When it comes to detecting smoke, time is of the essence, so we designed a test that simulates a real emergency.
The first thing you need to know is that these products are intended for two main types of fires: smoldering fires and flaming fires.
A smoldering (or slow burning) fire refers to combustion that occurs on the surface of a solid material. Smoldering fires produce large amounts of thick smoke, but cannot sustain a flame. Think of a cigarette that burns slowly, but never actually bursts into flame.
A smoldering fire can and will turn into a flaming fire if appropriate measures are not taken immediately.
We are most familiar with flaming fires. These are usually caused by the ignition of a combustible material, such as a flammable substance, wood, paper, etc. They generally produce less smoke than a smouldering fire, but are more destructive.
We built a room that housed the smoke detector being tested and piped the two different types of smoke into it through a chimney. We then timed the speed of the response. The faster the smoke detectors went off, the better.
To simulate the smoke coming from a smoldering fire, I lit 300g of charcoal and let it heat up for 10 minutes before adding the smoke. I wanted to make sure that most of the coals were lit and that the smoke was thick enough to set off the alarm.
It was easy to get a roaring fire going, just 30g of shredded paper was enough. Again, I made sure there was a substantial amount of smoke present before I directed it up the chimney and into the room.
Much to the inconvenience of my neighbors, who had to endure random smoke alarms going off, and to whom I sincerely apologize, I repeated the entire process three times for each smoke alarm model and type of smoke tested, for a total of six test runs per smoke alarm. Finally, I averaged the response times. You can see them in the graph below: