(3) Detectors D and E took substantially longer times to activate than the times
estimated by the computations. Detector G, however, activated sooner than the
computation indicated; it responded even before the ceiling jet temperature
reached the fixed rating temperature of the detector.
Additional tests with the identical settings also confirmed the same trend
mentioned above, i.e., excellent matches except the cases with detectors D, E, and G.
Detectors F and G were ‘‘rate compensated’’ fixed temperature detectors. They
respond at a pre-set fixed temperature when the surrounding temperature rises
relatively slowly. However, when the surrounding temperature rises fast, they tend to
activate sooner than the surrounding temperature reaches the pre-set fixed
temperature. The temperature measurement indicated that detector F might have
activated at the pre-set temperature. But it was clear that detector G was activated by
the rate of temperature rise rather than by the fixed temperature, which made the
estimated time using the RTI based on a fixed temperature rating irrelevant.
In order to see the behavior of the high-temperature rating detectors more closely,
the fire source was moved closer to the detectors. In the following tests, the center of
the pan was located 3.0m away in the radial direction and 3.0m away in the vertical
direction from the detectors. Figs. 23 and 24 show, respectively, the ceiling jet
temperatures and the ceiling jet velocities measured at velocity probes 1 and 2.
Fig. 25 shows a comparison of the measured and the computed response time of each
detector.