alarms were determined in this study. The alarm sound level measurements indicated that the sound level in many locations is likely to be too low to provide reliable notification, particularly for sleeping people, if smoke alarms are not installed in every room. In addition, changing to a lower frequency (520 Hz square wave) alarm would further aid effective notification of building occupants. The smoke alarm activation measurements showed that the time to detection (given a particular smoke source) was influenced by door position (open versus closed), the room in which the fire occurs, the location (room or hallway) of the detector, the type of detector and the smoke alarm manufacturer. Furthermore time to detection is also influenced by the type and form of the material that is burning. It was observed that photoelectric smoke alarms had the highest incidence of non-activation and when they did activate they, on average, took longer to activate than ionization and dual (ionization and photoelectric) smoke alarms over all smoke sources considered in this study. It is concluded that to achieve early detection and provide adequate notification, smoke alarms are necessary in every room and should be interconnected.